Search

As high explosive production capabilities modernize and performance models become more sophisticated, there is an increasing need for detailed microstructural information such as void size and morphology within pressed parts. Little is known of the relationship between the initial high explosive particle size distribution and the void size and morphology within a compact. Small-angle neutron scattering techniques were used to study the evolution of this relationship in TATB samples pressed from three distinct particle size distributions over the density range from 1.55 – 1.89 g/cc. In all samples studied the void volume is well described as ramified networks of voids. Details of the void network, including size and connectivity, will be discussed in terms of the pressed density and the initial TATB particle size distribution. [ABSTRACT FROM AUTHOR]

The microstructure of a high explosive (HE) material directly influences the reactive behavior under a variety of insults. Furthermore, the preparation method of polymer-bonded explosives (PBX), as well as the constituent material particle size distribution and consolidation parameters, will dominate the microstructural features that are observed. Here, we compare batches of PBX 9501 prepared by the typical wet slurry process and those prepared by resonant acoustic mixing. The "prills" produced by the wet slurry process leads to an inherently different microstructure than the coated powder produced by resonant acoustic mixing. We observed that the conditions required to consolidate (press) the material to the nominal PBX 9501 density varied depending on the preparation technique, which also dictates the observed binder distribution. Furthermore, X-ray microcomputed tomography and ultra-small-angle X-ray scattering analyses revealed distinct differences in the void structure and distribution. [ABSTRACT FROM AUTHOR]

Detonation waves that are nominally symmetric about a point or axis or plane nearly always exhibit a degree of accidental asymmetry. For many purposes one must eliminate and/or quantify this aberration in order to extract the desired information. We present symmetrization strategies and examples for several 1D and 2D detonation data sets. While none of basic principles are new to the world, we're unaware of any other tutorial compilation like this one. [ABSTRACT FROM AUTHOR]

Polymer-Bonded eXplosive (PBX) prills are spongy nuggets a few millimeters in diameter, comprising a spatially nonuniform distribution of HE crystals and binder. When pressed hotter than binder melting temperatures, binder is intended to uniformly disperse throughout the system. In reality it does so imperfectly, such that x-ray tomographic scans often look like a collection of mashed prills. Lore is that larger prills have more nonuniform binder distribution. If not, then larger prills still require binder to flow farther to disperse. The degree to which binder flows during pressing affects the void structure within pressed charges, which one suspects will affect material strength and likely shock sensitivity. Here we explore, for three PBX 9502 batches using the same TATB powder blend, how prill size correlates with ECOT test corner-turning distance. There is little effect for sufficiently small prills; however, for large prills corner-turning distance increases by an amount greater than experimental error. [ABSTRACT FROM AUTHOR]

Understanding the path through thermodynamic space that a high explosive experiences during shock to detonation transition (SDT), detonation propagation and failure are important for understanding the roles of these experiments for reactive flow model calibration. SDT, steady detonation propagation, corner turning (ECOT), multi-shock, isothermal and isentropic compression, shock release and re-shock, GapStick and other experiments may "overlap" with respect to the thermodynamic states attained. The focus of this study is to explore a methodology to examine where the reactants equation of state are taken by experiments. To that end, clearly a complete reactive flow model is needed to faithfully ascertain the dynamics of each experiment. For this exploration, an Arrhenius-Wescott-Stewart-Davis (AWSD) model is utilized for dynamic simulations of the tri-amino-tri-nitro-benzene (TATB) based plastic bonded explosive PBX 9502. Here, only an example of SDT will be given, as it serves as a validation of the more general methodology. [ABSTRACT FROM AUTHOR]

Plastic-Bonded Explosive (PBX) prills are agglomerates a few millimeters in diameter composed of High Explosive crystals non-uniformly distributed in and surrounded by a polymer binder. When pressed at elevated temperatures at or above the binder melt temperature the polymer flows to uniformly coat the HE crystals to create a well-consolidated compaction. In reality it does so imperfectly, such that x-ray tomographic scans often look like a collection of prills mashed together. It is speculated that the larger the prill, the more non-uniform the binder distribution, but this has never been proven. Even if that were not so, the larger the prill the farther binder must flow toward the middle to homogenize during pressing. Thus, the degree of homogenization depends in part on prill size. The degree to which binder flows during pressing in turn affects the void distribution within pressed charges, which one suspects will affect shock sensitivity and material strength. In this paper we explore, for ∼30 PBX 9502 formulation batches using the same TATB powder lot, how prill size affects the compressive mechanical properties. [ABSTRACT FROM AUTHOR]

