Vernon T. Bechel

Effect of stacking sequence on micro-cracking in a cryogenically cycled carbon/bismaleimide composite

An apparatus was developed to thermally cycle coupon-sized mechanical test specimens to −196 °C. Using this device, IM7/5250-4 carbon/bismaleimide cross-ply ([0/90]2S and [90/0/90/0/90/0/90/0/90]) and quasi-isotropic ([0/45/−45/90]S) laminates were submerged in liquid nitrogen (LN2) and returned to room temperature 400 times. Ply-by-ply micro-crack density (transverse cracks), micro-crack span, laminate modulus, and laminate strength were measured as a function of thermal cycles. The composite micro-cracked extensively in the surface plies followed by sparse micro-cracking of the inner plies. The tensile strength of the two blocked lay-ups (lay-ups with adjacent plies of the same orientation) decreased by 8.5% after 400 cycles. Sectioning of the samples revealed that the micro-cracks in the surface plies spanned the full width of the sample while many of the micro-cracks observed on the edge of the inner plies did not extend to the center of the samples, implying that a rectangular specimen with exposed free edges may result in a significantly different micro-crack density than a sample without free edges.

Evaluation of interfacial normal strength in a SCS-0/epoxy composite with cruciform specimens

In this study, the cruciform geometry is utilized to evaluate the interfacial normal strength in unidirectional SCS-0/epoxy composites by using single-fiber specimens. These model specimens are incrementally loaded in tension to failure with a specially built loading device mounted on the straining stage of a microscope. The initiation and location of interfacial debonding is observed in situ by microscopic examination and inspection of photoelastic fringe patterns. The model specimens are also loaded to failure in an MTS machine while strain and acoustic emission activity are continuously monitored. Analytical modeling by the 3-D finite-element method reveals that the radial stress at the interface is the dominant stress component and remains fairly constant (within 10%) over two-thirds of the central loading region. As a result, interface debonding always occurs in the interior of the sample (region initially free of stress singularities), thus avoiding the influence of free-edge effects on measured debond strengths.

A series solution of the volume integral equation for multiple-inclusion interaction problems

Abstract A modification of the volume integral equation (VIE) method is proposed for the solution of elastostatic 2-D problems in unbounded solids containing interacting multiple fibers. The efficiency of the VIE method developed by Lee and Mal (Lee, J., Mal, A. A volume integral equation technique for multiple inclusion and crack interaction problems. Journal of Applied Mechanics – Transactions of ASME 1997;64:23–31) where a weakly singular integral equation is involved has been improved by the use of a modification in the spirit of a subtraction technique for a single inclusion. The accuracy and efficiency of the method are examined through comparison with results obtained from finite element analysis.

The effect of residual stresses and sample preparation on progressive debonding during the fiber push-out test

Abstract Iterative finite-element analyses have been conducted to determine the debond length as a function of force during progressive debonding in fiber push-out tests. This procedure was applied to data from push-out tests on a steel/epoxy model composite. During push-out testing this composite debonded from the bottom (support) face of the sample. The modeling allowed debonding from the bottom and included pre-existing debonds caused by residual stresses and/or sample preparation. The predicted debond lengths were within 10% of the measured debond lengths. Evidence is also shown that cutting and polishing can cause debonds that are longer than one third of the sample thickness in samples with a relatively small diameter steel fiber.

Fiber–Matrix Interfacial Failure Characterization Using a Cruciform-Shaped Specimen

In this study, a cruciform-shaped test specimen is utilized to characterize the fiber–matrix interface under transverse and combined (tensile and shear) loading. We first present an overview of past references of how the cruciform geometry is optimized to promote interfacial failure. We then discuss a modification of the cruciform specimen where face-sheets are adhesively bonded to reinforce the sample. These face-sheets serve a twofold purpose, namely, to prevent premature failure in the fillet region and to encourage debond initiation at the center of the gage length. Finally, an off-axis cruciform geometry, in which the wings of the cruciform sample are inclined at an angle with respect to the loading direction, is introduced to characterize the fiber–matrix interface under combined transverse and shear loading. Using the measured value of applied stress at debond initiation, and the evaluated stress concentration factor at the fiber–matrix interface, a mixed-mode failure envelope is then constructed in the normal-shear stress space, and a quadratic failure criterion is proposed.

