CP620, Shock Compression of Condensed Matter - 2001
edited by M. D. Furnish, N. N. Thadhani, and Y. Horie
2002 American Institute of Physics 0-7354-0068-7
DETONATION INITIATION IN PRESHOCKED LIQUID EXPLOSIVES
Andrew J. Higgins1, Francois X. Jette1, Akio C. Yoshinaka1, John H.S. Lee1, and Fan Zhang2
McGill University, Department of Mechanical Engineering, 817 Sherbrooke St. W, Montreal, Quebec, Canada
2
Defence Research Establishment Suffield, P. O. 4000 Stn. Main, Medicine Hat, Alberta, Canada
Abstract. The initiation of detonation in a homogenous liquid explosive by the reflection of a strong shock
from a high impedance anvil is investigated. By transmitting a sub-critical shock through a test sample of
sensitized nitromethane and then reflecting it normally off a steel plate bounding the explosive, detonation
can be initiated in the pre-shocked medium. The initiation of detonation is observed via fiber optics
monitored by photodiodes and by manganin pressure gauges mounted on the steel plate. The initiation of
detonation by the reflected shock is inferred from the appearance of intense luminosity and an increase in
pressure at the explosive/steel interface, both appearing about 1 jas after shock reflection. The manganin
gauge measurements indicate that the critical pressure for incident initiation by a 100 mm diameter shock is
4-5 GPa, while the critical pressure for reflected shock initiation is 7 GPa.
propagation of detonation to be observed.
The present study examines the initiation of
detonation in a homogeneous liquid explosive by
reflecting a subcritical incident shock off a high
impedance plate located underneath the test explosive.
The charge is much larger than the critical diameter of
the test explosive, permitting detonation initiation and
propagation in a shock-compressed explosive to be
observed. These experiments seek to identify and
compare the critical pressure to initiate detonation in
the incident versus reflected mode of initiation, with an
ultimate goal of understanding the limits to shock
compression of homogenous energetic materials.
INTRODUCTION
Since the early work of Chaiken [1] and Campbell et
al. [2], the initiation of detonation in homogenous
explosives by a planar shock has been intensively
studied. However, the actual mechanism of initiation
in scenarios such as gap tests can be considerably more
complex, involving multiple shock interactions and
reflections. The use of a stiff anvil to reflect the
incident shock back into the test explosive, for
example, may result in initiation for conditions where
the incident shock by itself would not.
Recent experiments by Winter et al. [3] and Tarver et
al. [4] examined the reflection of shock waves
transmitted through plastic-bonded explosives off high
impedance backing plates. Initiation of reaction and
energy release without detonation as well as detonation
initiation by shock reflection were reported. As for
homogeneous explosives, attempts by Presles et al. [5]
to initiate reaction by reflecting a sub-critical shock off
an aluminum plate back into pure nitromethane did not
result in any measurable decomposition reaction upon
reflection. This was attributed to the relatively low
reflected-shock temperature (750 K). Extensive work
in recent years by Gruzdkov, Winey, and Gupta [6]
involved multiple shock reverberations between two
stiff anvils to compress nitromethane to pressures of up
to 19 GPa. The onset of reaction is observed at
temperatures above 940 K, suggesting a thermal
mechanism of initiation in nitromethane. While the
reverberating shock wave technique provides a means
to study the onset of chemical reaction in off-Hugoniot
(quasi-isentropically compressed) states, the samples
used are too small to permit the initiation and
EXPERIMENTAL DETAILS
The experimental charge configuration is similar to a
conventional gap test, with a point-initiated donor
charge of 100 mm diameter and 200 mm length
(Fig. 1).
The donor explosive is nitromethane
sensitized with 10% diethylenetriamine (DETA). The
donor explosive transmits a shock though an attenuator
of inert material (gray PVC plastic). PVC was used
rather than the more typical PMMA attenuator because
of its compatibility with nitromethane. The thickness
of the attenuator is varied to control the strength of the
shock transmitted into the test charge. The test charge
is contained in a PVC capsule of a larger diameter
(200 mm) than the donor. This configuration was used
to eliminate shock interactions with the capsule walls
that can result in initiation at anomalously low
pressures as compared to ideal shock initiation [7-9].
For the experiments examining incident initiation, the
1023
PVC
38.1 mm
Mylar window
base plate Steel for reflected shock initiation experiments
MDF for incident shock initiation experiments
FIGURE 1. Experimental charge configuration.
0
bottom of the charge was sealed with a sheet of
0.25 mm Mylar mounted on a medium density fiber
board (MDF). For experiments with reflected shocks,
the test capsule was prepared directly on a 12.7-mmthick mild steel plate. The capsule height was varied
between 12.5 and 50 mm. This range of heights was
chosen to ensure that the shock reflection from the
bottom plate occurred before the incident shock
reached the capsule side walls, so that initiation would
not occur due to interaction with the side walls and the
incident shock would not interfere with the side-on
fiber optics observing the reflection through a Mylar
window.
The luminosity generated by detonation was
observed via fiber optics connected to photodiodes.
