Wednesday, October 22, 2008

Dr. William Uhland-Covidien

Appendix B
Radiation Damage and Curing
The transparency of lead􏰀glass decreases as it is irradiated􏰁 Radiation damage
of lead􏰀glass leads to a yel lowing or darkening of the lead􏰀glass􏰁 The coloring of the
lead􏰀glass is caused by creation of color centers or by the development and growth
of absorption bands􏰁 The absorption bands decrease the amount of 􏰂
Cerenkov light which reaches the photomultiplier tube and􏰃 therefore􏰃 decrease the signal􏰁
A spectrophotometer which measures the transmittance or absorption of ma􏰀
terials has been used to monitor radiation damage􏰁 Several F􏰄 lead􏰀glass sample
pieces􏰃 􏰞 􏰅􏰆 􏰟 􏰄 􏰟 􏰄 cm􏰍 with only the two small end faces parallel and polished􏰃 have
been used to monitor the radiation damage􏰇exposure of the central calorimeter􏰁 The
calorimeter􏰈s lead􏰀glass blocks have not been used directly since they are contained
within the wedge shells and are inaccessible􏰁 Using the smaller lead􏰀glass sample
pieces􏰃 the spectrophotometer is used in its normal con􏰉guration 􏰊not as described
in Appendix A􏰋􏰁 In addition to the lead􏰀glass samples placed about the calorimeter
during the running􏰃 two pieces have been placed near the debuncher ring􏰈s injection
kicker magnet for a short period of time􏰁
Placement Azimuthal Distance
Comment Angle 􏰊cm􏰋
Upstream 􏰏x􏰀axis 􏰌
Upstream 􏰏y􏰀axis 􏰑􏰆
Upstream 􏰓x􏰀axis 􏰅􏰒􏰌
Upstream 􏰓y􏰀axis 􏰄􏰒􏰆
Downstream 􏰏x􏰀axis 􏰍􏰎􏰌
Downstream 􏰏y􏰀axis 􏰔􏰌
Downstream 􏰓x􏰀axis 􏰅􏰑􏰌
Downstream 􏰓y􏰀axis 􏰄􏰐􏰌
Table B􏰁􏰅􏰕 The azimuthal angles and radial distance from beam pipe of the lead􏰀glass
samples about the calorimeter􏰁
B􏰁􏰅 Placement of Lead􏰀glass Samples
The lead􏰀glass samples have been present constantly about the central calorime􏰀
ter whenever there has been beam in the 􏰖
p source􏰁 Twenty􏰀four samples have been
separated into sets of three and placed at eight locations about the calorimeter􏰁 Each
lead􏰀glass sample is wrapped individually in black paper􏰁 The locations have been
essentially along the experiment􏰈s positive and negative x and y axes 􏰊the z􏰀axis is
along the 􏰖
p direction and the y􏰀axis is pointing upward􏰋 upstream and downstream
of the calorimeter􏰁 Table B􏰁􏰅 shows the approximate azimuthal angles and the radial
distance from the beam pipe for the eight sets􏰁 The approximate polar angles at the
upstream end near the interaction point are 􏰔􏰆
for the sets along the x􏰀axis and
between 􏰒􏰄
and 􏰒􏰌
for the sets along the y􏰀axis􏰁 The downstream samples have
been connected to the central calorimeter cable support apparatus located 􏰄􏰐􏰌 cm
