Note: HTML Version. The original was published in the Proceedings of the 23rd International Cosmic Ray Conference, 2, 524 (1993). Please make your references to those conference proceedings, as I doubt that a URL reference will be acceptable in any printed publication. In any case, this URL is subject to change.

**Note further that, as reported at the conference, we found that the actual index of refraction for the aerogel blocks averaged 1.043+/-0.002, at the centers of the blocks. This was discovered after submission of the paper.

Silica Aerogel Cherenkov Counters for the Isotope Matter Antimatter Experiment

A.W. Labrador [1], R.A. Mewaldt [1], S.M. Schindler [1], E.C. Stone [1], L.M. Barbier [2], E.R. Christian [2], J.W. Mitchell [2], R.E. Streitmatter [2], S.J. Stochaj [3] I.L. Rasmussen [4]

[1]California Institute of Technology, Pasadena, CA 91125, USA
[2]NASA/Goddard Space Flight Center, Greenbelt, MD 20771, USA
[3]Particle Astrophysics Laboratory, New Mexico State University, Las Cruces, NM 88003, USA
[4]Danish Space Research Institute, Lyngby, Denmark


We describe the design, development, and performance of the large area aerogel Cherenkov counters employed in the Isotope Matter-Antimatter Experiment (IMAX), a balloon-borne, magnet spectrometer. IMAX incorporated two Cherenkov counters containing silica aerogels of refractive index n=1.055** for velocity measurement and electron/muon background rejection. We describe aerogel baking techniques which were found to improve measured light output from 30% to more than 100% for individual aerogel blocks, as well as mounting techniques that preserved the integrity of the aerogel radiators during shipping, balloon flight, and parachute landing. The instrument flew successfully from Lynn Lake, Manitoba, Canada, on 16 July 1992, remaining at float for ~16 hours. Laboratory measurements and analysis of pre-flight data on the counters' performance are presented.


The Isotope Matter-Antimatter Experiment (IMAX) is a balloon-borne superconducting magnet spectrometer designed to measure the galactic cosmic ray fluxes of protons, antiprotons, and the isotopes of hydrogen and helium. It uses precise measurements of rigidity, charge, and velocity in combination to yield unambiguous particle identification (Mitchell, et al, 1993). The experiment is a collaboration between the California Institute of Technology, the Goddard Space Flight Center, New Mexico State University (NMSU), the University of Siegen, Germany, and the Danish Space Research Institute. IMAX included two aerogel Cherenkov counters which contribute to electron and muon background rejection at magnetic rigidities below 2.7 GV and isotope velocity measurement at energies above 2 GeV/nuc. Each counter contained three radiators, each 3 cm thick.

Aerogels are of particular interest for Cherenkov detectors because they are solid radiators available in a range of refractive indices from ~1.01 to 1.25 (Poelz, 1986, and Rasmussen, 1989). The aerogels used in IMAX were manufactured by the Airglass Company in Sweden. As received from Airglass, they have a nominal index of 1.055** and approximate dimensions of 55 cm x 55 cm x 3 cm, considerably larger in area than the 14 cm x 14 cm blocks used in previous balloon-borne instruments (Rasmussen, et al, 1983). The aerogels are hard but brittle, having the appearance of semi-transparent "solid smoke." The blocks are easily chipped or cracked at the edges, requiring the use of special mounting and cutting procedures as described below.


2.1 Counter Design

The two IMAX aerogel Cherenkov counters, designated C2 and C3, consist of light collection boxes each housing three silica aerogel radiators, viewed by 14 and 16 photomultiplier tubes (PMT's), respectively. The counter design allowed the stacking of any number of aerogels, up to the limits imposed by space restrictions in the payload. (See Figure 1.) The boxes are constructed of dipŠbrazed 6061 aluminum, lined with millipore filter paper (specifications: 0.1 micron pore size, type VCWP) rather than barium sulfate paint, because the paint is known to outgas solvent which we have found degrades the aerogel performance with time. Measurements show that the millipore used has ~93-94% reflectivity at the Cherenkov wavelengths of interest. The PMT's are 3 inch diameter Hamamatsu R1848's, which are able to provide the single-photoelectron resolution required for low light yield measurements. They were also chosen because of their low profile, good quantum efficiency, and their dynode construction (box and grid plus mesh), which is relatively insensitive to low magnetic fields. Magnetic shielding for the PMT's is accomplished with multilayer shields composed of steel tubing and high-µ material of different thicknesses customized for each PMT location. The PMT's are recessed at varying depths within their mounting tubes to allow for more efficient magnetic shielding; recessing the PMT's lowers measured light yield by ~10%. The mounting tubes are lined with aluminized mylar, which provides ~90% reflectivity at the wavelengths of interest.

Figure 1: C3 with three stacked aerogels. The aerogels are mounted in individual frames and bolted directly to the top of the counter. Space restrictions in the payload required that some PMT's be mounted at various angles.

2.2 Aerogel Mounting

The raw aerogels were cut to size with a diamond wire saw. After baking (see Section 3 below), the aerogels were individually mounted in frames made of four aluminum channels, 4.13 cm wide, welded into a square (53.34 cm x 53.34 cm interior dimensions) with top and bottom faces ground flat and parallel.

Figure 2: A) Aerogel Mounting Frame, B) Aerogel Mounting Cross-Section

Mounting screws were inserted along the edges of the frame to hold the aerogel centered for potting. Once an aerogel is in its frame, a potting fixture was attached to each face of the frame to provide a 1 cm deep well between frame and aerogel. Sylgard 184, a clear silicone elastomer manufactured by Dow Corning, was then injected into the potting well to provide approximately 0.6 cm overlap of the elastomer beyond the edge of the aerogel (see Figure 2).