We know that microstructure and density affect the shock initiation of an insensitive high explosive (IHE), such as PBX 9502 (95% TATB). We suspect that those factors also affect some performance and propagation metrics, especially curvature effects and the explosive's ability to turn corners. A recently developed hydrodynamic burn model, πSURF, provides insight and predictive capability for the initiation regime of shock response based partly on a statistical characterization of the microstructure. Here we examine the ability of the πSURF model to predict the effects of microstructural features on the propagation regime characteristics of curvature and corner turning of the IHE PBX 9502. Specifically, we explore the hypothesis that informing the model about the presence or absence of intra-granular porosity (the 'nano-scopic' voids, as opposed to the larger, inter-granular 'micro-scopic' voids), as revealed by a void size distribution, will predict the corner turning behavior observed in the Enhanced COrner Turning (ECOT) experiments for these two classes of PBX 9502. Indeed, we learned that the model does show the expected behavior. We also learned that the model smoothly predicts, without adjustment, both the initiation and propagation regimes of shock response (other models require mathematical or code switching between the regimes). [ABSTRACT FROM AUTHOR]

Simultaneous ultra-high-speed framing and electronic streak photography are applied to the working surface of cylindrical, detonator-scale explosives to quantitatively assess detonation-wave breakout kinematics. The framing camera captures the evolution of two-dimensional cross-sections of the detonation wave exiting a detonator output pellet, and the streak camera records one-dimensional, space-time breakout curvature with sub-nanosecond resolution. Both data sets are analyzed to assess the geometrical characteristics of the detonation wave. An existing and idealized model equation for hyperbolic breakout of cylindrical high-explosive charges has been parameterized for nominal detonator-explosive performance and used to initially inform the geometry of the detonation wave within a functioning detonator pellet. Accurate extraction of output pellet detonation velocity is attempted using the idealized model equation, and uniqueness results are discussed. [ABSTRACT FROM AUTHOR]

Traditional slow detonation velocity (Dv) explosives components, such as Baratol (76% Ba(NO3)2 and 24% TNT), rely on dilution of a traditional explosive with a dense relatively inert material, while some utilize Ca(NO3)2, ZnO or BaCO3. However, our applications require solely CHNO-based formulations that exhibit slow Dv near theoretical maximum density. Given a target Dv of 6.5 mm/µs, ammonium picrate was chosen as a convenient explosive to be diluted with a high binder level (14-20%). Thermal equilibrium calculations were performed to determine the binder level to provide the desired Dv. Two formulations were produced, a molding powder with a polystyrene/dioctyl adipate binder, Dv = 6.45 mm/µs and a cast-cure using hydroxy-terminated polybutadiene/bis-(2,2-dinitropropyl)acetal-formal/MDI binder, Dv = 6.58 mm/µs. All formulations showed no sensitivity response or compatibility issues. The cast-cured formulation was chosen for further analysis, and cylinder expansion was performed followed by JWL parameterization. Detonation tests confirmed that cast-cured formulation would initiate and propagate unconfined at thicknesses above 12 mm. The ultimate test configuration of the formulation was a sweeping initiation from PBX 9502, with Proton Radiography (pRad) imagery to demonstrate the leading detonation front in the PBX 9502 and a lagging detonation front in the slower cast-cured formulation, as predicted by simulation. [ABSTRACT FROM AUTHOR]

Making high explosives that possess insensitivity on par with TATB-based plastic bonded explosives (PBXs), while outperforming them, has proven to be a difficult challenge. Many molecules that have challenged TATB have fallen short in either small-scale sensitivity (impact, friction), thermal stability, or possess a shock sensitivity that is either too high or too low. Recently, an alternative approach to single-molecule-based PBXs has been blending and/or cocrystallizing explosive molecules to address shortcomings of individual components. With this approach in mind, characterization of formulations of 1,1-diamino-2,2-dinitroethene (DADNE or FOX-7) or 3,3'-diamino-4,4'- azoxyfurazan (DAAF) with 3-nitro-1,2,4-triazole-5-one (NTO) were investigated. Two mono-molecular explosives and mixtures with binders were also evaluated, 5,7-diamino-4,6-dinitrobenzofuroxan (CL-14) and 1-nitroso-3,5-dinitro-1,3,5- triazacyclohexane (M-RDX or mononitroso-RDX) A new method for the production of M-RDX was discovered. The previous procedure was improved by economizing to a one-pot method involving two high yielding steps. The overall yield for this reaction was 80% with minimal RDX contamination. Half-scale (12.7 mm inner diameter) copper cylinder expansion testing was used to investigate the detonation performance of DAAF/NTO, FOX-7/NTO and CL-14 formulations. Small scale sensitivity tests were also performed on CL-14, M-RDX and NTO based formulations. [ABSTRACT FROM AUTHOR]