downstream of the interaction point􏰁
Two samples were placed near the debuncher ring􏰈s injection magnet􏰃 with no
shielding􏰃 during 􏰖
p source studies conducted in December 􏰅􏰑􏰑􏰆􏰁 In between the de􏰀
buncher and accumulator rings near the E􏰒􏰐􏰆 experimental area􏰃 concrete blocks
provide shielding for the detector􏰁 The placement of the shielding intercepts particles
coming from the 􏰖
p target area via the transport line and shields against radiation
produced at the nearby kicker magnets􏰁 Previous to the start of the experiment􏰃
these were the expected radiation sources for E􏰒􏰐􏰆􏰁
B􏰁􏰄 Radiation Exposure
Several radiation monitors have also been placed about the calorimeter to try
to measure the radiation dosage􏰁 The radiation monitors vary in type but all are
connected to a visual readout system in the E􏰒􏰐􏰆 counting room􏰁 Only two of the
radiation monitors have been continually attached to the central calorimeter􏰁 The
p source has operated for three di􏰗erent periods over two years with the central
calorimeter present􏰘 the radiation dosages received by the calorimeter according to the
two monitors are shown in table B􏰁􏰄􏰁 The two 􏰉xed target time periods had di􏰗erent
shielding arrangements and di􏰗erent injection line collimator settings􏰘 additionally􏰃
the injection line tuning improved with time􏰁 During the 􏰖
p source studies􏰃 the E􏰒􏰐􏰆
detector was removed from the beam line and moved as far away as possible within
the experimental area from the debuncher and accumulator beam lines􏰃 however􏰃 the
shielding was also removed and the collimators were opened wide􏰁 During data taking􏰃
Time Period Dosage 􏰊rad􏰋
Fixed target run 􏰅􏰑􏰑􏰆 􏰊May 􏰓 September􏰋 􏰞 􏰒􏰆 rad
Studies of the 􏰖
p source 􏰊December 􏰅􏰑􏰑􏰆􏰋 􏰞 􏰍􏰆 rad
Fixed target run 􏰅􏰑􏰑􏰅 􏰊June 􏰓 January􏰋 􏰞 􏰔􏰆 rad
Table B􏰁􏰄􏰕 The radiation dosages􏰃 according to monitors􏰃 that the central calorimeter
received in the 􏰖
p source􏰁
when the gas jet is operating􏰃 the radiation monitors showed small radiation doses as
compared to doses during stacking 􏰊the collection of 􏰖
A separate radiation monitor was placed near the kicker magnet when the two
samples were in place during the December 􏰅􏰑􏰑􏰆 􏰖
p source studies􏰁 This monitor􏰃 a
propane gas ion chamber􏰃 was operated in what is called neutron mode􏰃 which is a
factor of ten more sensitive to charged particles than neutral particles􏰘 the neutron
mode assumes that the radiation does not consist of charged particles􏰁 The nature
of the radiation is unknown and therefore the dosage determined from the monitor
may be an over estimate􏰁 The monitor measured a dosage of 􏰅􏰄􏰆􏰆 rad before failing􏰁
