https://doi.org/10.15388/vu.thesis.

122
https://orcid.org/0000-0003-4831-0641

VILNIUS UNIVERSITY
CENTER FOR PHYSICAL SCIENCES AND TECHNOLOGY

Dominykas
JUKNELEVIČIUS

Illuminating pyrotechnic compositions

and their specific features

DOCTORAL DISSERTATION

Natural sciences,
Chemistry (N 003)

VILNIUS 2020
This dissertation was written between 2016 and 2020 at Vilnius University.
Research was supported by Vilnius University, Lithuanian Research Council,
DAAD (German Academic Exchange Service) and Erasmus+ programs.

Academic supervisor – Prof. Habil. Dr. Arūnas Ramanavičius (Vilnius

University, Natural sciences, Chemistry – N 003).

This doctoral dissertation will be defended in a public meeting of the

Dissertation Defence Panel:
Chairman – Prof. Dr. Artūras Katelnikovas (Vilnius University, natural
sciences, chemistry – N 003).
Members:
Prof. Dr. Konstantin Karaghiosoff (Munich University, Natural sciences,
Chemistry – N 003);
Prof. Habil. Dr. Gediminas Niaura (Center for physical sciences and
technology, Natural sciences, Chemistry – N 003);
Prof. Habil. Dr. Rimantas Ramanauskas (Center for physical sciences and
technology, Natural sciences, Chemistry – N 003);
Doc. Dr. Ramūnas Skaudžius (Vilnius University, Natural sciences,
Chemistry – N 003.

The dissertation shall be defended at a meeting of the Dissertation Defense

Panel at 16:30 on 18th December 2020 at inorganic chemistry (NChA) lecture
hall, Naugarduko st. 24, Vilnius and also via MS Teams platform.
Tel.: 8620 27800; e-mail: dominykas24@gmail.com

The text of this dissertation can be accessed at the library of Vilnius University
and Center for Physical Sciences and Technology libraries as well on the
website of Vilnius University: www.vu.lt/lt/naujienos/ivykiu-kalendorius
 https://doi.org/10.15388/vu.thesis.122
https://orcid.org/0000-0003-4831-0641

VILNIAUS UNIVERSITETAS
FIZINIŲ IR TECHNOLOGIJOS MOKSLŲ CENTRAS

Dominykas
JUKNELEVIČIUS

Apšviečiameji pirotechniniai mišiniai ir

jų specifinės savybės

DAKTARO DISERTACIJA

Gamtos mokslai,
Chemija (N 003)

VILNIUS 2020
Disertacija rengta 2016 – 2020 metais Vilniaus universiteto Chemijos ir
Geomokslų fakultete. Mokslinius tyrimus rėmė Vilniaus universitetas,
Lietuvos mokslo taryba, DAAD (Vokietijos akademinė mainų tarnyba),
Erasmus+ programa.

Mokslinis vadovas – prof. habil. dr. Arūnas Ramanavičius (Vilniaus

universitetas, gamtos mokslai, chemija – N 003).

Gynimo taryba:
Pirmininkas – Prof. dr. Artūras Katelnikovas (Vilniaus universitetas,
gamtos mokslai, chemija – N 003).
Nariai:
Prof. dr. Konstantin Karaghiosoff (Miuncheno universitetas, gamtos
mokslai, chemija– N 003);
Prof. habil. dr. Gediminas Niaura (Fizinių ir technologijos mokslų centras,
gamtos mokslai, chemija –N 003);
Prof. habil. dr. Rimantas Ramanauskas (Fizinių ir technologijos mokslų
centras, gamtos mokslai, chemija –N 003);
Doc. dr. Ramūnas Skaudžius (Vilniaus universitetas, gamtos mokslai,
chemija– N 003).

Disertacija bus ginama viešame Gynimo tarybos posėdyje 2020 m. gruodžio

mėn. 18 d. 15:00 val. Neorganinės chemijos auditorijoje (NChA), Naugarduko
g. 24, Vilniuje ir MS Teams platformoje.
Tel.: 8620 27800; el. paštas dominykas24@gmail.com.

Disertaciją galima peržiūrėti Vilniaus universiteto, Fizinių ir technologijos

mokslų centro bibliotekose ir VU interneto svetainėje adresu:
https://www.vu.lt/naujienos/ivykiu-kalendorius
 List of original papers by PhD candidate

Paper 1 D. Juknelevicius, A. Hahma, R. Webb, T. M. Klapötke, A.

Ramanavicius, An experimental comparison of selected blue
flame pyrotechnics, Propellants Explos. Pyrotechn., 2020,
(accepted)

Paper 2 D. Juknelevicius, P. Alenfelt, A. Ramanavicius, The

Performance of Red Flare Pyrotechnic Compositions modified
with Gas Generating Additives, Propellants Explos. Pyrotechn.,
2020, 45: 671-679.

Paper 3 D. Juknelevicius, T. M. Klapötke, A. Ramanavicius, Blue Strobe

Pyrotechnic composition based on Aminoguanidinium Nitrate,
Propellants Explos. Pyrotechn., 2019, 44: 1466-1471.

Paper 4 D. Juknelevicius, A. Dufter, M. Rusan, T. M. Klapötke, A.

Ramanavicius, Study of Pyrotechnic Blue Strobe Compositions
Based on Ammonium Perchlorate and Tetramethylammonium
Nitrate, Eur. J. Inorg. Chem., 2017, 7: 1113-1119.

Papers published before the start of the PhD studies

Paper 5 D. Juknelevicius, R. Kubilius, A. Ramanavicius, Oxidizer Ratio

and Oxygen Balance Influence on the Emission Spectra of a
Green-Colored Pyrotechnic Flames, Eur. J. Inorg. Chem., 2015,
33: 5511-5515.

Paper 6 D. Juknelevicius, E. Karvinen, T. M. Klapötke, R. Kubilius, A.

Ramanavicius, M. Rusan, Copper(I) Bromide: An Alternative
Emitter for Blue-Colored Flame Pyrotechnics, Chem. Eur. J.,
2015, 21: 15354–15359.

Paper 7 D. Juknelevicius, L. Mikoliunaite, S. Sakirzanovas, R. Kubilius,

A. Ramanavicius, A Spectrophotometric Study of Red
Pyrotechnic Flame Properties Using Three Classical Oxidizers:
Ammonium Perchlorate, Potassium Perchlorate, Potassium
Chlorate, Z. Anorg. Allg. Chem., 2014, 640: 2560-2565.
 Author’s contribution to the original papers
relevant to the PhD

Paper 1 Conceptualization, investigation, data analysis, writing - original

draft.

Paper 2 Conceptualization, methodology, investigation, data analysis,

validation, writing & editing.

Paper 3 Conceptualization, methodology, investigation, data analysis,

writing - original draft.

Paper 4 Supervision, conceptualization, methodology, investigation,

review & editing.

6
 TABLE OF CONTENTS
List of abbreviations ....................................................................................... 8
Introduction .................................................................................................... 9
1. Literature review ................................................................................ 12
1.1. Pyrotechnic compositions .................................................................. 12
1.2. Illuminants (White, Red, Green, Yellow) .......................................... 12
1.3. Blue signals ........................................................................................ 14
1.4. Blue strobes ........................................................................................ 16
2. Experimental ...................................................................................... 18
2.1. Chemicals ........................................................................................... 18
2.2. Preparation of flare bodies, pellets, rods ............................................ 19
2.3. Spectrometer, lux meter, camera ........................................................ 21
2.4. Characterization ................................................................................. 22
3. Results and discussion........................................................................ 23
3.1. Evaluation of the Mg-rich system for red illuminants ....................... 23
3.2. The influence of additives .................................................................. 26
3.3. Influence of guanidinium nitrate on flare performance...................... 30
3.4. Properties of chlorine-free composition ............................................. 33
3.5. Blue illuminants ................................................................................. 35
3.6. Blue strobes ........................................................................................ 42
3.7. Humidity test ..................................................................................... 45
3.8. Differential thermal analysis (DTA) .................................................. 47
3.9. Origin of the bright blue flash ........................................................... 47
Conclusions .................................................................................................. 51
References .................................................................................................... 53
Supplements ................................................................................................. 56
SANTRAUKA ............................................................................................. 63
Acknowledgement ........................................................................................ 71
Curriculum Vitae .......................................................................................... 73
Original papers ............................................................................................. 75

7
 LIST OF ABBREVIATIONS

5-AT 5-aminotetrazole (anhydrous)

AGN Aminoguanidinium nitrate
AP Ammonium perchlorate
BCC Basic copper carbonate (malachite)
CP Chlorinated paraffin
CR Chlorinated rubber
Cu Copper powder (electrolytic)
CuOCl Copper oxychloride, Cu2Cl(OH)3
Dex Dextrin
DTA Differential thermal analysis
GN Guanidinium nitrate
HCB Hexachlorobenzene
Iv light intensity
IR Infrared region
Lsp Specific luminous intensity or efficiency (cdsg-1)
NC Nitrocellulose powder
MgAl Magnalium powder (Mg and Al alloy 50:50)
pe Color purity or saturation
PVC Polyvinyl chloride
RH Relative humidity
TMAN Tetramethylammonium nitrate
VIS Visible region
Ω Oxygen balance

8
 INTRODUCTION

The pyrotechnic illuminants are known for their ability to burn producing
brilliant colored flames. With brightness sometimes exceeding 50000 cd and
having a distinct flame color, such pyrotechnic compositions have been
applied in several utilitarian fields throughout the years.
For safety of life at sea several varieties of pyrotechnic articles are used
to attract attention in the case of an accident. Such articles are cheap, reliable
and produce a bright flame that is visible from a long distance. The importance
of a bright flame becomes even more pronounced at extreme weather
conditions such as storm, rain or fog.
In fireworks bright color flames of high purity are used extensively. The
first colored flame fireworks became available since the discovery of KClO3
in the late 18th century. However, they weren’t particularly bright, since only
later on metallic fuels such as Mg and Al were adopted in pyrotechnics.
Nowadays, a variety of colored flame compositions are used that yield
virtually any color and some that are modified for strobe burning or an
additional spark effect.

Aim of this study

To increase the efficiency of illuminating pyrotechnic compositions, while

maintaining high flame color purity, chemical stability and compatibility.

Objectives of this study and scientific novelt y

1) To investigate how far the amount of Mg in a hand-flare composition

can be increased, while maintaining a low burn rate of <2,1 mm/s without a
loss in color purity or light intensity and to asses other additives, which can
increase the flare’s performance.

Novelty: While some materials that increase the performance of a flare

composition have been reported, there was still no work done in order to
design an efficient and slow burning composition.

2) To identify the best additive, which can increase the performance of a

Mg-rich flare composition and to determine critical amounts of additive that
can be tolerated.
9
Novelty: The three-component flare composition of Sr(NO3)2/Mg/PVC is
known for a long-time, and only few of its modifications are used. The main
goal was to determine other additives that at higher ratios (at over 10 wt.-%)
can increase the performance of the Mg-rich flare composition.

3) To overview and to identify the best blue flame compositions presented

in the literature that are known for their high color saturation and good overall
performance. To compare the best blue flame compositions with respect to
their flame light intensity, efficiency, saturation, emission spectra and CIE
coordinates and dominant wavelength.

Novelty: This kind of experimental examination has not been done to date,
wherein a rather diverse range of blue flame pyrotechnic compositions are
examined quantitatively using a spectrometer. Very few researchers in
general have reported CIE coordinates with pe, Iv, Lsp values for blue flames.
In addition, some old-fashioned ingredients were tested to demonstrate that
arsenic-containing ingredients (Paris Green) offer no benefit in the color
purity, and other, more novel, compositions show better performance.

4) To evaluate what kind of pyrotechnic systems are suitable for a blue

strobe pellet that would function and possess moderate chemical stability and
to analyze some other strobe moderators and catalysts.

Novelty: Blue strobes are a notoriously rare pyrotechnic effect that has not
been much examined due to very limited number of compositions suitable to
produce these strobes and their expensiveness, however, they are very desired
for the fireworks. Such systems are complex and very few compositions are
presented that work and are chemically stable at the same time.

10
 Statements of defense

1. Combination of rounded Mg powder and gas generating additives

yields a high-performance flare composition (>22 kcd s/g) that has
a distinct red flame (pe >90%) and burn rate that is <2.1 mm/s.
2. Guanidinium nitrate (GN) is a suitable additive for red flare
compositions, when used in the range of 0-15 wt.-%. Resulting
mixture can reach up to 60% higher efficiency (Lsp) in comparison
to a similar GN-free composition.
3. Blue flame compositions with relatively high color saturation (pe
>50%) can be obtained without the use of arsenic-containing
compounds that have been used in the past, such as “Paris green”.
4. Blue strobe composition can be designed with the use of
aminoguanidinium nitrate as a strobe moderator.

11
 1. LITERATURE REVIEW

1.1. Pyrotechnic compositions

A pyrotechnic composition is meant to describe at least two component

containing heterogeneous solid mixture, that is capable of reacting in a self-
sustained exothermic reaction[1]. The reaction produces heat, light (glow,
flame, sparks), sound, smoke (physical and chemical) that has found its use in
many utilitarian fields[2]. By this definition, pyrotechnics are distinguished
from explosives and propellants, which are molecular compounds that have
fuel and oxidizer functional groups in the same molecule, for example, nitro
ester (oxidizer) groups in nitro glycerine.
The complexity of pyrotechnic reactions lies in the chaotic behavior that is
associated with them. There are three states of matter always involved in any
pyrotechnic reaction. First the reacting surface is pre-heated by the reacting
composition. Some components are melted before reacting. Then at the
ignition point decomposition occurs generating heat. Products are formed,
which can be gases, solids and liquids. In case the ionization potential is
reached, some elements get ionized as well. Both kinetic and thermodynamic
factors affect the reaction equilibrium.
The 20th century authors G.W. Weingart[3], B.E. Douda[4], A.A. Shidlovski[1],
J.H. McLain[5], A. Hardt[6], R. Lancaster[7], T. Shimizu[8–10], J.A. Conkling[11],
have contributed significantly with their books on pyrotechnics from
scientific, technical and artistic standpoints.
In more recent years, specific areas of pyrotechnics have been investigated.
One of which is illuminants.

1.2. Illuminants (White, Red, Green, Yellow)

Illuminating pyrotechnics stretches into a specific branch of such reactions

where the chemical reaction produces exclusively strong electromagnetic
radiation in comparison to other pyrotechnic reactions. The main two areas of
practical use are VIS radiation for signaling and entertainment (fireworks,
special effects) and IR for military countermeasures.
The sources of such strong emissions come from excitation of atomic
species, ionic species (plasma), molecules and incandescence emission by a
high temperature flame that is usually in the range of 1200 – 3600 °C.

12
 Illuminating compositions are mainly comprised of an alkali or alkaline
earth metal nitrate, which serves as both oxidizer and emitter source, a metallic
fuel such as Magnesium, Aluminum, Zirconium, and additional components
such as chlorinated organics, binders, gas generators, burn rate moderators,
stabilizers (Fig. 1).

Fig. 1. The main components of an illuminating flare composition.

Magnesium is an important component used since the late 19th century, but
it was not until the period between the World Wars that price and availability
enabled a more extended production and development of high temperature,
magnesium fueled, illuminating pyrotechnic products[12].
Before and during the Second World War products like hand-flares, signal
ammunition and parachute-flares were improved to a similar level of
efficiency as we see today. A lot of work was done to deepen the
understanding of the underlying phenomena during 1950s to 1960s[4], though
without affecting much of the products already in use. The main development
driving force in recent years has been focused on eliminating certain unwanted
ingredients and to a certain degree improving flame color depth[13–23].
Although having problems with corrosion induced by moisture[24] and
being a relatively expensive ingredient, Mg is absolutely indispensable for the
production of illuminating flares and the emission of high intensity colored
light in both military and civil applications. The reason for this is a couple of
unique chemical properties of Mg that distinguishes it from all other possible
fuels for high temperature colored light production in pyrotechnics. First of
13
all, the oxidation of magnesium produces large amounts of heat, which
directly relates to a high number of excited molecules in the flame [25]. As a
result, the flame is very luminous and spectral emitters like SrCl(g), SrOH(g),
BaCl(g), BaOH(g), Na(g) can produce flames that have high color saturation.
Secondly, Mg has a low vaporization temperature, 1107 °C, which is much
lower than its oxide’s decomposition temperature of ~3200 °C. This satisfies
the Glassman criteria for metal vapor phase combustion[26], which results in
extended vapor phase diffusion combustion zone, which allows for a large
radiative area.
Generally, fundamental demands for a good illuminating flare composition
are: a) burning with luminous flame with maximum candlepower; b) having a
distinct flame color of purity higher than 80% (e.g. red distress signal); c)
preferably creating a minimum of slag, d) being chemically stable and
containing chemicals of minimum toxicity. Slag formation is unwanted since
long-lived burning sparks and glowing slag is always a fire hazard and could
make a handheld flare dangerous to use and burn off the vanes to the canopy
in a parachute suspended flare. Even more important is that a flare producing
lots of slag is probably not delivering its best pyrotechnic performance
(maximum light output in the case of illuminants).

1.3. Blue signals

A pyrotechnic flame with a high color saturation or color purity (pe), is one of
the biggest challenges in pyrotechnics especially for the blue color (Fig. 2).
While several suitable pyrotechnic compositions can be formulated with a
reasonable effort for red and green at high color purity and high luminous
intensity (Iv), there are much fewer examples for blue illuminants available.

Fig. 2. A relatively dim blue flame appears above the burning surface of a 6
mm, 0.45 g blue firework star (pellet). Oxidizer: KClO4.

14
 An excellent historic perspective on colored flame development is given
in the thesis by Sturman[27]. Sturman discusses the evolution of colored flames
throughout the years, even from the period before the introduction of
potassium chlorate into pyrotechnic compositions.
One of the first spectroscopic investigations that reported an analysis of
blue flames was carried out by Barrow and Caldin[28]. Barrow and Caldin
identified the emitting species as CuCl(g). Interestingly, mercury(I) chloride
was used in these compositions as chlorine source.
Dr. Bernard E. Douda has contributed significantly to the understanding
of colored flames, that includes blue flames with copper[4].
A systematic study of compositions, which are generating blue flames,
was performed by Shimizu[29]. He used a self-made spectrograph containing a
sample holder, a 0.04 mm slit, water prisms, lenses and a photographic plate.
The proliferation of relatively low-cost spectrometers has been helpful
for studying pyrotechnic flames. Pyrotechnic compositions are assessed in a
"static" way, whereby the sample is burned, and the smoke is removed well
enough to ensure a free line of sight to the sample. In 2003, Brian Ingram
investigated the spectra of red, green and blue pyrotechnic flames [30].
Meyerriecks and Kosanke studied the principal emitters in colored flames [31].
They utilized solutions of various chemicals in combination with a nebulizer
and an oxygen/propane/acetylene flame.
Several papers have been published about the desired emitter in
pyrotechnic flames comprising a copper and a chlorine source. Dolata
proposed the formation of a trimer of CuCl[32]. This was quickly followed by
Sturman, who provided disproof of this hypothesis[33].
While static measurements are a sound and reproducible method for
characterizing compositions, such measurements are not representative of the
true environment in the most of applications. In actual use, pyrotechnic
compositions must burn at various airspeeds. Consequently, mixing of
ambient air will alter the flame stoichiometry and temperature. Furthermore,
pyrotechnic stars may extinguish in flight[25,34].
Few publications are available about "Round Robin trials" of spectral
color measurements. Douda's[35] review compares results from different
pyrotechnic flare testing facilities. He used 81 mm mortar flares as test objects.
The test tunnels were located at the former company E.H.D. Hendrickx in
Belgium, at the former government facilities at Fort Halstead, UK and at the
German armed forces facilities in Meppen, Germany.

15
 Recent efforts have been made to identify alternatives to the well-known
copper-chlorine based systems. Some of us have published work on copper(I)
bromide, and compared it to copper(I) chloride[13]. The work includes the 1931
CIE coordinates for the isolated spectra of CuCl and CuBr. In the same year
Koch compared all four copper(I) halides[15]. Koch has provided values of the
1931 CIE coordinates, as well as the dominant wavelength and color purity.
He has concluded that copper(I) bromide provided good efficiency, even
outperforming the classical copper(I) chloride. The copper(I) fluoride and the
copper(I) iodide-based system were found to be inferior to the bromide and
chloride. Koch briefly states that the experimental alternative blue
compositions are "unsuitable for technical use". The compositions were
optimized for the formation of K2SO4, to enable spectral measurements
minimizing the interference of potassium halides.
Recently, indium has been investigated. The first results do not indicate
that indium is, even from a purely technical point of view, a viable alternative
emitter[17].

1.4. Blue strobes

A strobe composition burns in an oscillatory manner with a smoulder reaction

occurring at all time and flash reaction occurring periodically (Fig. 3). The
smoulder reaction produces relatively a low amount of heat and forms a slag
at the burning front. When sufficient heat is generated, the semi-reacted slag
is ignited, and a flash occurs. Then the smoulder reaction continues further to
the depth of the pyrotechnic column[36].

