A Plastic Jug of Heavy Water

a plastic jug of heavy water

When a neutrino observatory needed a 40-foot acrylic vessel to hold tons of D₂O, the builders had to do their homework.

This article was prepared by staff writers in collaboration with outside contributors.

While most people are inclined to roll their eyes upward at the mere mention of studying solar neutrinos, the Sudbury Neutrino Observatory has its detector more than a mile below the Earth's surface. The depth protects it from the noise of background radiation. Designing and building the neutrino detector was no easy task because it relies on an acrylic sphere four stories high to hold 1,000 metric tons of heavy water. The heavy water, worth about $300 million (Canadian), is a key to detecting neutrinos. It is on loan from Atomic Energy of Canada Ltd.

It was impossible to perform complete physical tests on the vessel, given its size and location. Consequently, the builder responsible for the acrylic vessel, Reynolds Polymer Technology Inc., commissioned William Jones, director of the MSC.Software Expert Solutions Group in Santa Ana, Calif., to provide a simulation study for optimizing the design.
Neutrinos are one of the fundamental building blocks of nature and a cornerstone of the standard theory of elementary particles. Three types are known to exist: the electron neutrino, the muon neutrino, and the tau neutrino. Electron neutrinos are emitted in vast numbers by the nuclear reactions that energize the sun. They are notoriously hard to detect and require the use of massive, sophisticated devices such as the Sudbury detector.

The neutrino detector was built to help resolve a 30-year-old mystery regarding the nature of neutrinos: Why do physicists see only one-third the number of neutrinos from the sun that is predicted by current theories? Initial data provided by the Sudbury observatory reveal new and important properties of neutrinos, and carry implications for a fuller understanding of the universe.

At the heart of the Sudbury Neutrino Observatory, an acrylic vessel holds more than 2 million pounds of heavy water.

The neutrino detector identifies the interaction between a solar neutrino and a molecule of heavy water inside the acrylic vessel. According to Eugene Beier, a professor of physics at the University of Pennsylvania, the neutrino reacts specifically with the neutron in the nucleus of the deuterium in the heavy water. Common hydrogen, as found in light water, contains no neutron.

Each interaction releases a burst of light and waiting to record it is an array of almost 10,000 photomultiplier tubes. The tubes, mounted on a geodesic sphere surrounding the vessel, are light detectors that can sense a single photon.

Photons interact with a thin film inside the tube to eject an electron. This electron in turn strikes high-voltage plates to cause a cascade of electrons producing a pulse, which is directed to data recorders.

Although trillions of neutrinos pass through the vessel, only about 20 interactions with the heavy water in the sphere are expected to occur each day.

The neutrino detector includes a 40-foot-diameter acrylic sphere 2.2 inches thick with a 25-foot-tall, 6-foot-diameter chimney for filling and draining. The sphere is made of 120 rectangular panels that are curved by thermo-forming and machined to a tolerance within 0.005 inch. As far as the people involved with the detector have been able to learn, the sphere is the largest structure ever made entirely from acrylic.

The acrylic vessel containing more than 2 million pounds of D₂O is suspended in ultrapure light water, a refinement of the familiar H₂O, in a rock chamber 6,800 feet below ground level in the Creighton nickel mine operated by Inco Ltd. near Sudbury, Ontario. The light water helps balance the pressure from within the sphere.

Because extremely low radiation background levels are required to detect the neutrino reactions, the entire detector is constructed from materials selected for their low radioactivity content. The acrylic material had to be very pure, since trace impurities could be a source of radioactive particles, which would create false readings.

"The basic design issue was to keep the walls of the acrylic vessel thin enough for the photomultiplier tubes to sense the light resulting from neutrino interactions, while ensuring that the vessel did not fail, causing the heavy water to commingle with the light water," Jones said. "The thinner we made the walls, the more transparent it was and the less contamination the vessel contained, which minimized radioactive material. All these things were driving us to thinner walls. We were working under the most extreme of conditions to make the vessel work, while satisfying safety requirements and securing the heavy water."

Stress analysis and other testing by MSC.Software helped shape the final design of the plastic vessel. The sphere is anchored by tethers at its equator and rests suspended in light water.

