CHANGES IN THE MECHANICAL PROPERTIES OF ENERGETIC MATERIALS WITH AGING

CHANGES IN THE MECHANICAL PROPERTIES OF ENERGETIC
MATERIALS WITH AGING

Donald A. Wiegand
Energetics and Warheads Division, TACOM-ARDEC
Picatinny Arsenal, New Jersey 07806-5000




The mechanical properties of energetic materials were
studied as a function of actual age and accelerated aging
produced by heat treatment at elevated temperatures for
several months. All measurements were made in uniaxial
compression at one or more of three temperatures, i.e. –45,
25 and 65 C at strain rates between 1 and 10/sec. For TNT
that was in the stockpile for up to 40 years no change in the
compressive properties were found. Single base
nitrocellulose propellants that were subjected to 65 C for
18 months also gave no change in the compressive
strength. In addition, the compressive strength of a plastic
bonded explosive, PAX2A, were not changed after six
months at 60 C and 12 months at 50 C. In contrast, after
six months at 70 C the compressive strength and modulus
of octol were reduced by up to about 40%. The
compressive strengths of double and triple base
nitrocellulose propellants were also decreased by 50 to
70% by the same heat treatment given to the single base
propellants. These changes for octol and the propellants
are a function of measurement temperature. Other changes
in most of these material were observed and are also
discussed. Probable reasons for the changes in mechanical
properties are presented.

INTRODUCTION
changed with time in the field.1

Changes with aging of plastic bonded
Changes in performance and safety
explosives (PBX’s) and TNT based
characteristics of explosive and
explosives properties are also of
propellant formulations with aging are a
considerable interest. The effect of
continuous concern. For example, the
thermal conditioning at elevated
degradation of nitrocellulose with time is
temperatures to simulate aging on the
well known and there are very active
properties of PBX 9501 have been and
surveillance programs to monitor the
continue to be the object of studies and
properties of nitrocellulose base
changes with this type of treatment have
propellants with aging. In addition,
been found.2,3 Similar studies have been
there have been programs to determine if
made for LX-14.4,5,6 The stability of the
the properties of explosive fills have
properties of one of the newest PBX’s,

---

PAX 2A, developed recently by the
and two. The end faces of all samples
Army is also of interest.7 The
and the loading platens were lubricated
composition of PAX 2A is similar to that
to minimize frictional effects between
of PBX 9501 since it has the same
the samples and the platens. Samples
explosive and plastizer but a different
were conditioned at temperatures
polymer.2,7 The stability of TNT and
between –45, 25 and 65 C for at least
the TNT based explosive octol are also
two hours before mechanical
of interest. Therefore, the aging of both
measurements and were then
PAX2A and octol were studied
compressed along the cylinder axis to
experimentally. Because TNT has been
obtain engineering stress and
in the stockpile for many years it was
engineering strain. In most cases three
possible to obtain data as a function of
samples of PAX2A and TNT were
actual age. The work reported here on
measured at each condition, four or in
mechanical property aging is part of
most cases five samples of octol were
more general studies to determine if
measured at each condition and between
several properties of PAX 2A, TNT and
three and nine propellant samples were
octol change with aging.1,8 However,
measure for each condition.
rather than wait many years before

initiating the study of aging, the work
Samples of PAX2A were prepared
with PAX2A and octol was done by
by pressing at 71 C to the approximate
accelerating the aging processes by
size and then machining to final
conditioning samples at elevated
dimensions. Samples of octol were
temperatures but shorter periods.
prepared by casting into molds six

inches in length and one inch in
Some
gun
propellants
have
diameter, and then cutting and
compositions similar to those of plastic
machining to the final dimensions.

bonded explosives.9 Therefore, it is
Samples of TNT were obtained by
appropriate to present results on the
sectioning shells which were brought in
effect of the same type of accelerated
from the stockpile, cutting and then
aging on the mechanical properties of
machining to the final dimensions.

some of these materials.9,10
Precautions were taken to insure that the

cylinder end faces of all these samples
EXPERIMENTAL
were flat and parallel. In contrast, the

propellant samples were cut from
Stress-strain
data
in
uniaxial
propellant grains containing
compression were obtained using an
perforations. The densities of all
MTS servo-hydraulic system operated at
PAX2A and octol samples were
a constant strain rate of 1.0/sec for PAX
measured and were in a narrow range
2A, octol and TNT, while the strain rate
close to the maximum theoretical (zero
for the propellant samples was
porosity) density. A range of densities
approximately 10/sec. All samples were
was obtained for the TNT samples. The
in the form of cylinders. The samples of
densities of the propellant samples were
PAX 2A and octol were 1.90 cm in
not determined. The densities of
length and diameter and the samples of
PAX2A and octol before and after
TNT were 1.27 cm in length and
thermal conditioning but before
diameter. The propellant samples had
deformation were determined by
length to diameter ratios between one
weighing in air and measuring the
dimensions. The densities after

---

deformation of PAX2A were determined
only one group is discussed here. The
by weighing in air and in water. The
propellant samples were held at 65 C for
water density was corrected to the
varying time periods between one and
measurement temperature.
eighteen months.