Fig. 3. Combustion process of a 8 mm white strobe pellet. Flash frequency:

6Hz. Composition: Ba(NO3)2/KClO4/KNO3/MgAl (MX51 Ecka
werke)/GN/Polivinylbutyral/NC/Cu (200mesh atomized) in ratio of
56.3/4.7/4.7/26.4/2.5/1/2.5/1.9 (Courtesy of P. Alenfelt).

While colored strobes are created without significant difficulties using a

nitrate oxidizer or ammonium perchlorate (AP) combined with Mg, or
magnalium (MgAl) powder[37,38], a blue strobe is more difficult to make, since
16
the high temperature flash reaction involving oxidation of Mg or MgAl can
destroy the temperature sensitive CuCl(g) blue flame emitter.
Therefore, a blue strobe composition calls for a different chemical
composition. A few of such compositions have been described by Jennings-
White, who uses the AP/tetramethylammonium nitrate (TMAN)/Cu system[39]
and by McCaskie who described guanidinium nitrate (GN) based blue strobe
compositions[40]. Also, similar compositions as mentioned above have been
analyzed in our recent work[41]. However, TMAN is difficult to obtain and
possibly an expensive material. In our experience, such TMAN based
compositions have difficulties of sustaining combustion, which can lead to a
low wind resistance in practical applications.

17
 2. EXPERIMENTAL

CAUTION! The mixtures described herein are potential explosives, which are
sensitive to mechanical stimuli, such as impact, friction, heat, and electrostatic
discharge. Although we encountered no problems in the handling of these
materials, appropriate precautions and proper protective measures (safety
glasses, face shields, leather coats, Kevlar gloves, and ear protectors) should
be taken when preparing and manipulating them.
In previous work AGN was found to be extremely dangerous when mixed
with copper bromate, Cu(BrO3)2 and copper iodate, Cu(IO3)2. The sensitivity
to friction was extreme and small amounts of experimental mixtures self-
ignited during storage. Also, a side product of AGN reaction with BCC (black
in color, obtain under acidic conditions) had self-ignited while being wet on a
filter paper.

2.1. Chemicals

For flares, Rounded Mg powder 45-100 µm (max 5% >100 µm and max 5%

<45 µm, purity of 99.0-99.5%) and Sr(NO3)2 (99.5 % with <0.4 % barium
nitrate, <0.05 % calcium nitrate and <0.04 % sodium nitrate) were received
from Jengo S.P.A., Chlorinated rubber in the form of Pergut B10 (finer than
250 µm) from Bayer. Sr(NO3)2 was milled in a hammer mill until about 50 %
of the particles were smaller than 50 µm before the experiment. 2-
cyanoguanidine, Azodicarbonamide, Calcium resinate, Shellac, Strontium
oxalate, Lactose, Dextrin, Phenolic resin, Lignin sulfonate, GN, and Sulfur (all
very fine powders, purity 97-99%) were only passed through a 250 µm sieve
to break up lumps. Melamine, 5-aminotetrazole (5-AT), Urea and Hexamine
(purity 97-99%) were ground with a mortar and pestle to a fine powder and
particles bigger than 150 µm were removed with a sieve. The linseed oil used
was boiled type. The following formulas were applied for calculating oxygen
balance (ΩCO2) for variable Mr compounds: Linseed oil C57H92O6, Calcium
resinate C40H58CaO4, Shellac C16H26O4, Phenolic resin C53H45O7, Lignin
sulfonate C20H24O10CaS2, Chlorinated rubber C10H11Cl7.
For blue flames and strobes AP that was employed was synthesized in
Prof. Klapötke’s energetic materials research group (LMU, München) by
neutralizing perchloric acid with ammonia solution. A fine -50 mesh powder
of the fuel TMAN was used as received. Hexamine was purchased from
18
SIGMA ALDRICH. Tetramethylammonium perchlorate (TMAP) was
synthesized according the following protocol: perchloric acid (5.79 g, 0.0576
mol), was diluted with distilled water and tetramethylammonium hydroxide
(5.25 g, 0.0576 mol) was cooled down separately in an ice bath while stirring.
Merging the two solutions resulted in the flocculation of the white metathesis
product. Water was evaporated at 150°C to form a crystalline product which
was then placed in a drying cabinet. The fuels ethylenediammonium nitrate
(EDN) and tetramethylethylenediammonium nitrate (TMEDN) were prepared
as well by reacting ethylenediamine or tetramethylenediamine with an excess
of nitric acid in ethanol. The precipitate products were filtered and washed
with ether.
AGN was synthesized by reacting nitric acid with aminoguanidinium
bicarbonate. The final product was dried in a desiccator before use and water
solution of AGN had pH 7-8. Copper (II) benzoate was synthesized by
reacting potassium benzoate with Cu(II) chloride dihydrate in aqueous media.
Copper powder (<150 µm) was from Grüssing. Basic copper carbonate
(malachite), guanidinium nitrate and PVC powder were from Sigma Aldrich.
Nitrocellulose (NC) with a nitrogen content of 13.25% was from Nitrochemie
Aschau GmbH. Sulfur, lactose, dextrin, phenolic resin (BLIKAS LTD.,
Lithuania), and red gum were used as received. All chemicals used were
ground and sieved through a 50 mesh screen before conducting experiments.

2.2. Preparation of flare bodi es, pellets, rods

Three main types of test units were prepared.

1. 23 mm flare bodies (red flares)
2. 13-16.8 mm dry pressed pellets (for blue illuminants and strobes)
3. 5 mm rods (blue strobes)

For flares, the compositions were hand mixed, sieved through a 30 mesh
screen, cured for 4 h at 80 °C and then sieved again through 20 mesh screen to
break lumps. Test compositions were pressed into metal tubes (ID = 23 mm)
containing waxed paper liners (OD = 23 mm, ID = 21.6 mm). Hence, the
pyrotechnic composition was compacted into a column of ID = 21.6 mm and
70-85 mm in height (Fig. 4). Flares were pressed two at a time at 2 tons with
4 increments of 11.5 g (scale accuracy 0.05 g). 2 g charge of igniter
composition was pressed on top of each flare for ignition.

19
 Fig. 4. Preparation of 23 mm flare bodies: Top: Sieving, pressing, bottom:
composition & igniter (black), flare cut view.

The strobe compositions were mixed using a mortar and pestle (Fig. 5)
Pellets of 2 g (13 mm in diameter, 5-10 mm in height, ρ ~1.6 g·cm-3) were
pressed in one increment by a consolidation dead load of 2 tons. In some cases,
for TMAN based compositions, an igniter composition was pasted on top of
the pellet (~0.1 g).

Fig. 5. Preparation of 13 mm pellets: Weighing, mixing, storing, pressing.

For extruding of 5 mm diameter rods the following procedure was

followed: 2-4 g of composition was moistened with either MEK or Acetone to
swell or dissolve the binder material (PVC or NC). Then a 5 mm ID plastic
20
syringe with a cut end was used as a simple pump to press 10-25 mm long rods
(~0.4 g each, ρ ~1,15 g·cm-3). The syringe was filled with the moist strobe
composition and while facing a hard surface the plunger was pressed down by
hand to eject entrapped air and to consolidate the compositions as much as
possible. Finally, the compressed rod was extruded and left to dry overnight
at room temperature.

Fig. 6. Composition (50 g batches), pressing tool, pressed pellets and

composition 4 from up close.

Blue flame compositions were compacted into 16.8 mm pellets, 25-30 mm

height, ρ = 1.5-2.5 g·cm-3 using a hydraulic press (Fig. 6).

2.3. Spectrometer, lux meter, camera

For flares the experiment was performed in a wind tunnel to reduce smoke
obscuration. The wind speed in the chamber, where flares were burnt, was 1
m/s. Illuminance was measured with a calibrated lux meter (LMT, I(ph)–meter
1500, data processed with DasyLab software) and emission spectra were
recorded with a spectrometer (Stellar net, EPP 2000, CXR-SR; irradiance
calibrated for 300–1100 nm range). The detector-sample distance was 10 m.
The flares were also filmed with a video camera (Casio Exilim ex-f1) at
automatic settings for the evaluation of shape, spark and slag formation.
For strobes and blue flames measurements were carried out using a
HR2000+ES spectrometer with an ILX511B linear silicon CCD-array detector

21
controlled by software from OCEAN OPTICS. The integration time for
recording the emission spectra or strip charts were set according to the signal
intensity. It was in the range of 5 – 500 ms. The detector-sample distance was
0.5 - 1 m. At least three samples were measured for strobes and 3-5 samples
for blue flames. The performance of each composition was evaluated with
respect to flame color, smoke generation and the amount of solid residue.
The combustion process was filmed with a digital video camera recorder
(SONY, DCR-HC37E), Visario G2 1500 Weinberger speed camera for
filming at 1000 fps and SONY RX10 III for 100 fps and 500 fps, respectively.

2.4. Characterization

The DTA curves were measured with a 552-Ex differential thermal analyzer
from OZM at heating rates of 5 °Cmin-1. The decomposition points for TMAN
based strobe compositions were measured with a LINSEIS PT10 DSC
apparatus at heating rates of 5 °Cmin-1. The impact and friction sensitivities
were determined using a BAM drophammer and a BAM friction tester. The
ESD test was performed with Xspark 10 instrument from OZM. The
sensitivities of the compounds are indicated according to the U.N.
Recommendations on the Transport of Dangerous Goods (+): impact:
insensitive >40 J, less sensitive >35 J, sensitive >4 J, very sensitive 4 J;
friction: insensitive >360 N, less sensitive <360 N, very sensitive <80 N,
extreme sensitive <10 N.

22
 3. RESULTS AND DISCUSSION

3.1. Evaluation of the Mg -rich system for red illuminants

The first experiment was orientated to explore the effectiveness of the well-
known strontium nitrate, magnesium, and chlorine donor tri-component
system that contained a rather high percentage of Mg (40–60%). Typical red
flare compositions described in the literature contain around 30–40% of Mg
powder, 55–35% of Sr(NO3)2, 12–20% of chlorine donor and 4–10% of other
additives such as additional oxidizers, binders, stabilizers, burning rate
retardants etc. (Table 1). An amount of over 50 % Mg powder is more common
in yellow (white) flares based on NaNO3, where atomic Na(g) is the dominant
light emitter.

Table 1. Chemical compositions typically used in red flares.

Chemical composition (wt.-%)
Sr(NO3)2 43 55 48
KClO4 9
Mg (50-100 mesh) 29 30
Mg (30-50 mesh) 33
PVC 12 15 15
Laminac 4116/Lupersol 7 4
Source [2] [1] [22]

Having a high percentage of Mg in the composition of course requires

lowering the amounts of other components which, in turn, can result in lower
emitter concentration, more rapid combustion, and stronger incandescent
emission. It is, however, a potential approach to gain extra candlepower from
the flare composition.
A series of test compositions were prepared, see Table 2. Rounded Mg
powder was chosen over “cut” Mg for its ability to burn slower. An amount of
4% of linseed oil was added to each composition as a binder. Two flares of
each composition were pressed for evaluation.

23
Table 2. Chemical composition of 18 formulations tested in the first
experiment.
Compositon # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Sr(NO3)2 39 35 31 42 38 34 30 41 37 33 29 40 36 32 28 39 35 31

Mg rounded (45-100 µm) 40 44 48 40 44 48 52 44 48 52 56 48 52 56 60 52 56 60

Chlorinated rubber 17 17 17 14 14 14 14 11 11 11 11 8 8 8 8 5 5 5

Linseed oil 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4

Density [g·cm-3] 1.56 1.53 1.49 1.60 1.56 1.53 1.49 1.61 1.56 1.52 1.49 1.59 1.55 1.51 1.48 1.59 1.56 1.51

Burning time [s] 35.0 41.0 48.8 29.0 32.6 35.0 41.0 26.8 28.7 31.0 35.6 24.9 25.6 27.6 29.6 22.4 22.8 23.0

Burning rate [mm·s ] -1 2.24 1.94 1.68 2.63 2.40 2.28 2.00 2.84 2.72 2.59 2.31 3.08 3.08 2.92 2.78 3.44 3.44 3.51

Light intensity [kcd] 17.6 14.4 9.2 25.4 23.6 21.3 14.3 31.0 30.6 28.9 23.2 39.3 43.9 39.5 34.0 48.8 53.4 55.3

L(sp) [kcd·s·g-1] 13.4 12.9 9.8 16.0 16.7 16.2 12.7 18.1 19.1 19.5 17.9 21.3 24.3 23.7 21.8 23.7 26.4 27.7

2 -1
L(sp)/BR [kcd·s ·g ·mm ] -1 6.0 6.6 5.8 6.1 6.9 7.1 6.4 6.4 7.0 7.5 7.8 6.9 7.9 8.1 7.8 6.9 7.7 7.9

An important parameter that describes the performance of a pyrotechnic

illuminating composition is specific luminous intensity Lsp (Eq. 1).

𝐼𝑣 𝑡
𝐿(𝑠𝑝) (𝑘𝑐𝑑 ∙ s ∙ g −1 ) = Eq. 1
𝑚

Equation 1 presented in normalized units describes how much visible light

in kilo candelas is produced during one second from one gram of pyrotechnic
composition[16–18]. For Mg containing illuminants, this value is usually
reported to be in the range of 2-7 kcd∙s∙g-1 for green colored flares with barium
nitrate[42], around 15-50 kcd∙s∙g-1 for illuminating flares with sodium
nitrate[16,17,20] and around 14-16 kcd∙s∙g-1 for red colored flares with strontium
nitrate[21].
A drawback is that Lsp does not take into account burning area of the flare
and the geometry of the flame. Even though the choice of the pyrotechnic
composition, the ratio of its components and its density after compaction are
the major factors that determines size and shape of the flame envelope, both
emission and absorption take place in the envelope of the flame and absorption
is a function of distance (flame width). Also, heat loss is different for flares of
different diameters and, therefore, different shaped columns or pellets that
contain exactly the same composition will often have different Lsp values.
Thus, the results from this work can only be compared to other flares of similar
diameter and density.

24
 A pyrotechnic composition that possesses a high Lsp value is often desired,
since then a lower amount of pyrotechnic composition is needed to achieve a
certain effect, for example, a desired burning time of a parachute flare. And,
secondly, the pyrotechnic unit can be made smaller and lighter which in turn
saves space and volume for practical applications.
As a complement, the ratio of Lsp and burning rate (Lsp/BR, Table 2) was
used to give guidance in determining which compositions burned both bright
and slow.

Fig. 7. Lsp distribution plotted in a ternary diagram of measured compositions

with high Mg content. Small numbers at the end of each contour line
represent Lsp values.

The combustion of all 18 flares was monitored with a lux meter. Since the
distance between the photo diode and flare was known, a software (DasyLab)
was used to calculate and plot light intensity in kilocandelas (kcd) over time.
The Lsp values were calculated and their distribution was depicted in a ternary
diagram (Fig. 7). The trend observed was that there is a significant increase of
Lsp with decreasing amount of chlorinated rubber. The main purpose of
chlorinated rubber is to serve as a chlorine donor to enhance SrCl(g) emitter
formation, that supplements emission from SrOH(g). This shifts the flame color
from reddish orange to red. Additionally, the use of chlorine donors is one of
few possible ways to quench the MgO(s/l) incandescent emission via reaction
with Cl2(g) or HCl(g) that yields more volatile MgCl(g) [25]. The color purity of
chlorine-rich compositions 1-3 was measured to be 97-99% with dominant
wavelength (DW) = 613-614 nm, while chlorine-low compositions, 16-18,
had lower color purity of 79-87%, at DW = 605-609 nm.
25
 Moreover, it was determined that Lsp values and BR are directly connected.
Generally fast burning compositions result in more luminous flames with high
Lsp values. Distribution of BR of the same system is depicted in Fig. 8 for
comparison. However, from a practical perspective, high BR is only desired
in tracer compositions, whereas hand flares, parachute flares etc. must contain
a slow burning, yet luminous flame producing composition, which is the real
challenge.

Fig. 8. Lsp and flare’s light intensity dependence on burning rate for
compositions from Table 2.

3.2. The influence of additives

After the evaluation of 18 compositions, the composition 6 was chosen as a

base for further investigation, because it was slow burning, highly luminous
(determined by Lsp/BR value) and had a decent color purity of 90%. To
improve the compositions, it was considered essential to evaluate additives
that at least theoretically could enhance flame size by the formation of gas or
contribute positively by being an effective secondary fuel.
The use of gas generating materials has been reported back since the
1970s[44], and in recent years have been used in various illuminating
pyrotechnic compositions as an energetic additive for smoother combustion,
more luminous flames, and compositions emitting less smoke[19,20]. In the
present work several non-exotic, gas generating materials were chosen to
26
investigate how they affected the performance of the magnesium-rich system
In addition to the gas generating materials, calcium resinate, shellac, strontium
oxalate, phenolic resin, lignin sulfonate (with sodium impurities) and sulfur
were also tested as additives for comparison. Additives were evaluated in
“Composition 6”, which was modified by replacing 5% chlorinated rubber, by
5% additive (Table 3).

Table 3. Modified composition 6 for testing additives.

Compositions 19-33
Sr(NO3)2 34
Mg rounded (45-100 µm) 48
Chlorinated rubber 9
Additive 5
Linseed oil 4

Table 4. Performance parameters of compositions 19-33 with different

additives.
Comp. 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 6
Azodicarbonamide (ADCA)

Guanidinium nitrate (GN)

5-amino tetrazole (5-AT)

Chlorinated rubber
2-cyanoguanidine

Strontium oxalate
Calcium resinate

Additive
Lignin sulfonate
Phenolic resin

Hexamine
Melamine

Lactose

Dextrin
Shellac

Sulfur
Urea

C 29 29 21 75 68 14 40 44 80 45 14 20 10 51 0 29
Additive chemical consistency (organic

H 5 5 3 9 9 0 7 6 6 5 4 7 5 9 0 2

O 0 0 28 10 23 36 53 49 14 30 0 27 39 0 0 0
elements, %)

N 67 67 48 0 0 0 0 0 0 0 82 47 46 40 0 0

O,N amount [%] 67 67 76 10 23 36 53 49 14 30 82 73 85 40 0 0

Density [g·cm ] -3 1.62 1.65 1.69 1.63 1.62 1.71 1.66 1.63 1.62 1.64 1.65 1.76 1.67 1.64 1.68 1.53

Burning rate [mm·s-1] 2.39 2.51 2.23 2.27 2.17 3.11 2.26 2.31 2.15 2.63 2.56 1.99 2.39 2.40 2.92 2.28

L(sp) [kcd·s·g ] -1 17.4 18.3 19.7 8.8 10.7 20.3 17.6 19.2 10.9 22.5 21.0 20.1 22.7 15.3 17.9 16.2

Light intensity [kcd] 25.4 28.6 28.0 12.3 14.1 40.5 24.9 27.3 14.3 36.6 33.5 26.4 34.2 22.8 33.0 21.3

L(sp)/BR [kcd·s2·g-1·mm-1] 7.3 7.3 8.8 3.9 4.9 6.5 7.8 8.3 5.1 8.6 8.2 10.1 9.5 6.4 6.2 7.1

When testing compositions 19-33, it was noted that certain compositions

produced more luminous flames, while others burned very rapidly. And it was
clear that not only nitrogen-rich materials were successful. For example,
27
lactose and dextrin were almost as good as 2-cyanoguanidine and melamine
when comparing light intensity values, even though they did not contain any
nitrogen. Further on, experimental data was plotted in several ways to
determine which factor determines an efficient composition. Finally, a trend
was observed when plotting the Lsp values against the total amount of nitrogen
and oxygen (in moles or grams) in the additive used (Fig. 9; Fig. 10) and
against oxygen balance (Fig. 11).

Fig. 9. Lsp dependence on the amount of oxygen and nitrogen (in moles)
present in the additive.

Fig. 10. Lsp dependence on the amount of oxygen and nitrogen (in grams)
present in the additive.

28
 Fig. 11. Lsp dependence on the oxygen balance of the composition.

Several additives (calcium resinate, strontium oxalate, lignin sulfonate,

sulfur and chlorinated rubber) contained other elements besides C, H, N, O;
therefore, those were marked differently in the graphs. They cannot be
evaluated together with fully organic additives, since some contained metal
ions, i.e. Na (lignin sulfonate), Ca, Sr that are potential emitters or emitter
formation enhancers. Hence, some of them, notably lignin sulfonate and
strontium oxalate, show quite high Lsp values due to additional emitters
present.
The presented trends support the fact that high nitrogen materials favor the
production of more luminous flames[21], but it is evident that influence of
oxygen is equally as important. Apparently, formation of N2(g) and H2O(g) in
such Mg-rich flare compositions have a similar effect on the increasing Lsp.
A general explanation of the observed phenomenon is that the conversion
of chemical energy to light is more efficient in illuminating compositions that
contains gas generating materials[2]. When gas generating materials are
introduced into a composition instead of a non-gas generating material, such
as chlorinated rubber, a relatively large flame envelope can result due to gas
flame expansion. That, in turn, has a higher emissive area, which results in an
increased light emission intensity. Other effects could be the change in the
flame’s absorbance and emissivity factors that directly determine the Lsp.
When gaseous molecules replace less volatile or incandescent species, the
inner absorbance in the flame envelope is decreased resulting in a more
luminous flame.
29
 According to the NASA CEA[45] code estimation of compositions (6, 31,
38) at atmospheric pressure (1 atm), N2(g) is in equilibrium with Mg3N2 at
lower temperatures (1000-1500 K), however above 2000 K, N2(g) formation is
predicted. H2O(g) rapidly reacts with excess Mg(g) to yield H2(g) and MgO(s/l).
MgO(s/l) is then in equilibrium reaction with soot[9] that yields CO(g) and Mg(g).
Given that Mg-containing compositions reach temperatures above 2000 K[9]
and all water vapor (and oxygen) reacts with excess Mg, then N2(g), H2(g), CO(g)
molecules are formed and assist in increasing Lsp in such Mg-rich flare
compositions.