The main objective of the analysis was to determine if the structure would maintain integrity when filled with its load of heavy water, which Jones for his analysis calculated at 2.3 million pounds. The sphere is supported by 10 pairs of polymer ropes attached to its middle ring of panels. The analysis of the sphere was conducted in three stepsÑa scoping study that broadly defined the optimum design, a high discontinuity stress study, and a seismic loading analysis.

Although the tensile strength of the acrylic is quite high, in the vicinity of 10,000 psi, the allowable stress for this material is very low. An allowable stress in tension of 580 psi was used for the parent material to preclude long-term crazing. Crazing, a phenomenon that occurs in acrylic materials, is characterized by the formation of extremely small surface cracks, which can eventually reduce the transparency of the material or cause it to fail.

To reduce the tensile stress in the sphere, the vessel was underfilled with heavy water, a condition that actually produced a state of compression because the outside pressure of the light water was greater than that of the D₂O inside.

However, the acceptable amount of compressive stress is also quite low because the sphere's very large ratio of diameter to thickness (40 feet to 2.2 inches, or more than 200:1) raises the possibility of buckling. The final design balances tensile stress and compressive stress to achieve the largest margin of safety possible, while minimizing the wall thickness.

For the scoping study, a relatively simple finite element model was developed with 2-D shell elements, enabling the evaluation of a number of variations to the basic design. A true optimization study was run using the 2-D axisymmetric shell element, in which a set of design values minimizing the volume of acrylic in the sphere was automatically selected. "The spherical shape was selected because it provided the least amount of surface area required to bound a given volume," Jones, an ASME fellow, said. "This minimized the amount of acrylic material necessary to contain the D₂O, which reduced foreign materials and therefore background radiation."

The acrylic vessel at Sudbury, which was designed and built by Reynolds Polymer, was dry-fit on the surface first, then assembled using a new bonding process at a depth of 6,800 feet underground.

Stresses in the acrylic material were minimized by floating the vessel in ultrapure H₂O. The spherical shape was also ideal for reducing the material stresses from internal pressure. How-ever, because D₂O is about 11 percent heavier than the H₂O providing the buoyant force, the membrane stresses in the vessel were not completely eliminated at all elevations in the sphere.

After determining the optimum design with the shell model, the next step needed a more detailed representation of the sphere constructed with a 2-D axisymmetric solid element. This technique provided a representation of the intersection of the sphere and chimney, including local fillet radii and tapers from the thicker T section down the nominal 2-inch-thick spherical shell, which makes up most of the sphere.

This model was used to investigate areas of high discontinuity stress, as well as to evaluate small design changes for reducing the calculated stress to the allowable material. Once the final details of the sphere/chimney intersection were determined, a series of runs calculated the effects of changing water level to investigate the sensitivity of the design to filling the sphere and cavity initially. Jones also calculated the stress distribution in the sphere when completely dry and hanging under its own weight and the reaction loads.

The maximum stress levels in the sphere were determined to be relatively insensitive to variations of water levels. The filling procedure producing the lowest stresses on the vessel, as it went from empty to full, was determined by assuming a series of D₂O levels, while finding the level of H₂O that produced the least stress in the vessel.

"Balancing the amount of heavy water in the vessel with light water in the cavity was a critical aspect of ensuring the integrity of the vessel," Jones said. "Our analysis indicated that if both heavy and light water levels were matched, stress levels during filling would be well below the allowable short-term stresses of the acrylic material and that filling would not present a problem."

An eye to catch the invisible: The completed acrylic vessel with its phototdetector tube array in place is designed to interact with solar neutrinos passing through Earth and to record each event.

The third step was a seismic loading analysis, which was based on a seismic acceleration caused by blasting operations and rock burst phenomena, which are endemic to deep mines such as this. The analysis used a time history trace of a typical rock burst or blast measured in the mine. A detailed finite element model simulated the spherical shell and the light and heavy water.

Jones conducted a transient dynamic analysis of the induced pressure wave to calculate the stresses in the acrylic sphere. He compared these results with the short-term stress limits of the acrylic material.

