Combustion, Explosion, and Shock Waves, Vol. 38, No. 4, pp. 463 469, 2002
Performance of Emulsion Explosives
S. Cudzilo,1 P. Kohlicek,2 UDC 662.215
V. A. Trzcinski,1 and S. Zeman2
Translated from Fizika Goreniya i Vzryva, Vol. 38, No. 4, pp. 95 102, July August, 2002.
Original article submitted September 18, 2001.
Some performance of a number of emulsion explosives containing glass micro-baloons
were studied experimentally and theoretically. For each of the explosives, detona-
tion velocity was measured and calculated and ballistic mortar tests and cylinder
expansion tests were carried out. The results obtained enables a comparison of the
usefulness of both testing methods. The influence of some metal nitrates contained in
the emulsion matrix on the performance and detonation parameters of the explosives
was examined.
Key words: explosives, glass micro-balloons, nitrates of metals, detonation param-
eters, performance.
INTRODUCTION tion behavior is almost ideal, unlike other ammonium
nitrate based composite fuels. This is manifested, for
example, in the fact that the experimental values of det-
Emulsion explosives belong to commercial explo-
onation parameters of emulsion explosives measured at
sives. They were developed in the early 1960s [1, 2] and
large charge diameters nearly coincide with theoretical
have become quite important blasting means because
values obtained under the assumption of complete am-
of their comparatively high detonation parameters and
monium nitrate decomposition at the detonation front.
exceptionally good safety characteristics [3 6].
In the work described here, we studied the deto-
Emulsion explosives were prepared by agitation of
nation and performance properties of five emulsion ex-
a supersaturated aqueous solution of some oxidizers and
plosives containing glass micro-balloons. The emulsion
a fuel doped with an emulsifying agent. The composi-
matrix was prepared using aqueous solutions of pure
tion thus prepared is extremely insensitive to initiation,
ammonium nitrate or its mixture with sodium, calcium,
and, hence, it has to be sensitized by appropriate addi-
nickel, and cobalt nitrates. The contents of other ingre-
tives (usually, polymeric or glass micro-balloons). The
dients were changed insignificantly in order to maintain
basic oxidizer used in emulsion explosives is ammonium
the oxygen balance within the range of Ä…2%. For each
nitrate (AN), and sodium and calcium nitrates are fre-
explosive, the detonation velocity was measured and
quently added to modify properties of the oxidizer so-
calculated, and ballistic mortar and cylinder expansion
lution. The fuel phase consists of various mineral oils,
tests were carried out. The results obtained enabled a
waxes, and (in some cases) polymers.
comparison of the efficiency of both methods for testing
A distinguishing feature of emulsion explosives is
explosive performance. They were also used to analyze
that in the aqueous solution, the oxidizer is present in
the effect of metal nitrates on the performance of the
the form of tiny droplets covered with a very thin fuel
explosives and to choose the formulations characterized
layer. Consequently, the interfacial surface of the emul-
by the best detonation and performance characteristics.
sions is very large, as a result of which their detona-
1
Department of Explosives and Physics of Explosion,
Military Technological University, Warsaw, Poland;
[e-mail].
2
Department of Theory and Technology of Explosives,
University of Pardubice, Pardubice, Czech Republic.
0010-5082/02/3804-0463 $27.00 © 2002 Plenum Publishing Corporation 463
464 Cudzilo, Kohlicek, Trzcinski, and Zeman
1. TESTED EXPLOSIVES
The compositions of the emulsion matrices are
given in Table 1. The emulsions were prepared using
a simple facility consisting of a thermostat and a con-
tainer equipped with a stirrer. The solution of oxidiz-
ers was heated to 120ć%C and then slowly added to the
container, in which a preheated (95ć%C) mixture of fuels
(oil and polybutadiene) with the emulsifier was agitated
with the stirrer at a speed of 900 rpm. After adding the
entire amount of the oxidizer, the agitation was contin-
ued for about two minutes to obtain fine particles of the
emulsion.