3.3. Influence of guanidinium nitrate on flare

performance

The performance of guanidinium nitrate (GN) in place of chlorinated rubber

was quite impressive. Having almost the same burning and mass consumption
rates, GN (comp. 31) possesses 40% higher Lsp value and 60% higher light
intensity in comparison to comp. 6 in this system. Urea was also very efficient,
and was noted for its low BR, see composition 30. However, urea is not useful
in practice as it is rather hygroscopic in combination with strontium nitrate
and could, therefore, be unstable together with Mg. Thus azodicarbonamide
(ADCA) is a great alternative (comp. 21), which is far less hygroscopic and
well performing in comparison to urea. 5-Aminotetrazole (5-AT) (and related
N-rich compounds) have received much attention these days for their unique
combination of high nitrogen content and high stability due to aromaticity[20].
Comp. 29 contained 5-AT and performed quite well, but with a slightly higher
BR than comp. 6.
Due to excellent performance as an additive, GN was chosen for further
investigation. Another advantage over other additives explored herein, is that
GN by itself is an energetic material and contains an oxidizing moiety – the
nitrate anion. Therefore, it can support combustion better in a strongly oxygen
deficient pyrotechnic systems such as the ones of our focus. For the same
reason, GN was sometimes used in the past as an oxidizer alone in illuminating
compositions[46] and gas generators[2]. Therefore, the Mg-rich illuminating
system could possibly tolerate even higher amounts than 5% of GN, which
would be difficult for other well performing fuel-like additives from Table 4.
In fact, tests of using ADCA in the 10% range were performed, but the
combustion process became uneven and pieces of burning composition were

30
shooting out of the flare chimney as it was burning. This phenomenon is likely
to occur when the gas generating material, present in the flare composition,
starts to decompose in the pre-reaction zone of the composition column. In
this case by “pre-reaction zone” we refer to the compacted flare composition
just a short distance below the combustion surface. As a result, a piece of solid
unburnt composition detaches and ejects through the flare chimney disturbing
the combustion performance.
Compositions were designed with a higher amount of GN (10-25 wt.-%
range). In this series of tests GN mainly replaced some of chlorinated rubber
and Sr(NO3)2, while Mg amount was kept in 40-50% range (Table 5). The
experiments proved that GN can be used in rather large amounts as an additive
in Mg-rich compositions for low BR and higher color purity and Lsp.

Table 5. Formulations with 10-25% of GN and their burning characteristics.

Compositon # 34 35 36 37 38 39 40 41 42 43

Sr(NO3)2 36 31 26 36 31 26 21 31 26 21

Mg rounded (45-100 µm) 40 40 40 45 45 45 45 50 50 50

GN 15 20 25 10 15 20 25 10 15 20

Chlorinated rubber 6 6 6 6 6 6 6 6 6 6

Linseed oil 3 3 3 3 3 3 3 3 3 3

Density [g·cm ]-3

1.71 1.67 1.64 1.69 1.66 1.62 1.58 1.63 1.60 1.57

Burning time [s] 39.0 47.8 54.4 32.7 41.0 50.4 57.0 33.0 43.5 54.9

Burning rate [mm·s ] -1

1.83 1.53 1.37 2.22 1.80 1.50 1.36 2.27 1.75 1.42

Light intensity [kcd] 17.9 14.5 11.8 25.5 21.0 16.1 13.5 30.9 22.1 15.4

L(sp) [kcd·s·g-1] 15.1 15.0 14.0 18.1 18.7 17.6 16.7 22.1 20.9 18.3

L(sp)/BR [kcd·s2·g-1·mm-1] 8.3 9.8 10.2 8.2 10.4 11.8 12.3 9.7 11.9 12.9

Flame color purity [%] 85 92 94 83 91 97 99 89 95 97

Spark formation* 0 1 1 0 0 1 2 0 1 2

* defines the relative amount of sparks generated outside the flame envelope.
0 – no sparks; 1 – some sparks; 2 - lots of sparks.

Sparks or glowing slag were observed to form during combustion only in

certain compositions. In the first experiments, compositions with the highest
Mg content had high burning rates of >3 mms-1. These compositions produced
fine sparks that were ejected upwards and were not observed falling down.
Some GN modified compositions from Table 5 were rather slow burning
(<1.8 mms-1). These tended to produce larger sparks or glowing slag pieces
that fell down. None the less, certain slow burning compositions possessed

31
high color saturation (39, 40, 43), and some others (37, 38, 41) burned
efficiently and clean (Table 5, Fig. 12)
When compositions from Table 5 were tested, an interesting phenomenon
was observed when evaluating the emission spectra. In Fig. 13 emission
spectra of compositions 37-40 are compared. The incandescence emission
seemed to have decreased with decreasing burning rate and increasing GN wt.-
%. The lower burning rate can affect the proportion of how much solid
particles remain deposited on the flare’s metal chimney and how much that
enters the flame. Also, lower burning rate can affect the kinetic parameters,
which determine how big spark particles are formed. Perhaps the explanation
behind comp. 40 performance (lots of sparks, yet a high purity red flame) is
that spark particles that formed were relatively big and not as hot to give strong
incandescence. Fine solid particles that easily can be dispersed and heated up
possibly emit stronger incandescence emission, which could explain
composition 37 performance (higher burning rate and lower color purity).
Additionally, low burning rate can cause more solid particles to be deposited
on the walls of the chimney, because reaction products travel slower from the
burning surface, which provides more time for solids to settle. That can also
cause the flame color purity to increase.

Fig. 12. Combustion of flares with composition 11 (A) producing some

sparks and glowing slag, and 38 (B) producing very few sparks.

Additionally, not a single composition from Table 4 shot out burning

pieces of solid composition, which was a rare, but possible case for
compositions 1-33. The uniform and slow burning is likely achieved due to
GN property to melt prior to decomposition, which seals the pores in the
composition, before generating gases and combusting.

32
 Fig. 13. Flame emission spectra of compositions 37-40. A decrease of
incandescence emission results in increasing color purity and a dominant
wavelength (DW) shift.

3.4. Properties of chlorine -free composition

By further tuning compositions with GN and 6% of chlorinated rubber, and an

inspiration from recent papers[14,22], a chlorine-free, Mg-rich composition was
evaluated. Even though it was similar to compositions 34-43, it possessed
slightly different spectral characteristics. Since SrCl(g) emission was absent,
the emission spectrum was solely composed of strong SrOH(g) emission
together with atomic Na(g), K(g) emissions (Fig. 14). Chlorine is known to
quench incandescence emission of solid particles such as MgO(s/l), SrO(s/l), by
reacting with them, and forming more volatile chlorides MgCl(g), SrCl(g).
Therefore, with absence of Cl, incandescence emission of the chlorine-free
composition was stronger and resulted in a decreased color purity of around
79-82%.
The final comparison of compositions that possess similar BR (in groups)
is presented in Table 6, together with the chlorine-free composition 44.

33
Fig. 14. Flame emission spectrum of chlorine-free, Mg-rich composition 44.
Na(g), K(g) peaks are present due to impurities.

Table 6. Comparison of two groups of compositions from this work, that

possessed similar burning/mass consumption rates and Cl free composition 44.
A B CCl -
Group
free
Comp. No. 6 21 31 3 39 44

Sr(NO3)2 34 34 34 31 26 34
Mg rounded (45-100) 48 48 48 48 45 48
Chlorinated rubber 14 9 9 17 6
Linseed oil 4 4 4 4 3 4

ADCA 5
GN 5 20 13

Magnesium stearate 1

Burning rate [mm·s ] -1

2.24 2.23 2.39 1.68 1.50 1.71
Lsp [kcd·s·g ]-1
16.7 19.7 22.7 9.8 17.6 16.2

Light intensity [kcd] 21.5 28.0 34.2 9.2 16.1 17.4

Lsp /BR [kcd·s ·g ·mm ]

2 -1 -1
7.4 8.8 9.5 5.8 11.8 9.5
Flame color purity [%] 89 88 87 92 97 80

34
 3.5. Blue illuminants

Pyrotechnic blue illuminant formulations were gathered from different

sources. Ten of them having the highest color saturation (purity) based on the
information in the literature were selected for this experiment. All the selected
blue flame compositions contain four main components: an oxidizer, a fuel, a
copper source and a chlorine source. The variety of chemical components was
another criterion for the compositions being as different one from another as
possible. This can provide more information about the best compositions and
the best chemical components for generating blue flames with a high color
saturation. The selected compositions are presented in Table 7 and their
combustion characteristics are given in Table 8.

Table 7. Experimental compositions: chemicals, ratios, sources.

No. Source Composition & ratio Ratio Comments
Optimized for
maximum
1 Hahma AP/Cu/CR/Hexamine/CP 62/14/4/10/10
[HCl+Cu]/[CuO]
formation
2 McGriffen[47] AP/Cu/Stearic acid/Paraffin 74/11/11/4 Ashless blue flare
3 Naud [48]
AP/5-AT/BCC 47.5/47.5/5 N-rich blue
Classic AP/Cu
4 Dumont[49,50]* AP/Copper benzoate/Dex 79/17.5/3.5
benzoate blue
Copper
KClO3/Paris green/Stearic
5 Hardt[6] 62/21/8/4/5 acetoarsenite-
acid/HCB/Dex
containing
KClO3/CuOCl/Lactose/CP/D Chlorate-lactose
6 Ofca[51] 65/13/13/5/4
ex blue
KClO4/CuO/Red Firework star with
7 Veline[52] 53/14/9/14/6/4
Gum/CR/MgAl/Dex MgAl
Chinese blue for
8 Stanbridge[53] KClO4/CuO/HCB/S/Dex 39/37/6.5/15/2.5
small pellets
KClO4/CuO/PVC/Hex/Red
9 Naud[54] 61/17/10/6/3/3 Naud Ref. blue
Gum/Dex
Perchlorate-
10 Pihko** KClO4/CuO/Hexamine/CR 62/13/10/15
Hexamine blue
Private communication: *Modification by Arno Hahma (2012), ** Petri Pihko
(1988)

50 g of each composition (except 25g of No. 5) were prepared. Five 10 g

pellets (16.80 mm diameter) were pressed out of each composition. The values
presented in Table 7 are an average over 5 parallel measurements. It must be
35
noted that even though pellets were ignited from the top, not all of them burned
evenly (cigarette burning). For compositions 5-10, that were based on
potassium chlorate and perchlorate, the surface flame propagation was more
pronounced than that to the depth of the pellet. For that reason, the recorded
burn rate data presented in Table 7 is not as precise as it was expected.
The first composition burned with a uniform tall flame, especially during
the first seconds. It was one of the very few compositions passing the 50% pe
threshold and had DW of 445 nm, which is the lowest of all 10 compositions.
This composition also possesses the highest Iv and Lsp values among AP based
compositions in this study.
The second composition was different from the first one due to wax and
stearic acid as fuels. Occasionally, large yellow spots appeared in the blue
flame envelope due to soot formation. This can be addressed to the largest
oxygen deficit resulting in a reduced pe. The burn rate was also the lowest
among the 10 compositions studied.
The third composition was unique because of its high nitrogen content,
which was achieved by using 5-AT as the main fuel. The poor fuel properties
of this compound resulted in a composition, which could not sustain
combustion on its own. However, the composition could be burned by holding
a glowing sparkler wire on the burning surface. It should be noted that this
composition is a derivative from the original presented by Naud at al[48].
The composition 4 was comprised of AP, which was an oxidizer and
chlorine source and copper benzoate, a fuel and a copper source. This well-
balanced composition (Ω = -5.4) passed the 50% pe threshold while its Iv and
Lsp values remained on the average level.
The composition 5 was unique due to the use of copper acetoarsenite. An
interesting note is stated by Shimizu[10,25]: “copper acetoarsenite was used in
Japan in almost all blue compositions in 1980’s as it produces a very pretty
blue”. Nowadays, such compositions are used less due to toxicity and
environmental concerns. This composition produces more smoke compared to
AP compositions (1-4) due to the presence of potassium, which creates solid
particles.
The composition 6 had the highest oxygen balance of Ω = +3.7. The
ingredient CP was difficult to distribute evenly throughout the composition,
which in addition to the positive Ω may have contributed to the lower
performance of the composition. Also, a decrease in color saturation was
observed during the combustion of each pellet.

36
 The composition 7 is a popular one in fireworks. The metallic fuel MgAl
increases the flame temperature and light output accordingly but results to a
reduced color saturation. With pe of 30%, the composition 6 had the highest Iv
and Lsp of the tested compositions.
The composition 8 had the highest pe of 55.7%. This composition is unique
for its high copper oxide content (37%), low potassium perchlorate content of
39% and for the absence of an energetic fuel such as wax, hexamine, etc.
Sulfur is used as the main fuel instead. Sulfur helps to scavenge potassium in
the flame resulting in an increased chlorine concentration promoting CuCl (g)
emitter formation[15,25]. HCB acts as a chlorine donor.
The composition 9 was chosen as a reference from Naud[54]. It was similar
to composition 10 with a similar pe of 41% It had only a slightly lower Lsp and
Iv values.
The composition 10 was an efficient blue flame composition yielding pe of
39% with Lsp of 414 cd·s·g-1. It also produced the largest amounts of glowing
slag on the test plate, where pellet was fixed.

Table 8. Experimental results: Oxygen balance (Ω), Burn time, average pellet length,
burn rate, TMD - theoretical maximum density TMD % (indicates the fraction of
measured density divided by TMD), Iv, Lsp, and color coordinates are presented. The
mass of all pellets was in the range 9.5-11.7 g, except composition 5, which was
limited to 4.8-5 g for each pellet. The white point was set at x = 1/3, y = 1/3.

l BR TMD TMD I Lsp pe DW CIE CIE

No. Ω [%] t [s] ± ± ±
[mm] [mm·s-1] [g·cm-3] [%] [cd] [cd·s·g-1] [%] [nm] x y

1 -20.2 17.5 23.2 1.3 1.98 96% 91 6.6 162 12.1 51.6 ±1.2 453.7 0.240 0.173

2 -26.3 42.5 26.0 0.6 2.01 85% 22 2.5 94 4.1 30.4 ±4.1 556.2 0.328 0.251

3 -15.1 27.5 32.7 1.2 1.75 83% 10 2.3 26 3.2 38.1 ±1 432.8 0.269 0.210

4 -5.4 12.5 26.0 2.1 1.75 95% 79 18.3 98 2.5 52 ±1.6 463.2 0.233 0.179

5 -12.7 9.0 11.3 1.3 2.17 90% 58 0.8 107 4.3 45.3 ±1.9 469.6 0.240 0.208

6 3.7 8.0 23.5 3.0 2.18 86% 46 3.2 37 1.8 28.6 ±3.4 462.8 0.278 0.250

7 -12.8 11.6 24.7 2.1 2.16 83% 388 27.6 454 22.9 29.8 ±2 464.1 0.276 0.246

8 -3.2 6.9 19.8 2.9 3.02 73% 108 21.7 75 8 55.7 ±2.6 467.9 0.221 0.175

9 -5.5 8.0 22.9 2.9 2.31 82% 267 19.7 218 15.9 40.8 ±1.6 477.3 0.242 0.240

10 -5.8 11.1 23.1 2.1 2.30 82% 365 48.4 414 35.8 38.6 ±0.2 481.1 0.242 0.261

37
 Fig. 15. Lsp (A) and Iv (B) dependence on the oxygen balance.

Fig. 16. The distribution of pe.

The Ω of all compositions lies in the range of -26 to +4%. The compositions
at the extreme ends (6 and 2) had low Lsp and Iv values, which may be related
to the unbalanced system. Too low Ω results in soot formation, while too high
Ω may cause the emitter species to be oxidized. The Lsp and Iv dependence on
Ω are depicted in Figure 15.
The measured pe lies in the range of 28 – 56% (Fig. 16). The lowest pe of
~30% was measured for compositions 2,6,7 and the highest exceeding 50%
for compositions 1,4,8. Deep blue flame compositions with high pe are usually
observed to burn with a relatively dim flame compared to the bright ones
having a low pe. In this work, composition 8 flame was indeed the least
luminous of the KClO4 compositions (107 cd, 74 cd·s·g-1). However, pe
reached 56% on average from 5 pellets, with one pellet reaching 57.7%. By
observing the pe vs time graphs some pellets of the composition 8 peaked at
38
65% pe during the first seconds of combustion, before the pe dropped to 50-
57%. Possibly the reduction of the color purity is associated with slag
formation on the burning surface. The composition 8 is designed to be used as
small stars in fireworks. Therefore, it possibly did not deliver the best
performance when burned as a 16 mm diameter, 10 g pellet.
The emission spectra were recorded for each of 10 compositions. The main
focus of interest was to observe how the CuCl(g) emissions compare to the
CuOH, CuO, Na and black body emissions (K atomic emission at 767 nm is
outside the visible spectrum). In figure 17A KClO3 compositions 5,6 are
compared. Composition 5 has a less pronounced grey body radiation and
strong CuCl(g) emissions at 400-470 nm; hence, it possesses a higher color
purity than composition 6. In Figure 17B, the spectra of compositions 8 and
10 are compared. The CuOH emission from composition 10 appears much
more pronounced compared to composition 8. This emission decreases the
color purity of the composition 10; however, the latter is much brighter
burning. In Figure 18, the raw spectra of composition 6 are depicted at selected
times. The intensity decreased as a function of time. This phenomenon was
not observed with other compositions, at least not to such a strong extent. This
effect may have resulted from both (i) the CP being unevenly distributed in
the composition and (ii) also the slag formation.
Finally, the CIE x/y coordinates plotted in the CIE 1931 color diagram (Fig.
19) are located in the blue and blueish-white region of the chromaticity
diagram. The compositions 1,4,8 show a noticeable shift towards the blue
region of the chromaticity diagram.
Video captures of the burning pellets are collected in Figure 20. From these
photos the flame size, shape, smoke and slag formation can be estimated. The
AP based compositions 1,2,3,4, tend to burn with little smoke and no slag
formation. The flame, especially in the beginning of the combustion, is tall
and narrow. Later on, it evolves to a shape seen in Figure 20 (1,2,3,4). The
KClO3 and KClO4 compositions (5-10) produce some slag and a significant
amount of smoke. Some of them had higher Iv, Lsp and pe values than AP
compositions, which are often considered superior blue flame compositions.
The reason for the higher light output can be associated with smoke reflection
during the measurements. If the smoke in the measurement chamber reflects
some of the emitted light from the flame towards the spectrometer, the
reflected light adds to the recorded data. This effect could not be completely
avoided, because it is not possible to extract the smoke at the edge of the flame
completely. On the other hand, the eye will also perceive this, and the
pyrotechnic composition will appear brighter than it actually is. In that sense,
no error is produced, when the back reflection is ignored.
39
 NASA CEA2 code was used to estimate the transient species present in the
flames of tested compositions (supplements). Only composition 5 was
omitted, as there was no data on arsenic species. The adiabatic flame
temperatures range from 1840 K for composition 8, to 2690 K for composition
10. Besides the high concentration of typical combustion products i.e. H2O,
CO2, CO, H2, N2, a substantial amount of HCl (0.15 mol-%) was produced for
AP compositions 1 – 4. Moreover, both KCl, HCl were present in KClO3 and
KClO4 based compositions 6 - 10. The target species for this experiment was
CuCl. While most compositions had around 0.03-0.05 mole-% of CuCl
produced, compositions 8 had 0.07 mole-% of CuCl, which is in relation to
the high performance of this composition described earlier. This can be
associated with very high concentration (37 wt.-%) of CuO used in the
compositions, that shifts the equilibrium towards CuCl formation.

A B

Fig. 17. Emission spectra of compositions 5,6 (A), 8,10 (B).

Fig. 18. The composition 6 pellet flame’s raw emission spectra recorded at
2.2, 4.7, 6.2, 8.4 s after ignition. A decrease in the intensity is observed. The
strongest emission in the blue region is observed after the ignition and the
least intense emission is observed in the last seconds of combustion process.

40
 Fig. 19. Zoomed in chromaticity diagram including the evaluated
compositions with the full diagram in the upper left corner.

Fig. 20. Video captures of each burning composition (3 s after ignition).