The results for samples subjected to
As noted above three aged samples of
accelerated aging by conditioning at
PAX2A, four or five aged samples of
elevated temperatures, the aged samples,
octol, and the same number of control
are compared to control samples from
samples were compressed at each of
the same lot and measured under the
three temperatures, -45, 25 and 65 C. In
same conditions. The aged samples of
addition, a few samples of the aged and
PAX2A were held in air at 60 C for 6
control groups of each explosive
months or at 50 C for 12 months while
formulation were measured at
the aged samples of octol were held in
intermediate temperatures. While most
air at 70 C and relative humidity of 30%
measurements were at a strain rate of
for six months. Control samples were
1.0/sec., a few aged and control samples
held at ambient temperature in air for 6
were compressed at a strain rate of
or more months. Because the results for
0.001/sec. All of the TNT samples were
the two groups of aged samples of
compressed at 25 C and the propellant
PAX2A are very similar, only the
samples were compressed at –45 and 23
samples held at 60 C for six months are
C.
discussed here. There were also two

groups of aged and control samples of
The main components of the
octol produced by difference energetic materials considered here are
manufacturers. These two groups of
given in Table 1.
samples also gave similar results and so

TABLE 1. PRINCIPAL COMPONENTS OF ENERGETIC MATERIALS
STUDIED

PAX2A
HMX - 85%; BDNPA/F - 9%; CAB - 6%

Octol
HMX – 75%; TNT – 25%

NACO
NC – 93.61%; Ethyl Centralite – 1.15

M26
NC – 66.1%; NG - 25.8%; Ethyl Centralite – 6.35%

M30
NC - 27.61%; NG – 22.67%; NQ – 47.96%; Ethyl Centralite – 1.49%

HMX – Cyclotetramethylene Tetranitramine; BDNPA/F – Bis(2,2-
Dinitropropl)AcetaL/Formal; CAB – Cellulose Acetate Butyrate; TNT –
Trinitrotoluene; - NC Nitrocellulose; NG – Nitroglycerine; NQ – Nitroguanadine.


RESULTS
The stress-strain curves in uniaxial

compression for all of the energetic
materials considered here are as follows:

---

With increasing strain from zero, the
softening. However, in some cases the
stress initially increases linearly with
stress decreases abruptly due to cracking
strain and Young's modulus is obtained
and fracture. This abrupt decrease of the
from the slope of this linear portion of
stress occurs primarily at lower
the curve. With further increases in the
temperatures and or higher strain rates.
strain the stress curves over and passes

through a maximum. The maximum
Changes in the initial part of the
stress, the compressive strength, is taken
stress-strain curve with aging, i.e
as a measure of the failure stress and the
changes in the compressive strength,
strain at the maximum is taken as a
Young's modulus and the failure strain
measure of the failure strain. For further
are the primarily focus of this paper. In
increases in strain the stress decreases
addition, changes in the work softening
continuously. In most cases the stress
region are discussed for PAX2A and
decreases gradually in this region of the
octol. Changes in the brittleness are also
stress-strain curve due to work or strain
discussed.

PAX 2A and Octol

TABLE 2. PERCENTAGE CHANGES OF PAX2A AND OCTOL MECHAN ICAL
PROPERTIES AFTER ACCELERATED AGING

Measurement
-45
C
25
C
65
C

Temperature

PAX2A

Compressive
Unchanged*
Unchanged
Unchanged
Strength

Young's
Modulus
-10%*
15
14.

Strain
at
maximum
16%*
-5%
-17%
Stress

Octol

Compressive
-38%
-36%
-8%
Strength

Young's
Modulus
-44%
-18%
-26%

Strain
at
Maximum
13%
Unchanged
16%
Stress

Damage
Modulus
-38%
-63%
-70%

---

* Below the ductile to brittle transition temperature, which is about –30 C for PAX2A
for a strain rate of 1.0/sec., the mechanical properties of interest here are extremely
variable from sample to sample. Therefore, the values given at –45 C are very tentative
because of this variability and because the number of samples available for
measurements at each temperature was limited.