Characterization of interfacial failure using a reflected light technique

A simple optical method of observing and tracking interface failure in transparent matrix composites is developed to extend the range of material systems that can be tested with the cruciform test for measuring the normal strength of a bi-material interface. This technique of detecting debonds using reflected light is demonstrated on two model material systems with fibers 140 and 15 μm in diameter. The reflected light technique is more reliable and provides information about the interface that cannot be obtained with previous methods of debond detection, such as photoelasticity, surface strain gages, and acoustic emission. Using this technique, debond initiation, location, length, and shape are measured as a function of applied load; thus, all the required parameters for calculating the normal strength of the interface, as well as the interfacial energy release rate are acquired.

Bismaleimide/preceramic polymer blends for hybrid material transition regions: Part 1. Processing and characterization

In this article, initial steps were taken in the study of processing mixtures of a polymer and a preceramic polymer to produce a hybrid material that contains a region of gradual transformation from a polymer to a ceramic. It is thought to be a first step toward the eventual development of a hybrid that is graded from a polymer matrix composite to a ceramic matrix composite. Thermal analysis, elemental analysis, and morphology characterization of blends of RD-730 preceramic polymer, which converts to silicon carbide upon pyrolyzation, and Matrimid A/B polymer which is a bismaleimide were carried out as a function of cure cycles. The cure cycles were chosen to vary the resin viscosity at specific times during processing in order to affect the amount of phase separation of the two resins. These results were then used to make conclusions on how processing parameters affected the miscibility of the two resins and the likelihood of producing a hybrid having mechanical and thermal properties that fall between those of the two constituents. High RD-730 loadings were achieved without phase inversion, the glass transition temperature (Tg) of the blended material was shown to be significantly influenced by a post-cure, and a spatial gradient in RD-730 concentration was noted.

Bismaleimide/preceramic polymer blends for hybrid material transition regions Part 2. Incorporating compatibilizers

The problem of phase separation in polycarbosilane (Starfire resin)–Matrimid thermoset blend structures, designed for thermal protection applications, was addressed in this study. The structural design of the compatibilizers synthesized via a hydrosilylation–deprotection sequence was primarily based on a polymer structure with a nonpolar segment consisting of a carbosilane unit and a polar segment with a bisphenol-type moiety. Besides the homopolymer, block copolymers with varying contents of the polycarbosilane unit and the bisphenol A unit were synthesized and characterized. Curing experiments on the blends of Starfire resin–Matrimid incorporating 1–5% of the compatibilizers were performed followed by morphological characterization of the cured resin plaques. The effectiveness of the compatibilizers was assessed by the StarFire domain size and the distribution of cured specimens was examined in the scanning electron micrographs. Results indicated that the compatibilizers that were more effective in minimizing phase separation were relatively more polar and comprised a larger fraction of the phenolic functional units. In fact, the blend composition incorporating 5% of the homopolymer compatibilizer with the largest fraction of the polar bisphenol A unit was highly effective in providing a homogeneous distribution of very small StarFire domains, that is, fairly uniform domain sizes of the order of 5 µm or less.

Damage and the Ignition of PMCs Mechanically Impacted in Liquid Oxygen

The ASTM D2512 liquid oxygen (LOX) impact test is often one of a group of tests used in combination to determine the tendency for a material to ignite in the presence of LOX. Polymer matrix composites (PMCs) with epoxy matrices typically have difficulty passing it. In this work, experiments were done to build on past investigations into the relationship between damage formation and ignition rate when using the D2512 test. An impacting apparatus was fitted with a photodiode for detecting flashes due to ignition events and a load cell for recording the mechanical response of the sample to the impact. The extra instrumentation allowed the link between the time when a flash occurred and various features of the force versus time data during impact to be measured in order to relate damage events to the initiation of an ignition event. Both an epoxy matrix composite and a polyimide matrix composite were tested using multiple lay-ups. In this way, results could be compared across materials and across samples that had large differences in their response to the first impact during multiple impact tests. Among several results, it was found that quite often small ignition events occurred which were not visible to the naked eye in both materials and that ignitions usually took place only after the force versus time data indicated that damage was likely to be forming or already present in the sample.