Shock initiation of powder mixtures of

aluminum with dense metal oxides

Cite as: AIP Conference Proceedings 1426, 279 (2012); https://doi.org/10.1063/1.3686273
Published Online: 29 March 2012

Francois-Xavier Jetté, Sam Goroshin, David Frost, et al.

ARTICLES YOU MAY BE INTERESTED IN

A thermochemical model for shock-induced reactions (heat detonations) in solids

The Journal of Chemical Physics 92, 1839 (1990); https://doi.org/10.1063/1.458066

Impact-induced initiation and energy release behavior of reactive materials

Journal of Applied Physics 110, 074904 (2011); https://doi.org/10.1063/1.3644974

In-situ measurements of the onset of bulk exothermicity in shock initiation of reactive powder
mixtures
Journal of Applied Physics 109, 084905 (2011); https://doi.org/10.1063/1.3553861

AIP Conference Proceedings 1426, 279 (2012); https://doi.org/10.1063/1.3686273 1426, 279

© 2012 American Institute of Physics.

 SHOCK INITIATION OF POWDER MIXTURES OF ALUMINUM
WITH DENSE METAL OXIDES

F.X. Jetté1, S. Goroshin1, D.L. Frost1, and F. Zhang2

1
McGill University, Dept. of Mech. Eng., Montreal, Canada
2
DRDC-Suffield, Ralston Canada

Abstract. Strong and dense structural reactive materials may be produced by mixing aluminum
powders with heavy metal-oxide powders (such as Bi2O3, PbO, Pb3O4, I2O5, etc.). The addition of
certain additives to such mixtures, such as V2O5 and B2O3, can lower the softening point of the oxide
mixture below the melting point of aluminum. This could lead to the fabrication of dense and non-
porous aluminum-metal oxide structural materials. The shock sensitivity of aluminum-metal oxide
mixtures was investigated in this work. The minimum shock initiating pressure was obtained for
various porous and non-porous aluminum-metal oxide mixtures using the shock recovery technique.
Since most reactions of Al in metal oxide mixtures produce little pressure and material velocity
changes but large increases in temperatures, thermocouples were used to observe the bulk reaction
onset, which relates to the overall reaction rate, in those mixtures. The mixtures tested were found to
be very sensitive to shock initiation and their reaction rates were found to be very fast, compared to
other types of reactive powder mixtures. Finally, the addition V2O5 and B2O3 additives or the addition
of liquid heptane (to fill the pores) did not significantly lower the sensitivity or reaction rates of the
mixtures investigated.
Keywords: shock initiation, dense glass, thermocouple, recovery capsules, thermite
PACS: 47.40.Nm, 62.50.+p, 82.40.Fp.

INTRODUCTION viscosity becomes small near 700ºC) can be

In order to achieve a dense reactive structural
material, it is proposed to combine aluminum with
a mixture of dense metal oxides, such as Bi2O3 and
PbO. Such reactive powder mixtures are dense,
highly exothermic, and as will be shown in this
study, very sensitive to shock initiation.
The density of those powder mixtures can be
further increased if the pores are filled with a
liquid, such as liquid heptane. It may be possible to
prepare even denser samples via the melt-casting
method described in [1], provided the metal oxide
mixture can be melted without thermally initiating
reactions with the aluminum. Metal-oxide mixtures
whose softening point is below 300ºC (and whose
prepared by adding small amounts of V2O5 and
B2O3 to mixtures of dense oxides. The thermal
initiation sensitivity can also be decreased when
nickel-clad aluminum is used instead of pure
aluminum.
The objective of the current project was to
determine how sensitive to shock initiation various
powder mixtures of aluminum and metal oxide are,
as well as the rate at which shock-initiated
reactions proceed. To this end, the initiation of
reaction was monitored in samples contained in
instrumented steel recovery capsules and shocked
using a charge of high explosive detonated in close
proximity with the steel capsule.

Shock Compression of Condensed Matter - 2011

AIP Conf. Proc. 1426, 279-282 (2012); doi: 10.1063/1.3686273
© 2012 American Institute of Physics 978-0-7354-1006-0/$0.00