For incident initiation, both end-on and side-on fiber
optics, mounted in brass light pipes, were used. For
reflected initiation, only side-on fiber optics were used
so as not to interfere with the shock reflection off the
bottom steel plate.
As discussed below, the
photodiodes were only sensitive enough to determine
the onset of detonation but not chemical reaction. The
arrival time of the shock at the test explosive/steel
interface was determined by a shock pin centered on
the plate. Select experiments were performed with
manganin gauges (Dynasen MN4-50-EK) embedded in
the attenuator to determine the critical incident
pressure or mounted on the steel plate to determine the
critical reflected shock pressure for initiation.
The nitromethane used for the test mixture was
commercial grade sensitized with 5% DETA by mass.
The experiments were always performed within a few
hours of mixture preparation and were fired at ambient
conditions (5-22 °C). Considerable care was taken in
5
10
15
Time (^is)
FIGURE 2. Luminosity signals for incident initiation.
final charge assembly to ensure no air bubbles were
trapped on the attenuator/test explosive interface or on
the steel plate.
RESULTS
Incident Initiation
Before experiments to examine the initiation of
detonation by shock reflection were performed, the
critical shock pressure for incident shock initiation for
this scale of experiment (100-mm-diameter donor) was
identified. The results with a 21.5-mm-thick PVC
attenuator are shown in Fig. 2. The entry of the shock
into the test mixture is time zero. Note the photodiodes
observing the charge via fiber optics detect luminosity
beginning about 1 j^s after shock entry. The side-on
fiber optics clearly show a detonation propagating
down the charge, and the end-on fibers show that
initiation occurred on the central axis of the charge.
If the attenuator thickness is increased to 35 mm, no
luminosity is observed at all and only a decaying shock
propagates through the test charge. The fact that the
photodiodes detect no luminosity in a slightly
subcritical case (where reactions are present prior to
being quenched by lateral and rear-generated
rarefactions) suggests that they are not sensitive
enough to observe shock-initiated reaction and can
1024
only be used to determine the presence of detonation.
Repeating these experiments with different thicknesses
of attenuator showed that the critical gap thickness for
incident initiation was in the range of 30-35 mm.
Based on manganin measurements of shock pressure in
the attenuator and the average shock velocity over this
range, the critical shock pressure for incident initiation
is estimated to be 4-5 GPa from impedance matching
calculations.
22.5 mm
steel plate
a) NM + 5% DETA
tI
Reflected Initiation
If the experiments described above were repeated
with a steel plate on the bottom of the test charge,
initiation of detonation could be observed upon
reflection. Shown in Fig. 3a are the photodiode and
manganin gauge traces of a subcritical shock (as
transmitted by a 42-mm-thick attenuator) propagating
through a 22.5-mm-thick capsule and then reflecting
off the steel bottom plate. The time of shock reflection
was 7 (is, as indicated by the shock pin and manganin
gauge. Within 1.5 |j,s after shock reflection, the
appearance of intense luminosity was detected by the
photodiodes. A pressure signal as measured by a
manganin gauge mounted on the center of the bottom
steel plate is also shown in Fig. 3a. The post-reflectedshock pressure of 7.0 GPa is nearly constant in
amplitude, until the appearance of a "hump" 1.2 j^s
after shock reflection. The appearance of this hump is
simultaneous with the start of luminosity as detected
by the fiber optics/photodiodes and is apparently
associated with the onset of detonation. The amplitude
of this hump (8.5 GPa) is below the CJ pressure, so it is
unclear if this is the record of the establishment of a
self-sustained detonation (retonation wave) or a result
of reaction without complete detonation.
The
manganin gauge used is also only rated to 12.5 GPa, so
the fact that the signal did not exceed this value cannot
be taken as a conclusive indication of the pressures
reached during reflected initiation. It is clear from the
photodiode traces, however, that a detonation did
propagate from the charge axis to the wall of the test
capsule after shock reflection.
Shown in Fig. 3b are manganin traces of an
experiment similar to that in Fig. 3a, except that pure
nitromethane as opposed to sensitized nitromethane
was used. The photodiode traces (not shown here) in
this case did not detect any luminosity, and the
explosive is believed to have remained inert
throughout the experiment. The signal of manganin
gauge A mounted on the steel plate can be compared to
the signal from Fig. 3a. Manganin gauge B (suspended
in the liquid 3 mm above the steel plate) clearly shows
the record of incident and, as the shock passes the
gauge a second time, reflected shock as well. The
amplitude of the incident shock (3.5 GPa) and the
reflected shock (7.0 GPa) are in good agreement with
impedance matching calculations, and also match the
pressures in the experiment with sensitized
"c
ic
•3
b) Neat NM
03
Q_
O
2
Q_
FIGURE 3. (a) Reflected shock pressure and luminosity for
reflected initiation in sensitized NM (b) Incident and reflected
shock pressure in pure (nonreacting) NM.
nitromethane in Fig. 3a. This agreement suggests that
the reflected shock in sensitized nitromethane is
initially nonreacting. The results in pure nitromethane
show a slow decay in pressure and lack the distinctive
"hump" associated with the reflected initiation of
Fig. 3a.
