Peter Hobson Brunel University, UK3 July 2003

1
VPT faceplate testing

• Samples of faceplate glass are irradiated to
  approximately 20 kGy gamma ray dose in our
  facility at Brunel.
• At least 3 faceplate samples per glass batch are
  tested.
• Samples are irradiated in the dark, at a
  temperature around 20C, at a dose rate of abour
  300 Gy per hour.
• Induced absorption is measured within two hours
  of irradiation.
• Irradiated samples are stored in the dark at
  201C.
• Spectrophotometer is recalibrated every 12 months.

2
VPT faceplate testing

• We use the latest measured PbWO4 emission
  spectrum appropriate for end-cap crystals
• This is convolved with the induced absorption
  over the wavelength range 350 nm to 610 nm
• The integrated transmission after irradiation
  must be gt 90 of the unirradiated transmission
  for all sample in one batch.

3
VPT faceplate testing

• Irradiated causes the growth of an absorption
  band centred around 330 nm. This band is well
  known in the literature for borosilicate glasses.
• Even after 20 kGy dose there is no induced
  absorption at 800 nm. This allows a measurement
  consistency check.
• The induced absorption shape is very similar in
  different batches, peak height does however vary.

4
VPT faceplate testing

• Recent faceplates (June 2003) were generally very
  good quality, but note batch 29992.

5
VPT faceplate testing

• One sample from batch 29992 showed very poor
  radiation tolerance.

6
VPT faceplate testing

• Long term annealing (in dark at 20C) shows
  increase in absorption in the band centres at 330
  nm.

7
Summary

• Pre-testing of faceplate glass is an essential
  part of ensuring that production tubes will be
  able to operate over the full lifetime of CMS.
• We have a standard protocol for determining which
  glass batches can be used.
• A new Hitachi U4100 spectrophotometer, used for
  all glass batches in the production run, has
  improved our measurement accuracy repeatability.
• Recent glass batches have been very good.
• The anomalous result for one sample of a batch
  (29992) is unique all other batches have shown
  very little variation between samples.