Long term sealing ability of butyl o-rings.

Long term sealing ability of butyl o-rings

The Galileo spacecraft will perform a series of scientific measurements on Jupiter to increase our knowledge of the planet. The spacecraft, consisting of an orbiter and a probe, will conduct the first "in situ" measurements of Jupiter's atmosphere. After entering the atmosphere at a speed of approximately 48 km/sec, the probe will decelerate, deploy a parachute and descend. Various instruments will measure the atmospheric temperature and pressure profile, size and structure of clouds, planetary thermal radiation, chemical composition of atmosphere, and lightning and radio emissions. The data is transmitted from the probe to the orbiter, then relayed to Earth.

A delay in the space shuttle program caused the Galileo launch date to slip from May 1986 to November 1989. Furthermore, a change in the spacecraft booster rocket increased the Earth/Jupiter transit time from approximately 2-1/2 to 6 years. Due to the resulting delay of the Galileo spacecraft arrival at Jupiter, a study was conducted to identify shelf life limitations which might impact mission performance.

One of the shelf life concerns identified during this study was the long term sealing ability of butyl o-rings. The probe's RF exciter units use butyl o-rings to seal the internal electronics from the Jupiter atmosphere up to a design pressure of 2 MPa. The failure of this seal would have serious implications.

It is well known (ref. 1) that butyl rubber undergoes a chemical stress relaxation involving molecular bond scission and crosslinking in an atmosphere containing oxygen. Bond scission is dominant over crosslinking in butyl rubber. Furthermore, the rate of bond scission is relatively constant for oxygen pressures greater than 100 Pa, and decreases substantially for oxygen pressures below this. The internal pressure of the unit will decrease during the transit from Earth to Jupiter. However, unless the measured leak rate of 6 x [10.sup.-6] std atm cc/sec increases dramatically, the oxygen pressure in the unit will never drop below 100 Pa. Therefore, the o-rings will continue to age (and bond scission will occur) throughout the cruise portion of the mission.

The unit leak rate is required to be less than 2 x [10.sup.-4] std atm cc/sec during descent through the Jupiter atmosphere. Therefore, to accommodate the backup launch date, the unit leak rate must increase no more than a factor of 33 over a 17 year lifetime. To establish the o-ring sealing ability over such an extended time period, a series of accelerated aging tests was conducted. Originally, a set of measurements was to be performed on o-ring segments to determine the appropriate temperature for an accelerated aging of the engineering model exciter simulating 17 years. Leak tests were to be performed during and following the exciter test. Later, however, the exciter test was dropped and testing of the o-ring segments was expanded. A discussion of the tests performed follows.

Accelerated aging tests were begun in April 1986 and continued for more than a two month period, using butyl o-rings from the same lot used in construction of the Galileo flight hardware. These o-rings, cured in the last quarter of 1979, were made from compound B318-70 and have a diameter of 2.62 mm. The o-rings were cut into segments approximately 3 cm in length. Grooved fixtures were fabricated to compress the o-ring height from 2.62 to 1.75 mm, as in the exciter o-ring groove. The width of the fixture groove was approximately 1 cm, allowing the o-ring segment to expand freely in this direction. This permits the o-ring to take a compression set at least as large as it would in the exciter o-ring groove.

The fixtures, with o-rings installed, were maintained at temperatures ranging from room temperature up to 87 [degrees] C. The o-rings were periodically removed from the fixtures to perform compression set measurements. Under compression, the resistive force of the o-ring decreases with time due to both physical and chemical stress relaxation. As a result, the segments assume a new flattened shape. The change in shape resulting from physical stress relaxation is recoverable (memory effect); the change resulting from chemical stress relaxation produces permanent change (compression set).

Since the original goal of this test sequence was to determine the acceleration of chemical degradation as a function of temperature, an attempt was made to remove the physical compression set by returning the o-ring alone to its oven for a period of 15 minutes after removal from the fixture. Then a compression set measurement was performed on the "recovered" o-ring segment. It was later shown that this technique removed most, but not all, of the physical compression set. Subsequent measurements included both the compression set immediately upon removal from the fixture and the technique outlined above.

The data recorded from o-rings compressed at the various temperatures are shown in figures 1 and 2. It was expected that the o-ring compression set would exponentially approach a value of 100%. Furthermore, it was hoped that the compression set would exhibit an Arrhenius type rate/temperature dependency:

[k.sub.1] = Ae -[E.sub.a]/kT, where [k.sub.1] = aging rate at temperature T

A = proportionality constant

[E.sub.a] = activation energy 

k = Boltzmann's constant The data do not support either of these expectations.

The o-ring compression set (CS) clearly does not follow the form:

CS = 1 - [e.sup.-Alphat]. Instead, the measured compression set is found to be roughly proportional to the log of the exposure time:

CS [Alpha] log t. Only the 87 [degrees] C data, taken over a 70 day time span, shows a departure from this time dependence. The departure begins when the compression set reaches approximately 80%. This functional dependence is not well understood but may be partially due to the long time constant for the physical compression set relaxation.

Furthermore, the aging rate/temperature behavior proved to be more complex than can be described by a simple Arrhenius treatment. This presumably arises because the physical and (two or more) chemical mechanisms contributing to the measured set each have different rate/temperature behavior. In other words, at higher temperatures the dominant mechanism changes, making accelerated aging tests at these temperatures nonrepresentative of the aging which the butyl experiences at the projected long term use temperature (i.e., room temperature).

Based on extrapolation from the room temperature data only, the predicted compression set after 17 years is approximately 23%. Superposition of the room temperature and 47 [degrees] C data yields a predicted compression set of 30% after 17 years. The superposition curve generated is also consistent with the 28% measured compression set of the engineering model exciter o-ring which had been installed for seven years.

A cross-sectioned squeeze of at least 9% is required for an initial o-ring design (ref. 2). The design of the exciter o-ring seal is such that even with a compression set of 75%, it will experience a 9% cross-sectioned squeeze of the "set" cross section.

Considerable margin exists between the predicted compression set after 17 years and the 75% value based on the initial squeeze requirement. Furthermore, the manufacturer indicates that the o-ring should maintain its sealing ability for compression set exceeding 90%, increasing the margin even more. The Galileo exciter butyl o-ring seal was therefore judged to be acceptable for a time period in excess of 17 years.

PHOTO : Figure 1 - compression set vs. time for o-ring segments under compression at room temperature, 47 [degrees] and 57 [degrees] C

PHOTO : Figure 2 - compression set versus time for o-ring segments under compression at 65 [degrees] and 87 [degrees] C

References

1. A.V. Tobolsky, "Properties and structure of polymers," Wiley, 1960. 2. Military specification G5514F.

Acknowledgments

"Adhesives for bonding cathodically protected rubber to metal devices," is based on a paper presented at the Fall 1990 Rubber Division meeting "Adhesion and wetting: Similarities and differences," is based on a paper presented at the Fall 1990 Rubber Division meeting "Low temperature sealing capabilities of fluoroelastomers," is based on a paper presented at the 1991 SAE exposition and is reprinted with permission 1991, SAE, Inc. "Long term sealing ability of butyl o-rings," was based on a paper presented at the Spring 1988 Rubber Division meeting.