The final explosive mixtures were prepared by mix-
ing the emulsion matrices with glass micro-spheres (pro-
duced by 3M; mean size H"80 µm). Mixing was per-
formed carefully (using the same stirrer but at lower
speed) until the distribution of the micro-spheres in
the emulsion matrix became uniform. In all exper-
iments, the mass fraction of micro-spheres was 3%
Fig. 1. Relative performance (R) versus the param-
(above 100%). The initial density of the explosives was
eter Á0D2.
determined from the weight and volume of samples in
plastic or copper tubes used in detonation velocity mea-
surements and in cylinder expansion tests.
Theoretical calculations of detonation parameters
(Dideal, [AN]inert, and ng) were performed using the
thermodynamic code CHEETAN [8] with the following
2. BALLISTIC MORTAR TESTS coefficients of the BKW equation of state recommended
in [9]: Ä… = 0.5, ² = 0.298, ć = 10.5, and ¸ = 6620.
The relative performance of the explosives was These values were optimized using a very large product
measured by the procedure described in [7]. Ten- database and measured Chapman Jouguet properties
gram samples placed in thin tubes of 24 mm diameter compiled by Hobbs and Baer [9]. The basis for deter-
were detonated with a cap inside the explosive chamber mining the amount of ammonium nitrate that did not
(283 cm3 in volume) of the mortar. The magnitude of reacted in the detonation wave was the conformity of
the pendulum deflection was recorded and used to ob- experimental and theoretical values of detonation ve-
tain (from a calibration curve) the relative performance locity.
defined as the ratio of the mass of the blasting charge An analysis of the data in Table 2 shows that
of the tested explosive to the mass of dynamite that the formulation containing only AN as the oxidizer
produces the same pendulum deflection. Values of the has the highest performance and detonation parame-
relative performance obtained in this way are listed in ters. Partial substitution of nickel nitrate (AN NiN)
Table 2, which also gives experimental values of density or sodium nitrate (AN NaN) for ammonium nitrate
of (Á0), detonation velocity (Dexp), and some calculated causes a substantial decrease in performance. This can
characteristic for the explosives (Dideal is the ideal det- be explained by the fact that these explosives generate
onation velocity, [AN]inert is the fraction of ammonium smaller amounts of gaseous detonation products.
nitrate that remained inert in a detonation wave, and ng The data of Fig. 1 show how the introduction of
is the number of moles of gaseous detonation products). metal nitrates into the emulsion matrices affects both
The detonation velocity was measured by ioniza- the detonation and performance characteristics. From
tion sensors. The explosive charges were confined in the results presented in the figure it follows that the use
plastic tubes 250 mm long with an inner diameter of of calcium nitrate (AN CaN) as an additional oxidizer
32 mm. In each charge there were measuring courses deteriorates the parameters considered most insignifi-
50 mm long, and the first sensor was at 90 mm from cantly, whereas the addition of sodium nitrate (AN
the booster. These values were large enough to enure a NaN) to the mixture gives an explosive with compara-
steady detonation and determination of the detonation tively high detonation parameters but its performance
velocity with an accuracy of H"1%. is rather poor. The absence of correlation between
Performance of Emulsion Explosives 465
TABLE 1
Composition of Emulsion Explosives
Mass fraction of components, %
NH4NO3 H2O Emulsifier Mineral Polybu- Ca(NO3)2 Ni(NO3)2 Co(NO3)2 NaNO3
Explosive
(AN) oil tadiene · 4H2O · 6H2O · 6H2O
(CaN) (NiN) (CoN) (NaN)
AN" 84.0 10.0 1.9 3.6 0.5    
AN"" 82.0 12.0 2.2 3.3 0.5    
AN CaN 65.6 13.0 2.0 3.6 0.8 15.0   
AN NiN 67.3 12.0 2.0 3.4 0.3  15.0  
AN CoN 67.3 12.0 2.0 3.4 0.3   15.0 
AN NaN 65.7 13.0 2.0 3.6 0.7    15.0
Notes. Here and below AN" is the composition used only in ballistic mortar tests and AN"" is the composition used only
in cylinder expansion tests.