41
 3.6. Blue strobes

The idea of using a AP/GN base for blue strobes has been already investigated
by McCaskie[40]. The high percentage of GN requires a lot of energy to melt it
and, even with catalytic help of copper compounds, it decomposes slowly.
Therefore, GN based strobe compositions are possible; however, their
performance does not surpass TMAN based compositions. Some TMAN
based compositions have been tested containing 55% AP, 30% TMAN, 15%
of copper sources, that include Cu, CuO, Cu(II) benzoate, BCC, and up to
additional 2% of additives that are NC or sulfur. Such compositions possessed
medium sensitivity to friction (120-360 N) and shock (4-15 J). Such
compositions burn fiercely with a strobing behavior and a re-erupting blue
flame. However, when burnt as 13 mm pellets the flashes were overlapping
and some compositions did not sustain combustion. In addition to that, TMAN
is a rather expensive material for the use in pyrotechnics and fireworks;
therefore, AGN was suggested as a possible replacement that could surpass
GN in its reactivity and have an advantage over TMAN being more easily
available and cheaper option. A series of strobe compositions containing AP,
GN, AGN, different copper sources (Cu, CuO, Cu2O, BCC) were tested.
BCC was chosen over copper and copper oxides as it could produce a blue
flame flash of higher color purity and a better flash separation. However, CuO
can also work. PVC was also employed as recommended by E. McCaskie.
PVC seems to serve well as a low-energetic fuel that possibly assists the
smolder reaction[55]. After several experiments of testing different additives
and varying the oxygen balance of such compositions led to a narrow area of
a working blue strobe compositions. Such working compositions possessed a
distinct strobing behavior, sustainable combustion and a seemingly good blue
color of the strobe flashes. In Table 9 three different blue strobe compositions
(GN, AGN, TMAN based) are compared. In Fig. 21 and 22 their strobing
patterns are presented, that were obtained by monitoring the light intensity at
450 nm in time. Two types of samples were tested: 5 mm rods (0.4 g sample)
and 13 mm pellets (2 g). The 5 mm rod samples were made with an intention
of having a minimal energetic feedback, which could help lowering the strobe
frequency.

42
Table 9. Blue strobe chemical compositions of a GN base (A), AGN base (B),
TMAN base (reference). Strobe frequency, linear burn rate, chemical stability
and sensitivity parameters are present.
Compositions A B Ref.
AP 30 25 55
GN 50
AGN 55
TMAN 30
PVC 5 5
CuCO3·Cu(OH)2 (BCC) 15 15
Cu powder (40-100 mesh) 15
NC powder +1%
Frequency, Hz 13 mm 3.5±2 8±2 >10
5 mm rods const. 3.8±2 >10
Linear burning rate 1.4 1.1 5.2*
(5 mm rod), mm/s
CIE coordinates** X 0.262 0.249 0.236
Y 0.296 0.275 0.289
Color purity,** % 27 32 35
Chemical stability (DTA), ⁰C 247 179 245
Sensitivity tests Impact (J) 15 10 6
Friction (N) >360 >360 240
ESD (mJ) 42 33 51
*unusually high combustion rate resulted from the enhanced surface flame
propagation of this particular composition. The linear burning rate is expected
to be lower.
**light produced by the flash reaction

43
 1000
500
A - 5 mm (cont. burn) B - 5 mm f = 3,8 Hz Ref. - 5 mm f >10 Hz
2500

(a.u.), int.t. - 10 ms
800
400
2000

300 600
1500

200 400
1000
I [a.u.]

100
200 500

0 0
0

6000
[a.u.] int.t. - 20 ms

1600
A - 13 mm f = 3,5 Hz B - 13 mm f = 8 Hz Ref. - 13 mm f >10 Hz
4000
5000

1200
3000 4000

3000
800
2000
I (a.u.),

2000

400
1000
1000

0 0
0

0 1 2 3 0 1 2 3 0 1 2 3
Time [s]
(s)

Fig. 21. Light intensity at 450 nm vs time.

3600 B 13 mm

2400

1200
(a.u.)
I [a.u.]

900 B 5 mm

600

300

0 5 10 15 20 25 30
Time (s)
[s]

Fig. 22. Consistent over time strobing pattern of composition B tested as

a 5 mm rod and 13 mm pellet. From the ignition point the rod was consumed
in 25 s at a linear burn rate of 1,1 mm/s. Light intensity was registered at
450 nm.

44
 The composition A with GN produced the strobing effect (13 mm pellet),
even though the light intensity of the flashes was rather weak. The 5 mm rod
of A burned constantly with a very weak re-appearing blue flame that was not
registered by spectrometer.
The reference composition with TMAN burned with overlapping blue
flashes as seen in Fig. 21.
The B composition performed quite well producing stable and distinct (Fig.
22) blue flashes when tested as both 5 rods and 13 mm pellets. Due to good
performance as a blue strobe the composition B was analyzed further to
evaluate its chemical stability and strobe mechanism.

Fig. 23 Chromaticity diagram indicating color points of the flash reactions

of A, B and reference compositions.

The chromaticity diagram (Fig. 23) shows that the blue strobe light has a
color purity in the range of 27% to 35%, when the Illuminant C point is set as
a reference.

3.7 Humidit y test

Blue strobes are generally known to degrade over time due to possible ageing
reactions that can occur. For example, AP can react with metallic copper, salts
can undergo double exchange reactions, acid-base as well as complex
formation reactions are also possible.

45
 Therefore, a humidity test was performed in order to estimate how moisture
sensitive is the composition B at room temperature (Table 10). Each test was
performed in a desiccator with a saturated salt solution over a 3-day period.
Three samples of ~0.6 g were weighed with an analytical scale before and after
exposure to enhanced relative humidity.

Table 10. Exposure of composition B to different relative humidity generated

by saturated salt solutions at room temperature.
Salt K2CO3 NaBr NaCl KNO3
Relative humidity at 25 °C, % 43 57.6 75.3 93.6
Pellets Weight change 0 0.0004 -0.0005 -0.0246
Color change/cracks no no no yes*
Powder Weight change 0.001 -0.002 -0.0115 -0.0375
Greyish violet
Color change no (lite) Greyish violet deep purple**

*In 6 h purple spots appear

**In 3 h turns grey-violet, in 6h purple, after three days - deep purple

After exposure to RH 75%, the first significant change in sample’s weight

was registered. Moisture induces a reaction that releases gases, which causes
the weight loss and the color change.
The color change was also registered in several sample compositions that
were stored before pressing pellets. However, well dried compositions did not
cause such reaction for up to 2 months of storage at room temperature.
It must be noted that dry and acid-free AGN was used in our experiments
(water solution shows pH of 7-8 on universal indicator paper). In our
experience, compositions with non-acidic AGN have significantly greater
shelf life.
When 3 g of acid-free AGN is dissolved in water and 1 g of finely ground
BCC is added to the solution, the color of undissolved BCC changed first to
light violet and then to purple. Bubbling is observed. After stirring the
suspension and letting react overnight, 2.8 g of purple precipitate is formed,
which is likely to be a Cu aminoguanidinium complex. When dried and ignited
as loose powder, the complex material burns fiercely producing a green flame
that comes from CuOH(g) emissions. The chemical stability this complex
species is quite unclear, and as mentioned before, AGN and BCC reaction’s
side product had self-ignited one time during drying. Therefore, long term
stability tests at elevated temperatures are suggested before using in practice.

46
 3.8 Differential thermal anal ysis (DTA)
DTA measurement was performed for composition B and certain
compositions with AGN. It was found that BCC catalyses the decomposition
of AGN and significantly changes the DTA curve of AGN (Fig. 24). The B
curve is very similar to AGN/BCC with the same decomposition point at
179 °C, meaning that AP and PVC do not participate in this process.

Fig. 24. DTA curves of pure AGN, AGN/BCC (55/15 wt.-%. ratio) and
composition B.

Sometimes during the incremental heating of the DTA sample, the test
composition B (and very similar ones) showed a high-order deflagration
(likely at 238 °C) that were powerful enough to rupture the mini glass test tube
of the DTA instrument. The sample weight was ~40 mg.
Generally, the AP/TMAN based strobe compositions were thermally stable
up to 245 °C.

3.9 Origin of the bright blue flash

The high-speed camera videos give interesting information on the strobe

reactions observed in this work. There have been a few curious phenomena
registered that are described; however, it is somewhat difficult to draw a clear
conclusion from the latter that would help to explain the strobe reactions that
occur in AGN/AP/BCC system. However, they give some clues.
During the smolder phase gases are generated from the reaction surface
(Fig. 25). Around 50% of the smolder reaction time small blue-glowing lines
of gases appear at the reaction surface as well as micro blue flashes (Fig. 26A).
47
 Fig. 25. Smoldering surface of composition B 5 mm rod (A); 13 mm pellet
(B) (capture from the high-speed camera footage at 1000 fps). Heavy boiling
and gas generation is observed on the molten reaction surface. Whitish
boiling reactants can be observed in the center of both samples.

Fig. 26. High speed camera footage of the oscillatory burning B composition. A)
Glowing gases appear at the reaction surface at all time (500 fps); bright flame
spot appearing prior to the flash reaction can be seen in the fifth frame; in the
last frame flash reaction starts. B) and C) capture series of the occurring blue
flash at 100 fps and 500 fps respectively. The blue color is associated to the
emission of CuCl(g) and reddish tip to CuO.

48
 As the glowing gases eject from the burning surface, sometimes, especially
few milliseconds before the flash reaction, a significantly brighter flame spot
appears (Fig. 26A, the fifth frame) that travels upwards with the flow of gases.
Perhaps this is a region of higher temperature or a piece of molten reactive
composition being ejected from the pellet‘s surface, that has a delayed ignition
and burns to produce a significantly bright blue flash/spot, that is associated
to the same brightness and color of the flash reaction (Fig. 26A).
The flash reaction seems to be a result of (i) rapid exothermic reaction that
causes the ejection of flammable gases from the boiling surface or (ii) an
ignition of gases that have already been formed above the pellet (Fig. 26B).
Possibly it can be a combination of the two as well.
The first assumption is made from our previous work with TMAN strobes
and the evidence from the current DTA measurement, as the exothermic
decomposition at ~240 °C tends to be quite energetic often breaking the mini
test tube of the sample.
The second assumption followed observing the ash scaffold formation
(Fig. 27) and analyzing the high-speed camera footage. Most test samples
burned rather clean without significant ash formation, however, few samples
of a 5 mm rods burned leaving a thin scaffold of ash that was almost as tall as
the 20 mm long test sample (Fig. 27). Also, the ash scaffold was observed to
be glowing red during combustion, indicating the surface temperature of 600-
900 °C.

Fig. 27. The ash scaffold is formed due to a strong heat loss to the
surrounding atmosphere during the combustion of 5 mm rod (comp. B).
Partially red glowing ash scaffold in the last frame is observed.

Interestingly, being thin and fragile as the scaffold is, it did not fall apart
due the strobe reaction that would make one expect to have a certain pressure
fluctuation at the surface that is accompanied by the popping sound observed
49
every time when testing compressed and uncompressed B composition. For
example, a classical white strobe composition based on AP/MgAl/BaSO4 [10]
burns to produce flashes that are more similar to a small portions of flash
powder deflagrating. This produces pressure fluctuations, that can be seen
when such composition is compact into a cardboard casing, that eventually
burns off and the remaining’s of the paper casing are blown away by the
pressure waves of the flashes. However, in the case of rod, the ash scaffold
remained still, indicating that there are no significant pressure fluctuations at
the burning surface, what supports the second assumption.
Moreover, in Fig. 26B it can be seen, the erupted blue flash did not create
pressure that would blow away the smoke cloud that had been formed from
the beginning. This would indicate that the flame had spread through the
flammable semi-reacted smoke above the pellet, consuming it, however not
creating any significant pressure in the flame envelope. This also supports our
personal observations during measurements. Visually, the combustion of the
B comp. rod produced a rather steady flow of gases from the rod’s surface that
only flashed rapidly, and the flashes did not disturb the uniform flow of the
gases. Also, the popping sound was very distinct and observed all time for
tests of comp. B.

50
 CONCLUSIONS

1) Oxygen and nitrogen-rich additives were noted for their ability to

increase the specific luminosity (Lsp) of a Mg-rich flare composition. The
resulting flares can exceed the 22 kcd s/g efficiency, with a burning rate
of 2.1 mm/s. This way, pyrotechnic flares can be made brighter, with
more saturated color flames.
2) Guanidinium nitrate showed the best performance as an additive in the
tested system. While maintaining low BR, it produced 40% higher Lsp and
60% higher light intensity in comparison with a similar GN-free
composition. The preferred amount was in the range of 0-15 wt.-%.
Changes in flame’s emissivity and absorbance factors, as well as the total
emissive area, have probably had the greatest influence on the increase of
flare’s illuminating performance. Different from conventional
compositions, a rather small amount of chlorine donor (4-6 wt.-%) was
found to be sufficient to achieve a deep red flame when used in
combination with gas generating additives and high amount of Mg. Mg-
rich and Cl-free composition can also be obtained using GN as an
energetic additive. The resulting flame’s color purity is in the 80% range.
3) Even though highly regarded in older literature, Paris green-containing
composition did not surpass the other top compositions in color purity
nor in light intensity. While it contains both Cu and As, only copper
seemingly contributed to the emission spectrum and no As emission was
registered. While Paris green-containing composition 5 had purity pe =
45.3% and produced light of Iv = 58 cd and Lsp = 107 cd·s·g-1, a well-
balanced, As-free composition 10 produced much stronger emission of
Iv = 365 cd, Lsp 414 cd·s·g-1 and had a similar, perhaps only slightly lower
pe of 38.6%. The highest average purity (pe) of 55.7% (and up to 58% for
a single pellet) was observed from the composition 8 with 108 cd,
75 cd·s·g-1. The peak pe = 65% was registered during the first 3 seconds
of combustion. The dominant wavelength of this composition was 468
nm. KClO4/S system used in the composition 8 was found useful for
producing a blue flame with high purity. It also proves that KOH
continuum and K(g) do not interfere significantly with desired CuCl(g)
emissions.
4) AGN and TMAN were proved to be a suitable strobe moderator in
combination with AP and BCC for producing a blue strobe pyrotechnic

51
pellet. The strobe pellets were observed to be very sensitive to any
deviations of the composition stoichiometry and geometrical parameters
of the pellet. The amount of additive (NC, sulfur) was only tolerated in
1-2% range. Blue strobe composition in powder form showed sensitivity
to moisture at RH = 58-75%, and better resistance with RH = 75-94%
when the pellet was bound with PVC/MEK.

52
 REFERENCES

[1] A. A. Shidlovsky, Fundamentals of Pyrotechnics, CFSTI, 1965.

[2] H. Ellern, Military and Civilian Pyrotechnics, Chemical Publishing
Company, 1968.
[3] G. Weingart, Weingart’s Dictionary and Manual of Fireworks,
Chemical Publishing Company, 1937.
[4] B. E. Douda, Theory of Colored Flame Production, U.S. Naval
Ammunition Depot, Crane, Indiana, 1964.
[5] J. H. McLain, Pyrotechnics: From the Viewpoint of Solid State
Chemistry, Franklin Institute Press, 1980.
[6] A. P. Hardt, Pyrotechnics, Pyrotechnica Publications, Idaho, USA,
2001.
[7] R. Lancaster, R. E. A. Butler, T. Shimizu, Fireworks: Principles and
Practice, Chemical Publishing Company, 1998.
[8] T. Shimizu, Fireworks from a Physical Standpoint, Pyrotechnica,
Austin, 1989.
[9] T. Shimizu, Selected Pyrotechnic Publications of Dr. Takeo Shimizu
Part 3: Studies on Fireworks Colored Flame Compositions, Journal Of
Pyrotechnics Inc., 1999.
[10] T. Shimizu, Fireworks: The Art, Science, and Technique, Pyrotechnica
Publications, Austin TX, 1996.
[11] C. Mocella, J. A. Conkling, Chemistry of Pyrotechnics: Basic
Principles and Theory, Third Edition, CRC Press, 2019.
[12] C. E. VUONO, Military Explosives, Headquarters Department Of The
Army, Washington, D.C., 1984.
[13] D. Juknelevicius, E. Karvinen, T. M. Klapötke, R. Kubilius, A.
Ramanavicius, M. Rusan, Chem. - A Eur. J. 2015, 21, 15354–15359.
[14] J. Glück, T. M. Klapötke, M. Rusan, J. J. Sabatini, J. Stierstorfer,
Angew. Chemie Int. Ed. 2017, 56, 16507–16509.
[15] E.-C. Koch, Propellants, Explos. Pyrotech. 2015, 40, 799–802.
[16] J. Gluck, T. M. Klapötke, M. Rusan, J. Stierstorfer, Chem. Eur. J. 2014,
20, 15947–15960.
[17] J. Glück, T. M. Klapötke, T. Küblböck, Zeitschrift für Anorg. und Allg.
Chemie 2020, 646, 133–137.
[18] T. M. Klapötke, B. Krumm, M. Rusan, J. J. Sabatini, Chem. Commun.
2014, 50, 9581–9583.
[19] J. J. Sabatini, Propellants, Explos. Pyrotech. 2017, 43, 28–37.
[20] G. Steinhauser, T. M. Klapötke, Angew. Chem. Int. Ed. 2008, 47,
3330–3347.
[21] J. D. Moretti, J. J. Sabatini, J. C. Poret, Chem. Eur. J. 2014, 20, 8800–
8804.

53
[22] J. J. Sabatini, E. C. Koch, J. C. Poret, J. D. Moretti, S. M. Harbol,
Angew. Chemie - Int. Ed. 2015, 54, 10968–10970.
[23] E.-C. Koch, Propellants, Explos. Pyrotech. 2004, 29, 67–80.
[24] P. Alenfelt, Pyrotechnica 1995, XVI, 44–49.
[25] K. Kosanke, Pyrotechnic Chemistry, Journal Of Pyrotechnics inc.,
2004.
[26] I. Glassman, R. A. Yetter, Combustion, Academic Press, 2008.
[27] B. Sturman, The Chemical Transformation of Fireworks in the 19th
Century, Monash university, 2017.
[28] R. F. Barrow, E. F. Caldin, Proc. Phys. Soc. 1949, Section B,.
[29] T. Shimizu, Pyrotechnica 1980, VI, 5.
[30] B. Ingram, J. Pyrotech. 2003, 17, 1.
[31] W. Meyerriecks, K. L. Kosanke, J. Pyrotech. 2003, 18, 1–22.
[32] D. P. Dolata, Propellants, Explos. Pyrotech. 2005, 30, 63–66.
[33] B. Sturman, Propellants, Explos. Pyrotech. 2006, 31, 70–74.
[34] K. L. Kosanke, B. T. Sturman, R. M. Winokur, B. J. Kosanke,
Encyclopedic Dictionary of Pyrotechnics, Journal Of Pyrotechnics,
2012.
[35] B. E. Douda, Survey of Military Pyrotechnics NWSC/CR/RDTR-595,
1991.
[36] J. M. Corbel, J. N. van Lingen, J. F. Zevenbergen, O. L. Gijzeman, A.
Meijerink, Angew. Chem. Int. Ed. Engl. 2013, 52, 290–303.
[37] C. Jennings-White, in Pyrotech. Chem. (Ed.: B.J. Kosanke), Journal Of
Pyrotechnics Inc., Whitewater, 2004.
[38] R. Cardwell, Pyrotechnica 1982, V, 6–24.
[39] C. Jennings-White, Pyrotechnica 1992, 14, 33–45.
[40] E. McCaskie, Pyrotechnica 1993, 15, 35–45.
[41] D. Juknelevicius, A. Dufter, M. Rusan, T. M. Klapötke, A.
Ramanavicius, Eur. J. Inorg. Chem. 2017, 2017, 1113–1119.
[42] J. D. Moretti, S. M. Harbol, D. E. Riegner, N. Carlucci, J. J. Sabatini,
J. C. Poret, J. Energ. Mater. 2014, 32, 293–299.
[43] Hansson PyroTech AB, Historic Product Portfolio, Sweden, 1990.
[44] C. Bernardy, Illuminating Pyrotechnic Composition Which Generates
Gases, 1978, US4078954.
[45] B. J. McBride, S. Gordon, NASA CEA, Glenn Research Center,
Cleveland, Ohio, 2015.
[46] C. Jennings-White, Pyrotechnica 1993, XV, 23–28.
[47] W. McGriffen, J., Ripley, Investigation of Visibility and Formulation
of Ashless Blue Flare, Naval Ammunition Depot Crane In., 1962.
[48] D. Chavez, M. Hiskey, D. Naud, J. Pyrotech. 1999, 10, 17-34.
[49] F. Drexler, K. Basse, B. Brauer, Pyrotechnic Light-Emitting
Composition, Useful e.g. in Fireworks, Contains Aromatic Caboxylic
Acid Salt, e.g. Strontium or Potassium Benzoate, as Color Former Also
Acting as Fuel, 2005, DE102005053849A1.
54
[50] J.-L. Dumont, Pyrotechnic Compositions for the Production of
Coloured Fireworks, 1986, EP0252803B1.
[51] B. Ofca, Bill Ofca’s Technique in Fire, Mastering Cut Stars the Easy
Way, Vol. 5, B & C Products, Hyde Park, NY, 1994.
[52] R. Veline, Self Publ. 1989.
[53] M. Standbridge, Self Publ. 2007.
[54] D. L. Naud, M. A. Hiskey, D. E. Chavez, Zeitschrift Fur Anorg. Und
Allg. Chemie 2013, 639, 702–706.
[55] T. Shimizu, Pyrotechnica 1982, 8, 5–28.