Estimating from other radiation monitors in the area􏰃 the two samples received an
additional 􏰅􏰆􏰆􏰆 to 􏰄􏰆􏰆􏰆 rad􏰁
B􏰁􏰍 Radiation Damage Analysis
The transmission spectra of the lead􏰀glass samples were made and recorded by
an HP􏰔􏰎􏰌􏰅A spectrophotometer before exposing them to radiation􏰁 One spectrum
is shown in 􏰉gure B􏰁􏰅 along with a spectrum from one of the 􏰌􏰆 cm long lead􏰀glass
Figure B􏰁􏰅􏰕 A lead􏰀glass sample􏰈s transmission spectrum and a 􏰌􏰆 cm long lead􏰀glass
block spectrum􏰁
blocks􏰁 The twenty􏰀four lead􏰀glass samples􏰈 transmittances have been re􏰀measured
after each time period as described in table B􏰁􏰄􏰁 None of the transmission spectra
show any change􏰃 outside of the spectrophotometer􏰈s intrinsic error􏰃 from the original
spectrum for the lead􏰀glass samples that were located about the central calorime􏰀
ter􏰁 The spectrophotometer􏰈s transmission measurement error for a 􏰄 nm wavelength
band interval is 􏰄􏰙􏰁 A 􏰄􏰙 change in the transmittance of a small lead􏰀glass sample
Figure B􏰁􏰄􏰕 The transmission spectra of one of the lead􏰀glass samples before and after
being placed near the debuncher ring􏰈s kicker magnet showing the a􏰗ect of radiation
corresponds to 􏰅􏰆􏰙 change in the transmittance of a lead􏰀glass block􏰁 The absorp􏰀
tion of the 􏰚ashlamp light would correspondingly change the 􏰚ashlamp response by
􏰄􏰆􏰙 since the light has to travel the block length twice􏰁 As stated in section 􏰎􏰁􏰎􏰁􏰅􏰃
there has been no measureble decrease of the 􏰚ashlamp response attributed to radia􏰀
tion damage􏰁 The lead􏰀glass samples about the calorimeter agree with the 􏰚ashlamp
However􏰃 the two samples which were exposed to the larger radiation dose have
shown radiation damage􏰁 Figure B􏰁􏰄 shows the radiation damage to the transmission
spectrum􏰘 the original transmission spectrum is shown for comparison􏰁 The radia􏰀
tion damage a􏰗ects the ultraviolet wavelengths more than the rest of the spectrum􏰁
Visually􏰃 the samples were darker and slightly brown􏰁
The radiation dose can also be estimated another way􏰁 Kirsebom and Sollie 􏰛􏰔􏰎􏰜
have parameterized the transmission properties of F􏰄 lead􏰀glass for absorbed doses􏰃
D􏰃 up to 􏰌􏰆􏰆􏰆 rad􏰁 The radiational absorption is de􏰉ned
a􏰊􏰢􏰘 D􏰋 􏰝 􏰅 􏰠 T 􏰊􏰢􏰘 D􏰘 x 􏰝 􏰅cm􏰋
To 􏰊􏰢􏰋 􏰘 􏰊B􏰁􏰅􏰋
where To 􏰊􏰢􏰋 is the non􏰀irradiated transmittance T 􏰊􏰢􏰘 D 􏰝 􏰆 rad􏰘 x 􏰝 􏰅 cm􏰋􏰁 The
parameterization found by Kirsebom and Sollie is
a􏰊􏰢􏰘 D􏰋 􏰝 􏰅 􏰠 e
􏰘 􏰊B􏰁􏰄􏰋
where 􏰖􏰊􏰢􏰋 is experimentally determined􏰁 The transmittance through x cm of lead􏰀
glass is
T 􏰊􏰢􏰘 D􏰘 x􏰋 􏰝 􏰛T 􏰊􏰢􏰘 D􏰘 x 􏰝 􏰅cm􏰋􏰜x 􏰊B􏰁􏰍􏰋
and substituting in appropriately the transmittance is