279
 EXPERIMENTAL PROCEDURE
The experiments were conducted using the
cylindrical configuration shown in Fig. 1. The
powder samples were hand-compacted into a
25.4 mm diameter cavity (10 mm deep) in a mild
steel recovery capsule (the bottom portion of the
capsule was 50.8 mm in diameter, while the cover
was 63.5 mm), similar to that of [3].
A strong shock wave was delivered to the
sample by detonating a charge of high explosives
in close contact with the test capsule. A shock-
attenuating disc of PMMA of varying thickness
was placed between the explosive and the capsule
in order to control the shock pressure – the shock is
strongest when this PMMA layer is thinnest. The
explosive charge was 454 g pentolite booster
(Orica Pentex AP, 50.8 mm in diameter) with 20 g
of C4 that provided a smooth surface between the
PMMA and the booster.
To prevent its motion, the test capsule was
held rigidly in place in a heavy steel block/anvil
using three strong set screws. After each
experiment the sample remained sealed within the
capsule, which was later cut open to allow post-
shock analysis of the sample.
Due to the relatively small volume of gases
produced by reactions in the mixtures of interest,
exothermic chemical reactions are typically
characterized by small changes in pressure and
particle velocity, large changes in temperature, and
intense luminosity. Hence, temperature or light-
measuring techniques provide the most direct
means to determine the onset of reaction.
Jetté et al. [2] showed that although thermo-
couples made from wires with a diameter of
0.8 mm (wire gauge AWG 20, Omega part #: 5TC-
TT-K-20-36) cannot be relied upon to measure
actual bulk temperature on short time scales, they
were appropriate to measure the onset of bulk
temperature changes on time scales of the order of
1.8 ms (which is the time required for the centre of
the junction to “feel” a temperature change at the
surface, estimated based on an analytically-
determined Fourier number of 0.028 and an
approximate value of the heat diffusivity of the
junction). Bulk reaction onset times presented in
the results section were generally longer than this
resolution. Thus the thermocouple technique is a
simple, reliable, robust, and accurate method to
280
obtain reaction rate information for various types
of reactive samples shocked in recovery capsules.
In each experiment, two thermocouples were
inserted into the sample cavity from the bottom of
the capsule and the holes for the wires were sealed
with steel epoxy, which also served to hold the
thermocouples in place. The 3.2 mm diameter
holes for the thermocouples were located 8.89 mm
from the centre at 120º from each other. The top
view of the sample cavity in Fig. 1 shows the
location of the thermocouples: one thermocouple
was located exactly at the centre of the sample,
whereas the other one was located approximately
8.7 mm from the centre. Both thermocouples were
laid against the steel at the bottom of the sample, as
shown in Fig. 1 to facilitate the compacting of the
powders, the thermocouples were placed at the
bottom of the test sample. Weight (small steel
beads) and steel epoxy were added in the cavity
under the capsules to hold the thermocouple wires
more solidly in place.
All electric signals were amplified (gain=100,
Omega OMNI AMP III amplifier) and recorded on
a 400 MHz bandwith digital oscilloscope. The
temperature rise resulting from the initial shock in
the thermocouple signals was used to trigger the
digital oscilloscopes.
The mixtures tested consisted of the following:
Mixture 1: Bi2O3 + aluminum
Mixture 2: PbO + aluminum
Mixture 3: oxide mix + aluminum
Mixture 4: oxide mix + aluminum + liquid heptane
Mixture 5: oxide mix + HAI Ni-Al
Mixture 6: oxide mix + Phelly Ni-Al
The oxide mix was constituted of the
following mass fractions: 61% PbO, 33% Bi2O3,
4% B2O3, and 2% V2O3. All oxide powders were
acquired through Alfa Aesar. Spherical Valimet H-
5 aluminum powder was used in mixtures 1-4. For
mixture 5, nickel-clad aluminum powders contain-
ing 2% nickel were manufactured at HAI Ad-
vanced Materials Specialists Inc. by depositing Ni
on -325 Al powders. For mixture 6, Phelly Mater-
ials Inc’s TS-4 Ni-clad Al powders (30%
aluminum) were used (-140+325 mesh). The
amount of aluminum was adjusted to stoichio-
metric proportions (assuming full conversion to
Al2O3), except for mixture 6, where only 75% of
the aluminum required to make a stoichiometric
mixture was used.
 The powder mixtures were prepared by mixing
the component powders in a bottle which was then
rolled with a roller mill for several hours (>4 hrs)
to uniformly mix the particles. The mixtures were
then compacted into the test capsules in increments
of approximately 2 g. For mixture 4, heptane was
added after each powder increment (resulting in a
15% heptane mass fraction). Densities were
calculated based on the empty volume of the
sample cavity in the steel capsule and the mass of
powder used in the experiment (note that the
thermocouples volume in the sample cavity was
not accounted for, so actual densities were slightly
greater than those presented in the table below).
After each experiment the sample was either
found to have reacted entirely to completion or
there was no evidence of bulk reaction within the
sample. If the sample reacted, the steel capsule was
too hot to handle by hand after the experiment,
while if it did not react, the capsule was only
slightly warmer than room temperature. After all
experiments, the capsule was cut open and visual
inspection of the sample was also used to
determine whether the sample was a compacted
powder or a solid reaction product.
Top view
thermocouple
detonator junctions
50.8 mm
pentolite
sample
steel
anvil
thermocouple
wire
Figure 1. Experimental setup. The capsule cover
thickness above the sample is 10mm.
RESULTS
Typical temperature signals are presented in
Figs. 2-4. In each figure, the shock wave caused an
initial temperature jump on the order of 50-200 ºC
when it reached the location of the thermocouples.
This temperature increase is in accordance with the
281
expected bulk shock heating for the shock
pressures and sample densities tested. It should be
noted however that the thermocouple technique is
not ideal for precisely measuring shock heating
since the temperature indicated by the
thermocouple at the instant the shock traverses it is
not dependent only on the temperature of the
medium it is in [4].
In experiments where the sample did not react
(Fig. 2), the bulk temperature decayed slowly
following the initial temperature jump, whereas in
experiments where the sample reacted exother-
mically (as in Fig. 3 and 4), the temperature signals
increased once again after a certain delay on the
order of a few ms. During this second rise, the
signals typically increased more slowly and
showed some noise (more noise than in samples
that yield non-gaseous products [2], which could
be due to the imperfect thermal contact between
the thermocouple and the products). While this
technique cannot be used to measure the exact
reaction time or peak temperature, the time of the
post-shock rise in temperature can provide an
indication of the timescale for the onset of bulk
exothermicity sufficient to heat the sample.
Figure 2. Temperature histories for Al + oxide mix
(46% TMD) with 53.8 mm PMMA (shot TC-78).
Figure 3. Temperature histories for Phelly Ni-Al + oxide
mix (48% TMD) with 17.8 mm PMMA (shot NG-2).
 DISCUSSION AND CONCLUSIONS