55
 SUPPLEMENTS

1. Chemical equilibrium calculations

The chemical equilibrium calculations were performed for compositions

from table 7 with NASA CEA code using „hp“ problem at p = 1.013 bar,
t = 2000 K.
Note: Composition 5 was impossible to model as the program lacks As species
in the database. Also ∆Hf of Paris Green was not found.
Heat of formation (∆Hf) values and formulae used for calculations:

Ingredient Formula ∆Hf (kJ/mol)

5-aminotetrazole CN5H3 324
Ammonium perchlorate NH4ClO4 -295.8
Chlorinated paraffin (64% Cl) C15.5H12.25Cl8.75 -765*
Chlorinated rubber C10H11Cl7 -395*
Copper Cu 0
Copper benzoate CuC14H10O4 -682
Copper carbonate, basic (malachite) Cu2CH2O5 -1067
Copper oxychloride Cu4Cl2H6O6 -1657
Copper(II) oxide CuO -156
Dextrin C6H10O5 -948.4†
Hexachlorobenzene C6Cl6 -131
Hexamine C6H12N4 124.1
Lactose monohydrate C12H24O12 -2519
Magnalium Al12Mg17 -3.7± 0.1
Paraffin C20H42 -682.5
Polyvinyl chloride C2H3Cl -94.6
Potassium chlorate KClO3 -398
Potassium perchlorate KClO4 -433
Red gum C30H30O13 -550**
Stearic acid C18H36O2 -947.7
Sulfur S 0

† values for starch

* estimated by Barry T. Sturman.
** estimated by DJ.

56
Compositions:

No. Composition Ratio

1 AP/Cu/CR/Hexamine/CP 62/14/4/10/10
2 AP/Cu/Stearic acid/Paraffin 74/11/11/4
3 AP/5-AT/BCC 47.5/47.5/5
4 AP/Copper benzoate/Dex 79/17.5/3.5
6 KClO3/CuOCl/Lactose/CP/Dex 65/13/13/5/4
7 KClO4/CuO/Red Gum/CR/MgAl/Dex 53/14/9/14/6/4
8 KClO4/CuO/HCB/S/Dex 39/37/6.5/15/2.5
9 KClO4/CuO/PVC/Hex/Red Gum/Dex 61/17/10/6/3/3
10 KClO4/CuO/Hexamine/CR 62/13/10/15

57
Calculations:

1 2 3
T, K 2320 T, K 2230 T, K 2690
MOLE MOLE MOLE
FRACTIONS FRACTIONS FRACTIONS

*CO 0.15548 *CO 0.17812 *CO 0.10709

*CO2 0.07437 *CO2 0.06787 *CO2 0.03205
*Cl 0.00392 *Cl 0.00148 *Cl 0.00925
*Cu 0.01353 *Cu 0.01086 *Cu 0.00681
CuCl 0.04957 CuCl 0.03253 CuCl 0.00398
CuO 0.00001 CuOH 0.00003 CuO 0.00001
CuOH 0.00004 Cu2 0.00003 CuOH 0.00002
Cu2 0.00003 *H 0.00272 *H 0.01831
*H 0.00355 HCl 0.12421 HCl 0.08358
HCl 0.17086 *H2 0.16456 *H2 0.11442
*H2 0.11057 H2O 0.33752 H2O 0.22921
H2O 0.29951 *NO 0.00003 *NO 0.00131
*NO 0.00009 *N2 0.0791 *N2 0.38207
*N2 0.11662 *O 0.00001 *O 0.00094
*O 0.00003 *OH 0.00091 *OH 0.01015
*OH 0.00177 *O2 0.00001 *O2 0.00076
*O2 0.00004

58
 4 6 7
T, K 2630 T, K 2060 T, K 2690
MOLE MOLE MOLE
FRACTIONS FRACTIONS FRACTIONS

*CO 0.09624 *CO 0.0033 AlCl 0.00001

*CO2 0.1661 *CO2 0.31845 AlOCl 0.00001
*Cl 0.02549 *Cl 0.00126 AlOH 0.00002
ClO 0.00003 *Cu 0.0035 *CO 0.25047
Cl2 0.00003 CuCl 0.05081 *CO2 0.16319
*Cu 0.00481 CuO 0.00007 *Cl 0.01077
CuCl 0.01125 CuOH 0.00012 ClO 0.00001
CuO 0.00004 Cu2 0.00001 *Cu 0.04349
CuOH 0.00003 Cu3Cl3 0.0001 CuCl 0.03124
*H 0.00778 *H 0.00006 CuO 0.00017
HCl 0.15268 HCL 0.01925 CuOH 0.00013
HOCl 0.00001 *H2 0.0007 Cu2 0.00009
HO2 0.00001 H2O 0.32143 *H 0.01001
*H2 0.03239 *K 0.0002 HCl 0.05631
H2O 0.36597 KCl 0.23203 *H2 0.03624
*NO 0.00262 KO 0.00001 H2O 0.15792
*N2 0.09345 KOH 0.00342 *K 0.00566
*O 0.00323 K2Cl2 0.00155 KCl 0.15482
*OH 0.02301 *O 0.0002 KH 0.00001
*O2 0.01482 *OH 0.00356 KO 0.00008
*O2 0.03995 KOH 0.00275
K2Cl2 0.00007
*Mg 0.00449
MgCl 0.00235
MgCl2 0.00908
*MgO 0.00127
MgOH 0.00059
Mg(OH)2 0.00069
*O 0.00194
*OH 0.01206
*O2 0.00341
MgAl2O4(l) 0.02085
MgO(cr) 0.01981

59
 8 9 10
T, K 1840 T, K 2430 T, K 2510
MOLE MOLE MOLE
FRACTIONS FRACTIONS FRACTIONS

*CO 0.00044 *CO 0.08588 *CO 0.11412

*CO2 0.17012 *CO2 0.27812 *CO2 0.23538
*Cl 0.00067 *Cl 0.00303 *Cl 0.00959
*Cu 0.00055 *Cu 0.04291 ClO 0.00001
CuCl 0.06493 CuCl 0.0489 Cl2 0.00001
CuO 0.00001 CuO 0.00026 *Cu 0.0232
CuOH 0.00001 CuOH 0.0003 CuCl 0.04582
Cu3Cl3 0.00848 Cu2 0.00021 CuO 0.00013
HCl 0.01108 *H 0.00211 CuOH 0.00013
*H2 0.00003 HCl 0.0265 Cu2 0.00005
H2O 0.0514 *H2 0.01306 *H 0.00344
*K 0.00001 H2O 0.25389 HCl 0.06661
KCl 0.19999 *K 0.00325 *H2 0.01678
KOH 0.00031 KCl 0.18185 H2O 0.21576
K2Cl2 0.00445 KO 0.00007 *K 0.00208
K2SO4 0.00001 KOH 0.00564 KCl 0.18507
*O 0.00002 K2Cl2 0.0002 KO 0.00004
*OH 0.00034 *NO 0.00079 KOH 0.00244
*O2 0.01204 *N2 0.03678 K2Cl2 0.00015
SO 0.00009 *O 0.00084 *NO 0.00115
SO2 0.34724 *OH 0.00867 *N2 0.05998
SO3 0.00038 *O2 0.00674 *O 0.00125
Cu2O(L) 0.1274 *OH 0.01025
*O2 0.00656

60
2. Emission spectra

These spectra are the sum of all spectra recorded for one pellet (out of five)
during combustion. The background is subtracted.

61
62
 SANTRAUKA

Įžanga ir naujumas

Pirotechniniu mišiniu yra apibūdinamas bent dvikomponentis kietafazis

heterogeninis mišinys, kuris geba reaguoti save palaikančia egzotermine
reakcija. Reakcijos metu išsiskiria sąlyginai didelis kiekis šilumos, taip pat
šviesos (liepsna, kibirkštys), garso, dūmų.
Pirotechninių mišinių sudėtingumas yra susijęs su jų chaotiškumu.
Faktiškai kiekvienos pirotechninės reakcijos metu išsiskiriant dideliam kiekiui
šilumos, yra stebimos visos trys agregatinės reagentų ir reaktantų būsenos,
kartais netgi plazma. Prieš užsidegant pirotechniniam mišiniui, dalis jo
paviršiuje esančių komponentų išsilydo. Tada, vykstant cheminei degimo
reakcijai, susidaro produktai: dujos, kietosios dalelės, skysčio lašeliai.
Įvertinant tai, kad temperatūra gali svyruoti drastiškai nuo maždaug 300 °C iki
3500 °C, tampa sudėtinga modeliuoti ir matuoti tokio tipo mišinius. Svarbiausi
registruojami tokių eksperimentų parametrai yra:
 Linijinis degimo greitis (mm/s)
 Liepsnos emisijos spektras (UV, VIS, IR)
 Šviesos stipris (cd)
 Specifinis šviesos stipris (cd·s/g)
 Parametrai registruojami filmuojant:
o Šlako susidarymas
o Kibirkštys
o Liepsnos forma, atmosferinio deguonies difuzija
o Degimo paviršius
o Degimo profilis (intensyvumo, šviesumo pokyčiai)
o Garsas
20-ojo amžiaus autoriai G.W. Weingart[3], B.E. Douda[4], A.A.
Shidlovski[1], J.H. McLain[5], A. Hardt[6], R. Lancaster[7], T. Shimizu[8–10], J.A.
Conkling[11] reikšmingai prisidėjo savo knygomis prie pirotechnikos
suvokimo iš mokslinės, techninės ir meninės perspektyvos.
Paskutiniais metais labiausiai buvo nagrinėjamos specifinės pirotechnikos
sritys t. y.: apšviečiamieji mišiniai, spalvotos liepsnos, pul-suojančio degimo
pirotechniniai mišiniai, IR spinduoliai, molekulės su dideliu azoto kiekiu,
termitiniai mišiniai.
Šiame darbe buvo tiriami apšviečiamieji mišiniai, spalvotos liepsnos ir
pulsuojančio degimo pirotechniniai mišiniai.
63
 Apšviečiamoji pirotechnika

Tokio tipo pirotechniniai mišiniai pasižymi specifine savybe: degdami

išskiria išskirtinai didelį kiekį spinduliuotės regimojoje arba UV, IR spektrų
dalyse. Tai pritaikoma kariniams tikslams, pvz., apšviesti karo lauką arba IR
liepsnoms, kurios naudojamos priešraketinėje gynyboje.
Dažniausiai, bet ne visada, intensyvi spinduliuotė pirotechniniuose
užtaisuose yra susijusi su aukšta liepsnos temperatūra, kuri padeda sužadinti
kuo didesnį kiekį atomų arba molekulių.
Tokio tipo mišinių pagrindinės sudedamosios medžiagos yra šarminių arba
šarminių žemių metalų nitratai, pvz., Sr(NO3)2, NaNO3, kurie veikia kaip
oksidatoriai ir spinduolių šaltiniai. Taip pat įeina metalinis kuras (dažniausiai
Mg, Al, Zr) bei mažesni kiekiai chlorintų organinių junginių (pvz., chlorintos
gumos) SrCl, BaCl spinduolių emisijai sužadinti, rišiklių (epoksidai), degimo
greičio modifikatorių (stearino rūgštis), stabilizatorių (virtas linų aliejus).
Nors ir turintis gana prastą atsparumą drėgmei, magnis yra nepakeičiamas
kuras tokio tipo užtaisuose. Jis degdamas su oksidatoriumi sukuria labai
aukštą liepsnos temperatūrą, dėl ko pasiekiama maksimali spinduolių šviesos
emisija.
Svarbiausi kriterijai geram UV-VIS liepsnos mišiniui yra: a) degti
maksimaliai ryškiai, b) turėti sodrią liepsnos spalvą, kurios sodrumas būtų
>80 % (pvz., raudonasis pavojaus signalas jūrai), c) degant išskirti kuo mažiau
šlako, d) būti chemiškai stabiliam ir mažai kenksmingam.

Spalvotos liepsnos mišiniai

Tokio tipo mišiniai yra labai panašūs į apšviečiamuosius, tačiau, jiems

svarbesnis faktorius už liepsnos ryškumą yra spalvos sodrumas. Tokio tipo
mišiniai yra taikomi fejerverkuose. Prie jų priskirtini ir mėlynos liepsnos
mišiniai, kurių liepsnos sodrumas, deja, nėra didelis (100-500 cd; 30-65 % pe).

Pulsuojančio degimo mišiniai

Pulsuojantis degimas pirotechnikoje yra apibrėžiamas kaip pirotechninio

mišinio degimas, kurio metu šviesa išsiskiria ne pastoviai, o cikliškai ir
dažniausiai pastoviu periodu viso degimo metu. Taip degimo metu stebimas
liepsnos mirksėjimas, kurio metu organoleptiškai galima išskirti atskirus
blyksnius ir tarp jų esančią „tamsiąją“ fazę.
64
 Šie mišiniai turi panašumo į apšviečiamuosius, nes jų degimo metu stebimi
blyksniai yra ne ką mažiau ryškūs už apšviečiamojo mišinio liepsną.
Galima išskirti kelis tipus literatūroje aprašytų pulsuojančio degimo
mišinių. a) NH4ClO4 / Mg pagrindu, b) Ba(NO3)2 / MgAl pagrindu, c)
NH4ClO4 / guanidinio nitrato / vario katalizatoriaus pagrindu.
Iš „a“ tipo mišinių nesudėtinga išgauti spalvotus blyksnius, kurie yra vieni
patraukliausių efektų fejerverkuose. Tačiau mėlynos spalvos pulsuojantis
mišinys yra labai sudėtingas, ir šiuo metu nerasta tokia sudėtis, kuri būtų gerai
veikianti ir chemiškai stabili. „c“ tipo mišinys yra iš esmės skirtas mėlynai
pulsuojančiai liepsnai išgauti.

Tikslas
Pagerinti apšviečiamųjų pirotechninių mišinių specifinį šviesos stiprį, ne-
prarandant liepsnos spalvos sodrumo bei mišinio cheminio stabilumo.

Uždaviniai

Šiai daktaro disertacijai buvo suformuluoti uždaviniai:

a) Ištirti ir nustatyti, kokį Mg kiekį gali toleruoti apšviečiamieji
pirotechniniai mišiniai, išlaikydami mažą (<2,1 mm/s) degimo greitį ir
neprarasdami liepsnos spalvos sodrumo bei šviesos intensyvumo. Ištirti
Sr(NO3)2/Mg/CR trikomponentę sistemą stebint šviesos intensyvumą,
efektyvumą, sordrumą. Ištirti priedus, kurie galėtų pagerinti mišinio
savybes.
b) Rasti tinkamiausią priedą ir jį ištirti, bandant didesnius jo kiekius.
Įvertinti, kokius priedo kiekius mišinys gali toleruoti, degdamas
stabiliai.
c) Optimizuoti mėlynos liepsnos mišinius. Apžvelgti geriausius mėlynų
liepsnų mišinius, jų įvairovę, specifines savybes ir medžiagas, kurios
padeda išgauti sodriausią liepsnos spalvą. Palyginti geriausius mišinius
matuojant jų liepsnų šviesos intensyvumą, efektyvumą, spektrą, CIE
1931 koordinates, sodrumą ir dominuojantį bangos ilgį.
d) Rasti chemiškai stabilų pulsuojančios mėlynos liepsnos mišinį.
Atkartoti jau paskelbtus tokio tipo mišinius. Išbandyti kitus pulsavimo
moderatorius. Išbandyti kitus vario katalizatorius.

65
 Ginamieji teiginiai

1. Naudojant apavalių grūdelių magnio miltelius ir kombinuojant mišinį

kartu su dujas generuojančiu komponentu, galima sukurti itin efektyvų
(>22 kcd·s/g) pirotechninį mišinį, kuris dega <2,1 mm/s greičiu.
2. Guanidinio nitratas yra tinkamas priedas apšviečiamiesiems mišiniams
(naudojant 0-15 %), padedantis padidinti efektyvumą (Lsp) net iki 60 %.
3. Pulsuojančios mėlynos liepsnos mišiniui gali būti naudojamas
aminoguanidinio nitratas.
4. Mėlynos liepsnos mišiniai su santykinai dideliu spalvos sodrumu
(pe >50 %) gali būti sukurti nenaudojant tokių arseno darinių kaip
„Paryžiaus žaluma“.

Eksperimentinė dalis

Šiems tyrimams naudoti reagentai buvo laboratorinio grynumo, išskyrus

kai kurias sunkiau prieinamas medžiagas, kurios buvo susintetintos arba
gautas techninio grynumo mėginys.
Mišinių ruošimui reagentai buvo sumalti iki tokio smulkumo, kad
nesunkiai būtų prasijojami per 0,6 mm sietelį. Kai kuriais atvejais buvo
aktyvuojamos rišamosios medžiagos tam, kad mišinys neišsisluoksniuotų jį
beriant arba presuojant. Kai kuriais atvejais buvo užtepamas dengiamasis
mišinys tam, kad mėginiai lengviau užsidegtų. Jo sudėtis buvo a) 75 % KClO4,
20 % beržo anglies ir 5 % šelako b) 65 % CuO ir 35 % Mg c) 80 % KNO3,
15 % anglies ir 5 % chlorintos gumos.

Buvo naudojami trijų tipų mėginiai:

1. 40 g mišinio, ir 2 g dengiamojo mišinio, supresuoti į 21,6 mm metalinę
tūtą (aukštis ~70 mm);
2. 13-16,8 mm sausos tabletės (mėlynoms liepsnoms ir pulsuojančiam
degimui), supresuotos iš 2-10,5 g mišinio (aukštis 5-30 mm);
3. 5 mm strypeliai (0,5 g mišinio, 25 mm aukščio).

40 g užtaisai buvo tiriami reguliuojamo greičio vėjo tunelyje, o

mažesnieji – traukos spintoje. Mėginiai buvo uždegami dujiniu degikliu arba
dagtimi. 40 g užtaisai buvo padegami nuotoliniu būdu, naudojant trintimi
aktyvuojamą degiklį.

66
 Šviesos intensyvumas buvo matuojamas šviesos stiprio matuokliu LMT,
I(ph)–meter 1500, arba kalibruotu OceanOptics spektrometru JAZ ULM,
kuris kartu registravo ir emisijos spektrus. Atstumas nuo mėginio iki
detektoriaus buvo nuo 0,3 m mažiems mėginiams (2-5 g) iki 10 m 40 g
užtaisams. Integravimo laikas buvo parenkamas pagal signalo stiprumą.
Degimo proceso analizei mišinys buvo filmuojamas su SONY DCR-
HC37E kamera. 1000 kadrų per sekundę greičio filmavimui buvo panaudota
Visario G2 1500 Weinberger „speed camera”, o SONY RX10 III buvo
naudojama filmuoti 100 ir 500 kadrų per sekundę greičiu.
Be degimo proceso registravimo, geriausiai veikiantys mišiniai arba jų
svarbūs komponentai buvo papildomai tiriami. DTA analizė buvo atlikama
552-Ex diferenciniu terminiu analizatoriumi iš OZM, kaitinimo sparta 5 °C
min-1. Jautrumas smūgiui ir trinčiai buvo matuojamas BAM krentančiu kūju ir
BAM trinties testeriu. Atsparumas elektrostatinei iškrovai buvo matuojamas
su Xspark 10 instrumentu iš OZM. Jautrumai buvo vertinami pagal „Jungtinių
tautų rekomendacijas pavojingų medžiagų transportavimui“:

smūgiui:
 nejautrus >40 J,
 mažai jautrus >35 J,
 jautrus >4 J,
 labai jautrus <4 J;

trinčiai:
 nejautrus >360 N,
 mažai jautrus <360 N,
 labai jautrus <80 N,
 itin jautrus <10 N.

Raudonų liepsnų mišiniai modifikuoti panaudojant dujas

generuojančius priedus su dideliu Mg kiekiu

Pirmojo eksperimento tikslas buvo ištirti gerai šių tyrimų srityje

pažįstamos Sr(NO3)2/Mg/PVC sistemos efektyvumą, kuri turėtų ypač didelį
kiekį Mg (40-60 %). Literatūroje aprašomus mišinius sudaro 30–40 % Mg,
55–35 % Sr(NO3)2, 12–20 % chlorintų organinių darinių, 4–10 % kitų priedų
kaip oksidatoriai, rišikliai, stabilizatoriai, degimo greičio modifikatoriai.
67
 Mg kiekio mišinyje didinimas gali turėti ir neigiamą įtaką kitiems liepsnos
parametrams, tokiems, kaip: a) sumažėjusi spinduolių koncentracija, b)
padidėjęs degimo greitis, c) padidėjęs foninis švytėjimas, kuris sumažina
liepsnos spalvos grynumą.
Eksperimento metu buvo ištirta 18 skirtingų mišinių, kurių sudėtis yra: 40-
60 % Mg, 28-42 % Sr(NO3)2, 5–17 % chlorintos gumos, 4 % virinto linų
aliejaus. Atsižvelgiant į degimo greitį ir specifinį šviesos stiprį, optimalus
mišinys buvo pasirinktas tolimesniam tyrimui.

