DRAFT DOCUMENT- V4
Irradiation of VPT faceplates - part 1
Peter R Hobson
Department of Electronic and Computer Engineering
Brunel University
Uxbridge
UB8 3PH
Samples
A number of faceplate samples were provided by Electron Tubes (UV glass), by Hamamatsu (borosilicate, UV glass, quartz, radiation-resistant glass) and Dmitri Seliverstof (C96-1 glass). In addition two samples of optical quality glass with added cerium oxide were provided by Schott, UK. All samples were flat and polished on both sides.
Source
All samples were irradiated with 60Co photons (E=1.173 MeV and 1.333 MeV) in the dark at a constant temperature of 19.5 C.
Schedule
The irradiations were performed at the dose rates given in table 1. The dose in air was measured using a Farmer dose meter (type 2502/3 from Nuclear Enterprises with a 0.6 cm3 cell). The doses quoted in the table were corrected from dose in air to a dose in Pyrex glass (80% SiO2). Not all the glass samples received the fifth irradiation step.
Irradiation / Rate (Gy.hr-1) / Dose (Gy) / Cumulative Dose (Gy)First / 1.66 / 121 / 121
Second / 11.0 / 512 / 633
Third / 11.0 / 1839 / 2472
Fourth / 11.0 / 1558 / 4030
Fifth / 11.0 / 140 / 4170
A second set of irradiations, at 11Gy.hr-1, was performed on three UV glass faceplates from Hamamatsu (figure 8)
Absorption measurement
Samples were measured using a Perkin-Elmer Lambda-9 spectrophotometer within one hour of the end of each irradiation. The data were taken at 1nm.s-1 with a 2.0 nm resolution and an integration time constant of 0.5 s. The measurement aperture (illuminated area on the sample) was 5 mm. The data plotted are the induced absorption, that is after subtraction of the original unirradiated value. Some baseline correction, using data at 650 nm, has been used due to small instrumental offsets that affect certain samples.
Figure 1 shows the absorbances, uncorrected for refractive index or sample thickness, before irradiation of all the glass samples discussed in this report. The absorbance shown here is dominated by Fresnel reflectance, at long wavelengths, or by a cerium absorption edge (glasses C96-1, BK7G18 and K6G05) at short wavelengths. Some instrumental effects have led to small baseline shift in EB01/2, whose unirradiated absorbance is too low.
Figure 1Uncorrected absorbance before irradiation of the samples discussed in this report
Results
Electron Tubes UV glass
Glass is Schott 8337B
Sample thickness was measured to be 2.41 mm. The results are shown in figure 2
Figure 2Radiation induced absorbance (normalised to a thickness of 1mm) for the UV glass sample provided by Electron Tubes
Hamamatsu UV glass
Glass is Schott 8337B
Sample thickness was measured to be 1.06 mm. The results are shown in figure 3. Subsequently three further samples were provided and tested, they are discussed below (figure 8).
Figure 3Radiation induced absorbance (normalised to a thickness of 1mm) for the first UV glass sample provided by Hamamatsu
Hamamatsu borosilicate glass
Glass is a standard Schott borosilicate used as the default choice for photomultiplier faceplates
Sample thickness was measured to be 1.06 mm. The results are shown in figure 4
Figure 4Radiation induced absorbance (normalised to a thickness of 1mm) for the standard borosilicate glass sample provided by Hamamatsu
Russian radiation resistant glass
This is a cerium oxide doped glass produced in Russia
Sample thickness was measured to be 1.61 mm. The results are shown in figure 5. The improvement in the absorbance with radiation as evidenced by the negative numbers in the UV and blue, although unexpected is not unheard of for glasses. There is no evidence of systematic measurement error.
Figure 5Radiation induced absorbance (normalised to a thickness of 1mm) for the cerium oxide doped radiation resistant glass sample provided by Dmitri Seliverstof.
Schott radiation resistant optical glass
Two samples of Schott optical glass which have had cerium oxide added were provided by the UK agents. These have excellent optical quality and would in general be very closely controlled batch to batch. Schott list a number of radiation resistant optical glasses, but only about four, including the two tested here, are readily available at present.
