Stewart Palmer

Deformation, strengths and strains to failure of polymer bonded explosives

This paper describes a study of the strength and failure properties of a range of polymer bonded explosives (PBXS). These are composite systems in which small (typically micrometre up to millimetre-sized) explosive crystals are bonded by a polymer (typically 2–10% (by mass)). In PBXS it is important to optimise the mechanical properties, while maintaining a low sensitiveness (i. e. the material is safe to manufacture, store and handle) and high explosiveness (i. e. reacts powerfully to a prescribed stimulus). The Brazilian test, in which a disc-shaped specimen is loaded diametrically, was chosen for the study. The advantages are that relatively small specimens of typically 10 mm diameter and 4 mm thickness can be used, and that the tensile stresses on the central axis are achieved by applying compressive stresses at the anvil so that complicated gripping arrangements are not required. The technique of double-exposure laser speckle photography was chosen to measure the in-plane displacement field. The technique can measure displacements to sub-micrometre accuracy and provide information over the whole specimen surface. These are distinct advantages over strain gauge methods that involve attaching gauges to the specimen and which only give pointwise information. The double-exposure speckle pattern records were interpreted using an automated Young’s fringes method. The PBXS were of three explosive types and those based on HMX were studied systematically for two crystal sizes and three different binder materials, of two different weight percents. In general, compositions based on micronized crystals were the strongest. Polishing techniques were developed to study the deformation of the individual crystals, the points of nucleation of failure and the fracture paths through the PBXS. The failure modes are discussed in terms of various theoretical models. The mechanical twinning which was shown in earlier work to be an important failure mode in β-HMX also takes place in PBXS based on HMX. The general applicability of the techniques developed in this research for other composite systems is emphasized.

Temperature-time response of a polymer bonded explosive in compression (EDC37)

The compressive strength of the energetic composition EDC37 has been measured at a temperature of 293 ± 2 K over a range of strain rates from 10−8 to 103 s−1, and at a strain rate of 10−3 s−1 over a range of temperatures from 208 to 333 K. The results show that failure stress is a monotonic function of applied strain rate or temperature, which is dominated by the relaxation properties of the polymeric binder; this is confirmed by dynamic mechanical thermal analysis performed on both EDC37 and its binder. Similarities between the compressive strain rate/temperature data sets can be understood by temperature–time superposition; data collected at a strain rate of 10−3 s−1 over a temperature range 208 to 333 K were mapped onto a plot of strain rate dependent strength at 293 K, using an empirically determined sensitivity of −13.1 ± 0.3 K per decade of strain rate. Sample size was noted to have a modest effect on the stress–strain behaviour; small length to diameter ratios gave results consistent with an increased degree of confinement. Samples taken to large strains exhibited strain localization in the form of shear bands.

White-light digital image cross-correlation (DICC) analysis of the deformation of composite materials with random microstructure

Abstract A sophisticated image cross-correlation algorithm (J. Appl. Opt. 33 (1994) 6667) has been used to measure both components of in-plane displacement at the surface of a deforming composite material. The natural random pattern present on the surface of a polished polymer bonded explosive (PBX) sample is photographed using a high-resolution digital camera mounted on an optical microscope. Frames are taken as the sample deforms and small subimages from each frame are compared with a base-frame recorded before loading. The resulting displacement information reveals the average motion present from the centre of each of the small subimages used in the analysis, relative to the original position. Currently, there is considerable interest in monitoring the safety of PBXs over the medium to long term and in response to accidental stimuli. Knowledge of the ways in which cracks are created and propagate within the material is a key aspect of understanding. Experiment shows that the accuracy of this technique is 0.1 pixels corresponding to 50 nm in the experimental arrangement used. The technique has wide application to many small to medium strain experimental situations due to the ease of data recording. In the experiment presented here, the natural random microstructure of the sample is used to make measurements, however, any manufactured random pattern applied to a sample would enable the method to be successful.

Impact Sensitivity of Propellants

This paper reports an experimental study of the rapid deformation and ignition behaviour of a number of cast double-base propellants both at room temperature and at temperatures below the glass transition of those compositions that were elastomer modified. A range of techniques were used to obtain stress/strain data on the materials. A drop-weight machine with transparent anvils was used in conjunction with a high-speed camera to observe the deformation behaviour during impact. The aim was to gain understanding of the key parameters leading to the impact initiation of propellants. The only propellants that ignited violently in solid disc form at room temperature were the conventional compositions, the most sensitive being a composite modified propellant containing aluminium and ammonium perchlorate. All compositions were sensitized if they contained gas spaces. In addition, all the elastomer modified compositions were sensitized by cooling them below their glass transition (ca.210K). The mechanisms leading to ‘hot spot’ formation and ignition of the propellants are discussed.

The mechanical response of a PBX and binder: combining results across the strain-rate and frequency domains.