In addition to the axisymmetric analyses, detailed three-dimensional models of the vessel were developed that included the thickened equator section and the fittings for the tethers. The analyses were reviewed by numerous agencies for completeness and accuracy. Reviewers included the Los Alamos National Laboratory in Los Alamos, N.M.; Carleton University in Ottawa, and the Sudbury Neutrino Observatory design team led by Robert Brewer of Agra-Monenco Ltd., a unit of Agra Inc. of Oakville, Ontario. Members of the Physics Department at Oxford University independently carried out pressure time history calculations.

Before the four-story-high container was built 2 km below ground, test calculations included resistance to seismic disturbance, crazing, and buckling.

Finally, a series of buckling analyses were done showing that the vessel would hold up under applied loads. Because the pressure inside the sphere increases more rapidly with depth than the pressure in the water, a net differential pressure across the sphere wall is always present. With a very high radius-to-thickness ratio, buckling of the sphere had to be considered as a possible failure mode. Jones said, "The pressure distribution on the vessel is quite complex because of the difference in density between the D₂O and H₂O, the geometry of the vessel, and the low differential water levels used to minimize stresses on the vessel."

Once the component panels of the observatory's acrylic vessel were built, each had to be cleaned and individually packaged for shipment to the mine. The entire laboratory site located at the 6,800-foot level of the mine is operated as a level 2000 cleanroom, making assembly and bonding more difficult.

Proper assembly of the vessel required tight tolerances of less than an inch over the 40-foot structure, no mine dust in the bonding syrup, and a new bonding process. According to David Duff, vice president of Reynolds Polymer in Grand Junction, Colo., "The assembly was dry fit and documented before shipping it to the bottom of the mine. We used a theodolite system that triangulates with infrared light to any point relative to any other point with submillimeter accuracy to determine exactly where panels had to be positioned in relation to each other."

Bonding Accuracy Is Vital

Bonding structural acrylic panels together is not like gluing a model together; it's more of a casting process so the light doesn't reflect off the bonds. The curing rate is important because air bubbles can get trapped in the bonding syrup if it hardens too quickly. With a tolerance on the bond gap of an eighth to a sixteenth of an inch, particularly when applied to a 40-foot-diameter sphere, the accuracy of the bonding process is critical.

"With conventional bonding, heat is used to cure the bond," said Duff, an ASME member. Heat distorts the bonds and can cause them to shift. "We developed a new room temperature curing process that cured the entire bond gap at the same time, producing linear shrinkage," he added. "This allowed us to control panel movement."

To visualize the assembly of the detector vessel, think of an igloo. A platform that moved inside the cavity was used to assemble the vessel. First, the chimney was positioned on the platform and bonded together. The finished chimney was then hoisted out of the way and suspended from the cavity's ceiling. Next, the 10 panels of the equatorial row were positioned on the platform and bonded together. Then each row above it was positioned and bonded, until the top row was bonded to the chimney.

Special Vectran ropes were then inserted into grooves in the 10 panels of the middle row specially machined to hold them. The structure was put into position and the whole upper hemisphere was suspended. The construction platform was then progressively lowered as the lower rows were positioned and bonded to the upper hemisphere.

"Even with all the special handling, in one area there was some contamination in the bonding syrup," Duff said. "Because the light emitted off this point, it actually was a positive, since it provides a zero point."

As a result of the successful assembly and subsequent two years of reliable operation of the Sudbury neutrino detector, scientists were able to report last June that electron neutrinos from the sun transform into neutrinos of other types during their flight to Earth. According to Art McDonald of Queen's University in Kingston, Ontario, who is director of the Sudbury project, the data, combined with results of earlier experiments conducted in Japan, have important implications for ideas about energy and matter.

One implication is that current theories concerning energy generation in the core of the sun are substantially correct, he said. The information on neutrino properties will be very valuable in extending the present Standard Model of Elementary Particles, which currently assumes that neutrinos do not transform their type. The information obtained on neutrino mass contributes to determining the influence of neutrinos produced in the original Big Bang on the evolution of the universe.

sidebar: new detector to monitor neutrinos