TABLE 3. OTHER CHANGES IN PAX2A AND OCTOL AFTER
ACCELERATED AGING

PAX2A

Very small decreases in average dimensions and weights < 0.5%

No change in average density within 0.5%

Change of color from white to pale yellow

Decreased stress at 65 C in the work softening strain range

Increased cracking after deformation at 65 C in the work softening strain range

Greater decrease in density due to deformation at 65 C into the work softening strain
range

Lower work/volume required for deformation at 65 C into the work softening strain range

No change in the ductile to brittle transition temperature

Octol

Small decrease in average weight, less than 0.6%

Small increase in average volume ~ 1%

Small decrease in average density ~ 1.5%

Change of color from tan to brown


In Table 2 the average percentage
in the effect of aging on the given
changes in the compressive strength,
properties of PAX2A and octol. There
Young's modulus and the failure strain
is no change in the compressive strength
are given as a function of the
of PAX2A at all three temperatures
temperature of compression for PAX2A
within experimental uncertainty, but a
and octol for the accelerated aging
decrease in the compressive strength of
conditions given above. The results of
octol at all temperatures. The results of
this table indicate significant difference
Table 2 also indicate an increase in

---

Young's modulus for PAX2A at the two
to 1.5%. There are also changes in color
higher temperatures, but a decrease in
of both explosives that indicate that
the modulus at all three temperatures for
chemical changes have taken place with
octol.
accelerated aging. These color changes

were observed on both external surfaces
In addition to the results given in
and on fracture surfaces, thus indicating
Table 2 after accelerated aging, several
changes throughout the volume.

other properties of PAX2A and octol
However, the density of chemical
were monitored with and without aging
species which are required to produce
and the results are summarized in Table
these color changes are most probably
3. Of particular interest are changes in
much smaller than the other percentage
the work softening region. For PAX2A
changes observed here.
at 65 C measuring temperature there is a

significant decrease in the stress in the
Propellants
work-softening region. In addition, at

this temperature there is an increase in
In Table 4, the compressive strengths
cracking, a larger density decrease due
of a single base, a double base and a
to deformation and a decrease in the
triple base nitrocellulose propellants are
work/volume to deform to a strain in the
given at a measuring temperatures of -
work-softening region. For octol at all
45 C and 23 C as a function of
temperatures of measurement straight
accelerated aging time at 65 C.9 Within
lines can be fitted to the stress-strain
the spread of the data at –45 C there is
curves in most of the work softening-
no change in the compressive strength of
region. Since the decrease of stress with
the single base propellant, but very
increasing strain is due to strain induced
significant decreases in the compressive
damage, the slope of this line is taken as
strengths of the double and triple base
a damage modulus, D, and the
propellants. One other single base and
percentage changes with aging are given
two other triple base propellants gave
in Table 2. Very significant decreases in
similar results to those of Table 4 at –45
D are observed. The stress-strain curves
C.9 Although the compressive strengths
for PAX2A have continuous curvature in
of the single base propellants did not
the work softening region so that a
change with aging, the nature of the
damage modulus as in the case of octol
failure did exhibit change. With
cannot be clearly defined. However,
increasing strain before the accelerated
these results for both PAX2A and octol
aging the stress decreased gradually after
suggest greater strain induced damage in
the maximum stress and then abruptly
the aged material.
fractured only at large strains. Failure

was in the form of cracking but not
The densities are also of interest. For
fracture. In contrast, after 18 months of
PAX2A there are small decreases in
accelerated aging the stress decreased
weight and dimension of less than or of
abruptly at or near the maximum stress
the order of 0.5%. These changes are
and the samples fractured. The stress as
such that there is no change in average
a function of strain for the double and
density within an experimental error of
triple base propellants at –45 C
about 0.3%. In contrast, for octol small
decreased in a moderately abrupt fashion
decreases in weight and small increases
after the maximum stress before aging,
in dimensions were found, thus giving
but decreased more abruptly after 18
decreases in densities of the order of 1%
months aging. Thus all three types of

---

propellants became more brittle after 18
also given in Table 4. The results
months of accelerated aging. The rather
suggest a small decrease in the
large experimental spread of some of the
compressive strength of the triple base
data of Table 4 at –45 C was found for
propellant, no change in the double base
propellants and also plastic bonded
propellant, and either no change or
explosives when the failure is brittle
possibly a small increase in the strength
(See comment at bottom of Table 3).
of the single base propellant. One other

single base propellant gave similar
The compressive strengths measured
results at 23 C.
at 23 C as a function of aging time are

TABLE 4. THE COMPRESSIVE STRENGTH OF PROPELLANTS IN MPa
AFTER ACCELERATED AGING FOR SEVERAL PERIODS

Measured At -45 C

Aging Period
Single Base
Double Base
Triple Base
Months NACO
M26
M30

0 217
+/-17
239
+/-41
234
+/-32

6 232
+/-18
154
+/-44
172
+/-47

12
201
+/-12
76
+/-21 177
+/-23

18
230
+/-32
70
+/-18 109
+/-44

Measured At 23 C

0 117
+/-1 61
+/-10 82
+/-5

18
134
+/-14
70
+/-4
59
+/-6


Between 5 and 9 samples were measured at each aging period for each propellant at –45
C and 3 samples were measured at each aging period for each propellant at 23 C. The +/-
is the standard deviation of the measurements in each case.