The mechanical response of a polymer bonded explosive (PBX) has been measured using a Split Hopkinson Pressure Bar at a strain-rate of 2000 s−1, across a range of temperatures from 173 to 333 K, with the aim of observing its behaviour in the glassy regime. The yield stresses increased monotonically with decreasing temperature and no plateau was found. The failure mechanism was found to transition from shear-banding with crystal debonding fracture to brittle failure with some evidence of crystal fracture. Similar experiments were performed on samples of its nitrocellulose-based binder material, at a strain-rate of 3000 s−1 across a temperature range 173–273 K. The failure stresses of the binder approach that of the composite at temperatures near −70 °C. The elastic moduli were estimated from post-equilibrium regions of the stress–strain curves, and compared with those obtained for the composite using 5 MHz ultrasonic sound-speed measurement, and powder dynamic mechanical analysis measurements and quasi-static behaviour reported in a previous paper. The moduli were plotted on a common frequency axis: a temperature shift was applied to collapse the curves, which agreed with the Cox–Merz rule.

Rate dependent strengths of some solder joints

The shear strengths of three lead-free solder joints have been measured over the range of loading rates 10−3 to ~105 mm min−1. Binary (SnAg), ternary (SnAgCu) and quaternary (Castin: SnAgCuSb) alloys have been compared to a conventional binary SnPb solder alloy. Results show that at loading rates from 10−3 to 102 mm min−1, all four materials exhibit a linear relationship between the shear strength and the loading rate when the data are plotted on a log–log plot. At the highest loading rate of 105 mm min−1, the strengths of the binary alloys were in agreement with extrapolations made from the lower loading rate data. In contrast, the strengths of the higher order alloys were found to be significantly lower than those predicted by extrapolation. This is explained by a change in failure mechanism on the part of the higher order alloys. Similar behaviour was found in measurements of the tensile strengths of solder joints using a novel high-rate loading tensile test. Optical and electron microscopy were used to examine the microstructures of interest in conjunction with energy dispersive x-ray analysis for elemental identification. The effect of artificial aging and reflow of the solder joints is also reported.

Fracture Studies of PBX Simulant Materials

Fracture studies have been performed on three inert PBX simulants; PBS 9501 which consists of sugar bound in Estane and is a PBX 9501 simulant. EDC1037 and EDC1032 which consist of barium sulphate and melamine bound in NC/K10 and Viton‐A respectively, and are simulants for EDC37 and EDC32. The effect of microstructure, geometry and testing rate are investigated, and through the application of elastic‐plastic fracture mechanics, energy release rates have been calculated. Such data are required for the development and validation of accurate failure models.

Microstructural strain analysis by high-magnification moire interferometry

A high magnification phase-stepping interferometer has been constructed which has a spatial resolution of the order of one micrometers and a sub-mm field of view. Laser illumination is delivered to the microscope head by polarization-preserving single-mode optical fibers. The head itself is a compact unit consisting of collimating optics, objective lens, CCD camera, and a separate white light source. Phase gratings are cast on the polished specimen surface by replication from a master grating, in either silicone rubber or epoxy resin. Subsequent evaporation of a thin layer of gold onto the grating increases the reflectivity and reduces the speckle noise in the images. By switching between the laser illumination and the white light unit, it is possible to view the underlying microstructure in exact registration with the measured displacement fields. The instrument is illustrated with several applications including the visualization of delamination cracks in graphite-epoxy composites and measurement of the strain-to-failure of polymer-bonded-explosives.

Thermodynamic work of adhesion measurements of polymer bonded explosive constituents via the Wilhelmy plate technique and their application to AFM pull-off experiments

A major strength limiting factor for polymer bonded explosives above their glass-transition conditions is the magnitude of adhesion that exists between the polymeric matrix binder-system and the filler particles. Experimental measurements of the components of the free surface energy of the binder KEL-F8OO have been made using the Wilhelmy Plate technique. These data can be combined with equivalent data on the filler particles to calculate the so-called Thermodynamic Work of Adhesion. This under-pinning quantity can be used to predict the levels of load (stress) required to cause debonding in different geometries. A simple geometry of interest is a spherical-cap of polymer debonding from a flat substrate. Experiments using this geometry have been performed with an Atomic Force Microscope pulloff technique to measure the critical loads (stresses) required for debonding. There is excellent agreement between the predicted values based on the Wilhelmy Plate data and the measured values from the Atomic Force Microscope. Experimental data and understanding are required for the development and validation of microstructural models of mechanical behaviour.

BRAZILIAN DISC TESTING OF A UK PBX APPROACHING THE GLASS TRANSITION CONDITION

Previous research at the Cavendish Laboratory has shown a change in failure mechanism, from that of intergranular to transgranular failure, when PBX Brazilian disc specimens are tested below the glass transition condition of their binder system. The current study builds on this earlier research, illustrating how the strengths of samples tested at two different strain‐rates change as a function of temperature. The increase in strength at low temperatures is identified with stiffening of the polymer binder as the glass transition condition is approached. The changes in strength due to temperature and strain‐rate can be reconciled by assuming a log‐linear interrelationship.