To whom it concerns:
====================

This is a summary of experiences with dismounting, disassembling, cleaning,
painting, glueing, filling, bubbling, reassembling and testing the EUROBALL
neutron detectors, collected in the period 1997 - 2003, written by

Hubert Grawe

GSI, April 2004
===========================================================================

1. Dismount
-----------

-- Dismount PM covers and voltage dividers.

-- Remove metshield from PM's.

-- Remove PM's from glass and optical grease carefully, don't break PM's.

-- Open and remove bellow blank-off flange.

-- Empty liquid into container, don't spill the liquid, use rubber gloves.

-- Leave open detectors under extractor hood for a day to let residual liquid
   dry.

2. Disassembly and Cleaning
---------------------------

-- Break glass windows mechanically (hammer!) and/or by heating the glass
   glue.

   Note: Ask Bicron what is the best procedure with their glue.
   -----

-- Remove remaining glass parts and glue in the groove by heating and with
   a portable grinding machine.

-- Soak inside with about 1 cm of acetone over night to soften the paint,
   which will be done mainly by the acetone vapor.

   Note: Cover cans with lid, acetone is agressive and dangerous, this
   ----- should possibly happen under an extractor hood.

-- Peel off paint as well as possible. There will be surfaces and edges and
   corners which need further treatment, as scratching surfaces and
   repeating the ''soaking'' procedure with acetone.

-- Close bellows with blank-off flanges and fix them with the stabilizing
   cylinders, protect ''delicate'' surfaces against sandblasting.

-- Sandblasting in the workshop. It will generally turn out that still
   some paint is left in corners and edges. Removing the glue does
   not seem to be a problem.

-- With repeated acetone exposure these little pieces of paint can be
   removed.

-- Repeated sandblasting will not be necessary, but a thorough cleaning
   procedure of the can is required to remove any sand from can and bellow
   ribs.

-- Generally by the chemical treatment the stabilizing cylinders will come
   off if glued. Experience shows that the bellow glueing will normally
   not survive either. Therefore the cylinders and bellows should routinely
   be removed after this procedure.

   Note:
   -----
   It should be considered and investigated, whether it is not possible to
   use an ultrasonic bath with an appropriate chemical solvent (acetone?)
   to remove paint and glue for bellows and stabilisers. There are
   certainly companies, who have capacious enough equipment. It must be
   checked and warranted, however, that the electro-braised joints survive
   the procedure.

3. Painting
-----------

-- The detectors should be carefully cleaned and rinsed with a chemical
   solvent before painting to remove fat, oil and sand remainders. If no
   sandblasting was done (see last Note), it is advisable before the final
   cleaning to sandblast the inner surfaces of the can to make the paint
   stick better to the Al walls. The normally oxidised Al surface does not
   accept the paint well.

-- Paint should be mixed and prepared according to the prescription of the
   provider (Bicron). Don't prepare too much of the paint as it should be
   applied in several thin layers.

-- Use a small flat brush less than 10 mm wide (or diameter) for the
   painting.

-- Apply the paint to the walls and bottom one by one only in the horizontal
   position to avoid dripping paint that will produce too thick a layer.

-- Before painting the next wall wait for some minutes to let the paint dry.

-- At last paint the PM side of the chamber with not too much paint and an
   even smaller brush. Don't forget to paint the inner circumference of the
   window opening and bellow holes in the base plate of the can.

-- Normally, depending on the consistency of the paint, two layers of paint
   are enough, if not otherwise recommended by the provider. Before
   applying the second layer some hours of drying under an extractor hood
   are recommended. After inspection in places where the paint is too thin
   a third layer may be added, again after drying. The inside of the can
   after painting should look uniformly white.

-- Let the paint dry for at least a day under an extractor hood.

   Note: It is recommended to perform the painting procedure under stable
   ----- temperature and humidity conditions in an air-conditioned room.