TABLE 2
Characteristics of Explosive in Ballistic Mortar Tests
Relative Á0, Dexp, Dideal, [AN]inert, ng,
Explosive
performance, % kg/m3 m/sec m/sec % mole/kg
AN" 74 1051 5100 5924 25.8 34.43
AN CaN 73 1073 4800 5716 33.2 31.78
AN NiN 65 1109 4410 5712 47.5 28.32
AN CoN 72 1063 4560 5559 36.8 31.24
AN NaN 67 1077 4740 5575 36.6 29.04
the relative performance and the parameter Á0D2 in- characteristics of explosives. For example, from the de-
dicates that there are other factors influencing these pendence of the radial displacement of the external tube
characteristics. One of these can stem from the fact wall on the axis coordinate, it is possible to determine
that the relative performance and detonation velocity the expansion velocity of the shell and then the Gurney
were measured for different samples and under differ- energy and detonation energy of the tested explosive
ent conditions. Therefore, we conducted cylinder ex- [15, 16].
pansion tests to determine the performance of the ex- In the present work, the cylinder expansion test was
plosive using some energy parameters obtained during employed to obtain some quantitative energy parame-
these tests [10]. ters of the emulsion explosives. Their experimental and
calculated characteristics are shown in Table 3. Here
the values of Dexp are much larger than those in Ta-
3. CYLINDER EXPANSION TEST
ble 2. This indicates that in cylinder expansion tests,
more ammonium nitrate reacts in the detonation wave
The so-called cylinder expansion test consists of
and there is better agreement between measured and
recording a chosen stage of acceleration of a shell driven
theoretical detonation velocities. A comparison of the
by expanding products of detonation grazing along the
data of Table 2 and 3 shows that the performance of
internal surface of the shell (copper tube). Expansion
the charge shell has a strong effect on detonation pa-
of the detonation products can be recorded by methods
rameters, which is manifested in nonideal detonation
used to measure fast processes, such as streak photogra-
properties.
phy, laser interferometry or flash radiography. Results
A notable feature of the quantity ng is that in cylin-
of recording the acceleration process are commonly em-
der expansion tests, compositions containing sodium
ployed to derive a semi-empirical equation of state of
and nickel nitrates yield a smaller amount of gaseous
detonation products [11 14]. Moreover, cylinder test
detonation products. However, the difference between
data can be used to determine some other important
466 Cudzilo, Kohlicek, Trzcinski, and Zeman
TABLE 3
Experimental and Calculated Characteristics of Explosive
Used in Cylinder Expansion Tests
Á0, Dexp, Dideal, [AN]inert, ng,
Explosive
kg/m3 m/sec m/sec % mole/kg
AN"" 1062 5460 5973 17.0 37.47
AN CaN 1065 5095 5680 22.5 34.78
AN NiN 1146 5240 5872 29.3 33.54
AN CoN 1093 5200 5687 21.4 35.58
AN NaN 1079 5040 5584 27.0 31.62
3.1. Accelerating Ability of Detonation
Products  Gurney Energy
One of the parameters that can characterize explo-
sive performance is the accelerating ability of detona-
tion products. Souers and Kury [17] proposed to eval-
uate this parameter using velocities of an accelerated
copper tube under cylinder test conditions at volume
expansion ratios of 2.2, 4.1, and 6.5. The accelerating
ability can also be described as the Gurney energy EG,
Fig. 2. Radiograph of a copper tube driven by detona-
which is the sum of the kinetic energy of a driven liner
tion products of the composition AN"".
and detonation products following it [18]. In the case
of a cylindrical liner, the Gurney energy is given by the
them and the other explosives is not as large as that in
equation
ballistic mortar tests (see Table 2).
The expansion of a copper tube driven by detona- 1 u2
L
EG = µ + , (1)
tion products was recorded by a SCANDIFLASH x-ray
2 2
setup [10, 13]. The copper tube filled with the ex- where µ is the ratio of tube mass to explosive mass and
plosive tested was placed at a distance of 2.7 m from
uL is the total velocity of the tube.
the x-ray source and at 0.5 m from the recording film.