Measurements of bulk temperature rises in

shock-initiated exothermic powder mixtures using
embedded thermocouples revealed that the delay
between arrival of the shock wave and onset of
bulk temperature rise was on the order of a few
10’s to a few 100’s of ms when the Ni-clad
aluminum powder containing 70% nickel was used,
but less than 10 ms with all other Al or Ni-Al
Figure 4. Temperature histories for HAI Ni-Al + oxide powders tested. Therefore, reaction rates are fastest
mix (54% TMD) with 25.5 mm PMMA (shot NG-11). when the amount of Ni is very small. This should
be expected since flame temperature (and burning
The delay times between shock arrival and velocity) is smallest when much Ni is present. The
temperature rise resulting from bulk reaction in the addition of liquid heptane also caused a slight
sample are compiled in table 1 for the two gauges increase in delay time.
used in each experiment. It is interesting to note The critical shock strength (or PMMA
that those onset times do not show a strong attenuator thickness) was found to be independent
dependence on incident shock pressure (or PMMA of the particular oxide mixture employed, but the
layer thickness). addition of a Ni coating on the aluminum powder
had a significant de-sensitizing effect, especially
TABLE 1 Summary of Delay Times when the coating was thick as in the Phelly
PMMA powders. The addition of liquid heptane also had a
%
layer τ1a τ2
Mix tested small de-sensitizing effect. Thus, initiation was
TMD (ms) (ms)
(mm) partially inhibited when no pores were present and
Al + 53.8 34.6 1.9c 2.2 when the reactants were separated by an inert layer.
Bi2O3 48.6 53.9 no reaction
44.6 28.1 2.4c 2.4 ACKNOWLEDGEMENTS
Al +
42.7 35.8 2.1c 2.1
PbO Rick Guilbeault, Alexandre Coderre-Chabot
43.7 53.9 no reaction
and Jean-Frédéric Ruel are thanked for their
47.7 37.5 3.8c 6.3
Al + assistance with the experiments.
45.6 44.0 1.1 4.2c
oxide mix
45.9 53.8 no reaction REFERENCES
Al + N/A 24.8 4.6 6.9c
oxide mix N/A 44.1 no reaction 1. Jetté, F. X., et al., “Experimental Investigation of
+ heptane N/A 53.7 no reaction Gasless Detonation in Metal-Sulfur Compositions,”
54.1 25.5 1.9c 2.0 Combust. Explo. Shock 45, pp. 211-217 (2009).
HAI 2. Jetté, F. X., et al., “In-Situ Measurements of the
50.1 25.5 2.6 3.9c
Ni-Al + Onset of Bulk Exothermicity in Shock Initiation of
48.3 34.5 no reaction Reactive Powder Mixtures,” J. Appl. Phys. 109,
oxide mix
47.7 51.5 no reaction 084905 (2011).
52.1 17.7 <5 3. Jetté, F. X., et al., “Shock Reactivity of Non-Porous
Phelly 48.1 17.8 80c 258 Mixtures of Manganese and Sulfur,” AIP Conf.
Ni-Al + 54.5 20.0 146 149c Proc. 955, pp. 1033-1036 (2007).
oxide mix 51.8 22.7 no reaction 4. Bloomquist, D.D., Duvall, G.E., “Electrical Re-
sponse of a Bimetallic junction to Shock Compres-
51.3 25.4 no reaction sion,” J. App. Phys. 50 pp. 4838-4846 (1979).
a
τ1 is the earliest bulk reaction onset time, τ2 is the latest
c
these gauges were located at the centre of the sample

282