Following G.I. Taylor and our own previous work based upon it, we continue to explore HE confiner motion as test diagnostic. A handful of authors (including us) have adapted Taylor's analysis to infer equation-of-state information from detonation cylinder tests. We have also successfully applied this class of analysis to deflagration cylinder tests (DFCTs). Here, we analyze a miniature DFCT diagnosed by multiple PDV probes. We spatially interpolate between PDV records to obtain a smooth function for lateral expansion vs. axial distance, z, and time, t. From this we may construct snapshots of the tube at any time; hence we may create animations of the tube motion and compute the product gas volume V vs. t. Combining Newton's law with simple stress theory, we estimate the gas pressure P vs. z and t. We then estimate the burned HE mass from the wall kinetic energy. Finally, we discuss some important test scaling issues. [ABSTRACT FROM AUTHOR]

The standard statistical hot spot model (SHSM) assumes that all hot-spot-triggered burn waves initiate simultaneously within homogenized volume elements. In reality the shock passes through such elements, lighting burn waves in a phased manner. Simple simulations are employed to illustrate the resulting shock-triggered heterogeneous reaction topology. These show that the conventional continuum prescription may not be satisfied. An alternative strategy yields a robust continuum description, and enables an extended analytic SHSM that reduces to the standard model as the burn-front to shock-speed ratio, β, approaches zero. [ABSTRACT FROM AUTHOR]

We present detonation copper cylinder test results on aged PBX 9404 explosive. The charges were newly pressed from 37.5 year-old molding powder. We compare these results to equivalent data performed on the same lot when it was 3.5 years old. Comparison of the detonation energy inferred from detonation speed to that inferred from wall motion suggests that the HMX energy is unchanged, the NC energy has decreased by ∼88%, and that ∼89% of the NC decomposition products are retained in the molding powder. [ABSTRACT FROM AUTHOR]

For explosive systems that exhibit gaps or cracks between their internal components (either by design or mechanical failure), measureable time delays exist for detonation waves crossing them. Reinitiation across such gaps is dependent on the type of explosive, gap width, gap morphology, confinement, and temperature effects. To examine this reinitiation effect, a series of tests has been conducted to measure the time delay across a prescribed gap within an “infinitely” confined PBX 9501 system. Detonation breakout along the explosive surface is measured with a streak camera, and flow features are examined during reinitiation near the gap. Such tests allow for quantitative determination of the time delay corresponding to the time of initiation across a given gap oriented normal to the direction of the detonation wave. Measured time delays can be compared with numerical calculations, making it possible to validate initiation models as well as estimate detonation run-up distances. Understanding this reinitiation behavior is beneficial for the design and evaluation of explosive systems that require precision timing and performance. [ABSTRACT FROM AUTHOR]

We determine the reaction progress function for an ideal hot spot model problem. The considered problem has an exact analytic solution that can be derived from a reduction of A. Nichols’ statistical hot spot model [1]. We perform numerical calculations to verify the analytic solution and to illustrate the error realized in real, finite systems. We show how the baseline problem, which does not distinguish between the reactant and product densities, can be scaled to handle general cases for which the two densities differ. [ABSTRACT FROM AUTHOR]

Recently there has been significant interest in evaluating the potential of many different non-ideal energetic materials to cause blast damage. We present a method intended to quantitatively compare the blast loading generated by different energetic materials through use of an explosively driven shock tube. The test explosive is placed at the closed breech end of the tube and initiated with a booster charge. The resulting shock waves are then contained and focused by the tube walls to form a quasi-one-dimensional blast wave. Pressure transducers along the tube wall measure the blast overpressure versus distance from the source and allow the use of the one-dimensional blast scaling relationship to determine the energy deposited into the blast wave per unit mass of test explosive. These values were measured for C4, ANFO, and two perchlorate explosives. Explosive equivalencies from these values were found to agree with prior theory and experiment. [ABSTRACT FROM AUTHOR]