Priedų įtaka
Bandant išgauti geresnes degimo savybes, buvo panaudoti priedai. Naujo
mišinio sudėtis buvo 48 % Mg, 34 % Sr(NO3)2, 9 % chlorintos gumos, 4 %
virinto linų aliejaus ir 5 % priedo.
Dujas generuojantys priedai pirotechnikoje jau yra žinomi nuo 1970 metų,
tačiau dar iki šių laikų jie yra bandomi taikyti apšviečiamojoje pirotechnikoje,
mažesnio dūmingumo užtaisuose ir dujų generatoriuose. Išbandytos
medžiagos buvo 2-cianoguanidinas, melaminas, azodikarbonamidas, kalcio
rezinatas, šelakas, stroncio oksalatas, laktozė, dekstrinas, fenolinė derva,
lignino sulfonatas, 5-aminotetrazolas, urėja, guanidinio nitratas, heksaminas,
siera.

Guanidinio nitratas
Išbandžius priedus paaiškėjo, kad guanidinio nitratas pasižymi puikiomis
savybėmis kaip priedas tokio tipo mišiniuose, todėl sėkmingai gali būti
dedamas vietoje chlorintos gumos. Su guanidinio nitratu pagamintas užtaisas
turėjo panašų degimo greitį kaip ir mišinys su chlorinta guma, tačiau jo šviesos
stipris buvo 60 % didesnis, o specifinis šviesos stipris – 40 % didesnis.
Vėliau buvo nustatytas guanidinio nitrato kiekis mišinyje, kuris yra
tinkamas su dideliu magnio kiekiu. Toleruojama riba siekė 25 %.

Mišinys be chloro
Degant chlorintiems organiniams junginiams, ypač žemesnėje
temperatūroje, padidėja polichlorintų bifenilų susidarymo tikimybė. Tokie
junginiai yra kancerogeniniai, dėl šios priežasties buvo tiriamas mišinys, kuris
pasižymėtų dideliu specifiniu šviesos stipriu bei neturėtų chloro. Buvo

68
surastas mišinys, sudarytas iš 48 % Mg, 34 % Sr(NO3)2, 13 % guanidinio
nitrato, 4 % virinto linų aliejaus ir 1 % magnio stearato.
Tokio tipo mišinys duoda liepsną, kurios pagrindinis spinduolis aukštoje
temperatūroje yra SrOH. Nors liepsnos spalovos sodrumas yra 15% mažesnis,
tačiau specifinis šviesos stipris vis tik lieka panašus į mišinio su chloru.

Mėl ynos liepsnos

Pirotechnikoje vienas didžiausių iššūkių yra didelio sodrumo spalvotos,
ypač mėlynos, liepsnos gavimas. Šio eksperimento tikslas buvo palyginti
geriausius žinomus mėlynos liepsnos mišinius spektrofotometru ir nustatyti
kurio mišinio liepsnos spalva yra sodriausia. Tam tikslui buvo surinktos
populiariausios mėlynų liepsnų mišinių formulės. Pagrindiniai oksidatoriai
buvo NH4ClO4, KClO3 ir KClO4. Mišiniai gana smarkiai skyrėsi deguonies
balansu, įnešamu vario ir chloro kiekiu, kurais.
Išmatavus šių mišinių liepsnas spektrofotometru, buvo nustatyta, kad
liepsnų sodrumo vertės yra 28-58 % intervale. Mišinys, sudarytas iš KClO4,
CuO, heksachlorbenzeno, sieros, dekstrino santykiu 39/37/6,5/15/2,5, davė
sodriausią mėlyną liepsną – 52 % (5 tablečių vidurkis). Maksimalus sodrumas
(65 %) buvo stebimas pirmąsias 3 degimo sekundes. Liepsnos šviesos stipris
buvo 108 cd, o specifinis šviesos stipris – 75 cd·s·g-1. Dominuojantis bangos
ilgis - 468 nm.

Mėl ynos liepsnos pulsuojančio degimo mišiniai

Pulsuojančio degimo pirotechniniai mišiniai yra naudojami fejerverkų
gamyboje. Tokio tipo baltos spalvos mišiniai yra žinomi jau ilgą laiką.
Raudonos, žalios ir geltonos spalvos pulsuojančio degimo mišiniai gana
nesunkiai gaunami, kombinuojant stroncio arba bario nitratą su MgAl
milteliais, tačiau mėlynos splvos pirotechninis pulsuojantis mišinys yra kur
kas sudėtingesnis. Šio tyrimo tikslas buvo ištirti mišinius, kurie naudoja a)
guanidinio nitratą (GN), b) aminoguanidinio nitratą (AGN),
c) tetrametilamonio nitratą (TMAN) kaip degimo moderatorius.
Visi tokio tipo mišiniai naudoja amonio perchloratą kaip oksidatorių (25-
55 %), degimo moderatorių arba energentinį kurą (30-55 %), vario druską arba
vario oksidą (6-20 %). Optimizuojant mišinius buvo gauta stabiliai pulsuojanti
mėlyna liepsna su aminoguanidinio nitratu. Jos pulsavimo dažnis buvo 4-8 Hz.
Šiuo atveju dažnis priklausė nuo degančio paviršiaus ploto. Su TMAN irgi
gaunama pulsuojanti liepsna, tačiau dėl skirtingo deguonies balanso mišinys
69
sunkiau dega ir dažniau užgęsta savaime. Tai nutinka, kai smilkimo reakcija
užtrunka per ilgai ir nepakanka šilumos įvykti blyksnio reakcijai. GN yra
mažiau reaktingas negu pastarieji, todėl jo mišiniai dega vibruojančia liepsna,
o ne pulsuoja.

Išvados

Medžiagos, turinčios didelį deguonies ir azoto kiekį, padėjo gauti didesnį

liepsnos šviesį apšviečiamuosiuose mišiniuose su dideliu magnio kiekiu.
Guanidinio nitratu modifikuojant mišinį, buvo gautas didžiausias, 60 %
didesnis už palyginamąjį bandinį specifinis šviesos stipris (efektyvumas)
naudojant 4-6 % chloro donoro mišinyje. Mišinys be chloro buvo sukurtas
naudojant guanidino nitratą, tačiau jo liepsnos sodrumas neviršijo 80 %.
Nors ir vertinamas senoje literatūroje, Paryžiaus žaluma (vario
acetoarsenitas) nepralenkė kitų vario junginių be arseno savo liepsnos
sodrumu (45 % pe lyginant su 50-55 % pe naudojant CuO ir Cu atitinkamai).
Naudojant KClO4/S bazinį mišinį, pavyko gauti labai didelio sodrumo mėlyną
liepsną, kuri degimo metu siekė iki 65 %. Tai taip pat parodo, kad KOH
švytėjimas liepsnoje yra nereikšmingas ir netrukdo CuCl(g) molekulių
sužadinimui, kurios lemia mėlyną spinduliuotę. Ryškiausią mėlyną liepsną
pavyko gauti kaip papildomą kurą naudojant MgAl. Toks mišinys degė 388 cd
šviesos stipriu ir 454 cd·s·g-1 efektyvumu. Tačiau šiuos parametrus pavyko
gauti prarandant liepsnos sodrumą, kuris buvo tik apie 30%.
Mėlynos liepsnos pulsuojančio degimo mišinys buvo sėkmingai sukurtas
degimo moderatoriumi naudojant AGN. Mišiniai su AGN ir su TMAN yra
labai jautrūs bet kokioms mišinio modifikacijoms. Kitų priedų kiekis mišinyje
negali viršyti 1-2 %.

70
 ACKNOWLEDGEMENT

First, I would like to express my gratitude towards my supervisor and

mentor Prof. Dr. Arūnas Ramanavičius for his helpful guidance, support and
encouragement during the entire PhD period.
I would like to thank the staff of Vilnius University, The Faculty of
Chemistry and Geosciences, Erasmus+ program, DAAD (German academic
exchange service), University of Munich (LMU) for omnifarious support with
regards to research and non-research related questions.
I would like to give a special thanks to dear colleagues and friends abroad
and in Lithuania, without whose support this work would not have been
possible.
Firstly, to Per Alenfelt from Hansson PyroTech AB, who provided the
opportunity for my internship in Sweden. For his supervision, tips, ideas and
support for many experiments at that time and afterwards. Per’s passionate
view on pyrotechnics continues to inspire me.
Rutger Webb from Clearspark B.V., for his friendly company that
resulted in many hours of skype conversations and debates covering various
aspects of pyrotechnic research.
Prof. Thomas M. Klapötke, for providing me an opportunity to do an
internship at his work group in Munich. Also, for the lab members there, that
have provided me help and good ideas in the path of research.
Dr. Arno Hahma from Diehl defense GmbH, who provided me an
opportunity to do a measurement campaign with him. For support and sharing
his impressive knowledge on pyrotechnics and science.
Also, Gaspard Angé for all the great time spent together debating much
more than only pyrotechnics.
Dr. Lina Mikoliūnaitė, my first scientific mentor, for her kind support not
only during PhD, but also during BS and MS period. Also, Aura, Modestas,
Ramūnas, Tomas, Inga, Edita for their great company at our department of
physical chemistry.
Monika Marija Kirsnytė for her friendly company and good laugh during
my time in Germany.
Reviewers Prof. Dr. Artūras Katelnikovas and Dr. Rytis Kubilius for their
suggestions and tips. Also Dr. Barry Sturman for his help with the
thermodynamic calculations.

71
 Last, but not least, I would like to thank my parents, Romualdas and
Jolita, for their calm yet constant support. I am very grateful they never doubt
my curiosity for pyrotechnics. Also, my dear family members and friends that
have been a tremendous help, love and support during these four years.

72
 CURRICULUM VITAE

EDUCATION

Vilnius University, Faculty of Chemistry and Geosciences:

1. 2016 – 2020 Doctoral studies in Chemistry.

2. 2014 – 2016 MSc degree in Chemistry.
3. 2010 – 2014 BSc degree in Chemistry.

WORK EXPERIENCE:

1. 2018 (7 months) Junior researcher at Vilnius University.

2. 2016 - 2017 (6 months) Intern at Nammo AB.
3. 2010 – 2014 (part time) Pyrotechnitian at UAB Blikas.

PROFESSIONAL TRAINING

1. DAAD research grant (Octorber 2018 – October 2019)

2. Erasmus+ practice (December 2016 – May 2017)
3. DAAD research grant (September 2015 - December 2015)
4. Erasmus+ practice (February 2015 – September 2015)
5. Erasmus intensive program “Eucheme” (Italy, 1-11 August 2013)

CONFERENCES

1. D. Juknelevicius, P. Alenfelt, A. Ramanavicius, The Properties of

Modified Red Flare Pyrotechnic Compositions, NTREM, 1-3 April 2020,
Pardubice, Czech Republic.
2. D. Juknelevicius, T. M. Klapötke, A. Ramanavicius, Blue strobe
pyrotechnic compositions, 50th PGI convention, 10-16 August 2019,
Gillette, WY, USA.
3. D. Juknelevicius, T. M. Klapötke, A. Ramanavicius, Blue strobe
pyrotechnic compositions with different additives, 17th International
Symposium on Fireworks, Puerto Vallarta, Jalisco, Mexico 2019.
4. D. Juknelevicius, P. Alenfelt, A. Ramanavicius, Magnesium Rich
Pyrotechnic Flare Compositions, Energetic Materials Synthesis,
Processing, Performance, 26-29 June 2018, Karlsruhe, Germany.

73
5. D. Juknelevicius, A. Dufter, M. Rusan, T. M. Klapötke, A.
Ramanavicius, Study of ammonium perchlorate and
tetramethylammonium nitrate based pyrotechnic blue strobe
compositions, 16th International Symposium on Fireworks, Ōmagari,
Japan 2017.
6. D. Juknelevicius, P. Genys, L. Mikoliunaine, M. Rusan, T. M. Klapötke,
A. Ramanavicius, Copper(I) Bromide: An Alternative Emitter for Blue-
Colored Flame Pyrotechnics, Advanced materials and technologies, 25-
28 August, Palanga, Lithuania 2016.
7. D. Juknelevicius, E. Karvinen, T. M. Klapötke, R. Kubilius, A.
Ramanavicius, M. Rusan, Blue-Colored Pyrotechnics using Copper (I)
Bromide as Emitter, 42th International Pyrotechnics Seminar, 10-15 July,
Grand Junction, Colorado, USA 2016.
8. D. Juknelevicius, T. M. Klapötke, R. Kubilius, A. Ramanavicius and M.
Rusan (2015), Ideas For Improving Colored Flame Compositions:
Analysis of the Influence of Oxygen Balance, 15th International
Symposium on Fireworks, Bordeaux, France 2015.
9. D. Juknelevičius, R. Kubilius, L. Mikoliūnaitė, A. Ramanavičius, A
Spectrophotometric Study of Green Pyrotechnic Flame Properties Using
Ammonium Perchlorate, Barium Nitrate and Shellac System, Chemistry
and chemical technology, Vilnius 2015.
10. M. Gicevičius, D. Juknelevičius, A. Ramanavičius, Tris(2,2'- bipiridino)
dichloroRutenio(II) heksahidrato elektrocheminės liuminescencijos
tyrimas ant skirtingų (Au, Pt, stikliškoji anglis) elektrodų paviršiaus,
Studentų moksliniai tyrimai 2013/2014, Vilnius.
11. D. Juknelevičius, L. Mikoliūnaitė, J. Voronovič, A. Ramanavičius,
Comparison Between Ruthenium Chelate, Luminol And N-(4-
Aminobuthyl)-N-Ethyl-Isoluminol Electroluminesence, Ecobalt 2013,
Vilnius.
12. D. Juknelevičius, L. Mikoliūnaitė, J. Voronovič, A. Ramanavičius,
Elektroliuminescncijos Tyrimas Spektrofotometriniais Metodais,
Studentų moksliniai tyrimai 2012/2013, Vilnius.
13. D. Juknelevičius, L. Mikoliūnaitė, J. Voronovič, A. Ramanavičius, Tris
(2,2’- Bipiridino) Dichlororutenio (II) Heksahidrato
Elektroliuminesencijos Tyrimas, Chemija ir cheminė technologija 2013,
Vilnius.

74
 ORIGINAL PAPERS

Paper 1

An experimental comparison of selected blue flame pyrotechnics

D. Juknelevicius, A. Hahma, R. Webb, T. M. Klapötke, A. Ramanavicius

Propellants Explos. Pyrotechn., 2020 (accepted)

Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced

with permission.

75
DOI: 10.1002/prep.202000114

An experimental comparison of selected blue flame pyrotechnics

Dominykas Juknelevicius,[a] Arno Hahma,[a,b] Rutger Webb,[c] Thomas M. Klapötke[d]* and Arunas
Ramanavicius[a]*

Dedication: In memory to Dr. Bernard E. Douda

Abstract: In this research 10 different pyrotechnic blue flame purity and high luminous intensity (I v), there are much fewer
compositions were designed and compared. Chromaticity and examples for blue illuminants available.
luminosity parameters of the flame were measured using Ocean An excellent historic perspective on colored flame
development is given in the thesis by Sturman1. He discusses
Optics JAZ-ULM VIS-Spectrometer equipped with a cosine corrector.
the evolution of colored flames throughout the years, even from
Color saturation, luminous intensity, specific luminous intensity, the period before the introduction of potassium chlorate into
oxygen balance, burn rate, actual and theoretical maximum density, pyrotechnic compositions.
color coordinates (X,Y) are presented and discussed. One of the first spectroscopic investigations that reported
an analysis of blue flames was carried out by Barrow & Caldin2.
They identified the emitting species as CuCl. Interestingly,
Key words: Blue flare pyrotechnic; Copper chloride; Fireworks; mercury(I) chloride was used in these compositions as chlorine
Pyrotechnics; Illuminating flame. source.
Blue flames have been studied by several researchers and
academics over the past decades. Douda has contributed
significantly to the understanding of colored flames3.
Abbreviations A systematic study of compositions, which are generating
blue flames, was performed by Shimizu4. He utilized a self-made
AP – ammonium perchlorate spectrograph containing a sample holder, a 0.04 mm slit, water
5-AT – 5-aminotetrazole prisms, lenses and a photographic plate.
BCC - basic copper carbonate (malachite) The proliferation of relatively low-cost spectrometers has
been helpful for studying pyrotechnic flames. Pyrotechnic
Dex – dextrin compositions are assessed in a "static" way, whereby the
HCB – hexachlorobenzene sample is burned, and the smoke is removed well enough to
CuOCl – copper oxychloride ensure a free line of sight to the sample. In 2003, Brian Ingram
CR – chlorinated rubber investigated the spectra of red, green and blue pyrotechnic
CP - chlorinated paraffin flames5. Meyerriecks and Kosanke studied the principal emitters
MgAl – powdered magnesium-aluminium 50/50 alloy in colored flames6. They utilized solutions of various chemicals
in combination with a nebulizer and an
PVC – polyvinyl chloride
oxygen/propane/acetylene flame.
Ω – oxygen balance Several papers have been published about the desired
Iv – light intensity (cd) emitter in pyrotechnic flames comprising a copper and a chlorine
pe – color saturation or purity source7. Dolata speculated about the formation of a trimer of
Lsp – specific luminous intensity (cd s / g) CuCl8. This was quickly followed by Sturman, who provided
disproof of this hypothesis9.
While static measurements are a sound and reproducible
method for characterizing compositions, such measurements
Introduction are not representative of the true environment in the most of
applications. In actual use, pyrotechnic compositions have to
A blue pyrotechnic flame color with high saturation or color purity burn at various airspeeds. Consequently, mixing of ambient air
(pe), is one of the biggest challenges in pyrotechnics. While will alter the flame stoichiometry and temperature. Furthermore,
several suitable pyrotechnic compositions can be formulated pyrotechnic stars may extinguish in flight.10,11.
with a reasonable effort for red, green and yellow at high color Few publications are available about "Round Robin trials"
of spectral color measurements. Douda has reviewed 81mm
mortar flares as test objects12.
[a] D. Juknelevicius, Prof. Habil. Dr. A. Ramanavicius Recent efforts have been made to identify alternatives to
Faculty of Chemistry and Geosciences the well-known copper-chlorine based systems. Some of us
Vilnius University have published work on copper(I)bromide, and compared it to
Naugarduko st. 24, 03225, Vilnius, Lithuania copper(I)chloride13. The work includes the 1931 CIE coordinates
E-mail: dominykas.juknelevicius@chf.vu.lt for the isolated spectra of CuCl and CuBr. In the same year
Koch compared all four copper(I)halides14. He provided values
[b] Dr. Arno Hahma
Diehl defence GmbH
of the 1931 CIE coordinates, as well as the dominant
Fischbachstraße 16, D-90552 Röthenbach an der Pegnitz, Germany wavelength and color purity. He has concluded that
E-mail: arno.hahma@diehl-bgt-defence.de copper(I)bromide provided good efficiency, even outperforming
the classical copper(I)chloride. The copper(I) fluoride and the
[c] Rutger Webb copper(I) iodide based system were found to be inferior to the
Vilnius University, NanoTechnas bromide and chloride. The formulations tested by Koch were
Naugarduko st. 24, 03225, Vilnius, Lithuania optimized for the formation of K2SO4, to enable spectral
rutgerwebb@gmail.com measurements minimizing the interference of potassium.
[d] Prof. Dr. T. M. Klapötke 76
Department of Chemistry, Ludwig-Maximilian University of Munich
Butenandtstr. 5–13 (D), 81377 Munich, Germany
Fax: +49 (0)89 - 2180 77492
E-mail: tmk@cup.uni-muenchen.de
1
 Recently, the flame color of pyrotechnics containing Composition #2 was different from #1 due to wax and
metallic indium has been investigated. These results do not stearic acid as fuels. Occasionally, large yellow spots appeared
indicate that indium is, even from a purely technical point of in the blue flame envelope due to soot formation. This can be
view, a viable alternative emitter15.
addressed to the largest oxygen deficit resulting in a reduced pe.
The goal of the present study was to compare the p e and
dominant wavelength of the best-known blue flame (star) The burn rate was also the lowest among the ten compositions
compositions. An overview of pyrotechnic blue illuminants helps studied.
to select the best compounds for achieving the highest color Composition #3 was unique because of its high nitrogen
saturation in a practical application. content, which was achieved by using 5-aminotetrazole (5-AT)
as the main fuel. The poor fuel properties of this compound
resulted in a composition, which could not sustain combustion
Results and Discussion on its own. The heat feedback from the flame wasn’t sufficient
enough to sustain combustion. However, the composition could
Pyrotechnic blue illuminant formulations were gathered from be burned by holding a glowing sparkler wire on the burning
different sources. Ten of them having the highest color surface. It should be noted that this composition is a derivative
saturation based on the information in the literature were from the original presented by Naud 17.
selected for this experiment. All the selected blue flame Composition #4 has the fewest number of components
compositions contain four main components: an oxidizer, a fuel, with AP as an oxidizer and chlorine source, copper(II) benzoate
a copper source and a chlorine source. The variety of chemical serves as a fuel and a copper source. This nearly oxygen-
components was another criterion for the compositions being as balanced composition (Ω = -5.4) passed the 50% pe threshold
different one from another as possible. This can provide more while its Iv and Lsp values remained on the average level.
information about the best compositions and the best chemical The composition #5 was unique due to the use of copper
components for generating blue flames with a high color acetoarsenite. An interesting note is stated by Shimizu 10,25 :
saturation. The selected compositions are presented in Table 1 “copper acetoarsenite was used in Japan in almost all blue
and their combustion characteristics are given in Table 2. compositions in 1980’s as it produces a very pretty blue”.
Nowadays, such compositions are nearly obsolete due to toxicity
Table 1. Experimental compositions: chemicals, ratios, sources. and environmental concerns. This composition produces more
No. Source Composition & ratio Ratio Comments
Optimized for
smoke compared to AP compositions (1-4) due to the presence
1 Hahma AP/Cu/CR/Hexamine/CP 62/14/4/10/10
maximum of potassium, which creates solid particles.
([HCl+Cu]/[CuO])
formation Composition #6 had the highest oxygen balance of Ω =
Ashless blue
2 McGriffen 16
AP/Cu/Stearic acid/Paraffin 74/11/11/4
flare +3.7. Chlorinated paraffin (CP) was difficult to mix evenly with
3 Naud17 AP/5-AT/BCC 47,5/47,5/5 N-rich blue other components the composition, which in addition to the
Classic AP/Cu positive Ω may have contributed to the lower performance of the
4 Dumont18,19* AP/Copper benzoate/Dex 79/17,5/3,5
benzoate blue
Copper
composition. Also, a decrease in color saturation was observed
KC/Paris green/Stearic
5 Hardt20
acid/HCB/Dex
62/21/8/4/5 acetoarsenite during the combustion of each pellet.
containing
Chlorate-lactose Composition #7 is a popular one in fireworks. The metallic
6 Ofca21 KC/CuOCl/Lactose/CP/Dex 65/13/13/5/4
blue
fuel MgAl increases the flame temperature and light output
KP/CuO/Red Firework star
7 Veline19,22 53/14/9/14/6/4
Gum/CR/MgAl/Dex with MgAl accordingly but results to a reduced color saturation. With pe of
Chinese blue for
8 Stanbridge23 KP/CuO/HCB/S/Dex 39/37/6,5/15/2,5
small pellets
30%, composition #7 had the highest Iv and Lsp of the tested
9 Naud24
KP/CuO/PVC/Hex/Red
61/17/10/6/3/3 Naud Ref. blue
compositions.
Gum/Dex
Perchlorate-
Composition #8 had the highest pe of 55.7 wt.-%. This
10 Pihko** KP/CuO/Hexamine/CR 62/13/10/15
Hexamine blue composition is unique for it’s high copper oxide content (37%),
Private communication: *Modification by Hahma (2012), ** Petri low potassium perchlorate content of 39 wt.-% and for the
Pihko (1988) absence of an energetic fuel such as wax, hexamine etc. Sulfur
is used as the main fuel instead. Sulfur helps to scavenge
50 g of each composition (except 25g of No. 5) were prepared. potassium in the flame resulting in an increased chlorine
Five 10 g pellets (16.80 mm diameter) were pressed out of each concentration promoting CuCl emitter formation 10,14. HCB acts
composition. The values presented in Table 2 are an average as a chlorine donor.
over 5 parallel measurements. It must be noted that even though Composition #9 was chosen as a reference from Naud24. It
pellets were ignited from the top, not all of them burned evenly was similar to composition #10 with a similar pe of 41% It had
(cigarette burning). For compositions 5-10, that were based on only a slightly lower Lsp and Iv values.
potassium chlorate and perchlorate, the surface flame Composition #10 was an efficient blue flame composition
propagation was more pronounced than that to the depth of the yielding pe of 39% with Lsp of 414 cd·s·g-1. It also produced the
pellet. For that reason, the recorded burn rate data presented in largest amounts of glowing slag on the test plate, where pellet
Table 2 is not as precise as it was expected. was fixed.
The first composition #1 burned with a uniform tall flame,
especially during the first seconds. It was one of the very few
compositions passing the 50% pe threshold and had dominant
wavelength (DW) of  = 445 nm, which is the lowest of all 10
compositions. This composition also possesses the highest Iv
and Lsp values among all ammonium perchlorate (AP) based
compositions tested in this study.
77
2
Table 2. Experimental results: Oxygen balance (Ω), Burn time, average pellet length, burn rate, TMD - theoretical maximum density
TMD % (indicates the fraction of measured density divided by TMD), Iv, Lsp, and color coordinates are presented. The mass of all
pellets was in the range 9.5-11.7 g, except composition #5, which was limited to 4.8-5 g for each pellet. The white point was set at x =
1/3, y = 1/3.