Usually, in the literature one can find final values
The charge was fixed vertically, and the line connecting
of the Gurney energy or Gurney velocity equal to the
the x-ray source and the film was perpendicular to the
square root of the doubled Gurney energy. The pa-
charge axis. The copper tubes were 300 mm long, and
rameters correspond to the velocity of a tube with ra-
their inner diameter and wall thickness were 25 mm
dius for which integrity of the tube material is still pre-
and 2.5 mm, respectively. The sensor triggering the
served [18]. Trzcinski [15] showed that the accelerating
x-ray pulse was at 20 mm from the end of the charge.
ability of an explosive can be estimated using the de-
This allowed us to observe continuously the expansion
pendence of Gurney energy [defined by relation (1)] on
of detonation products for H"40 µsec until the detona- the relative volume of expanding detonation products.
tion wave reached the charge section in which the sensor
From Eq. (1) it follows that to obtain this dependence,
was placed.
it is necessary to know the total velocity of the tube
Figure 2 gives a radiograph of a copper tube driven
(uL) driven by detonation products.
by detonation products of the composition AN"", which
A detailed description of the procedure of deter-
was used only in cylinder expansion tests. From the
mining tube velocity using the dependence of the posi-
photograph, the dependence of the outer radius of the
tion of the external surface of the tube on axial coor-
tube on the axial coordinate was determined using com- dinate recorded in a cylinder test is given in [15]. The
puter processing. The examined range of the axial co- position of the central cylindrical surface (rm) was first
ordinate was limited to the value for which the volume
determined. Under the assumption of incompressible
of detonation products was 9 10 times larger than the
tube material, the value of rm can be found from the
initial volume. This dependence along with data on
relation
detonation velocity (see Table 3) was used to determine
1
2 2
2
rm = re - re,0 - ri,0 , (2)
the explosive performance.
2
Performance of Emulsion Explosives 467
Fig. 4. Gurney energy versus relative volume of
detonation products.
Fig. 3. Radial displacement of the central surface of
the tube versus time: the points refer to the experi-
ment and the curves refer to Eq. (3).
An analysis of the curves in Fig. 4 shows that for
v/v0 > 6, the Gurney energy of the composition AN""
where re,0 and ri,0 are the initial radii of the external
practically does not increase, which is characteristic of
and internal surfaces and re and rm are the current
strong explosives. Thus, the values of EG for emul-
radii of the external and central surfaces of the tube at
sion explosives containing only ammonium nitrate are
a given distance x.
nearly ideal. If an explosive contains metal nitrates, the
Assuming that the motion of the detonation prod-
Gurney energy increases slowly but continuously even
ucts and the tube is stationary (i.e., x = Dt, where D
for v/v0 > 6. This shows that in this case, the energy
is the detonation velocity and t is time), we can replace
release behind the detonation front is decelerated. The
the dependence rm(x) by the dependence rm(t). In [15],
addition of nickel and sodium nitrates leads to the most
the position of the central surface of the tube in time is
considerable decrease in the detonation properties of the
approximated by the function
composition AN"".
rm = rm,0 + ai{bi(t - t0)
Figure 5 shows a curve of EG(Á0D2). For v/v0 =
i
2.5, it is close to that given in Fig. 1, and for v/v0 = 9,
-[1 - exp(-bi(t - t0))]}, (3)
the data for all explosives lie on one line. This implies
where rm,0 is the initial value of rm and ai, bi, and t0 that there is a linear relationship between explosive per-
are parameters. The thus obtained dependence rm(t) is formance and detonation parameters.
shown in Fig. 3. The data of Figs. 4 and 5 also suggest that in
Having determined the relations rm(t) and differ- charges with a strong confinement, chemical reactions
entiating Eq. (3), we can calculate the radial velocity in expanding detonation products occur during a long
of the tube period of time and the released additional energy is con-
verted to the expansion work of detonation products.
drm
um a" = aibi[1 - exp(-bi(t - t0))], (4)
dt
i
and, then, the total velocity
Åš um
3.2. Expansion Work of Detonation Products
uL = 2D sin , Åš = arctan . (5)
2 D
Using Eqs. (4) and (5) and the measured detona- In the literature (see, for example, [19, 20]), the
tion velocities (see Table 3), we determined values of expansion work of detonation products w is defined as
uL and then evaluated EG by Eq. (1). The calculated the work of the products expanding from volume at the
dependence of Gurney energy on the relative volume of Chapman Jouguet point (vCJ) to a volume v minus the
detonation products (v/v0) is shown in Fig. 4. energy of compression of the explosive:
468 Cudzilo, Kohlicek, Trzcinski, and Zeman
TABLE 4
Effective Poisson Adiabatic Exponent
Explosive Å‚
AN"" 2.86
AN CaN 2.85
AN NiN 2.80
AN CoN 2.84
AN NaN 2.83
Fig. 5. Gurney energy versus the parameter Á0D2.
v
w(v) = -ec + p dv. (6)
vCJ
To evaluate the expansion work, it is necessary to
know the isentrope of expansion of detonation products. Fig. 6. Expansion work of detonation products cal-
culated using the Poisson adiabat.
A simple isentrope is the Poisson adiabat
vCJ Å‚
p = pCJ , (7)
v
The dependence of the expansion work calculated
where Å‚ is the adiabatic exponent. To determine con-
by relation (8) on the relative volume of detonation
stants in Eq. (7), it is necessary to know values of two
products is given in Fig. 6. The figure also shows the
detonation parameters. The first of these is usually ex-
interval of volume of detonation products correspond-
perimental detonation velocity and the second is the
ing to the pressure range for rock blasting [22]. As soon
exponent of isentrope of detonation products. Most of-
as rock begins to fracture, cracks form and detonation
ten it is the so-called effective exponent of isentrope,
products penetrate into then, thus weakening the de-
which is not determined from the real isentrope in a cer-
structive effect of the explosive. Therefore, it is assumed
tain range of volume of detonation products rather than
that the values of expansion for v/v0 = 10, 15, and 20
from parameters at the Chapman Jouguet point. Ac-
can be used to compare the efficiency of explosion in
cording to the procedure proposed in [13, 21], its value
hard, strong, and weak rock [22].
can be determined from results of a cylinder expansion
As might be expected, the composition containing
test. In this method, an experimental profile of a cop-
only ammonium nitrate performed the greatest expan-
per tube is compared with that obtained by numerical
sion work (among the tested explosives). This result is
modeling of the expansion process. The state of detona-
obtained with the usage of both experimental methods.
tion products is described by Poisson s adiabat (7). As
However, for different explosives there is some difference
a result, a value Å‚ is chosen that ensures best agreement
in results obtained by different methods. It is difficult
between the calculated and experimental profiles of the
to reliably determine which of them is more accurate
tube. The effective exponent of isentrope of detonation
because the difference is rather insignificant. However,
products for the tested explosives are given in Table 4.
taking into account that the sample mass was larger
After integration using relation (7), from Eq. (6)
in cylinder tests and the duration of expansion in this
we obtain the following expression for the expansion
case was also much longer, it is concluded that the data
work:
of these experiments are more exact in comparing the
pCJvCJ Å‚ + 1 vCJ Å‚-1
w = - . (8)
blasting performance of explosives.
Å‚-1 2Å‚ v
Performance of Emulsion Explosives 469
CONCLUSIONS Detonation Symp., Boston, Massachusetts, July 12 16
(1993), pp. 741 748.
The results of the experiments described here show 6. K. Takahashi, M. Murata, Y. Kato, et al.,  Non-ideal
that the cylinder test can be used with advantage for detonation of emulsion explosives, J. Mater. Proces.
quantitative determination of explosive performance. Technol., 85, 52 55 (1999).
7. J. Kohler and R. Meyer, Explosives, VCH Publ., Wein-
Gurney energy is easy to determine from radiographs
heim, Germany (1993), pp. 24 25.
of the motion of the copper tube wall. Results of these
8. L. E. Fried,  CHEETAH 1.39  User s manual, Re-
experiments together with data of numerical modeling
port No. UCRL-MA-117541 (1996).
of the process allow one to obtain the equation of state
9. M. L. Hobbs and M. R. Baer,  Calibrating the
for detonation products and, hence, to estimate their
BKW-EOS with a large product species base and mea-
effective expansion work.
sured C J properties, in: 10th Int. Detonation Symp.,
The addition of metal nitrates to emulsion compo-
Boston, Massachusetts, July 12 16 (1993), pp. 409 418.
sitions deteriorates the detonation properties of explo-
10. S. Cudzilo and W. A. Trzcinski,  Application of a cylin-
sives containing only ammonium nitrate, which is man-
der test for determining some characteristics of explo-
ifested in a decrease in both detonation parameters and
sives (in Polish), Biuletyn WAT, 46, No. 5 (1997).
performance. The decrease in performance is most con-
11. H. Hornberg,  Determination of fume state parameters
siderable in the initial stage of expansion of detonation
from expansion measurements of metal tube, Propel-
products. With increase in the volume of detonation
lants, Explos., Pyrotech., 11, No. 1, 23 31 (1986).
products, these discrepancies decrease. This implies
12. L. E. Fried and P. C. Souers,  BKWC: an empirical
that chemical reactions also proceed in expanding deto-
BKW parametrization based on cylinder test data,
nation products and additional energy released in these
Propellants, Explos., Pyrotech., 21, 215 221 (1996).
reactions is converted to mechanical work. For this rea-
13. S. Cudzilo, R. Trebinski, and W. Trzcinski,  Determina-
son, detonation parameters and performance character-
tion of the effective exponent of isentrope for the detona-
istics determined in the initial stage of expansion de-
tion products of high explosives, Chem. Phys. Reports,
pend strongly on charge diameter and type shell type.
16, 1719 1732 (1997).
In this connection, the most realistic values of perfor-
14. R. Trebinski and W. A. Trzcinski,  Determination of
mance can be obtained in cylinder tests rather than in
an expansion isentrope for detonation products of con-
ballistic mortar tests.
densed explosives, J. Tech. Phys., 40, No. 4, 447 504
The authors are grateful to Jiri Tesitel (The Re-
(1999).
search Institute of Industrial Chemistry, Pardubice,
15. W. A. Trzcinski,  Application of a cylinder test for
Czech Republic) for his help in some of experiments
determining energetic characteristics of explosives,
and to Professor Andrzej Maranda (Military University
J. Tech. Phys., 42, No. 2, 165 179 (2001).
of Technology, Warsaw, Poland) for many discussions
16. S. Cudzilo and W. A. Trzcinski,  The application of the
and helpful suggestions.
cylinder test to determine the energy characteristics of
industrial explosives, Arch. Mining Sci., 46, No. 3,
291 307 (2001).
17. P. C. Souers and J. W. Kury,  Comparison of cylinder
REFERENCES
data and code calculations for homogeneous explosives,
Propel., Explos., Pyrotech., 18, 175 181 (1993).
1. R. S. Egly and A. E. Neckar,  Water resistant sensitizes
18. W. P. Walters and J. A. Zukas, Fundamentals of Shaped
for blasting agents, U.S. Patent No. 3,161,551 (1964).
Charges, Wiley-Interscience Publ., New York (1989).
2. N. E. Gehrig,  Aqueous emulsified ammonium nitrate
19. S. J. Jacobs,  Energy of detonation, Report NA-
blasting compositions containing nitric acid, US Patent
VORD  4366, United States Naval Ordnance Labo-
No. 3,164,503 (1965).
ratory (1956).
3. M. Yoshida, M. Iida, K. Tanaka, et al.,  Detonation
20. W. Fickett and W. C. Davis, Detonation, Univ. of Cal-
behavior of emulsion explosives containing glass micro-
ifornia Press, Berkeley (1979).
balloons, in: 8th Symp. (Int). on Detonation, Albu-
21. R. Trebinski and W. Trzcinski,  Method of determina-
querque, NM, July 15 19 (1985), pp. 993 1000.
tion of the effective exponent of isentrope for the det-
4. J. Lee and P. A. Persson,  Detonation behavior of emul-
onation products of condensed explosives (in Polish),
sion explosives, Propellants, Explos., Pyrotech., 15,
Biuletyn WAT, 45, No. 3 (1996).
No. 5 (1990).
22. P. A. Person and R. F. Chiappetta,  Shock waves and
5. M. M. Chaudhri and L. A. Almgren,  Detonation be-
detonation, explosive performance, in: Proc. NIXT 94,
havior of  water-in-oil type emulsion explosive contain-
Pretoria, RSA (1994).
ing glass micro-balloons of selected sizes, in: 10th Int.