4. Glueing
-----------

-- For glueing the glass window follow the instructions given by the
   provider (Bicron) to prepare the glue. Clean carefully the groove in the
   can base plate and make sure that the glass is totally free of fat
   (and of course dust etc.).

-- Before glueing the glass disk make sure the inside of the can is
   completely clean of dirt and dust. This is the last chance for access!

-- Apply a not too thick layer of glue in the groove and after insertion of
   the glass disk make sure, the glue is also uniformly distributed at the
   circumference between glass and Al. Experience shows that this provides
   the best stability against strain by temperature and pressure
   variations. Use a small spatula (less than 10 mm wide) for applying the
   glue.

-- Follow the same procedure for the bellows.

-- Avoid glue from dripping into the inside of the can.

-- Let the glue harden and dry as long as the provider recommends (usually
   some days at room temperature).

5. Filling, Bubbling and Closing
--------------------------------

-- Before starting make sure that you have all the equipment, material,
   tools, bits and pieces available that you (may) need. Be prepared to
   work fast if necessary. These are:
   -- rubber gloves, as you may come in contact with the scintillator
      liquid;
   -- cleaning tissue and waste can;
   -- scintillator liquid, preferably not in a big can or vessel, but in
      bottles; transfer liquid from can to bottle by an appropriate hand
      driven pump and siphon;
   -- funnel that fits into bellow inner diameter;
   -- stainless steel tubes (~6 mm diameter, ~200 mm long), tube connections
      to reducing valves, reducing valves mounted to a KF16 vacuum
      distribution system; a sixfold channel array allows to bubble 2
      detectors simultaneously;
   -- Ar gas bottle with reducing valve; gas purity should be good enough
      to use it in glove boxes;
   -- CF16 Al gaskets to close the bellows; 6 screws/nuts per flange;

-- The following procedures should be done under an extractor hood,
   preferably by two persons.

-- Fill liquid into detector chambers through funnel and bellow (two people
   needed to hold the bellow).

-- The liquid level in the can should be well below the base plate window
   side (~ 1cm) to avoid spilling in the bubbling procedure.

-- Before bubling make sure that the tube connections are tight and sealed.

-- Start bubbling immediately after filling 1-2 detectors. Adjust bubbling
   speed to warrant a smooth continuous flow without spilling liquid to
   outside the can. There will be splashes to the glass window which due to
   evaporation will cause stains, which will disappear after complete
   filling.

-- Against manufacturers recommendation bubbling for 12 hours is necessary.

-- After bubbling fill up the can with fresh liquid (can also be bubbled in
   a bottle). The filling level should at least be at half the bellow
   length, no big gas bubbles should be left in the can (test by tilting
   the can).

-- Two methods have been tried to close the bellows and were found to work
   equally well. The closing needs two persons.
   -- Compress the bellows as far as possible and fill liquid to the
      highest possible level. Then seal the flange as tight as possible.
      This makes sure there is no air left in the can and provides a small
      low pressure. Part of the volume will be filled by Ar gas solved in
      the liquid during bubbling. This is the dirty method.
   -- During sealing provide a little Ar gas flow to replace air in the
      bellow. For this procedure one needs not to compress the bellow. This
      was found to be the more convenient method.

-- Clean the outside of the detector carefully from spilled scintillator
   liquid with a "mild" solvent (e.g. ethylene) as too rapid evaporation
   cools too fast and can cause stress on glass and glue.

-- Mount the bellow stabilisers.

-- Before continuing work leave the detectors for a few days and inspect
   them carefully for leakage by turning them upside down, shaking, etc.
   During this time Ar bubbles may develop inside the can, which by turning
   the detector should be displaced into the bellows. This gas was dissolved
   in the liquid during bubbling.

