Wavelength Shifters for Atmospheric
Cherenkov Telescopes
Matthew Teig
Iowa State University Physics Department
Gamma Ray Astronomy Group
July 26, 2000
Wavelength Shifters for Atmospheric
Cherenkov Telescopes
Matthew Teig
Iowa State University Physics Department
Gamma Ray Astronomy Group
July 26, 2000
Standard photomultiplier tubes used in atmospheric Cherenkov telescopes (ACTs) are limited in
their UV response below 350 nm. Most of the Cherenkov radiation caused by cosmic gamma
rays in the atmosphere, however, lies below the visible spectrum in the UV range less than 350
nm. Organic wavelength shifting dyes can absorb light in the UV range and reemit it at longer
wavelengths. They can be applied to a PMT window in order to increase the PMT’s response to
Cherenkov radiation. A computer simulation of Cherenkov radiation passing through the
atmosphere, telescope mirror, light concentrators and PMT windows with and without the
wavelength shifters is discussed.
i
Table of Contents
I. Introduction……………………………………………………………………... 1
II. Detecting VHE Gamma Rays………………………………………………….. 1
A. Atmospheric Transmission…………………………………………….. 1
B. Atmospheric Cherenkov Telescopes (ACTs)……….…………………..2
C. Mirrors and Light Concentrators………………………………………..2
D. Photomultiplier Tubes…………………………………………………..3
III. Wavelength - Theory………………………………………………………….. 3
A. Why Shift the Light?…………………………..………………………..3
B. Selecting a Wavelength Shifter………………………………………....4
C. Simulations………………………………………………………….…..4
IV. Preliminary PMT Testing………………………………………….………….. 7
A. Setup and Procedure………………………………………………….....7
B. Results………………………………………………………………….. 7
V. Conclusions……………………………………………………………………..8
References…………………………………………………………………………..9
ii
I. Introduction
TeV gamma ray astronomy has become a major branch in high-energy astrophysics.
When a high-energy gamma ray collides with an atomic nucleus in the Earth’s upper
atmosphere, it creates an electron-positron
pair, which create secondary gamma rays.
The gamma rays are so energetic that they
excite secondary particles to move at high
speeds, often exceeding the speed light
travels in the upper atmosphere, since the
refractive index of the atmosphere slows
the light down. This phenomenon of
particles exceeding the speed of light
creates a flash of electromagnetic radiation
known as Cherenkov radiation and is
shown schematically in figure 1.
The spectral distribution of the
light emitted can be described by a 1/ 2
distribution and lasts for only a few
nanoseconds. The Cherenkov radiation is
focused in an angle of .8 degrees and
represents mainly the longitudinal
development of the particle shower1.
VHE gamma rays are not the only
particles that initiate the air showers.
High-energy protons and other cosmic
Figure 1: Schematic diagram of a Cherenkov air shower. The VHE
gamma ray excites secondary particles to speeds faster than the speed of
rays also produce air showers. These
light in the surrounding medium1.
showers, however, create more muons
than gamma ray showers and can be differentiated from gamma ray showers by their differing
shape when imaged.
II. Detecting VHE Gamma Rays
A. Atmospheric Transmission
Before the Cherenkov radiation can reach any detector, it must first pass through the
earth’s atmosphere. However, in order to protect its inhabitants from potentially harmful
radiation from the sun and other cosmic sources, the earth’s atmosphere acts as a shield, filtering
out parts of the electromagnetic spectrum. The atmosphere’s transmission curve is shown in
figure 2. (Note: in all transmission/reflectivity plots, the vertical scale is arbitrary. This is
because it is for comparison, and some constants have been omitted to simplify calculations.)
Since the Cherenkov radiation is a 1/ 2 distribution, the blue light dominates the
spectrum. But one will immediately notice that below about 300 nanometers, the atmosphere
effectively blocks out incoming photons, and much of the Cherenkov radiation below that is lost.
This is an important factor when distinguishing between gamma ray showers and air showers
caused by cosmic rays. The air showers caused by cosmic rays can create muons much closer to
the surface of the earth than the Cherenkov radiation caused by gamma rays, meaning that light
1
with relatively short wavelengths
produced when these muons are
created can still reach the earth’s
surface and the detector.
B. Atmospheric Cherenkov
Telescopes (ACTs)
The
air
showers
emit
Cherenkov radiation high above the
earth’s surface at an altitude between
eight to ten kilometers. This radiation
passes through the atmosphere and is
detected by large ACTs.
These
telescopes have large optical mirrors
usually five or more meters
(approximately
fifteen
feet)
in
Figure 2: Atmospheric transmission. Notice that above 350 nanometers the
atmosphere has little effect on the incoming light
diameter. Some have even larger
mirrors, such as the Whipple telescope
in Arizona, whose mirror measure ten meters in diameter. These mirrors reflect the Cherenkov
light onto the focal plane of light concentrators, which focus the light onto the windows of the
photomultiplier tubes (PMTs), capturing an image of the particle cascade.
The most advantageous aspect of ACTs is their ability to differentiate between air
showers caused by VHE gamma rays and the much more common cosmic rays. They can
accomplish this by distinguishing between the different shapes of the images and can reject over
ninety five percent of cosmic rays1.
C. Mirrors and Light Concentrators
The mirrors and light concentrators on the telescopes also affect the light from the
Cherenkov radiation. The reflectivity curve of the mirror is shown below in figure 3. Notice
that the mirror does not affect the visible light much, and only light with a wavelength below
about 200 to 250 nanometers is absorbed significantly.
Figure 3: Reflectivity of the mirror. The mirror does not
have as great an effect on the Cherenkov light, especially
below about 400 nanometers.
Figure 4: Transmission of the Plexiglas.
nanometers is transmitted.
2
No light below 300
In order to focus the light from the Cherenkov radiation onto the PMT window, various
light concentrators are used. Most of these reflect some of the light back out, but Plexiglas cones
have shown a tendency of having total internal reflection: none of the light they collect is
reflected back out of the telescope. The Plexiglas cones do have limitations, though, as can be
seen in their transmission curve in figure 4. The Plexiglas cuts off all the electromagnetic
radiation below 300 nanometers. Figure 8 shows that the Cherenkov spectrum (after passing
through the atmosphere and the telescope mirror) peaks close to 300 nanometers, and as a result
a significant amount of light is emitted from the gamma ray particle shower below this
wavelength. It is this light that is lost when the Plexiglas concentrator is added.
D. Photomultiplier Tubes
Once the light from the Cherenkov radiation hits the window of the PMT, it is incident on
the PMT’s photocathode, where a photoelectron is emitted. However, not every incident photon
will cause a photoelectron to
be emitted. The quantum
efficiency of the PMT
determines
how
many
photons
of
a
given
wavelength will emit a
photoelectron. The quantum
efficiency curve for a
Hamamatsu R1398 one-inch
PMT is shown in figure 5.
For this particular PMT, 18%
of photons with a wavelength
of 300 nanometers incident
on the photocathode will emit
Figure 5: Quantum efficiency of a Hamamatsu R1398 one-inch PMT. Notice the peak
a photoelectron.
between 350 and 400 nanometers and the fairly steep drop off on either side2.
III. Wavelength Shifters - Theory
A. Why Shift the Light?
The main motivation behind shifting the wavelength of incoming Cherenkov radiation is
to increase the amount of light detected from gamma ray showers, local muons, and to increase
the effectiveness of Plexiglas cones. After passing through the atmosphere and the ACT’s
mirror, the Cherenkov light from a gamma ray shower peaks at about 325 nanometers, slightly
shorter than the PMT’s optimum wavelength, between 350 and 400 nanometers. By absorbing
this light and reemitting it at the optimum wavelength of the PMT, a wavelength shifting dye can
increase the amount of light detected. Since the light from local muons passes through only a
small part of the atmosphere, a much larger portion of light below the optimum range of 350 to
400 nanometers can reach the PMT, increasing the effect of the wavelength shifter.
Wavelength shifters can also be used to increase the amount of light detected when a
Plexiglas cone is used. The Plexiglas absorbs all of the light below 300 nanometers, which is
where much of the Cherenkov radiation for both gamma ray and muon showers occurs, as shown
in figures 8 and 12. Coating the Plexiglas cones with the wavelength shifter allows this light to
3
be absorbed and reemitted in a range not affected by the Plexiglas, increasing the amount of light
detected by the PMT.
B. Selecting a Wavelength Shifter
The two main things to consider when selecting a wavelength shifter are what light needs
to be absorbed and where this light needs to be reemitted. For each case described above –
gamma ray showers, muon showers, and Plexiglas cones – the reemission wavelength is the
same, since the PMT efficiency is the determining factor there.
The absorption considerations for the three cases are nearly identical. The Cherenkov
spectrum from gamma ray showers peaks around 325 nanometers, and a significant portion is
below 300 nanometers. Therefore, a dye with an absorption spectrum in this range would be
ideal. The light from the muon showers is dominated by the 1/ 2 distribution and peaks well
below 300 nanometers, getting weaker with increasing wavelength, so again a dye with an
absorption curve that peaks below 300 nanometers is ideal. The Plexiglas cones also have a
similar need in that they absorb all the light below 300 nanometers, again requiring a dye whose
absorption curve peaks below this mark.
Figure 6: Absorption and emission curves for p-quaterphenyl.
Notice the absorption peak between 250 and 300 nanometers and
the emission peak between 350 and 400 nanometers3.
Figure 7: Absorption and emission curves for p-terphenyl. As
in the p-quaterphenyl case, notice the absorption and emission
curves in the theoretically ideal places.
Two dyes were selected that fit these criteria: p-quaterphenyl and p-terphenyl, and their
absorption and emission curves are shown in figures 6 and 7. Both have significant absorption
between 250 and 300 nanometers and both have emission peaks in the 350 to 400-nanometer
range.
C. Simulations
In order to predict the effect that wavelength shifting dyes would have on incoming
Cherenkov photons, a simulation program was employed. The first step in this process was to fit
each of the transmission/reflectivity curves shown above with a polynomial function. This was
4
done using Matlab’s interpolation routine. Once this was accomplished, the various functions,
which were basically probability functions, were multiplied together to get a final probability, as
a function of wavelength, that a photon would excite a photoelectron from the PMT’s
photocathode. The first two of these resultant curves are shown in figures 8 and 9.
Figure 8: Cherenkov spectrum from gamma ray showers.
Notice the peak just below 350 nanometers, roughly the same
place as the peak of the PMT efficiency.
Figure 9: Cherenkov spectrum from gamma ray showers after
passing through a Plexiglas light concentrator. Notice the amount
of light lost below 300 nanometers as compared to figure 8.
Figure 8 shows the gamma ray Cherenkov radiation after it has passed through the
atmosphere and the telescope’s mirror, and figure 9 shows the same spectrum after it has in turn
passed through a Plexiglas light cone. The quantum efficiency of the R1398 PMT peaks at
roughly the same spot on the spectrum as the Cherenkov light before it is passed through a
Plexiglas light cone, so one would not expect a very large increase in efficiency by coating the
PMT with a wavelength shifter in this case. However, examining absorption curves from both
dyes shows that both could save a considerable amount of light once the Plexiglas cone is
Figure 10: Efficiency prediction for shifted and unshifted
Cherenkov light from gamma ray showers. Notice the only part
of the spectrum showing a slight increase is around 300
nanometers.
Figure 11:
Efficiency prediction for shifted and unshifted
Cherenkov light from gamma ray showers after passing through a
Plexiglas light concentrator. There is a much larger gain than there
is without the Plexiglas.
5
introduced.
This is indeed the case. Figure 10 shows both the shifted and unshifted light from the
gamma ray Cherenkov radiation without a Plexiglas cone. The shifted light curves lie practically
on top of the unshifted curve, with only a slight increase around 300 nanometers. This
corresponds to a 2.9% increase in both the p-terphenyl and p-quaterphenyl graphs. Figure 11
shows both the shifted and unshifted light from the gamma ray Cherenkov radiation with a
Plexiglas cone. As expected, a much larger increase is predicted, mostly below 300 nanometers.
This increase is 53.4% for p-quaterphenyl and 53.6% for p-terphenyl.
Figure 12: Cherenkov light from local muons. The 1/ 2
distribution dominates since the only filter is the mirror on the
telescope.