T 􏰊􏰢􏰘 D􏰘 x􏰋 􏰝
􏰕 􏰊B􏰁􏰎􏰋
Several values of 􏰖 for several representative wavelengths have been determined
by Kirsebom and Sollie􏰁 Using these values for 􏰖􏰊􏰢􏰋 and the transmission spectra
Wavelength 􏰖 Calculated Dose 􏰊rad􏰋
􏰊nm􏰋 􏰊􏰅􏰆
􏰋 Sample 􏰅 Sample 􏰄
􏰍􏰌􏰆 􏰎􏰆􏰁􏰒 􏰎􏰍􏰎 􏰎􏰑􏰆
􏰎􏰆􏰆 􏰍􏰅􏰁􏰔 􏰐􏰑􏰄 􏰐􏰄􏰄
􏰎􏰌􏰆 􏰄􏰅􏰁􏰐 􏰒􏰔􏰄 􏰐􏰐􏰆
􏰌􏰆􏰆 􏰅􏰍􏰁􏰍 􏰔􏰍􏰅 􏰐􏰒􏰆
􏰐􏰆􏰆 􏰌􏰁􏰅 􏰐􏰌􏰄 􏰌􏰎􏰎
Table B􏰁􏰍􏰕 The calculated radiation dose for the two lead􏰀glass samples using 􏰖􏰊􏰢􏰋
values determined by Kiresbom and Sollie 􏰛􏰔􏰎􏰜􏰁
before and after irradiation􏰃 a calculated radiation dose can be determined􏰕
D 􏰝 􏰅
􏰖􏰊􏰢􏰋x ln 􏰣
T 􏰊􏰢􏰘 D􏰘 x􏰋 􏰤 􏰘 􏰊B􏰁􏰌􏰋
where Tx 􏰝 􏰛To􏰜
􏰘 the measured spectra are always for the transmittance through
x cm of lead􏰀glass􏰁 Table B􏰁􏰍 shows the calculated doses at the various wavelengths
for the two lead􏰀glass samples􏰘 the 􏰖 parameter is also shown􏰁 The radiation dosage
appears to be between 􏰐􏰆􏰆 and 􏰒􏰆􏰆 rad and disagrees with the radiation monitor
determination􏰁 Several explanations are possible for the discrepancy of a factor of 􏰄
􏰊the radiation monitor result before failure􏰋 to 􏰌 􏰊the largest extrapolated dosage􏰋􏰁 A
basic reason could be that the F􏰄 lead􏰀glasses are not the same􏰘 however􏰃 one would
not expect such a large di􏰗erence􏰁 Another possible explanation is that the radiation
monitors have overestimated the dosage since the nature of the radiation is unknown􏰁
A last possible explanation is that instant bleaching􏰃 or curing􏰃 of the the lead􏰀glass
occurs when exposed to light􏰁 The lead􏰀glass samples were exposed to arti􏰉cial lights
while being unwrapped and placed in the spectrophotometer􏰈s sample compartment
before a transmission spectrum was taken􏰁 The most reasonable explanation is that
the radiation monitors have overestimated the radiation dosage􏰁
􏰖 􏰊􏰅􏰆
􏰢 Sample 􏰅 Sample 􏰄
􏰊nm􏰋 􏰅􏰄􏰆􏰆 rad 􏰄􏰆􏰆􏰆 rad 􏰍􏰆􏰆􏰆 rad 􏰅􏰄􏰆􏰆 rad 􏰄􏰆􏰆􏰆 rad 􏰍􏰆􏰆􏰆 rad
􏰍􏰌􏰆 􏰅􏰐􏰁􏰅 􏰑􏰁􏰐􏰌 􏰐􏰁􏰎􏰄 􏰅􏰐􏰁􏰐 􏰑􏰁􏰑􏰐 􏰐􏰁􏰐􏰎
􏰎􏰆􏰆 􏰅􏰔􏰁􏰍 􏰅􏰅􏰁􏰆 􏰒􏰁􏰍􏰎 􏰅􏰐􏰁􏰍 􏰑􏰁􏰒􏰐 􏰐􏰁􏰌􏰄
􏰎􏰌􏰆 􏰅􏰎􏰁􏰅 􏰔􏰁􏰎􏰌 􏰌􏰁􏰐􏰄 􏰅􏰅􏰁􏰑 􏰒􏰁􏰅􏰄 􏰎􏰁􏰒􏰌
􏰌􏰆􏰆 􏰑􏰁􏰅􏰐 􏰌􏰁􏰌􏰆 􏰍􏰁􏰐􏰒 􏰒􏰁􏰎􏰄 􏰎􏰁􏰎􏰐 􏰄􏰁􏰑􏰒
􏰐􏰆􏰆 􏰄􏰁􏰒􏰐 􏰅􏰁􏰐􏰐 􏰅􏰁􏰅􏰅 􏰄􏰁􏰍􏰅 􏰅􏰁􏰍􏰔 􏰆􏰁􏰑􏰄
Table B􏰁􏰎􏰕 The calculated 􏰖􏰊􏰢􏰋 values for three radiation doses􏰁
On the other hand􏰃 assuming that the radiation monitor results are correct􏰃 the
two sets of spectra can be used to determine 􏰖􏰊􏰢􏰋􏰁 Three values of radiation dose
have been used to determine 􏰖􏰊􏰢􏰋 and are presented in table B􏰁􏰎􏰁 The determined
􏰖􏰊􏰢􏰋 values are about a factor of three to four di􏰗erent than the values determined
by Kirsebom and Sollie 􏰊table B􏰁􏰍􏰋􏰁 One does not expect the lead􏰀glasses to be
that di􏰗erent􏰘 the dosage determined from the radiation monitors is probably an