No. Ω l BR TMD TMD I Lsp pe DW

[%] t [s] [mm] [mm·s-1] [g·cm-3] [%] [cd] ± [cd·s·g-1] ± [%] ± [nm] CIE x CIE y
1 -20.2 17.5 23.2 1.3 1.98 96% 91 6.6 162 12.1 51.6 1.2 453.7 0.240 0.173
2 -26.3 42.5 26.0 0.6 2.01 85% 22 2.5 94 4.1 30.4 4.1 556.2 0.328 0.251
3 -15.1 27.5 32.7 1.2 1.75 83% 10 2.3 26 3.2 38.1 1 432.8 0.269 0.210
4 -5.4 12.5 26.0 2.1 1.75 95% 79 18.3 98 2.5 52.0 1.6 463.2 0.233 0.179
5 -12.7 9.0 11.3 1.3 2.17 90% 58 0.8 107 4.3 45.3 1.9 469.6 0.240 0.208
6 3.7 8.0 23.5 3.0 2.18 86% 46 3.2 37 1.8 28.6 3.4 462.8 0.278 0.250
7 -12.8 11.6 24.7 2.1 2.16 83% 388 27.6 454 22.9 29.8 2 464.1 0.276 0.246
8 -3.2 6.9 19.8 2.9 3.02 73% 108 21.7 75 8.0 55.7 2.6 467.9 0.221 0.175
9 -5.5 8.0 22.9 2.9 2.31 82% 267 19.7 218 15.9 40.8 1.6 477.3 0.242 0.240
10 -5.8 11.1 23.1 2.1 2.30 82% 365 48.4 414 35.8 38.6 0.2 481.1 0.242 0.261

.
The Ω of all compositions lies in the range of -26 to +4%.
The compositions at the extreme ends (#6 and #2) had low Lsp
500 and Iv values, which may be related to the unbalanced system.
7
10
Too low Ω results in soot formation, while too high Ω may cause
400
the emitter species to be oxidized. The Lsp and I dependence on
Lsp [kcd·s·g-1]

300 Ω are depicted in Figures 1 and 2.

The measured pe lies in the range of 28 – 56% (Fig. 3).
200
9 The lowest pe of ~30% was measured for compositions #2,6,7
1 and the highest exceeding 50% for compositions #1,4,8. Deep
100 2 5 4 blue flame compositions with high p e are usually observed to
8 burn with a reletvely dim flame compared to the bright ones
3 6
0 having a low pe. In this work, composition #8 flame was indeed
-30 -25 -20 -15 -10 -5 0 5 the least luminous of the KClO4 compositions (107 cd, 74 cd·s·g-
 [%] 1
). However, pe reached 56% on average from 5 pellets, with one
Fig. 1. Lsp dependence on the oxygen balance pellet reaching 57.7%. By observing the pe vs time graphs some
pellets of the composition #8 peaked at 65% pe during the first
seconds of combustion, before the pe dropped to 50-57%.
Possibly the reduction of the color purity is associated with slag
500
formation on the burning surface. The composition #8 is
designed to be used as small stars in fireworks. Therefore, it
400 7
10 possibly did not deliver the best performance when burned as a
16 mm diameter, 10 g pellet.
300
9 The emission spectra were recorded for each of 10
I [cd]

compositions. The main focus of interest was to observe how

200
the CuCl emissions compare to the CuOH, CuO, Na and black
8
body emissions. In figure 4 KClO3 compositions 5,6 are
100 1
5
4 compared. Composition #5 has a less pronounced grey body
6
2 3 radiation and strong CuCl emissions at 400-470 nm, hence it
0
-30 -25 -20 -15 -10 -5 0 5 possesses a higher color purity than composition #6. In Figure 5,
 [%] the spectra of compositions #8 and #10 are compared. The
CuOH emission from composition #10 appears much more
Fig. 2. Iv dependence on the oxygen balance. pronounced compared to composition #8. This emission
decreases the color purity of the composition #10, however the
latter is much brighter burning. In Figure 6, the raw spectra of
composition #6 are depicted at selected times. The intensity
decreased as a function of time. This phenomenon was not
observed with other compositions, at least not to such a strong
extent. This effect may have resulted from both (i) the CP being
unevenly distributed in the composition and (ii) also the slag
formation.
Finally, the CIE x/y coordinates plotted in the CIE 1931
color diagram (Fig. 7) are located in the blue and blueish-white
region of the chromaticity diagram. The compositions #1,4,8
show a noticeble shift towards the blue region of the chromaticity
diagram.
Video captures of the burning pellets are collected in
Fig. 3. The distribution of pe. Figure 8. From these photos the flame size, shape, smoke and
slag formation can be estimated. The AP based compositions
78#1, 2, 3, 4, tend to burn with little smoke and no slag formation.
3
The flame, especially in the beginning of the combustion, is tall
and narrow. Later on, it evolves to a shape seen in Figure 8. The
KClO3 and KClO4 compositions (#5 - #10) produce some slag
and a significant amount of smoke. Some of them had higher I v,
Lsp and pe values than AP compositions, which are often
considered superior blue flame compositions. The reason for the
higher light output can be associated with smoke reflection
during the measurements. If the smoke in the measurement
chamber reflects some of the emitted light from the flame
towards the spectrometer, the reflected light adds to the
recorded data. This effect could not be completely avoided,
because it is not possible to extract the smoke at the edge of the
flame completely. On the other hand, the eye will also perceive
this and the pyrotechnic composition will appear brighter than it
actually is. In that sense, no error is produced, when the back
reflection is ignored.
NASA CEA2 code was used to estimate the transient
species present in the flames of tested compositions (supporting
information). Only composition #5 was omitted, as there was no
Fig. 6. The composition #6 pellet flame’s raw emission spectra
data on arsenic species. The adiabatic flame temperatures
recorded at 2.2, 4.7, 6.2, 8.4 s after ignition. A decrease in the
range from 1840K for composition #8, to 2690K for composition
intensity is observed. The strongest emission in the blue region
#10. Besides the high concentration of typical combustion
is observed after the ignition and the least intense emission is
products i.e. H2O, CO2, CO, H2, N2 substantial amount of HCl
observed in the last seconds of combustion process.
(0.15 mol-%) was produced for AP compositions #1 – #4.
Moreover, both KCl, HCl were present in KC and KP based
compositions #6 - #10. The target species for this experiment is
CuCl. While most compositions had around 0.03-0.05 mole-% of
CuCl produced, compositions #8 had 0.07 mole-% of CuCl,
which is in relation to the high performance of this composition
described earlier. This can be associated with very high
concentration (37 wt.-%) of CuO used in the compositions, that
shifts the equilibrium towards CuCl formation.

Fig. 4. Emission spectra of compositions #5 and #6. Fig. 7. Zoomed in chromaticity diagram including the evaluated
compositions with the full diagram in the upper left corner.

Fig. 5. Emission spectra of the compositions #8 and #10.

79
4
 cd·s·g-1 it is almost as bright as the composition #7. With a pe of
38.6%, it has one of the best brightness to pe ratios.
Hexamine was found to be a useful fuel in blue flame
compositions. It has a high energy density of 30 MJ/kg being a
good, energetic, non-metallic fuel burning with a non-sooty
flame. Moreover, it is a synthetic compound that can be obtained
in a high purity.
5-AT containing composition #3 burned with significant
difficulties. Most probably a fuel mixture of hexamine and 5-AT
would have been more useful for this composition.
Even though highly regarded in older literature, Paris
green composition did not surpass the other top compositions in
color purity nor in luminous intensity. Hence, Paris green and
HCB (which can be easily replaced by a non-toxic chlorine
source for composition #8) are discouraged from being used in
practise.
This kind of experimental examination has not been done
to date, wherein a rather diverse range of blue flame pyrotechnic
compositions are examined quantitatively with a spectrometer.
Very few researchers in general have reported CIE coordinates
with pe, Iv, Lsp values for blue flames. In addition, some old-
fashioned ingredients were tested to demonstrate that arsenic-
containing ingredients (Paris Green) offer no benefit in the color
purity, and other, more novel, compositions show better
performance.

Experimental Section

The following chemicals were used: Ammonium perchlorate,

potassium perchorate (d50 10 µm), potassium chlorate (20 µm)
were all reagent grade, 5-AT from Sigma 02312TE, Hexamine
B. Kraft 16932.5600, Stearic acid Merck 673, Paraffin wax
(C31H64) Te-ce-wax H994 powder, Dextrin Roth 6777.1, Lactose
monohydrate Roth 8921.1, Red gum (C6H6O2.6) Ilotulitus Oy,
Hyrylä, Finland, MgAl MX 077 Eckart Werke in Fürth, Sulfur JT
Baker 0335 USP, PVDC Solvin 910 Solvay, Belgium, ground to
20 µm, CP (C10H15Cl7) Leuna-tenside GmbH CP135,
Chlorinated rubber (C10H11Cl7) S10 Covestro 00549421, HCB
old sample, PVC Solvin 374MB, Cu powder electrolytic Ecka 71,
Eckart Werke, basic copper carbonate (Cu2(CO3)(OH)2
malachite) Sigma 20.789-6, Copper benzoate was synthesized
Fig. 8. Video captures of each burning composition.
from potassium benzoate and a soluble copper salt, Paris green
(Cu4As6C4H6O8 was synthesized from arsenic trioxide and
Conclusions
copper acetate, CuOCl (synthesized), CuO Omikron GmbH 10-
Ten different blue pyrotechnic flame compositions were 0334.
compared with the focus at the color saturation pe. The The compositions were prepared by mixing the dry
measured values lie in the range of 28 – 58%. The highest ingredients and passing through a 40-mesh sieve. The
average pe of 55.7% (and up to 58% for a single pellet) was homogeneous powder mixture was moistened with a solvent to
observed from the composition #8 with 108 cd, 75 cd·s·g-1. The activate the particular binder. The compositions #1, #6 were
peak pe = 65% was registered during the first 1 to 3 seconds of
combustion. The dominant wavelength of this composition was moistened with dichloromethane to dissolve the CP. The
468 nm. composition #2 was moistened with n-Hexane to dissolve the
KClO4/S system used in the composition #8 was found paraffin and all the others, except the composition #3, were
useful for producing a blue flame with high purity. It also proves moistened with water to dissolve dextrin. The composition #3
that KOH continuum and K(g) do not interfere significantly with was used without any binder.
desired CuCl emissions when sulfur is used as fuel.
The compositions were pressed to nominally 10 g pellets
Similar to composition #8, but brighter and more efficient
was the AP based composition #1 with 91 cd and 162 cd·s·g-1. having a nominal diameter of 16,80 mm at 1,15 kbar pressure.
The pe was just slightly under 52% with a very low dominant Five pellets were pressed for each composition and measured
wavelength of 453 nm. with a digital caliper at ±0.01 mm precision. The dimensions
Composition #7 lies at the other extreme with 388 cd, 454 were used for calculating the densities and percentage of the
cd·s·g-1 This was the brightest composition, but had a mere TMD as well as the burn rates of each pellet.
29.8% saturation.
The spectra were recorded with Ocean Optics JAZ-ULM
The composition #10 appeared to be well balanced, simple
and practical. Yielding a 365 cd bright flame and Lsp of 414 VIS-Spectrometer equipped with a cosinus corrector. The
80
5
spectrometer was run in its high-speed absolute calibration ammunition depot, Crane, Indiana, 1964; AD A951815.
mode and placed 0.5 meters apart from the pellet. The cosinus (4) Shimizu, T. Studies on Blue and Purple Flame Compositions Made
corrector has an angle view of ±60 degrees and could record the with Potassium Perchlorate. Pyrotechnica 1980, VI, 1.
entire flame despite of the short distance needed because of the (5) Ingram, B. Color Purity Measurements of Traditional Pyrotechnic
low light output of blue illuminants. The resulting spectra were Star Formulas. J. Pyrotech. 2003, 17, 1.
recorded already calibrated. The calibration was verified against (6) Meyerriecks, W.; Kosanke, K. L. Spectra of the Principal Emitters in
a NIST traceable calibration lamp. The verification confirmed the Colored Flames. J. Pyrotech. 2003, 18, 1–22.
factory calibration is correct and precise. Hence, no correction of (7) Contini, A. E. Blue Flame Pyrotechnic Compositions: A Concise
the raw data was necessary. Review. J. Pyrotech. 2010, 29.
The combustion times were recorded using a Casio Exilim
(8) Dolata, D. P. Reassessment of the Identity of the Blue Light Emitter
EX-FH20 camera at 210 Hz frame rate and VGA-resolution. This
in Copper-Containing Pyrotechnic Flames – Is It Really CuCl?
camera has a considerably higher time resolution and precision
Propellants, Explos. Pyrotech. 2005, 30, 63–66.
than the spectrometer, which often had to be run at 500 ms
(9) Sturman, B. On the Emitter of Blue Light in Copper‐Containing
integration. In addition, the video recordings remain as a
Pyrotechnic Flames. Propellants, Explos. Pyrotech. 2006, 31, 70–74.
reference for each pellet. The video recordings were
(10) Kosanke, K. Pyrotechnic Chemistry; Pyrotechnic Reference Series;
automatically edited to include one second before and after the
Journal of Pyrotechnics, Incorporated, 2004.
combustion. The combustion times were defined as the
(11) Kosanke, K. L.; Sturman, B. T.; Winokur, R. M.; Kosanke, B. J.
beginning and the end of the sum curve exceeding a threshold
level set at 10 % of the maximum intensity of the pellet. Encyclopedic Dictionary of Pyrotechnics; Journal of Pyrotechnics,

All pellets were measured in a free state without wind. The 2012.

pellets were lacquered with 30% solution of Synthesia E37 (12) Douda, B. E. Survey of Military Pyrotechnics NWSC/CR/RDTR-595,
nitrocellulose in ethanol:diethyl ether 1:2 to make them burn only Naval ammunition depot Crane, Indiana, 1991.
at the end face (cigarette burning). This way, the burn time (13) Juknelevicius, D.; Karvinen, E.; Klapötke, T. M.; Kubilius, R.;
reflects the true burn rate, when divided by the pellet length. We Ramanavicius, A.; Rusan, M. Copper(I) Bromide: An Alternative
did, however, encounter some problems with the lacquer film Emitter for Blue-Colored Flame Pyrotechnics. Chem. - A Eur. J.
peeling off due to poor adhesion. This may have caused some 2015, 21, 15354–15359.
loss of the composition, however the losses were less than 0.2 g (14) Koch, E.-C. Spectral Investigation and Color Properties of Copper(I)
for each pellet and were ignored. Halides CuX (X=F, Cl, Br, I) in Pyrotechnic Combustion Flames.
Propellants, Explos. Pyrotech. 2015, 40 (6), 799–802.
Acknowledgements (15) Glück, J.; Klapötke, T. M.; Küblböck, T. The Flame Emission of
Indium from a Pyrotechnical View. Z. Anorg. Allg. Chem. 2020, 646,
Diehl Defence GmbH is gratefully acknowledged for providing 133–137.
test facilities, equipment and reagents for the experimental work. (16) McGriffen, J., Ripley, W. Investigation of Visibility and Formulation
The DAAD one-year grant program and Vilnius University of Ashless Blue Flare; Naval ammunition depot Crane, Indiana,
doctoral studies program is acknowledged for a scholarship
1962.
(D.J.). The author is very grateful to Per Alenfelt from Hansson
(17) Chavez, D.; Hiskey, M.; Naud, D. High-Nitrogen Fuels for Low-
PyroTech AB for various inspiring discussions and Dr. Barry
Smoke Pyrotechnics. J. Pyrotech. 1999, 10.
Sturman for the help with thermochemical calculations.
(18) Drexler, F.; Basse, K.; Brauer, B. Pyrotechnic Light-Emitting

Contributions Composition, Useful e.g. in Fireworks, Contains Aromatic Caboxylic

Acid Salt, e.g. Strontium or Potassium Benzoate, as Color Former

D.J.: Conceptualization, investigation, data analysis, writing - Also Acting as Fuel. DE102005053849A1, Germany, 2005.
original draft. (19) Dumont, J.-L. Pyrotechnic Compositions for the Production of
A.H.: Conceptualization, methodology, investigation, writing - Coloured Fireworks. EP0252803B1, 1986.
review & editing, validation. (20) Hardt, A. P. Pyrotechnics; Pyrotechnica publications: Idaho, USA,
R.W.: Writing - original draft (introduction part), writing - review & 2001.
editing, validation. (21) Ofca, B. Bill Ofca’s Technique in Fire, Mastering Cut Stars the Easy
T.K.: Conceptualization, supervision, revision, funding Way, Vol. 5; B & C Products: Hyde Park, NY, 1994.
acquisition. (22) Veline, R. A Compatible Star Formula System for Color Mixing. Self
A.R.: Supervision, writing - review & editing, funding acquisition, Publ. 1989.
submission.
(23) Standbridge, M. Personal Communication. Self Publ. 2007.
(24) Naud, D. L.; Hiskey, M. A.; Chavez, D. E. Perchlorate-Free
Keywords: blue • flare • pyrotechnics • illuminating • flame •
Pyrotechnic Formulations Utilizing Energetic Chlorine Donors: 1-
copper chloride • fireworks CDN, 11-CDN, 13-CDN. Z. Anorg.Allg. Chem. 2013, 639, 702–706.
(25) Shimizu, T. Fireworks: The Art, Science, and Technique;
(1) Sturman, B. The Chemical Transformation of Fireworks in the 19th
Pyrotechnica publications: Austin TX, 1996.
Century, Doctoral thesis, Monash university, 2017.
(2) Barrow, R. F.; Caldin, E. F. Some Spectroscopic Observations on
Pyrotechnic Flames. Proc. Phys. Soc. 1949, Section B.
(3) Douda, B. E. Theory of Colored Flame Production; U.S. Naval
81
6
 Paper 2

The Performance of Red Flare Pyrotechnic Compositions modified

with Gas Generating Additives

D. Juknelevicius, P. Alenfelt, A. Ramanavicius

Propellants Explos. Pyrotechn., 2020, 45: 671-679

Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with

permission.