Sample thickness of the BK7G18 was 5.22 mm and the K5G20 was 5.19 mm
Figure 6Radiation induced absorbance (normalised to a thickness of 1mm) for the BK7G18 sample provided by Schott, UK
Figure 7Radiation induced absorbance (normalised to a thickness of 1mm) for the K5G06 sample provided by Schott, UK
Hamamatsu UV glass (second batch of samples)
Glass is Schott 8337B, these samples were sent at a later date than the first sample tested (figure 3). It is not known if they came from a different melt of 8337B
Sample thickness was measured to be 1.0 mm. All three samples showed no sign of radiation damage effects at either of the two doses. The induced absorption data for the worst sample are shown in figure8
Figure 8Worst case radiation induced absorbance (normalised to a thickness of 1mm) for the second set (three) of UV glass faceplates from Hamamatsu. ) for the first UV glass sample provided by Hamamatsu
Annealing
The change in induced absorbance of two of the faceplates made from Schott 8337B glass was measured after the samples had been kept for several months in the dark at normal, but uncontrolled, laboratory temperature. Both the sample from Hamamatsu (HB02) and that from Electron Tubes (EB01/2) showed annealing at all wavelengths of interest over the period of 280 and 287 days respectively (figure 9a and 9b).
Figure 9aAnnealing of induced absorbance after 280 days in the dark for glass HB02
Figure 9bAnnealing of induced absorbance after 287 days in the dark for glass EB01/2
Discussion and conclusions
The significant induced absorbance shown by the standard borosilicate glass faceplate (figure 4) clearly shows why a radiation-resistant glass must be chosen. The three cerium oxide doped glasses (C96-1, BK7G18 and K5G06) are all excellent but suffer from two problems. In all three cases the unirradiated absorbance is high at wavelengths below 400 nm[1] (the samples are all a pale yellow colour), and the thermal properties of the two Schott glasses are such that graded seals would be needed increasing the length and cost of a VPT. It is not known whether the Russian C96-1 glass suffers from this second problem.
The Schott 8337B UV glass faceplates are significantly more radiation resistant than standard borosilicate. This is in part due to the damage occurring in the mid-UV region so that in our application we only see the longer wavelength tails (the unirradiated cut-off is around 230 nm). The first two samples (figures 2 and 3) show very similar damage profiles, with the sample from Electron Tubes being systematically more resistant. However the second test, on three samples from Hamamatsu, showed essentially no damage at the maximum dose of 3600 Gy, at least within the noise levels of our measurements.
This variability in the performance of the 8337B glass is likely to be intrinsic to a technical glass whose characteristics are not tightly controlled in the way that they are for optical glasses such as BK7G18. The author is unaware of previous studied made on the 8337B glass so cannot comment on how typical these fluctuations might be. A third set of three UV faceplates from Hamamatsu will be irradiated in October 1999.
From this study one can conclude
- Cerium doped radiation resistant glasses are a good choice from the point of view of damage, but they will cut off some of the PbWO4 scintillation light even when undamaged and will require (for Schott glasses at least) expensive and space consuming graded seals
- Some useful self-annealing of both the damaged Schott 8337B UV glass has been observed.
- The Schott 8337B UV glass is a good choice for the faceplate even although it does show some damage. Making a pessimistic assumption about how the induced absorbance might scale with dose, Derek Imrie showed[2] that the maximum loss in scintillation light at the inner edge of the calorimeter would be less than 14% after ten years operation at high luminosity (see figure 10 reproduced from his figure 4)
Figure 10.Relative photoelectron response as a function of at an integrated luminosity of 5x105pb-1. The VPT faceplate is assumed to be made from the Electron Tube UV glass sample EB01/2.
- Further tests on UV faceplates up to a total integrated dose of 50 kGy (or higher) must be performed as a matter of urgency
- Batch testing of 8337B glass before faceplates are sealed to VPT tube assemblies is strongly recommended.
Page 112/02/19
[1] P R Hobson "Schott Radiation Resistant Glasses" 2 March 1998
[2] D C Imrie "VPT Faceplates Of UV-Transmitting Glass?" 3 October 1998