Assuming that 􏰖􏰊􏰢􏰋 values of Kirsebom and Sollie are correct􏰃 the expected
damage to the lead􏰀glass samples on the calorimeter can be calculated􏰁 A 􏰍􏰙 to
􏰌􏰙 change of the transmittances for the wavelengths between 􏰍􏰌􏰆 nm and 􏰎􏰄􏰆 nm
is expected after a 􏰅􏰆􏰆 rad dose􏰁 The spectrophotometer is capable of showing this
change in transmittance􏰁 A di􏰗erence has not been seen and possible explanations􏰃
as indicated above􏰃 are either 􏰊i􏰋 the radiation monitor has overestimated the dose or
􏰊ii􏰋 instant bleaching􏰁
Cycle Exposure Time 􏰊min􏰋
Number Period Cumulative
􏰅 􏰅􏰌 􏰅􏰌
􏰄 􏰅􏰌 􏰍􏰆
􏰍 􏰍􏰆 􏰐􏰆
􏰎 􏰐􏰆 􏰅􏰄􏰆
􏰌 􏰅􏰔􏰆 􏰍􏰆􏰆
􏰐 􏰅􏰔􏰆 􏰎􏰔􏰆
􏰒 􏰅􏰔􏰆 􏰐􏰐􏰆
􏰔 􏰍􏰐􏰆 􏰅􏰆􏰄􏰆
􏰑 􏰍􏰐􏰆 􏰅􏰍􏰔􏰆
Table B􏰁􏰌􏰕 The exposure time periods and cumulative time that the two lead􏰀glass
samples were exposed to sunlight􏰁
B􏰁􏰎 Curing
Physics practice has been to expose irradiated lead􏰀glass to sunlight for curing􏰁
The estimated recovery time from radiation damage for lead􏰀glass is estimated to be
􏰌􏰆 years 􏰛􏰔􏰌􏰜 􏰊from measurements over a period of 􏰅􏰁􏰌 years of lead􏰀glass not exposed
to light􏰋􏰁 During May 􏰅􏰑􏰑􏰅􏰃 the two most irradiated lead􏰀glass samples were exposed
to sunlight for di􏰗erent time periods􏰁 Two non􏰀irradiated lead􏰀glass samples were
also exposed to the sunlight as controls􏰁 Table B􏰁􏰌 shows the exposure time periods
and cumulative times􏰘 the uncertainty of the time of exposure is a few minutes􏰁 The
exposure periods were scattered over a few weeks whenever there was not a threat of
rain􏰘 some of the exposure periods were while the sky was overcast􏰁
The never􏰀irradiated lead􏰀glass samples spectra showed no changes and contin􏰀
ually showed that the reproducibility of the spectrophotometer results􏰁 Figure B􏰁􏰍
shows the radiation damaged transmission spectrum and the spectra after seven of the
Figure B􏰁􏰍􏰕 The transmission spectra after di􏰗erent exposure periods 􏰊di􏰗erent cu􏰀
mulative times􏰋 as the lead􏰀glass sample cures􏰁 The 􏰅􏰅 and 􏰄􏰍 hour cumulative time
spectra are nearly the same􏰃 and the symbols are not resolved􏰁
curing periods for one of the lead􏰀glass samples􏰁 The spectra after 􏰅􏰅 and 􏰄􏰍 hours
of cumulative exposure to sunlight are nearly the same􏰁 The spectra of 􏰉gure B􏰁􏰍
have been normalized by the transmission spectrum taken before radiation exposure􏰁
These normalized spectra􏰃 as shown in 􏰉gure B􏰁􏰎􏰃 show the loss of transmittance
due to the radiation damage and the subsequent recovery due to exposure to sun􏰀
light􏰁 Complete recovery would be a value of 􏰅􏰁􏰆 corresponding to no di􏰗erence from
Figure B􏰁􏰎􏰕 The loss of transmittance due to the radiation damage and the recovery
due to exposure to sunlight􏰁 The transmission spectra after di􏰗erent exposure periods
