An effective segmentation of the aerogel Cerenkov counter, matching the segmentation of the trigger scintillators, can be used to separate multiple tracks through the focal plane and will allow an additional element of selectivity and track sensitivity in the focal plane instrumentation. This means that specific sections of the focal plane can be physically disabled from the trigger, if the experimental conditions require it. It will also provide the capability of identifying and separating pions and protons traversing the focal plane trigger scintillators and the vertical drift chambers (VDCs) within the resolving time of the system (double hits). For example, in the offline analysis, the aerogel counter PMT with the highest number of photoelectrons can be matched with the trigger counter and VDC information to identify the actual path of a pion, thus separating it from a simultaneously detected proton, which has no Cerenkovsignature. Such a capability of double hit resolution is not possible with diffusion Cerenkovcounter designs, because the photon collection efficiency does not have a strong correlation with the incident particle track within the aerogel material.
The requirement for segmentation, in addition to supplementing the information
on the individual particle position along the focal plane, also couples well
with the desirability of increasing the active solid angle viewed by the PMTs
in the counter. Although the photon detection probability is not as directly
proportional to the solid angle covered by PMTs as in the case of a diffusion
box, clearly, the larger the effective coverage, the higher the probability
will be that a photon will end up on a PMT. Given the divergence of the beam
envelope incident on the aerogel, and the diffusion of the light in the low
region by the aerogel material, an increase in the area covered by
PMTs results in an increase in the number of photons detected. As a result, a
total of 26 PMTs are used in the counter, as shown in figure, with minimal
spacing between their -metal shields (2.8 mm). The total area covered
by the PMT photocathode windows comprises 72% of the area of the counter
opposite the planar parabolic mirrors. A cross sectional schematic of the
detector is shown in figure
, clearly illustrating the planar parabolic design
of the mirror surfaces and their relative orientation with respect to the PMTs,
and the orientation of the counter relative to the central axis of the
spectrometers.
The close spacing of the -metal shields, which is also shown in the
photograph of figure , creates dielectric breakdown problems. The -metal
shields are at cathode potential (-2950 V) to avoid the capacitive discharge
from a grounded -metal shield to the glass of the photocathode, which
would contribute to the noise level in the PMT, and adversely affect their
performance at high operating voltages. This necessitates extra precautions, in
order to avoid dielectric breakdown between adjacent shields, and between the
shields and the aluminum structure of the counter, which is at ground
potential. The solution was to wrap the outer surfaces of the -metal
shields with a high dielectric value (12,000 V/mm), thin (0.254 mm) Teflon
film [12]. In addition, the PMT housings consist of fiberglass-epoxy
composites, with added inner and outer skins of 0.0254 mm thick Tedlar
[12], with a further combined insulating value of 3,000 V. Such a
combination of insulating materials eliminates any breakdown or small leakage
current induced noise and, at the same time, satisfies all safety requirements.
The final construction of the counter, described in this report, is built
around the two sides of the main (PMT) section, each consisting of two pieces
of aircraft quality aluminum alloy, with stiffening aluminum rods formed
integrally on the top and bottom. The openings for the PMT housings were
machined on these structures using CNC milling machines to keep tolerances to
fine levels. The double walled structure, on both sides of the enclosure,
further increases the rigidity of the exoskeleton by forming a second ``outer''
wall on each side, very similar in configuration to the inner one, and attached
to the latter with crossbolt braces, as shown in the photograph of
figure.
Each end plate is made out of the same aluminum alloy as the side walls, and
also incorporates stiffening lips folded integrally to each plate, one at the
top and one at the bottom. Each end plate has been provided with inlet and
outlet gasline connections, which will be used to fill the counter enclosure
with dry CO2 gas to protect the silica aerogel from water vapour absorption.
figures
and
show the bottom (tray) sections, and main plus upper (mirror)
sections, respectively. The main (middle or PMT) section, in
figure
,
contains the PMTs and provides the strength and rigidity for the whole counter.
The one piece aluminum end plates are also shown in both photographs.