Figure 13: Cherenkov light from local muons after passing through
a Plexiglas light concentrator. Notice the sharp drop off at 300
nanometers.
Figure 12 shows the light curve for Cherenkov radiation from local muons without a
Plexiglas cone and figure 13 shows the same radiation after passing through a Plexiglas cone.
Since these photons are not emitted high above the earth, the atmosphere is not able to block any
light out, so a significant amount of light below 300 nanometers is incident on the PMT without
the use of a Plexiglas cone. This suggests that the wavelength shifters would have more of an
effect on the light from local muons than from gamma ray showers, and this is confirmed by the
plots shown in figures 14 and 15
Figure 14: Efficiency prediction for shifted and unshifted
Cherenkov light from local muons.
Figure 15:
Efficiency prediction for shifted and unshifted
Cherenkov light from local muons after passing through a Plexiglas
light concentrator. Notice the huge increase due to the Plexiglas
cone cutting off all light below 300 nanometers.
6
Since the atmosphere plays only a small part here, the wavelength shifter recovers almost
all the light below 300 nanometers. The p-quaterphenyl dye results in an increase of 4.1% and
the p-terphenyl dye results in an increase of 5.6% when the Plexiglas cone is not in use. When
the cone is in use, the p-quaterphenyl dye results in an increase of 111.7% while the p-terphenyl
dye has an increase of 110.9%.
IV. Preliminary PMT Testing
A. Setup and Procedure
The setup used for the experiment was very simple and straightforward. The entire
system was enclosed in a black box to block out any outside light. A voltage pulse with a period
on the nanosecond time scale was applied across an LED, which emitted a flash of light towards
two Hamamatsu R1398 one inch PMTs whose photocathodes were aimed at the LED. The light
from the LED triggered the emission of photoelectrons from the PMT, and the resulting current
was read on an oscilloscope. The integral of the signal was integrated over time for a varying
number of events, ranging from 3000 to 5000. The analyzing software could only integrate one
tube at a time, so the PMTs were analyzed individually and switched between the two positions
for each run. The two positions were included to account for any error arising from geometrical
effects.
The only procedure done in this part of the experiment was to test the PMTs for stability.
This was done by running the experiment described above six times for each tube. Three of the
times the tube was in position A and the remaining three times the tube was in position B. Plots
of the gain for each run of each tube, including error bars, are shown in figures 16 and 17.
Figure 16: Stability measurements for tube A. The tube is
very stable over time, with a chi squared of five for six
measurements.
Figure 17: Stability measurements for tube B. Like tube A, tube B
is very stable over time, again with a chi squared of five for six
measurements.
B. Results
A chi-squared analysis was done for each tube, and the result in each case was five.
Since only six runs were performed, this indicates that both tubes are indeed stable.
Furthermore, the relative position in the box and the amplifier used made no difference, as each
run was within two error bars or the other five.
7
V. Conclusion
The vast majority of the work done in this experiment was theoretical, but based on that
data, several conclusions can be drawn. The first is that the tubes are both stable enough to
perform future tests on without having any effects due to the aging of the tubes. The conclusions
on the predicted usefulness of the wavelength shifters vary with the situation. When the
Plexiglas cones are not used in combination with the Hamamatsu R1398 one inch PMT, the
increase in Cherenkov light from air showers is small. The Cherenkov radiation from both the
gamma ray showers and the muon showers peaked close to the peak of the quantum efficiency of
the tube, and as a result there was very little increase in the amount of light detected, the largest
increase being 5.6% for the p-terphenyl dye for local muons. However, when the Plexiglas
cones are introduced, the usefulness of the tubes increases greatly, to just over 50% for both dyes
in the case of gamma ray showers and to over 110% for both dyes in the case of Cherenkov
radiation from local muons.
8
References
1. Yodh, Gaurang; Very High Energy Gamma Ray Detectors; The Beam Line; Vol. 28, No. 3;
Stanford Linear Accelerator Center; Stanford, Ca.; Fall/Winter 1998.
2. Hamamatsu; Photomultiplier Tubes and Assemblies; 1998.
3. Berlman, Isadore B.; Handbook of Fluorescence Spectra of Aromatic Molecules; Academic
Press; New York; 1965.