Once the Sylgard 184 was cured, the holding screws were withdrawn to prevent damage to the aerogels from thermal expansion/contraction of the aluminum frame. The aerogel is then held securely in the frame by the potting material, which provides both mounting capability and shock/vibration isolation. Tests indicate that Sylgard 184, in contact with the aerogel, does not degrade light yield with time. This mounting technique was used successfully in the IMAX instrument; the mounted radiators were recovered fully intact after parachute termination and landing. Prior to flight, the aerogels were stored under a contaminant-free dry nitrogen gas flow, with the relative humidity kept below 15% to reduce water absorption.


During the development phase of the counters, we conducted tests to determine the light yield characteristics of the aerogels and to maximize the light yield of the counters. Laboratory measurements of light yield with muons were made with a particle telescope and Camac system employing the same PMT's, amplifiers, and analog-to-digital converters used in flight. Signals from the PMT's were individually digitized. With this system, it was possible to resolve the 0, 1, and, in some cases, 2 photoelectron (pe) signals for individual PMT's. For a single PMT, the average light yield, in pe's, was obtained from histograms of the PMT response and calculated by the "fraction of zeros" equation
where C is the average light yield, p0 is the number of events with 0 photoelectrons, and p is the total number of events in the histogram. Equation (1) assumes that the light yield follows a Poisson distribution. Average light yield for a given counter is obtained by summing Equation (1) over all the PMT's in the counter.

3.1 Aerogel Baking

Methanol is a normal by-product of the aerogel manufacturing process which should have been removed during the drying phase of the process. Water could also be absorbed from the air. However, an unexpectedly low light yield was initially obtained for the radiators, and this, coupled with a yellowish discoloration, led us to believe that residual interstitial methanol and possibly water remained in the blocks. We determined that a secondary baking process employing a large oven could remove both methanol and water and ultimately improve the light yield (Henning, 1979). The procedure involved a slow ramp-up from room temperature to 500 C over 14 hours, a dwell period at 500 C for 3 hours, and a slow ramp-down from 500 C to room temperature over 14 hours. The methanol was oxidized to formaldehyde and formic acid, both of which were rapidly removed by forced air circulation throughout the procedure. Any residual water in the aerogels was removed by evaporation.

Prior to baking, some aerogel blocks had a yellow coloration in reflected light. However, after baking, these blocks had a more uniform bluish tint. The improvement in light yield varied for each manufacturing run at Airglass, ranging from a 30% improvement to more than 100% improvement for individual blocks.

3.2 Laboratory tests

The post-bake improvement in light yield was most striking in the laboratory tests in which we measured light yield as more aerogel layers were added. If we include attenuation of Cherenkov light by scattering or absorption in the radiator, the total light emitted by a charged particle passing through a radiator can be represented by
where L(t) is the total light emitted by a Z=1, beta=1 particle passing through a radiator of index n and thickness t, and lambda is interpreted as an effective transmission length characteristic of the radiator and the counter geometry. K is a constant.

Figure 3: C3 radiator light yield, in photoelectrons, vs. radiator thickness, for unbaked and baked aerogels. All blocks have an index of refraction of n=1.055**. (Uncertainties are smaller than the plot symbols.)

Curve fits on Figure 3 show that the baked aerogels have lambda = 4.7 cm in C3, while the aerogels prior to baking had lambda = 1.7 cm. While individual results varied from block to block, these results were typical of the aerogels used for flight. From lab tests, we predicted a fraction of zeros average response for Z=1, beta=1 particles of 13 pe's in the counter designated C3, and 10 pe's in the counter designated C2.


The instrument was launched from Lynn Lake, Canada, on 16 July 1992. The payload achieved an average float altitude of 36 km for a period of about 16 hours. Analysis of in-flight performance is currently under way. Current maps of counter response have been made using pre-flight muon data. Light yields are obtained on an event-by-event basis with photoelectron scales calculated from the 0 and 1 pe response per PMT. The maps indicate that, for Z=1, beta=1 particles, C2 yielded approximately 13 pe's at the center, and C3 yielded approximately 15 pe's. Each map is an array of 5 cm x 5 cm bins. In the central 40 cm x 40 cm regions of the counters, response is reasonably flat, varying by less than ~+/-10% of the total yield in each counter. Response maps will be refined with flight data.


We thank Bob Golden and the NMSU-BBMF crew for providing overall payload support. Glen Albritton and Caltech's Central Engineering Service were instrumental in the construction of the counters. This research was supported in part by NASA grant NAGW-1919. One of us (AWL) thanks the NASA Graduate Student Researchers Program for its support.


Henning S., Thesis, University of Lund, Lund, Sweden, 1979.
Mitchell J.W., et. al., "IMAX: The Isotope Matter-Antimatter Experiment" in these proceedings.
Poelz G., "Aerogel in High Energy Physics", Aerogels. Ed. J. Fricke. (Springer-Verlag, Berlin, 1986) p. 176
Rasmussen I.L., et. al., Proc. 18th International Cosmic Ray Conference, Bangalore, 8, 77 (1983)
Rasmussen I.L. "Use of SiO2 Aerogels with n=1.05--1.25 as Cerenkov Detectors", Revue de Physique Appliquee, 1989-C4, p. 221-227.