.
TNT
modulus and the failure strain. In

addition, the mode of failure did not
Measurements of the compressive
appear to change with age. Over 100
mechanical properties of samples of
samples were measured and these were
TNT taken from shells that had been in
taken from a number of shells of three
the stockpile for 20, 30 and 40 years
different diameters. The average
indicates no statistically significant
compressive strength of all samples
change of the compressive strength with
measured is 13.6 +/- 2.3 Mpa. This
age. The same is true for Young's
number can be compared with the

---

average compressive strength of 12.8 +/-
region of the strain-strain curve at higher
1.2 Mpa for a group of TNT samples
temperatures, e.g. 65 C. There is a
prepared for laboratory studies. This
marked decrease of the stress at strains
latter group of samples was measured
in this region and so a smaller
within one year of preparation.
work/volume to deform the material in

this region. Thus, for a constant energy
DISCUSSION
of deformation, e.g., by impact, the final

strain and so the degree of cracking (see
A discussion of each energetic
below) will be greater in the aged
material or group of materials is
material than non-aged material at 65 C.
followed by a brief discussion of the
There also is a greater density decrease
relative stability of the various materials
due to deformation of the aged material
against changes with age.
than the non-aged material for equal

amounts of strain into this region of the
PAX2A
stress-strain curve. This density

decrease is accompanied by increased
The compressive strength of PAX2A
cracking in the aged samples. This
was found to be stable for the
increased cracking and so larger density
temperatures and time periods of
decrease can be attributed to thermally
accelerated aging and the conditions of
activated crack growth in the aged
measurement used in this study (see
material which does not occur in the
Table 2). However, Young's modulus
non-aged material.12,13 The crack
was found to increases somewhat (~15
growth in the non-aged or reference
%) and the failure strain was found to
samples was found to be approximately
decrease somewhat for the two higher
independent of temperature while the
temperatures of Table 2. For no change
additional crack growth in the aged
in the compressive strength but an
samples increases with temperature such
increase in Young's modulus, there must
that the additional crack growth is not
be a decrease in the failure strain as
observed at 25 C but is observed at 65 C.
observed if there are no other significant
There are other differences in the
changes in the shape of the stress-strain
properties of this additional crack
curve in this region. Because there is an
growth at higher temperatures and this
increase in the modulus but no change in
whole matter is discussed in greater
the compressive strength, the
detail elsewhere.12 The increased
proportionality between the compressive
cracking in the aged samples could be
strength and the modulus found
associated with clusters of defects, e.g.,
previously is not valid for these
clusters of plasticizer molecules, formed
experimental conditions.12 The increase
during the aging process.
in the modulus may be due to a

segregation of plasticizer with aging. A
Because the kinetics and associated
more complete discussion of the
parameters are not established, the
properties of PAX2A with this
stability under other conditions of aging
accelerated aging is given elsewhere.12,13
can not be determined at this time.

However, because there are two aging
The most significant changes in the
periods at two temperatures, it is
mechanical properties of PAX2A due to
possible to obtain an activation energy
this aging occur in the work softening
by assuming first order kinetics. In this
manner a activation energy of about 0.6

---

ev is obtained. This value is close to the
before aging, E and σ are the values for
value obtained using first order kinetics
an increase in porosity ∆P, and α and β
for PBX 9501.3 Since both materials
are constants. ∆P is taken as the
have HMX and BDNPA/F, the similarity
fractional decrease in density. α is not
of activation energies suggest that it is
equal to β because the averaging
associated with one or both of these
processes are different for the strength
components. This activation energy
and the modulus. Assuming that all of
may then be associated with the
diffusion of NO
the decreases of E and σ after aging are
2 molecules as proposed
for PBX 9501.3
due to porosity, values of α in the range

of 13 to 37 and values of β in the range
Octol
of about 8 to 30 are obtained for a