To prevent initiation upon reflection, one can
increase the attenuator thickness (thus lowering the
pressure of the incident shock) or increase the height of
the test capsule (thus further attenuating the shock in
the test liquid before shock reflection occurs). In this
study, both approaches were taken. Figure 4 shows the
results of approximately 20 such experiments with
shock reflection in NM+5% DETA. Experiments with
attenuators thinner than 35 mm (the shaded region in
1025
The critical pressures reported here refer only to
peak pressure. Since the use of point-initiated charges
and attenuators involving lateral and rear-generated
rarefactions will result in a shock followed by an
expansion gradient, these values of critical pressure
cannot be directly compared to square-wave loading.
Nonetheless, the fact that unambiguous initiation of
detonation in a pre-shock-compressed explosive has
been obtained provides a new means to examine the
mechanisms of shock initiation in homogeneous
explosives.
•
reflected initiation
O
no initiation
- - - - total attenuation = 70 mm
50
:..:.
ACKNOWLEDGEMENTS
30
40
50
60
The authors would like to thank Massimiliano
Romano, Leo Nikkinen, and David Hanna for their
expertise with the photodiagnostics and Oren Petel for
assistance in conducting the field trials. Steve Kacani
and Charles Dolan are thanked for timely fabrication of
the charges. The dedicated assistance of the technical
staff of the Defense Research Establishment Suffield is
generously acknowledged.
70
Attenuator Thickness (mm)
FIGURE 4. Results for shock reflection off a steel plate as a
function of the attenuator thickness and the test capsule height.
Fig. 4) will typically result in initiation on incident
shock, and therefore are not plotted on this figure.
Experiments in which luminosity traces identical to
those shown in Fig. 3a were obtained are shown as
solid symbols. Experiments in which no luminosity
was observed at all during the test time are shown as
open symbols. The reflected initiation events were
very reproducible, with the onset of luminosity
consistently occurring 1.2-1.5 ^is after shock reflection
in the case of initiation.
Since the acoustic speed, shock impedance, and
Hugoniots of nitromethane and PVC are similar, it
would be expected that the strength of the reflected
shock is determined by the total distance of shock
travel in either material prior to reflection. Indeed, a
line denoting a total propagation distance of 70 mm
appears to bound the region of "initiation upon
reflection" from "no initiation" in Fig. 4. Based on the
manganin gauge measurement shown in Fig. 3 and
additional pressure measurements made 70 mm in
PVC, the pressure of the shock in sensitized
nitromethane prior to reflection in the critical case is
3.5 GPa, giving a reflected shock pressure of 7.0 GPa.
REFERENCES
1. Chaiken, R.F, J. Chem. Phys. 33, 760-761 (1960).
2. Campbell, A. W., Davis, W. C., and Travis, J. R.,
Phys. Fluids 4, 498-510 (1961).
3. Winter, R.E., Taylor, P., and Salisbury, D.A., "Reaction
of HMX-Based Explosive Caused by Regular
Re/lection of Shocks, " llth Symp. (Int.) on Detonation,
1998, pp. 649-656.
4. Tarver, C.M., Cook, T.M,. Urtiew, PA Tao, W.C.,
"Multiple Shock Initiation of LX-17," 10th Symp. (Int.)
on Detonation, 1993, pp. 696-703.
5. Presles, H.N., Fisson, F., and Brochet, C., Acta Astro. 7,
1361-1377(1980).
6. Gruzdkov, Y.A., Winey, J.M., and Gupta, Y.M, "Use of
Time-Resolved Optical Spectroscopy to Understand
Shock-Induced Decomposition in Nitromethane," 11 th
Symp. (Int.) on Detonation, 1998, pp. 521-524.
7. Travis, J.R., "Experimental Observations of Initiation
of Nitromethane by Shock Interactions at
Discontinuities," 4 Symp. (Int.) on Detonation, 1965,
pp. 386-393.
8. Seely, L.B., Berke, J.G., and Evans, M.W., AIM J. 5,
2179-2181 (1967).
9. Jette, F.X., Yoshinaka, A.C., Romano, M.,. Higgins,
A.J., Lee, J.H.S., Zhang, E, "Investigation of Lateral
Effects on Shock Initiation of a Cylindrical Charge of
Homogeneous Nitromethane," 18th Int. Colloquium on
the Dynamics of Explosions and Reactive Systems,
Seattle, WA, 2001.
DISCUSSION
The reflection of a strong subcritical shock off a
steel plate and back into the test explosive can result in
the initiation of detonation. The results obtained here
suggest that the critical pressure for incident initiation
of detonation in NM + 5% DETA at the 100 mm scale
is 4-5 GPa, while the pressure obtained upon reflection
must be 7 GPa for initiation. The higher pressure
required for reflected initiation is likely a result of the
fact that a single shock compression results in greater
shock heating than for the case of two (or more)
successive shocks to the same final pressure.
1026