6. Reassembling PM's
--------------------

-- Before starting make sure you have all equipment, tools, bits and pieces 
   like :
   -- PM tubes taped and wrapped with 3M insulating tape holding at least 
      2 kV voltage;
   -- PM tube light entrance windows cleaned from grease with ethylene;
   -- metshield cylinders prepared to put around the PM cup and shaft 
      (two different diameters); the cylinders must also be insulated on the 
      outside towards the PM tube covers;
   -- optical grease;
   -- soldering equipment to connect voltage dividers;
   -- distance pieces to press PM tube against grease covered glass window;
      they should be insulated, too;
   -- O-rings of various types for several purposes : 
      -- as distance spacer between PM tube cup / metshield cylinder and
         PM cover;
      -- in the O-ring groove in the window opening to seal the cover mount 
         against light;
      -- in the O-ring groove between cover lid and cover cylinder;   
   -- all sorts of necessary screws and nuts of proper type and length;

-- Place proper amount of optical grease in the middle of the glass window; 
   if it is too little, the contact is bad and glass maybe scratched or even
   damaged under applied pressure; if it is too much it will spill out on the 
   side after applying pressure.

-- Remove all air bubbles from the grease. There are various types of grease. If
   it is more liquid it will spread by itself over the window and one can insert 
   the PM tube. If it is more like a paste one can spread it by inserting the PM
   tube and pressing it against grease and window. Avoid turning the PM tube in 
   only one direction, one should turn it back and forth. This avoids smear and 
   flowmarks in the optical grease, which reduces light transmission. A gentle
   warm-up of the grease with a hair-dryer will help to remove air bubbles and 
   assist uniform spreading.

-- Insert metshield cylinders and PM cover and check carefully for HV-to-ground
   contact. This is the most common reason for failure and/or bad PM tube 
   performance.

-- Solder electrical lid-feedthrough connections to voltage divider and insert
   PM cover lid.

-- Adjust distance pieces to press PM against window, but not too much.

-- Before testing and applying high voltage wait for a day for decay of 
   fluorescence light emission from the PM cathode stimulated by exposure to 
   light during mounting in order to reduce the dark current.

7. Testing
----------

-- Apply high voltage and watch anode current for possible discharges.

-- Add a 60Co source and watch anode output, at around 1.5 kV one should see 
   pulses.

-- Adjust anode output pulses by HV to about 1 Volt (this is what I remember, ask
   Marek Moshinski for details).

-- To test the photo-electron yield follow the ''Marek Moshinski Method'' :
   -- connect anode output to spectroscopy amplifier, and the amplifier output to
      an ADC of a DAQ system or multichannel analyser;
   -- without gamma source increase amplifier gain to see the "single 
      photo-electron" peak, which should be clearly separated from the noise;
      typical gain value is x100;
   -- with a 137Cs gamma source and reduced gain (typically x1-10) look for Compton
      edge of 661 keV gamma line;
   -- from ratio of channel numbers, gain values the number of photo-electrons per
      MeV (normalise for 661 keV) can be calculated, which for a good detector
      and our multipliers should be between 1000 and 1500 phe/MeV for hexagons.

-- With a neutron source, a plastic reference detector and full electronics test 
   neutron-gamma discrimination in 2-dim zero-crossing vs. tof spectrum. 

8. Material, bits and pieces
----------------------------
Aside from the above mentioned equipment, material, bits and pieces the following 
items are needed :

-- BC501 liquid scintillator; provider : Bicron
-- BC ?  paint and hardener; provider : Bicron (old type BC622)
-- BC ?  glue and hardener; provider : Bicron
-- BC ?  optical grease; provider : Bicron
-- CF16  aluminum gaskets; (British) provider : Vacuum Generators, Maunsell Road,
   Castleham Industrial Estate, Hastings, East Sussex TN 38 9NN, England,
   Tel: +44-(0)1424-851291
-- metshield Vitrovac 6025, 50 mm x 0.025 mm, ~80 m needed;
   (German) provider : GELING GmbH, Regensburger Str. 215, D-90478 Nuernberg,Germany
                       Tel: +49-911-400028
======================================================================================