porosity change of 1.5%. While these
In contrast to the results for PAX2A,
values of α and β are large compared to
for octol both the compressive strength
the predictions of calculations, values of
and Young's modulus are decreased and
this magnitude have been reported for
the failure strain is either increased or
Composition B (39.5% TNT, 59.5
not changed for all three measurement
RDX% and 1% Wax) and other
temperatures after accelerated aging
materials.15 Therefore, the changes of
(Table 2). However, the octol samples
E, the modulus, and σ, the compressive
were aged at 70 C and so 10 C higher
strength, at all three measuring
than the highest temperature for PAX2A
temperatures after aging can be
sample aging. This difference is
attributed to porosity introduced during
discussed further below.
the aging process. However, the low

percentage decrease of the compressive
A decrease in both the compressive
strength at 65 C and so the small value
strength and the modulus is expected if
of β suggests that probably the failure
the porosity is increased by the aging
process has changed somewhat at this
process.15 However, if some or all of
temperature. The ratio of the percentage
the porosity is in the form of micro-
change of the modulus to the percentage
cracks, calculations indicate that the
change of the compressive strength is
compressive strength could increase or
much higher at 65 C than at the other
decrease but the modulus is still
two temperatures. Perhaps the role of
decreased.15 For specific porosity not in
porosity in the form of micro-cracks is
the form of micro-cracks calculations
different at the higher temperature.
indicate that the relationships between

the strength, the elastic modulus and
TNT and HMX are not expected to
porosity are approximately of the form
have any significant thermal

decomposition at 70 C for 6 months. In

E = E1 exp(-α∆P)
(1)
addition, the measurements given here

show that the compressive mechanical
and


properties of TNT are not changed with

age up to forty years. This result also
σ = σ
suggests that TNT does not significantly
1 exp(-β∆P)
(2)

decompose during actual storage
E
conditions. However, chemical analysis
1 and σ1 are Young's modulus and the
compressive strength at a porosity P
indicates an increase in the percentage of

---

HMX in octol after treatment at 70 C for
aging at 23 C the double and triple base
two and four months.16 Although the
propellants of

percentages of TNT and impurities were
Table 4 and other triple base propellants
not measured, this increase in the
have lower compressive strengths than
percentage of HMX indicates that either
the single base propellant of Table 4 and
the percentage of TNT and or the
other single base propellants.9 From
percentage of impurities has decreased.
these observations it is concluded that
Therefore, the observed small weight
the properties of nitrocellulose primarily
loss could be due to the loss of TNT and
determine the strength of the single base
or impurities. The effect of these loses
propellants, but the lower values of
on the observed volume expansion and
strength for the double and triple base
density decrease are not known.

propellants at 23 C are due to the
Impurities could decompose and cause
plasticizing effect of nitroglycerine. It is
internal strain. Impurities could also
also concluded that the addition of a
diffuse to the surface and either stay on
relatively large percentage of solids in
the surface or escape. Thus the small
the form of nitroguanidine has little or
volume expansion and small density
no effect on the compressive strengths at
decrease could be associated with
23 and –45 C.
impurities. Sublimation of TNT might

also take place since the aging
It is also important to know if these
temperature of 70 C is close to the
propellants are primarily in a glass or a
melting temperature of TNT at 81 C.
rubber state at –45 and 23 C as a

function of aging time. While
In summary, the decreases of the
measurements have not been made to
compressive strength and Young's
clearly determine this as a function of
modulus can be attributed to porosity
aging, some measurements have been
increases during the accelerated aging.
made on relatively freshly prepared
The observed volume expansion and
propellant. A ductile to brittle transition
density decrease support this was found to take place in the non-aged
interpretation. A decrease in the
triple base propellant (M-30) between 0
percentage of TNT and or impurities and
and –15 C for a strain rate of 10/sec., the
the observed weight decrease indicate
same strain rate as used here.17
that TNT and or impurities were lost
Therefore, it appears that this propellant
during the aging. The porosity increase
is in a glass state and a brittle state at –
may be associated with this loss.
45 C but in rubber state at 23 C before

aging. While a glass transition
Propellants
temperature of –57 C was observed for

this propellant, this was measured at a
Before
aging
the
compressive
much lower strain rate.17 The glass
strengths of the single, double and triple
transition temperature increases with
base propellants of Table 4 and other
rate. Because of the type of failure and
single and triple base propellants at –45
the composition, the double base
C are the same within experimental
propellant is most likely also in the glass
uncertainty.9 It is therefore concluded
state at –45 C and in the rubber state at
that the compressive strengths of these
23 C. For the single base propellant
propellants at this temperature are
before aging the situation is not clear.
determined primarily by the properties
At –45 C the failure mode is partly
of nitrocellulose. In contrast, before
ductile and partly brittle. Therefore, at

---