A physics-based empirical model is developed to characterize the time varying temperature profile from post-detonation
combustion. Fourier-transform infrared signatures are collected from field detonations of RDX-based aluminized high
explosives surrounded by an aluminized plastic-bonded spin-cast liner. The rate of change of temperature in the postdetonation
combustion fireballs are modeled using a radiative cooling term and a double exponential combustion source
term. Optimized nonlinear least-squares fit of the numerical solution of the empirical model to the temperature data
yields peak temperatures of 1290-1850. The observed heat released in the secondary combustion is well correlated with
the high explosive and liner heat of combustion with an average efficiency of 54%.
Several homemade explosives (HMEs) were manufactured and detonated at a desert test facility. Visible and infrared
signatures were collected using two Fourier transformspectrometers, two thermal imaging cameras, a radiometer, and a
commercial digital video camera. Spectral emissions from the post-detonation combustion fireball were dominated
by continuum radiation. The events were short-lived, decaying in total intensity by an order of magnitude within
approximately 300ms after detonation. The HME detonation produced a dust cloud in the immediate area that
surrounded and attenuated the emitted radiation from the fireball. Visible imagery revealed a dark particulate (soot)
cloud within the larger surrounding dust cloud. The ejected dust clouds attenuated much of the radiation from the
post-detonation combustion fireballs, thereby reducing the signal-to-noise ratio. The poor SNR at later times made
it difficult to detect selective radiation from by-product gases on the time scale (~500ms) in which they have been
observed in other HME detonations.
Continuum emission is predominant in fireball spectral phenomena and in some demonstrated cases, fine detail in the
temporal evolution of infrared spectral emissions can be used to estimate size and chemical composition of the device.
Recent work indicates that a few narrow radiometric bands may reveal forensic information needed for the explosive
discrimination and classification problem, representing an essential step in moving from "laboratory" measurements
to a rugged, fieldable system. To explore phenomena not observable in previous experiments, a high speed (10μs
resolution) radiometer with four channels spanning the infrared spectrum observed the detonation of nine home made
explosive (HME) devices in the < 100lb class. Radiometric measurements indicate that the detonation fireball is well
approximated as a single temperature blackbody at early time (0 < t ⪅ 3ms). The effective radius obtained from absolute
intensity indicates fireball growth at supersonic velocity during this time. Peak fireball temperatures during this initial
detonation range between 3000.3500K. The initial temperature decay with time (t ⪅ 10ms) can be described by a
simple phenomenological model based on radiative cooling. After this rapid decay, temperature exhibits a small, steady
increase with time (10 ⪅ t ⪅ 50ms) and peaking somewhere between 1000.1500K-likely the result of post-detonation
combustion-before subsequent cooling back to ambient conditions . Radius derived from radiometric measurements
can be described well (R<sup>2</sup> > 0.98) using blast model functional forms, suggesting that energy release could be estimated
from single-pixel radiometric detectors. Comparison of radiometer-derived fireball size with FLIR infrared imagery
indicate the Planckian intensity size estimates are about a factor of two smaller than the physical extent of the fireball.
A suite of instruments including a 100 kHz 4-channel radiometer, a rapid scanning Fourier-transform infrared spectrometer,
and two high-speed visible imagers was used to observe the detonation of several novel insensitive munitions being
developed by the Air Force Research Laboratory. The spectral signatures exhibited from several different explosive
compositions are discernable and may be exploited for event classification. The spectra are initially optically thick, resembling
a Planckian distribution. In time, selective emission in the wings of atmospheric absorption bands becomes
apparent, and the timescale and degree to which this occurs is correlated with aluminum content in the explosive formulation.
By analyzing the high-speed imagery in conjunction with the time-resolved spectral measurements, it may be
possible to interpret these results in terms of soot production and oxidation rates. These variables allow for an investigation
into the chemical kinetics of explosions and perhaps reveal other phenomenology not yet readily apparent. With an
increased phenomenological understanding, a model could be created to explain the kinetic behavior of the temperature
and by-product concentration profiles and thus improve the ability of military sensing platforms to identify explosive
types and sources.