New aircraft are expected to use more composite materials. This drives the need for nondestructive evaluation (NDE) of composite materials. Current aircraft are flown beyond their design lives, and have to be inspected for fatigue and other types of damage. This necessitates a need to understand the NDE signal-material relationship to evaluate the component state.

Atomic force microscope (left) and ultrasonic force microscope (right) images of PMR-15 Resin fatigued to fracture. Images show random microstructure. The “X” shape is a polishing artifact.

High-temperature thermoset monomers have three or more reactive sites to form bonds, unlike thermoplastic monomers, which have only two reactive sites. These extra sites enable crosslinking and network formation, as opposed to individual long chains in thermoplastics. When the reaction starts, the thermoset monomers have a low viscosity, which allows them to move about freely. Gradually, large groups of reacted monomers are formed (gel balls) in a solution of monomers and small groups of reacted monomers (oligomers). As the size of the gel balls grows, the viscosity increases, preventing the remaining monomers and oligomers from moving freely. In a cured thermoset polymer, this leads to regions of high crosslink density (gel balls) and regions that could not reach high crosslink density due to diffusion-limited motion (regions between the gel balls).

It is well known that crosslink density determines the physical properties of a polymer, with higher crosslink density resulting in higher elastic modulus (E), and higher glass transition temperature, Tg. It is expected that thermosets with regions of higher and lower crosslink density would behave like a two-phase material with one phase being stiffer (gel balls), and the other being more compliant (regions between the gel balls).

Thermoset polymers are viscoelastic materials. The two-phase nature of the material and the two-component nature of the model allow one to correlate the behavior of each phase to one of the components in the model. The gel balls being highly crosslinked and more stiff will behave more like the spring, while the less stiff between regions will behave more like the dashpot. The spring is assumed to be lossless and behave linearly with the stress, while the dashpot has a lossy behavior.

Under cyclic loading, the deformation of gel balls and the between regions will be different. Due to higher crosslink density, the gel balls are stiffer and will deform less compared to the between region. Depending on the load conditions, the polymer chains might be stretched and disentangled to different extents. Since the between region is softer than the gel balls, there is a possibility of rearrangement and alignment along the loading or other preferred direction, similar to slip in single crystalline materials. In tensile-tensile fatigue when the cyclic loading is interrupted, it is expected that at the peak stress, the material stretches, disentangling and elongating the polymer chains. Similarly, at minimum stress, the material relaxes, but not fully, owing to the viscoelastic nature of the material and the fact the material is still under some tensile stress.

PMR-15 resin is a high-temperature thermosetting polyimide used for aerospace composites where high temperature strength is required. This resin is rarely used by itself, and usually is used as the matrix material of a composite with glass or carbon fibers. In this work, PMR-15 resin was used alone so that hysteretic heating effects would not be complicated by contributions from both the resin and a fiber if a composite had been tested. To observe the changes in the behavior of hysteric heating under cyclic heating, dogbone-shaped specimens with 25.4 × 12.5 × 3 mm gage section were used. They were fatigued with a servo-hydraulic test machine with a stress ratio of 0.1, and a frequency of 3 Hz. Fatigue was stopped every 75,000 cycles for 30-60 minutes to allow the sample to return to ambient temperature. An infrared camera was used to record the sample heating behavior, and it was calibrated to have a high sensitivity about room temperature. A computercontrolled data acquisition system recorded the sample temperature.

To observe the changes in the microstructure due to fatigue, AFM and UFM measurements were performed on samples extracted from a fatigue fractured sample. The fatigue testing was interrupted every 75,000 cycles until the sample fractured. Between the interruptions, the sample was allowed to cool down to room temperature (30-60 minutes). In each of the interruptions, a clear transition in slope from a steeper initial slope to a shallower one is observed. This is indicative of a change in the heating behavior of the material. This experimental observation supports the two-phase deformation and hysteretic heating behaviors where the material effectively becomes stiffer once the polymer chains have been untangled and extended.

This work was done by J.T. Welter, E.A. Lindgren, and R. Hall of the Air Force Research Laboratory; and S. Sathish, G.P. Tandon, N. Schehl, M. Cherry, and V. Nalladega of the University of Dayton. AFRL-0213


This Brief includes a Technical Support Package (TSP).
Thermo-Elastic Nondestructive Evaluation of Fatigue Damage in PMR-15 Resin

(reference AFRL-0213) is currently available for download from the TSP library.

Don't have an account? Sign up here.