􏰊di􏰗erent cumulative times􏰋 as the lead􏰀glass sample cures􏰃 normalized by the non􏰀
irradiated spectrum􏰁 The 􏰅􏰅 and 􏰄􏰍 hour cumulative time spectra are nearly the
same􏰃 and the symbols are not resolved􏰁
Figure B􏰁􏰌􏰕 The transmission spectra of a lead􏰀glass sample before irradiation􏰃 radi􏰀
ation damaged and after sunlight exposure 􏰊cured􏰋􏰁
the non􏰀irradiated transmission spectrum􏰁 After 􏰄􏰍 hours of exposure to sunlight􏰃
the lead􏰀glass sample􏰈s transmittances have not shown any more recovery􏰁 The non􏰀
irradiated􏰃 radiation damaged and cured transmission spectra are shown in 􏰉gure B􏰁􏰌􏰁
The sunlight curing does not cause a complete recovery of the transmittance􏰁
The 􏰉rst hour of exposure to sunlight appears to have the greatest e􏰗ect􏰁
Figure B􏰁􏰐 shows the transmittance normalized by the non􏰀irradiated transmittance
Figure B􏰁􏰐􏰕 The transmittance of several wavelengths while curing as a function of
cumulative sunlight exposure time􏰁
as a function of exposure time to sunlight for 􏰉ve wavelengths􏰁 The curing rate does
not appear to be a function of wavelength􏰁
B􏰁􏰌 Discussion
The spectrophotometer has been used to monitor radiation damage to the cen􏰀
tral calorimeter􏰁 The lead􏰀glass samples placed about the central calorimeter have
not shown any degradation of the transmission spectrum􏰁
Two lead􏰀glass samples were irradiated about a factor of 􏰅􏰆 more than the lead􏰀
glass samples about the calorimeter􏰁 The two lead􏰀glass samples showed radiation
damage􏰁 There appears to be a discrepancy between the radiation dose received when
using radiation damage parameters from a published experiment and the radiation
monitors􏰁 The radiation monitors may be overestimating the dose􏰃 due to uncertainty
about the type of radiation􏰃 by a factor of 􏰄 to 􏰌􏰁
The two most irradiated lead􏰀glass samples have been partially cured by ex􏰀
posure to sunlight􏰁 The sunlight curing does not cause the transmission spectrum
to recover fully􏰘 the recovery leads to transmittances 􏰑􏰎􏰙 of before􏰀irradiation at
ultraviolet wavelengths to nearly 􏰅􏰆􏰆􏰙 for the green part of the spectrum􏰁 A single
central calorimeter counter Monte Carlo simulation 􏰊Appendix C􏰋 shows that the
photoelectron signal should decrease by 􏰅􏰔􏰙 for a 􏰅 GeV incident gamma ray when
using Kirsebom and Sollie􏰈s parameterization of the radiation damage and assuming
a 􏰅􏰆􏰆 rad dose􏰁 Since no large decrease in signal output has been observed􏰃 it is
probable that the radiation monitors associated with the central calorimeter have
overestimated the radiation dose􏰁

No comments: