Thermal analysis of black powder

Article abstract of DOI:10.1023/A:1025181424038

While there is abundant literature describing the factors affecting the performance and the mechanical sensitivity of black powder, only a few papers are devoted to its thermal properties. Previous work indicated that no exothermic reactions were observed below 300°C in an inert gas environment. In the present work a variety of thermal techniques (DSC, TG, simultaneous TG-DTA-FTIR-MS, ARC, HFC) has been used to study the thermal decomposition of black powder. Exothermic reactions were observed at temperatures as low as 230 and 140°C in inert and oxidizing atmospheres, respectively. The latter exothermic reaction is due to sulfur oxidation.

Full text of DOI:10.1023/A:1025181424038

Journal of Thermal Analysis and Calorimetry, Vol. 73 (2003) 105–118
THERMAL ANALYSIS OF BLACK POWDER
*
R. Turcotte , R. C. Fouchard, A.-M. Turcotte and D. E. G. Jones
Canadian Explosives Research Laboratory, Natural Resources Canada, 555 Booth St., Ottawa,
Ontario K1A 0G1, Canada
(
Received March 3, 2003)
Abstract
While there is abundant literature describing the factors affecting the performance and the mechanical
sensitivity of black powder, only a few papers are devoted to its thermal properties. Previous work indi-
cated that no exothermic reactions were observed below 300°C in an inert gas environment. In the pres-
ent work a variety of thermal techniques (DSC, TG, simultaneous TG-DTA-FTIR-MS, ARC, HFC)
has been used to study the thermal decomposition of black powder. Exothermic reactions were ob-
served at temperatures as low as 230 and 140°C in inert and oxidizing atmospheres, respectively. The
latter exothermic reaction is due to sulfur oxidation.
Keywords: ARC, black powder, DSC, FTIR, HFC, mass spectrometry, TG, TG-DTA, thermal de-
composition
Introduction
Black powder is the most ancient explosive known and the first documented ener-
getic material. It has been used for over 800 years [1] and is still extensively used in
the fireworks and pyrotechnic industries and for historical weapons. Black powder is
a mixture of potassium nitrate, sulfur and charcoal. While the standard composition
typically contains 75% potassium nitrate, 10% sulfur, and 15% charcoal, many varia-
tions have been used in order to obtain specific properties [2]. Although the combus-
tion properties of black powder have been extensively studied [3, 4], much less infor-
mation is available on its thermal behavior during the early stage of reaction. This
low-temperature chemistry is important in determining the thermal hazards and the
long-term stability of the material.
Perhaps the most comprehensive study on the early chemical processes occur-
ring in the decomposition of black powder was that of Blackwood and Bowden [5].
They studied its thermal decomposition, below its ignition temperature, by monitor-
ing the rate of gas evolution as a function of temperature. Chemical analysis of the
gaseous and condensed products was also carried out in the same work. Their essen-
*
Author for correspondence: E-mail: riturcot@nrcan.gc.ca
1
388–6150/2003/ $ 20.00
Akadémiai Kiadó, Budapest
©
2003 Akadémiai Kiadó, Budapest
Kluwer Academic Publishers, Dordrecht

1
06
TURCOTTE et al.: BLACK POWDER
tial conclusions were that the initial reaction was between molten sulfur and the
oxhydrocarbons (OHC) present in the charcoal, starting at about 150°C, but that the
main reaction responsible for run-away was the oxidation of charcoal by KNO3.
Campbell and Weingarten [6] carried out a thermal analysis study of black pow-
der and its constituents using an in-house DTA apparatus with sample sizes up to 2 g.
In these measurements, the DTA curves of black powder were characterized by a
broad endotherm between approximately 95 and 200°C, followed by an exothermic
pre-ignition reaction starting around 250°C and ignition taking place at about 300°C.
They also carried out TG experiments on the same materials and measured ignition
delay times using the method of Henkin and McGill [7]. In these experiments, virtu-
ally all the mass loss was observed in a single and extremely rapid step around 250°C.
Their data support a reaction between S and KNO as the pre-ignition reaction.
3
More recently, Brown and Rugunanan [9] have performed simultaneous
TG-DTA studies on black powder in nitrogen while Hussain and Rees [8] published
the results of DSC and TG/DTG experiments. In both cases, binary mixtures of its
constituents were also investigated. Their main conclusions were that, while no reac-
tion is observed between sulfur and charcoal, the energy release from the oxidation of
organic material might enable the reaction of S and KNO below the melting point of
3
KNO . While they regarded the latter as the ‘pre-ignition’ reaction, the strongly exo-
3
thermic oxidation of the charcoal by KNO , which takes place between 350 and
3
5
50°C, is the one that appeared to propagate the combustion of black powder.
This paper presents results obtained using a variety of thermal techniques: dif-
ferential scanning calorimetry (DSC), thermogravimetry (TG), simultaneous thermo-
gravimetric and differential thermal analysis (TG-DTA) combined with gaseous
product detection by Fourier transform infrared spectrometry (FTIR) and mass spec-
trometry (MS), accelerating rate calorimetry (ARC), and heat flux calorimetry
(
HFC). By using this combination of methods to generate complementary informa-
tion, it is possible to obtain a much more complete understanding of the system being
studied than would be possible with a single experimental technique. Some results
obtained at the early stage of this study have already been presented [10].
Experimental
DSC
A TA Instruments 2910 DSC was used for these measurements. The instrument was
3
purged at a rate of 50 cm min with dry and oxygen-free nitrogen. A heating rate of
–1
–
°C min was employed. Open aluminum pans and sealed glass ampoules were used
1
5
for the measurements. In the first case, the sample size was approximately 5 mg while
in the second case it was 0.5 mg. Temperature and heat flow calibrations were per-
formed prior to the experiments according to ASTM procedures [11].
J. Therm. Anal. Cal., 73, 2003

TURCOTTE et al.: BLACK POWDER
107
TG
A TA Instruments 2950 TG instrument was used. In these experiments, approxi-
mately 5 mg of sample material was held in a platinum pan, in a flow of dry nitrogen
3
100 cm min , split 60:40 between the furnace and the balance). Again, a heating
–1
(
–
1
rate of 5°C min was employed. The TG instrument was calibrated for mass and tem-
perature. The standard reference mass used for mass calibration was checked against
a Mettler M3 microbalance with a precision of ±1 µg. The temperature was calibrated
using the ASTM Curie point magnetic procedure with SRM nickel [12].
Simultaneous TG-DTA-FTIR-MS
Simultaneous thermogravimetry - differential thermal analysis was carried out using
a TA Instruments SDT 2960 instrument, with evolved gas analysis using a Bomem
MB100 FTIR spectrometer and a Balzers Thermostar GS300 quadrupole mass
spectrometer. The TG-DTA-FTIR-MS data were acquired simultaneously in the
same run. For these experiments, approximately 5 mg of sample material was placed
in an alumina pan. The DTA reference material consisted of about 5 mg of platinum
3
–1
foil. The sample and the reference were purged with a 50 cm min flow of dry air or
1
–
helium. The temperature was ramped from ambient to 500°C at a rate of 5°C min .
The FTIR interface consisted of a 5 mm i.d. Teflon tube and a Pyrex cell (10 cm
3
long, 50 cm internal volume) equipped with KBr windows. A quartz microfiber filter
–
1
was placed at the FTIR inlet. The acquisition rate of the FTIR was 1 scan min and
the time delay from the SDT furnace outlet to the FTIR cell was about 10 s. Bomem
Grams 32 software was utilized for data acquisition and analysis.
The mass spectrometer data were acquired using the Balzers Quadstar 422 soft-
ware. The heated quartz capillary interface was placed near the sample pan, in the
SDT furnace. Data were acquired using bar graph scan, programmed to acquired data
–
1
from 1 to 120 amu (atomic mass unit) at a speed of 0.2 s amu .
TG mass together with DTA baseline and temperature calibrations (indium and
tin) were performed prior to the experiments.
It should be pointed out that, while chemical analysis of solid residues would
have been useful to identify solid decomposition products, such work could not be
performed in the present study.
ARC
The experiments were carried out using one of the ARCs available at the Canadian
Explosives Research Laboratory (CERL). This ARC is an automated adiabatic calo-
rimeter distributed by Arthur D. Little Inc. and is used for the purpose of assessing the
thermal hazard potential of energetic materials [13]. Adiabatic conditions are main-
tained in this ARC provided that the rate of temperature increase does not exceed
–
1
about 10°C min . Experiments were carried out using standard CERL procedures al-
ready described in earlier publications [14–16].
J. Therm. Anal. Cal., 73, 2003

1
08
TURCOTTE et al.: BLACK POWDER
For these experiments, samples of 0.2 to 2 g of black powder were placed in
lightmass spherical titanium vessels and experiments were performed with the ARC
manifold closed in order to maintain any pressure resulting from vaporization or de-
composition of the sample. Experiments were conducted both in an inert atmosphere
(
Ar) and air.
For all experiments, the sample was heated quickly and equilibrated at the cho-
sen initial temperature. It was then subjected to the standard ARC procedure of
heat-wait-search’. In this study, the temperature of the system was raised in 5°C in-
‘
crements and the system was maintained adiabatic during the ‘wait’ period (30 min
for dissipation of thermal transients) and the ‘search’ period (20 min looking for an
exotherm). An exotherm was defined as a self-heating rate (R) greater than the
–
1
pre-selected threshold value of 0.02°C min . Since exotherms are not detected dur-
ing the ‘heat’ or ‘wait’ modes, the ‘true’ value of the onset temperature for thermal
run-away (T ) is obtained by extrapolating to R=0. Whenever an exotherm is detected
0
during the search mode, the heat-wait-search procedure is automatically interrupted,
and the calorimeter tracks the time, temperature, rate (calculated) and pressure. Data
collection was programmed to stop if the pressure exceeded a set value (usually
–
1
6
.89 MPa or 1000 psia), or R exceeded a certain value (usually 5°C min ).
HFC
A Setaram C-80 heat flux calorimeter was used for these experiments. Modifications
to this instrument with high-pressure manifold and cells have already been described
in detail elsewhere [17]. Two experimental methods were employed. In the first one,
samples of 0.1 to 1.0 g were loaded directly into the high-pressure cell. Dry air or ar-
gon could then be introduced to vary the nature and initial pressure of the initial atmo-
sphere. In the second method, less than 0.1 g of sample material was sealed in a glass
3
ampoule with an internal volume of approximately 0.7 cm . The gas contained within
the ampoule was air, but helium was used to provide good thermal contact between
the ampoule and the walls of the cell. The heating rate for these HFC experiments was
–
either 0.1 or 0.3°C min .
1
Table 1 Characteristics of black powder samples used in the present work
Sample number
Grade
Grain size/mm
Manufacturer
1
2
3
4
5
6
7
8
2FA
2FA
2.0–6.4
2.0–6.4
1.5–2.5
1.5–2.5
0.5–1.3
0.5–1.3
0.5–1.3
<0.2
Goex, USA
Pernambuco, Brazil
Goex, USA
3FA
3FA
Pernambuco, Brazil
Goex, USA
5FA
5FA
Pernambuco, Brazil
Kemijiska Industrija, Slovenia
Goex, USA
5FA
XF Meal
J. Therm. Anal. Cal., 73, 2003

TURCOTTE et al.: BLACK POWDER
109
Materials
A number of different samples of black powder was used during the course of this
work. Table 1 summarizes the data on their grades, grain sizes, and origin.
Results
DSC and TG
Figure 1 shows typical DSC (bottom) and TG (top) curves obtained for 5FA black pow-
–1
der, with a heating rate of 5°C min . Rather than the single rapid mass loss step around
50°C observed by Campbell and Weingarten [6], the TG curve in inert gas environment
N ) obtained in the present work shows two distinct mass loss steps. This is most proba-
2
(
2
–1
bly due to the faster heating rate they were using (15°C min ). In the first step, occurring
below 200°C, 10% of the mass is lost, which corresponds to the vaporization of sulfur. In
this temperature range, the corresponding DSC curve (open Al pan) shows the rhombic
to monoclinic sulfur phase transition at 108°C (1), the melting of sulfur at 116°C (2), and
the KNO rhombic to trigonal phase transition at 133°C (3). The second mass loss step
3
occurs between 200 and 450°C. In the latter, approximately 33% of the mass is lost. In
this temperature domain, the features of the open Al pan DSC curve are a gradual in-
crease in exothermic heat flow starting around 200°C, a sharp endothermic peak indicat-
ing the melting point of KNO (330°C, (4)), and a broad exothermic peak, between 350
3
and 450°C, corresponding to the oxidation of the charcoal by KNO3.
TG experiments were also performed in air and the results are also shown in
Fig. 1. Careful inspection of the two TG curves reveals that the second mass loss step
in air took place at lower temperature than in nitrogen.
Fig. 1 Typical TG (top) and DSC (bottom) results for 5FA black powder (note that
DSC baselines were shifted for clarity)
J. Therm. Anal. Cal., 73, 2003

1
10
TURCOTTE et al.: BLACK POWDER
DSC experiments with 0.5 mg of black powder in a sealed glass micro-ampoule
were also carried out and these results are compared to those obtained in an open Al
pan in Fig. 1. It is seen that, while the KNO phase transition is still observed, the on-
3
set of the charcoal/KNO reaction occurs at a much lower temperature (285±3°C), so
3
that the peak corresponding to the melting of KNO has disappeared, masked by the
complex exothermic structure.
3
Simultaneous TG-DTA-FTIR-MS
The results of a typical simultaneous TG-DTA run on 5FA black powder in an inert at-
–1
mosphere (He), at a heating rate of 5°C min , are shown in Fig. 2 (top). The TG curve is
in close agreement with that of the previous section with 10% mass loss in the first step
and about 33% mass loss in the second step. The DTA curve displays the same
endotherms associated with melting and phase transitions for S and KNO already seen in
3
the DSC experiments. Additionally, the oxidation of charcoal by KNO is observed to
3
start at a slightly higher temperature than in the corresponding DSC experiment.
Fig. 2 Simultaneous TG-DTA (top), FTIR (center), and MS (bottom) results for 5FA
black powder in helium (note that FTIR baselines have been shifted for clarity in
the inset)
A number of products were observed in the gas phase. The FTIR data in Fig. 2 (cen-
ter) are shown as absorbances, while the MS data (bottom) are displayed as ion currents.
Although the FTIR and MS data are not calibrated for quantitative measurements, ap-
proximate relative concentrations for some of the gaseous species can be obtained using
the scaling factors of Brill et al. [18]. The latter can be used to convert absorbance inten-
sity of equal partial pressures of gases to concentrations relative to CO . The results of
2
J. Therm. Anal. Cal., 73, 2003

TURCOTTE et al.: BLACK POWDER
111
this conversion are shown in Table 2. It is clear that CO , NO and N O were the most
2
2
abundant gaseous products observed. In helium, their evolution was observed to start
around 200°C, at the beginning of the second mass loss step, and to increase sharply
around 350°C. It is important to notice that no trace of H S could be detected in the pres-
2
ent work. Traces of two sulfur compounds, COS and CS , were observed at temperatures
2
below the melting point of KNO , while the evolution of water vapor and nitrogen oxides
3
was observed to coincide with CO production. Traces of SO were also detected. In this
2
2
case, two maxima were observed around 350°C, where the production of CO starts to in-
2
crease sharply, and around 450°C, close to completion of the second mass loss step.
Table 2 Relative concentration of some gaseous species observed by FTIR
3 –1
Helium (50 cm min )
3 –1
Dry air (50 cm min )
Frequency/
cm
Scaling
factor
Product
–1
*
**
*
**
T
max /°C
C
max
T
max /°C
C
max
3
4
00
20
0.49
1.00
CO
2
2349
2224
1903
1621
1.0
1.86
32.1
2.0
420
420
1.00
0.16
N
2
O
420
0.08
3
4
00
20
0.10
0.46
300
420
0.10
0.16
NO
NO
2
420
–
420
0.01
*
*
Temperature at maximum of absorbance peak
*
Relative concentration at maximum of absorbance peak
Fig. 3 Simultaneous TG-DTA (top), FTIR (center), and MS (bottom) results for 5FA black
powder in air (note that FTIR baselines have been shifted for clarity in the inset)
J. Therm. Anal. Cal., 73, 2003

1
12
TURCOTTE et al.: BLACK POWDER
Very significant differences were observed when the same experiment was car-
ried out in air, as seen from Fig. 3. While the TG curve is quite similar, less mass loss
was found to occur in the second step (28.7 % in air compared to 32.9% in helium). In
addition to the usual endotherms corresponding to melting and phase changes of S
and KNO , a new broad exotherm was observed around 230°C. As seen from the
3
FTIR and MS data, this exotherm coincides with the appearance of SO in the gas
2
phase. Another broad exotherm, extending from 275 to 350°C and partially masking
the KNO melting endotherm, was also observed to coincide with a first peak in the
3
CO evolution. These features were not observed in the experiment under inert atmo-
2
sphere. Only the second intense CO peak in the 350 to 450°C range, corresponding
2
to the oxidation of the charcoal by KNO , was observed in both cases. In the experi-
3
ment performed in air, the latter peak coincides with trace amounts of N O, NO, NO
2
2
and H O. The relative concentrations of the first three species, computed using the
2
scaling factors of Brill et al. [18], are also shown in Table 2.
ARC
Experiments on 5FA black powder using ARC were carried out both in ambient and
elevated pressure atmospheres of argon and air. Typical results are shown in Fig. 4.
In argon at ambient initial pressure, several experiments were performed with
sample masses from 0.2 to 2 g, using various initial temperatures (35, 100 and
2
00°C). No systematic dependence of T upon these two variables was observed. It
0
should be mentioned that when a 2 g sample was used, the run-away reaction was vio-
lent enough to explode the ARC sample vessel. Again, when argon at elevated initial
pressure (1.4 MPa) was used, no significant difference in the value of T was found. A
0
sample mean of all the measurements performed under an inert atmosphere yielded
the value of 225±14°C for T . In all cases, the Arrhenius plots of the exotherm data
0
(
Fig. 4, bottom) were found to be linear, except for the early stage of the exotherms.
As for the onset temperature, the activation energies (E) obtained from linear fits to
the ln(Rate) vs. 1/T data showed no systematic dependence on the various physical
parameters mentioned above. A sample mean from all the linear fits yielded the value
–
1
E=252±48 kJ mol .
Several experiments were also performed with 0.5 g samples in air at ambient pres-
sure and using various initial temperatures. Again no significant differences in T and E
0
–1
values were observed and sample means of T =234±7°C and E=268±19 kJ mol were
0
obtained.
When experiments were carried out with 0.5 g samples in air at elevated initial
pressure (1.4 MPa), a very different behavior was observed. The system was found to
go into run-away at much lower temperature (T =141±5°C). Also, in this case, the
0
Arrhenius plot was observed to be much more linear (Fig. 4, bottom) with a much
–
1
higher activation energy (E=601±11 kJ mol ).
J. Therm. Anal. Cal., 73, 2003

TURCOTTE et al.: BLACK POWDER
113
Fig. 4 Typical rate vs. temperature (top) and Arrhenius (bottom) plots from ARC ex-
periments on 5FA black powder in argon and air (note that, for clarity, curves
have been shifted along the vertical axis for the top graph and along the horizon-
tal axis for the bottom graph)
HFC
Experiments using HFC were carried out under a variety of experimental conditions and
on black powders of various grain sizes and origins. The results obtained for 5FA black
powder in open cells and under an ambient pressure argon atmosphere are summarized in
Fig. 5 (top). The same endothermic peaks observed below 200°C in the DSC experiments
(
phase transition of sulfur (1), melting of sulfur (2), and phase transition of KNO (3)) are
3
–1
also seen here. Using a 0.1 g sample and a heating rate of 0.3°C min , the exothermic re-
actions are observed at about 230°C. In order to obtain the full HFC curve, the heating
rate was reduced from 0.3 to 0.1°C min . In this case, it is seen that the position of the
exothermic peak shifted to lower temperature and that the reaction started at about
–1
–1
2
10°C. These results also yielded a value of ∆H = –1.4 kJ g . In both cases the observed
shift in the baseline is believed to result from the vaporization of sulfur. The sample size
was also increased to 0.5 and to 1 g. In these cases, ignition was observed even in an inert
gas environment. This behavior is also illustrated in Fig. 5 (top) for a 0.5 g sample. It is
–1
seen that, even with a heating rate of 0.3°C min , the exothermic reaction is detected at
an even lower temperature (190°C).
J. Therm. Anal. Cal., 73, 2003

1
14
TURCOTTE et al.: BLACK POWDER
Fig. 5 HFC results in argon (top) and in various configurations (bottom) for 5FA black
powder (note that baselines were shifted for clarity)
Samples were also loaded in sealed glass ampoules at ambient pressure of air. In
these cases, the ampoule was introduced in the sample cell, which was then loaded
with helium. In order to prevent rupture of the ampoule and thus contain the whole re-
action in the latter, the sample size had to be decreased to 0.05 g. A typical result ob-
–
1
tained with a heating rate of 0.3°C min is compared to that described above for a
0
–
.1 g sample in argon (0.1°C min ) in Fig. 5 (bottom). About twice the exothermic
1
–
1
energy (in excess of 2.4 kJ g ) is observed under the confinement of the ampoule.
Also, even though a faster heating rate was used with the ampoule, the exothermic re-
action is observed at approximately the same temperature (230°C).
HFC experiments were also carried out with the sample loaded directly in the
high pressure sample cell at ambient pressure of air. The results of an experiment us-
–
1
ing a sample size of 0.1 g and a heating rate of 0.1°C min are also shown in Fig. 5
bottom). Two observations can be made when these results are compared to those
(
obtained under the same conditions in argon: (i) the exothermic reaction is observed
at a slightly lower temperature in air (212°C in air vs. 230°C in argon), and (ii) a weak
exothermic peak is observed between 150 and 180°C. While the latter is rather weak
–
1
(99 J g ), it provides evidence of sulfur oxidation at low temperature even at ambient
pressure of air.
J. Therm. Anal. Cal., 73, 2003

TURCOTTE et al.: BLACK POWDER
115
Fig. 6 HFC results in argon for black powders of various grain sizes and origins (note
that baselines were shifted for clarity)
It is well known that the performance of black powder is affected by the particle
size. In order to determine if the latter also has an effect on its thermal behavior, HFC
studies were carried out on samples of various grain sizes and origins (Table 1). A
–
1
sample size of 0.1 g and a heating rate of 0.3°C min were used for these experi-
ments. The results are compared in Fig. 6. It is seen that, while the peak temperatures
are not a strong function of the particle size, the exothermic reactions for some larger
grain size samples (2FA and 3FA from Brazil) are more complex as shown by the left
shoulder in the corresponding main exotherms. It also appears that the origin of the
sample plays an important role in the observed thermal behavior, as the two 2FA
samples from different origin (USA and Brazil) display a very different shape for the
main exotherm.
Discussion
The TG and DSC results for the open pan system reported here are similar to those of
Brown and Rugunanan [9]. When the system is confined, as in the glass micro-
ampoules, it is seen that onset temperatures well under 300°C are observed. This is
probably due to the fact that the reaction gases do not escape, and these gases acceler-
ate the reaction. This confinement effect was also confirmed by the HFC experi-
ments, in which much higher exothermic energy release was observed with sealed
glass ampoules. The presence of air in the ampoule may also promote oxidation of or-
ganic material, which, in turn, may enable more sulfur oxidation by KNO [8].
3
J. Therm. Anal. Cal., 73, 2003

1
16
TURCOTTE et al.: BLACK POWDER
The simultaneous TG-DTA-FTIR-MS results in a helium atmosphere (Fig. 2)
show that no gas evolution was detected in the first mass loss step (T<200°C). This
may be due to the fact that the volatile organic species condense before reaching the
FTIR cell. In contrast to the data obtained by Blackwood and Bowden [5], no H2S
production is observed in the early stage of decomposition. This is probably due to
the fact that, in the present work, the system was constantly purged with helium or
dry air while they were using a closed system. In the present case, the vaporized S
would constantly be swept away while the volatile organic species might be condens-
ing rapidly, leaving insufficient residence time for the formation of a detectable con-
centration of H S. This interpretation is consistent with earlier work reported [6, 8]. It
2
also suggests that, if there were a contributing reaction between S and OHC, it would
have to occur in the vapor phase rather than the molten state.
The formation of traces of COS and CS early in the second mass loss step is
2
probably occurring in the gas phase as both S and CO are evolved at that stage. From
2
Table 2, it is seen that CO , NO, and N O are the most important observed gaseous
2
2
species. While CO and N O have only one major peak with a maximum at 420°C,
2
2
NO displays another peak with a maximum at about 300°C. This early NO production
is consistent with the following reaction:
S+2KNO → K SO +2NO
(1)
The latter was first proposed by Campbell and Weingarten [6] as the pre-ignition
reaction.
3
2
4
Various other reactions have been proposed for the oxidation of charcoal by
KNO , which is responsible for propagation [6, 19]. Earlier observations [8, 19] re-
3
port K SO and K CO as the most abundant solid products from the combustion of
2
4
2
3
black powder and CO , CO, and N as the major gaseous products from the oxidation
2
2
of C, taking place between 350 and 550°C. However, the present work clearly identi-
fies CO , NO and N O as the major gaseous species evolved in this temperature range
when the reaction occurs in helium. While NO production at this stage has also been
2
2
observed by Blackwood and Bowden [5], the production of N O may be explained by
2
a contribution from the reaction:
2
KNO +2C → 2CO +K O+N O
(2)
3
2
2
2
When the same experiments were repeated in air, the mass loss obtained in the
second step was found to be lower than in helium (28.7% compared to 32.9%). This
can be explained by the formation of even more K SO from the reaction
2
4
2
KNO +SO → K SO +2NO
(3)
3
2
2
4
2
as more SO formation would result from the presence of O in the gas phase. This is
2
2
confirmed by the FTIR and MS data, both showing a sharp increase in SO produc-
2
tion at 228°C (Fig. 3). Also, when the experiment is carried out in air, two clear peaks
are observed in the CO evolution. The first one, at about 330°C, may be due to the
oxidation of free carbon from the decomposition of the hydrocarbons contained in the
2
J. Therm. Anal. Cal., 73, 2003

TURCOTTE et al.: BLACK POWDER
117
charcoal also promoted by the presence of O in the gas phase. The second peak
2
would then be the result of oxidation of the charcoal by KNO3.
The temperatures for the onset of run-away reactions in black powder obtained
by ARC and HFC are much lower than those observed in DSC and DTA experiments.
This is believed to be due to the larger sample sizes used in the former, which can
lead to either an autocatalytic reaction or to quasi-adiabatic conditions in the central
area of the samples. With 0.5 g samples, T values as low as 190°C were observed in
0
HFC experiments, even in an inert atmosphere. Also, direct oxidation of sulfur in the
gas phase was observed at low temperature (160°C) in the HFC experiments in air at
ambient pressure. This is consistent with the increased early SO production noted
2
above in the FTIR and MS observations. While no such low temperature exotherm
could be observed in the corresponding ARC data, run-away is observed at a lower
temperature (140°C) in the latter when an elevated initial air pressure was used.
These observations may have important implications on the hazards of processing
black powder, particularly under conditions where long residence times may occur
with air intimately mixed with the powder, or under pressurized air atmosphere.
Significant variations in thermal behavior as a function of particle size were ob-
served in the present work. It appears that larger particles may lead to lower T val-
0
ues. However, the more complex exotherms displayed by samples with larger particle
size may also result from the different origins of the samples. This may be caused by
the use of different charcoal, for example. It is well known that variability in the char-
acteristics of charcoal can cause significant differences in the thermal behavior of
black powder [5, 20].
*
* *
The authors would like to thank Hands Fireworks Inc. for providing some of the samples of black
powder.
References
1
A. Marshall, ’Explosives – Their Manufacture, Properties, Tests and History’, P. Blakiston’s
Son & Co., Philadelphia 1915.
2
3
4
5
6
7
8
9
J. P. Cundill, ‘A Dictionary of Explosives’, The Royal Engineers Institute, Chatham 1889.
R. A. Sasse, Proc. Int. Pyrotech. Seminar, 9 (1984) 47.
D. J. Wood, Proc. Int. Pyrotech. Seminar, 10 (1985) 71.
J. D. Blackwood and F. P. Bowden, Proc. Roy. Soc. A, 213 (1952) 285.
C. Campbell and G. Weingarten, Trans. Far. Soc., 55 (1959) 2221.
Henkin and McGill, Ind. Eng. Chem., 44 (1952) 1391.
G. Hussain and G. J. Rees, Propellants, Explosives, Pyrotechnics, 15 (1990) 43.
M. E. Brown and R. A. Rugunanan, Thermochim. Acta, 134 (1988) 413.
th
1
0 D. E. G. Jones, H. Feng and R. C. Fouchard, 4 Int. Sym. Fireworks, 1 (1998) 155.
1 ASTM E 967, ASTM E 968, American Society for Testing and Materials, West Conshohocken,
PA, USA.
1
1
2 ASTM E 1582, American Society for Testing and Materials, West Conshohocken, PA, USA.
J. Therm. Anal. Cal., 73, 2003

1
18
TURCOTTE et al.: BLACK POWDER
1
1
1
1
1
1
3 D. I. Townsend and J. C. Tou, Thermochim. Acta, 37 (1980) 1.
4 H. T. Feng, K. J. Mintz, R. A. Augsten and D. E. G. Jones, Thermochim. Acta, 311 (1998) 105.
5 D. E. G. Jones, H. T. Feng, R. A. Augsten and R. C. Fouchard J. Therm. Anal. Cal., 55 (1999) 9.
6 D. E. G. Jones, H. T. Feng, K. J. Mintz and R. A. Augsten, Thermochim. Acta, 331 (1999) 37.
7 D. E. G. Jones, P. Handa and H. Feng, J. Therm. Anal. Cal., 53 (1998) 3.
8 T. B. Brill, H. Arisawa, P. J. Brush, P. E. Congwer and G. K. Williams, J. Phys. Chem.,
9
9 (1995) 1384.
1
2
9 I. Von Maltitz, J. Pyrotechnics, 14 (2001) 27.
0 G. Hussain and G. J. Rees, Propellants, Explosives, Pyrotechnics, 17 (1992) 1.
J. Therm. Anal. Cal., 73, 2003