We present a simple and inexpensive explosive line wave generator which has been designed using commercial sheet explosive and plane wave lens concepts. The line wave generator is constructed using PETN- and RDX-based sheet explosive for the slow and fast components, respectively, and permits the creation of any desired line width. A series of experiments were performed on a 100-mm design, measuring the detonation arrival time at the output of the generator using a streak camera. An iterative technique was used to adjust the line wave generator's slow and fast components, so as to minimize the arrival time deviation. Preliminary tests achieved a wavefront simultaneity of 100 ns with a 7.0 mm/μs detonation wave. Designs, test results, and concepts for improvements are discussed. [ABSTRACT FROM AUTHOR]

The classical thermal explosion theory of D. A. Frank-Kamenetskii (F-K) is analytic with respect to charge size but numerical with respect to charge shape. The numerical determination of criticality for a general shape is non-trivial, and is therefore not often performed in practice. For estimation purposes it would be useful to avoid computations by finding an approximate correlation relating the critical state to the relevant generic attributes of the charge shape. I present such a correlation that accurately does so for bodies of near-unity aspect ratio; however, the result breaks down for extended aspect ratio bodies. [ABSTRACT FROM AUTHOR]

The incidental effects of gaps in detonating explosives have been observed for over two centuries, yet the root cause of peripheral damage due to these features has been a partial mystery. To evaluate such damage for PBX 9501, a test series has been performed that examines single and multiply-directed detonations crossing gaps of varying widths and angles relative to the detonation wave fronts. Damage is evaluated with steel witness plates and quantified through trench profiling and photographic analysis. In addition, streak camera traces are used to track detonations along explosive material surfaces and across gaps. Such traces allow the quantification of timing delays due to the gap reinitiation process. For some reinitiation tests, a second detonation wave is directed to interfere at varying times with the post-gap initiation process of the first wave, thus allowing complex wave-wave interactions to be investigated in detail. With these cumulative observations, further insight into the mechanism of extrinsic damage due to gaps is gained. [ABSTRACT FROM AUTHOR]

This work predicts the critical conditions required for the onset of reaction runaway in a narrow high-explosive slot intended to simulate a crack. We review ongoing experiments where flames propagated through such slots at velocities up to 10 km/s, reaching pressures in excess of 1 kbar. A model is developed where slot pressurization is attributed to gas-dynamic choking at the slot exit. The combination of choking and a pressure-dependent reaction rate is shown to be capable of runaway reaction for a range of slot dimensions and pressures. This model agrees with experimental pressure measurements of reaction runaway in slots and provides a mechanism for the erratic burning observed with some explosives under high pressure. [ABSTRACT FROM AUTHOR]

Burning crack networks can play an important role in cookoff explosions of the HMX-based explosive PBX 9501. Combustion within cracks can potentially spread to connected porosity, which is believed to develop in the heated explosive. The burning surface area would thereby increase, causing a more violent outcome. I develop an analytic model that seeks to capture the essential physical mechanisms controlling how such a scenario runs away. The calibrated model behavior is consistent with Ref. 1 tests. © 2006 American Institute of Physics [ABSTRACT FROM AUTHOR]

A multidimensional implementation of DSD, formulated with the level set method, is applied to track the propagation of a detonation wave in a heterogeneous explosive consisting of an array of inert cylindrical obstacles with a liquid explosive in the interstitial space. With the Huygens assumption, the average detonation velocity through the explosive is less than that for the liquid explosive alone, due to the increased path length. When the normal detonation velocity is assumed to depend on front curvature, there is an additional, smaller reduction in the detonation velocity, which depends on the cylinder material. The detonation velocity deficits obtained in the computations are of the same order as those observed experimentally for a heterogeneous explosive consisting of a packed bed of spherical inert beads saturated with sensitized nitromethane. The DSD computations are relevant to the experimental results in the large-bead limit in which the pore dimension is large enough to support the propagation of discrete detonation wavelets in the interstitial liquid between the beads. [ABSTRACT FROM AUTHOR]