Reactive materials can be loosely categorized as composites of inert solid materials which, when subjected to a violent mechanical stimulus such as an impact, react exothermally with a rapid release of energy. This reaction, while aptly described as “explosive,” differs from a true detonation or deflagration in that it requires a mechanical stimulus to not only initiate the reaction but also to sustain it. Such materials can also be fairly robust mechanically and can serve as substantial structural components. Because of these properties, reactive materials have a number of potential ordnance applications. Various compositions have been investigated to tailor properties of reactivity, strength, and density to suit particular needs.
An aluminum/ polytetrafluoroethylene (Al/PTFE) formulation serves as a benchmark for current reactive material development. In addition to material development, work is underway to develop physics-based modeling capabilities for reactive materials. Basic constitutive models were developed for the inert behavior of this material. Compression tests were performed over a range of strain rates and temperatures relevant to the conditions present during low-speed impact. This data, presented in the following sections, is used to generate parameters for both the Johnson-Cook (JC) and Modified Johnson-Cook (MJC) constitutive models. Although not ideally suited to represent this material, these parameters are given primarily due to the widespread use of these models and their availability in the current suite of hydrocodes. An additional, and more appropriate, fit is given for the Zerilli-Armstrong (ZA) model for polymers.
The samples tested were supplied by General Sciences, Inc. (GSI) and were made from a material designated GSI-0017. It is a pressed and sintered mixture of aluminum and PTFE powders, 26.5% and 73.5% by weight, respectively. The initial powder sizes are 44 and 31 μm, respectively. The density of the compacted material, as measured by a buoyancy method based on Archimedes principle, is 2.29 g/cm3
Low-rate tests were performed with an Instron Model 1331 servo-hydraulic load frame. The load applied to the specimen was measured with a load cell, and the specimen deformation was measured using a linear variable differential transformer (LVDT) measurement of the cross-head displacement, and includes a correction for machine compliance. The specimens were cylindrical, nominally 6.35 mm in both diameter and length. Contact surfaces were lubricated with a heat-stable silicone lubricant. All of the low-rate tests were performed at room temperature (22 °C).
A 6.35-mm-diameter 7075-T6 aluminum Split Hopkinson Pressure Bar (SHPB) was used for the high-rate tests. A series of six tests was performed at rates from 600 to 8000/s, all initially at room temperature (22 °C). The specimens were cylindrical, 3.18 mm in diameter and length, and contact surfaces were lubricated with the same silicone lubricant used in the low-rate tests.
A final set of experiments was performed at elevated temperatures to quantify the thermal softening behavior of the material. These were performed at a consistent strain rate of 4000/s using the SHPB. Heating was accomplished by circulating heated air into a chamber that enclosed the specimen and the adjacent ~65 mm sections of the bars. Ideally, specimen temperature would have been monitored directly with a thermocouple glued to each specimen. However, this proved impractical because of the small sample size and also because of the difficulty in adhering gages to the specimen. Instead, specimen temperature was measured with a thermocouple probe placed within 1 cm of the specimen; i.e., the probe measures ambient air temperature and not the specimen temperature directly. The temperature in the chamber was allowed to equilibrate over a 20-minute period prior to each test to ensure that the specimen and relevant sections of the bars were allowed to reach the ambient temperature. The temperature gradient in the bars is believed to have negligible effects on the bar wave propagation and the strain gage measurements. Temp erature measurements made in this way are estimated to be accurate to within ±2 °C.