All internal surfaces of the detector, except the planar parabolic mirrors,
themselves, are lined with aluminized mylar [13] to increase the
overall reflectivity of the counter. The mirrors are made in moulded surfaces, formed in one rigid structure. The rigidity is provided
by two layers of carbon fiber epoxy composite backing, with a combined
thickness of 0.28 mm, and a single sheet of mylar with thickness 0.127 mm.
The special mylar material was obtained from exposed negative film used in the
cartographic industry, and is of high smoothness and uniformity. One side was
aluminized at CERN, while the other side remains in its exposed negative
(black) state, further adding to the successive light penetration barriers into
the enclosure. A representative reflectivity curve, as a function of
for these mirrors, is shown in figure.
The upper section of the counter containing the mirrors is mounted on its
own aluminum subframe, which is bolted to the main frame that houses the PMTs.
The upper section, on its own, is shown in the photograph of
figure, while
its configuration when mounted on the main section is shown in
figure
. The
light and gas sealing action is provided by continuous twin parallel rubber
strips along the joint area, and by Tedlar film of 0.025 mm thickness
covering the top of the outer planar parabolic area.
The third major component of the counter consists of a removable tray where the
silica aerogel is placed. The tray occupies the bottom part of the counter and
has inside dimensions of where the SiO2 silica aerogel
is placed. It is formed by a frame with twin aluminum panels, which, in turn,
secure the removable frame strung with fishing line in a criss-cross pattern
to hold the aerogel panels in place. This ``fishnet'' frame is secured by
screws and is easily removed without disturbing the aerogel panels or requiring
restringing. The bottom of the tray is formed out of a single layer of carbon
fiber epoxy skin (0.127 mm thick) and a layer of aluminized mylar of equal
thickness. Externally, it is covered by a single layer of Tedlar film to
assure integrity from light penetration; further environmental isolation is
provided by two parallel strips of rubber gasket seals enclosing the
circumference of the tray and containing the feed-through spacers for the
retaining bolts. The tray is equipped with SMA-type fiber optic feed through
connectors for the gain and timing monitor system, which utilizes fiber
optic cables. Each fiber illuminates two adjacent PMTs, except the last PMT
on either side (13T and 13B in figure), which have their own dedicated fiber.
The light is generated in a gas plasma discharge unit [14] and
duplicates the spectrum expected from Cerenkov radiation. In addition, the
fibers terminate beneath the silica aerogel, thus, the light reaching the PMTs
will have the absorption characteristics of real Cerenkov light produced
in the aerogel radiator.
Due to the nature of Cerenkovdetectors, where few photoelectrons (PEs) are emitted by the photocathodes in the PMTs, any extraneous light entering the enclosure is very troublesome. As a result of the small number of PEs expected, the PMTs operate either near to, or at, maximum high voltage, and, thus, at maximum gain. As such, they can suffer damage if a sudden light leak develops. In testing, we verified the extreme sensitivity to minute light leaks, even across the whole length of the structure, because of the mirrored surfaces inside the enclosure. With 26 PMTs operating at maximum gain, and viewing, effectively, a giant mirror, sealing the enclosure against single photon penetration requires extra care during initial testing and operations.
The PMTs chosen for the counter were Burle model number 8854, 127 mm
photocathode diameter [15]. The PMT amplification electronics have
been described in Refs. [16,17]. The dynode chain incorporated a
600 resistance between the cathode and first dynode, instead of the
nominal 300 . This generates a Vdyn=885 V across the cathode to
dynode gap, thus, increasing the photoelectron collection efficiency and peak
to valley (P/V) ratio. This modification has been proven successful in
increasing the PE collection efficiency and the single PE resolution. The
dynode amplification chain also incorporates a 11 resistor in series
with the -metal shield to eliminate the possibility of electric shock
through careless handling; this high impedance also limits the current drawn,
in the unlikely event of a complete dielectric breakdown between the shields
and the aluminum parts of the detector. A schematic diagram of the electronic
amplification chain is shown in figure.