76
Full Paper
DOI: 10.1002/prep.201900322

The Performance of Red Flare Pyrotechnic Compositions

Modified with Gas Generating Additives
Dominykas Juknelevicius,[a] Per Alenfelt,[a, b] and Arunas Ramanavicius*[a, c]
To Rev. Ronald Lancaster a.k.a. “Master blaster pastor”

Abstract: In this work, an effective, but still not well-ex- lated to the specific luminous intensity and color purity.
plored area of red flare pyrotechnic formulations was inves- Additives in the form of 2-cyanoguanidine, Melamine, Azo-
tigated. High magnesium content (40–60 % w.t.) and the dicarbonamide, Calcium resinate, Shellac, Strontium ox-
presence of conventional gas generating materials were alate, Lactose, Dextrin, Phenolic resin, Lignin sulfonate, 5-
the specific features of formulations explored herein. Such aminotetrazole, Urea, Guanidinium nitrate (GN), Hexamine
illuminant compositions possessed a low burning rate and and Sulfur were evaluated in Mg-rich flare compositions.
high specific luminous intensity. The investigated system The use of GN was studied more carefully as an exception-
was based on strontium nitrate, rounded magnesium pow- ally efficient additive that increased the flares light intensity
der, chlorinated rubber, binder and gas generating addi- by 60 %, in comparison to the standard GN-free composi-
tives. Emission spectra, light intensity and burning time tion, while maintaining high color purity of 90–97 %.
have been measured during the experiment and recalcu-
Keywords: Magnesium · Red Flare · Guanidinium nitrate · Fireworks · Pyrotechnics

1 Introduction luminous and spectral emitters like SrCl(g), SrOH(g), BaCl(g),

Secondly, Mg has a low vaporization temperature, 1107 °C,

BaOH(g) can produce flames that have high color saturation.
Magnesium containing pyrotechnic compositions have

perature of ~ 3200 °C. This satisfies the Glassman criteria for

been used since the late 19th century, but it was not until which is much lower than its oxide’s decomposition tem-
the period between the World Wars that price and avail-
ability enabled a more extended production and develop- metal vapor phase combustion [15], which results in an ex-
ment of high temperature, magnesium fueled, illuminating tended vapor phase diffusion combustion zone, that allows
pyrotechnic products [1]. for a large radiative area.
Before and during the Second world war products like Generally, fundamental demands for a good UV-VIS py-
hand-flares, signal ammunition and parachute-flares were rotechnic flare composition are: a) burning with luminous
improved to a similar level of efficiency as we see today. A flame with maximum candlepower; b) having a distinct
lot of work was done to deepen the understanding of the flame color of purity higher than 80 % (e. g. red distress sig-
underlying phenomena during the 1950s to 1960s [2], nal); c) preferably creating a minimum of slag, d) being
though without affecting much of the products already in chemically stable and containing chemicals of minimum
use. The main driving force in development in recent years toxicity. Slag formation is unwanted since long-lived burn-
has been focused on eliminating certain unwanted in- ing sparks and glowing slag is always a fire hazard and
gredients and to a certain degree improving flame color could make a handheld flare dangerously to use and burn
depth [3–12].
Although having problems with corrosion induced by
[a] D. Juknelevicius, P. Alenfelt, Prof. Dr. A. Ramanavicius
moisture [13] and being a relatively expensive ingredient, Department of Physical Chemistry
Mg is absolutely indispensable for the production of illumi- Vilnius University
nating flares and the emission of high intensity colored Naugarduko st. 24, 03225, Vilnius, Lithuania
light in both military and civil applications. The reason for [b] P. Alenfelt
this is a couple of unique chemical properties of Mg that Hansson PyroTech AB
distinguishes it from all other possible fuels for high tem- Box 54, 711 22, Lindesberg, Sweden
[c] Prof. Dr. A. Ramanavicius
perature colored light production in pyrotechnics. First of NanoTechnas – Centre of NanoTechnology and Materials Science
all, the oxidation of magnesium produces large amounts of Vilnius University
heat, which directly relates to a high number of excited Naugarduko st. 24, 03225, Vilnius, Lithuania
molecules in the flame [14]. As a result, the flame is very *e-mail: arunas.ramanavicius@chf.vu.lt

Propellants Explos. Pyrotech. 2020, 45, 671–679 © 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 671
83
Full Paper D. Juknelevicius, P. Alenfelt, Prof. A. Ramanavicius

off the vanes to the canopy in a parachute suspended flare. ilim ex-f1) at automatic settings for the evaluation of shape,
Even more important is that a flare producing lots of slags spark and slag formation.
is probably not delivering its best pyrotechnic performance
(maximum light output in the case of illuminants).
In this work, a novel area of flare formulations was in- 3 Results and Discussion
vestigated. First, a series of high Mg content compositions
were evaluated. Then nitrogen and oxygen generating ma- 3.1 Evaluation of the Mg-Rich System
terials were added in an unusually Mg-rich area of the for-
mulation. Finally, guanidinium nitrate (GN) was successfully The first experiment was orientated to explore the effec-
added to the pyrotechnic illuminant that induced an un- tiveness of the well-known strontium nitrate, magnesium,
usual low burning rate together with a high specific lumi- and chlorine donor tri-component system that contained a
nous intensity. rather high percentage of Mg (40–60 %). Typical red flare
compositions described in the literature contain around 30–
40 % of Mg powder, 55–35 % of Sr(NO3)2, 12–20 % of chlor-
2 Experimental Section ine donor and 4–10 % of other additives such as additional
oxidizers, binders, stabilizers, burning rate retardants, etc.
Rounded Mg powder 45–100 μm (max 5 % > 100 μm and (Table 1). An amount of over 50 % Mg powder is more com-
max 5 % < 45 μm, purity of 99.0-99.5 %) and Sr(NO3)2 (99.5 % mon in yellow (white) flares based on NaNO3, where atomic
with < 0.4 % barium nitrate, < 0.05 % calcium nitrate and Na(g) is the dominant light emitter.
< 0.04 % sodium nitrate) were received from Jengo S.P.A., Having a high percentage of Mg in the composition, of
Chlorinated rubber in the form of Pergut B10 (finer than course, requires lowering the amounts of other compo-
250 μm) from Bayer. Sr(NO3)2 was milled in a hammer mill nents which in turn can result in lower emitter concen-
until about 50 % of the particles were smaller than 50 μm tration, more rapid combustion, and stronger incandescent
before the experiment. 2-cyanoguanidine, Azodicarbona- emission. It is, however, a potential approach to gain extra
mide, Calcium resinate, Shellac, Strontium oxalate, Lactose, candlepower from the flare composition.
Dextrin, Phenolic resin, Lignin sulfonate, GN, and Sulfur (all A series of test compositions were prepared, see Ta-
very fine powders, purity 97–99 %) were only passed ble 2. Rounded Mg powder was chosen over “cut” Mg for
through a 250 μm sieve to break up lumps. Melamine, 5- its ability to burn slower. An amount of 4 % of linseed oil
aminotetrazole (5-AT), Urea and Hexamine (purity 97–99 %) was added to each composition as a binder. Two flares of
were ground with a mortar and pestle to a fine powder and each composition were pressed for evaluation.
particles bigger than 150 μm were removed with a sieve. An important parameter that describes the performance
The linseed oil used was boiled type. The following for- of a pyrotechnic illuminating composition is specific lumi-
mulas were applied for calculating oxygen balance (ΩCO2) nous intensity Lsp (Eq. 1).
for variable Mr compounds: Linseed oil C57H92O6, Calcium
resinate C40H58CaO4, Shellac C16H26O4, Phenolic resin Iv t
C53H45O7, Lignin sulfonate C20H24O10CaS2, Chlorinated rubber LðspÞ ðkcd � s � g�1 Þ ¼ (1)
m
C10H11Cl7.
The compositions were hand mixed, sieved through a Equation 1 presented in normalized units describes how
30 mesh screen, cured for 4 h at 80°C and then sieved again much visible light in kilo candelas is produced during one
through 20 mesh screen to break lumps. second from one gram of pyrotechnic composition [16–18].
Test flares were pressed into metal tubes (ID = 23 mm) For Mg containing illuminants, this value is usually reported
containing waxed paper liners (OD = 23 mm, ID = 21.6 mm). to be in the range of 2–7 kcd·s·g�1 for green colored flares
Hence, the pyrotechnic composition was compacted into a with barium nitrate [19], around 15–50 kcd·s·g�1 for illumi-
column of ID = 21.6 mm and 70–85 mm in length. Flares nating flares with sodium nitrate [16, 17, 20] and around 14–
were pressed two at a time at 2 tons with 4 increments of
11.5 g (scale accuracy 0.05 g). 2 g charge of igniter compo-
sition was pressed on top of each flare for ignition.
Table 1. Chemical compositions typically used in red flares.
The experiment was performed in a wind tunnel to re-
duce smoke obscuration. The wind speed in the chamber, Chemical composition (w.t.%)
where flares were burnt, was 1 m·s�1. Illuminance was
Sr(NO3)2 43 55 48
measured with a calibrated lux meter (LMT, I(ph)–meter KClO4 9
1500, data processed with DasyLab software) and emission Mg (50-100 mesh) 29 30
spectra were recorded with a spectrometer (Stellar net, EPP Mg (30-50 mesh) 33
2000, CXR-SR; irradiance calibrated for 300–1100 nm range). PVC 12 15 15
The flares were also filmed with a video camera (Casio Ex- Laminac 4116/Lupersol 7 4
Source [16] [17] [4]

672 www.pep.wiley-vch.de © 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Propellants Explos. Pyrotech. 2020, 45, 671–679
84
 The Performance of Red Flare Pyrotechnic Compositions Modified with Gas Generating Additives

16 kcd·s·g1 for red-colored flares with strontium nitrate

23,0

55,3
1,51

3,51

27,7
7,9
[18].
60
18
31

4
5
A drawback is that Lsp does not take into account the
burning area of the flare and the geometry of the flame.
22,8

53,4
1,56

3,44

26,4
7,7
56
17
35

Even though the choice of the pyrotechnic composition,

4
5

the ratio of its components and its density after compaction

22,4

48,8
1,59

3,44

23,7 are the major factors that determine size and shape of the
6,9
52
16
39

4
5

flame envelope, both emission and absorption take place in

the envelope of the flame and absorption is a function of
29,6

34,0
1,48

2,78

21,8
7,8

distance (flame width). Also, heat loss is different for flares

60
15
28

4
8

of different diameters and therefore, different shaped col-

umns or pellets that contain exactly the same composition
27,6

39,5
1,51

2,92

23,7
8,1
56
14
32

will often have different Lsp values. Therefore, the results

4
8

from this work can only be compared to other flares of sim-

25,6

43,9
1,55

3,08

24,3

ilar diameter and density.

7,9
52
13
36

4
8

A pyrotechnic composition that possesses a high Lsp val-

ue is often desired, since then a lower amount of py-
24,9

39,3
1,59

3,08

21,3
6,9
48
12
40

rotechnic composition is needed to achieve a certain effect,

4
8

for example, the desired burning time of a parachute flare.

35,6

23,2

And, secondly, the pyrotechnic unit can be made smaller

1,49

2,31

17,9
7,8
11
56
11
29

and lighter which in turn saves space and volume for prac-
tical applications.
31,0

28,9
1,52

2,59

19,5

As a complement, the ratio of Lsp and burning rate (Lsp/

7,5
11
52
10
33

BR, Table 2) was used to give guidance in determining

which compositions burned both bright and slow.
28,7

30,6
1,56

2,72

19,1
7,0
11
48
37

The combustion of all 18 flares was monitored with a

4
9

lux meter. Since the distance between the photodiode and

26,8

31,0
1,61

2,84

18,1

flare was known, a software (DasyLab) was used to calculate

6,4
11
44
41

4
8

and plot light intensity in kilo candelas (kcd) over time. The
Lsp values were calculated and their distribution was de-
41,0

14,3
1,49

2,00

12,7
6,4

picted in a ternary diagram (Figure 1). The trend observed

14
52
30

4
7

was that there is a significant increase of Lsp with a decreas-

ing amount of chlorinated rubber. The main purpose of
35,0

21,3
1,53

2,28

16,2
Table 2. Chemical composition of 18 formulations tested in first experiment.

7,1
14
48
34

chlorinated rubber is to serve as a chlorine donor to en-

4
6

hance SrCl(g) emitter formation, which supplements emis-

32,6

23,6
1,56

2,40

16,7

sion from SrOH(g). This shifts the flame color from reddish-
6,9
14
44
38

4
5

orange to red. Additionally, the use of chlorine donors is

one of few possible ways to quench the MgO(s/l) in-
29,0

25,4
1,60

2,63

16,0
6,1
14
40
42

candescent emission via reaction with Cl2(g) or HCl(g) that

4
4

yields more volatile MgCl(g) [14]. The color purity of chlor-

ine-rich compositions 1–3 was measured to be 97–99 %
48,8
1,49

1,68
9,2
9,8
5,8
17
48
31

4
3

41,0

14,4
1,53

1,94

12,9
6,6
17
44
35

4
2

35,0

17,6
1,56

2,24

13,4
6,0
17
40
39

4
1

Mg rounded (45–100 μm)

L(sp)/BR [kcd s2 g1 mm1]

Burning rate [mm s1]
Light intensity [kcd]
Chlorinated rubber

Burning time [s]

Density [g cm3]

L(sp) [kcd s g1]

Compositon #

Linseed oil

Figure 1. Lsp distribution plotted in a ternary diagram of measured

Sr(NO3)2

compositions with high Mg content. Small numbers at the end of

each contour line represent Lsp values.

Propellants Explos. Pyrotech. 2020, 45, 671–679 © 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.pep.wiley-vch.de 673
85
Full Paper D. Juknelevicius, P. Alenfelt, Prof. A. Ramanavicius

with a dominant wavelength of (DW) = 613–614 nm, while position 6”, which was modified by replacing 5 % chlori-
chlorine-low compositions, 16–18, had lower color purity of nated rubber, by 5 % additive (Table 3).
79–87 %, at DW = 605–609 nm. When testing compositions 19–33, it was noted that cer-
Moreover, it was determined that Lsp values and BR are tain compositions produced more luminous flames, while
directly connected. Generally, fast-burning compositions re- others burned very rapidly. And it was clear that not only
sult in more luminous flames with high Lsp values. The dis- nitrogen-rich materials were successful. For example, lac-
tribution of BR of the same system is depicted in Figure 2 tose and dextrin were almost as good as 2-cyanoguanidine
for comparison. However, from a practical perspective, high and melamine when comparing light intensity values, even
BR is only desired in tracer compositions, whereas hand though they did not contain any nitrogen. Further on, ex-
flares, parachute flares etc. must contain a slow-burning, perimental data was plotted in several ways to determine
yet luminous flame producing composition, which is the which factor determines an efficient composition. Finally, a
real challenge. trend was observed when plotting the Lsp values against
the total amount of nitrogen and oxygen (in moles or
grams) in the additive used (Figure 3; Figure 4) and against
3.2 The Influence of Additives oxygen balance (Figure 5).
Several additives (calcium resinate, strontium oxalate,
After the evaluation of 18 compositions, the composition 6 lignin sulfonate, sulfur, and chlorinated rubber) contained
was chosen as a base for further investigation, because it other elements besides C, H, N, O therefore, those were
was slow-burning, highly luminous (determined by Lsp/BR marked differently in the graphs. They cannot be evaluated
value) and had a decent color purity of 90 %. To improve together with fully organic additives, since some contained
the compositions, it was considered essential to evaluate metal ions, i. e. Na (lignin sulfonate), Ca, Sr that are potential
additives that at least theoretically could enhance the flame emitters or emitter formation enhancers. Hence, some of
size by the formation of gas or contribute positively by be- them, notably lignin sulfonate and strontium oxalate, show
ing an effective secondary fuel. quite high Lsp values due to additional emitters present.
The use of gas generating materials has been reported The presented trends support the fact that high nitro-
back since the 1970s [21], and in recent years have been gen materials favor the production of more luminous
used in various illuminating pyrotechnic compositions as an flames [18], but it is evident that the influence of oxygen is
energetic additive for smoother combustion, more lumi- equally as important. Apparently, the formation of N2(g) and
nous flames, and compositions emitting less smoke
[5, 11, 22, 23]. In the present work several non-exotic, gas
generating materials were chosen to investigate how they Table 3. Modified composition 6 for testing additives.
affected the performance of the magnesium-rich system In Compositons 19–33
addition to the gas generating materials, calcium resinate,
shellac, strontium oxalate, phenolic resin, lignin sulfonate Sr(NO3)2 34
Mg rounded (45–100 μm) 48
(with sodium impurities) and sulfur were also tested as ad-
Chlorinated rubber 9
ditives for comparison. Additives were evaluated in “Com- Additive 5
Linseed oil 4

Figure 2. Lsp and flare’s light intensity dependence on composi- Figure 3. Lsp dependence on the amount of oxygen and nitrogen
tion's burning rate for compositions from Table 2. (in moles) present in the additive.

674 © 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Propellants Explos. Pyrotech. 2020, 45, 671–679
86
www.pep.wiley-vch.de
The Performance of Red Flare Pyrotechnic Compositions Modified with Gas Generating Additives

According to the NASA CEA [24] code estimation of

compositions (6, 31, 38) at atmospheric pressure (1 atm), N2
(g) is in equilibrium with Mg2N3 at lower temperatures (1000-
1500 K), however above 2000 K, N2(g) formation is predicted.
H2O(g) rapidly reacts with excess Mg(g) to yield H2(g) and
MgO(s/l). MgO(s/l) is then in equilibrium reaction with soot
[25] that yields CO(g) and Mg(g). Given that Mg containing
compositions reach temperatures above 2000 K [25] and all
water vapor (and oxygen) reacts with excess Mg, then N2(g),
H2(g), CO(g) molecules are formed and assist in increasing Lsp
in such Mg-rich flare compositions.

3.3 Influence of Guanidinium Nitrate on Flare

Performance
Figure 4. Lsp dependence on the amount of oxygen and nitrogen
(in grams) present in the additive. The performance of guanidinium nitrate (GN) in place of
chlorinated rubber was quite impressive. Having almost the
same burning and mass consumption rates, GN (comp. 31)
possesses 40 % higher Lsp value and 60 % higher light in-
tensity in comparison to comp. 6 in this system. Urea was
also very efficient and was noted for its low BR, see compo-
sition 30. However, urea is not useful in practice as it is rath-
er hygroscopic in combination with strontium nitrate and
could, therefore, be unstable together with Mg. Thus azodi-
carbonamide (ADCA) is a great alternative (comp. 21),
which is far less hygroscopic and well-performing in com-
parison to urea. 5-Aminotetrazole (5-AT), (and related N-rich
compounds) have received much attention these days for
their unique combination of high nitrogen content and
high stability due to aromaticity [5, 18]. Comp. 29 contained
5-AT and performed quite well, but with a slightly higher
BR than comp. 6.
Due to excellent performance as an additive, GN was
chosen for further investigation. Another advantage over
Figure 5. Lsp dependence on the oxygen balance of the composi- other additives explored herein is that GN by itself is an en-
tion. ergetic material and contains an oxidizing moiety – the ni-
trate anion. Therefore, it can support combustion better in
a strongly oxygen-deficient pyrotechnic system such as the
H2O(g) in such Mg-rich flare compositions, have a similar ef- ones of our focus. For the same reason, GN was sometimes
fect on the increasing Lsp. used in the past as an oxidizer alone in illuminating compo-
A general explanation of the observed phenomenon is sitions [26, 27] and gas generators [16]. Therefore, the Mg-
that the conversion of chemical energy to light is more effi- rich illuminating system could possibly tolerate even higher
cient in illuminating compositions that contain gas generat- amounts than 5 % of GN, which would be difficult for other
ing materials [16]. When gas generating materials are in- well-performing fuel-like additives from Table 4.
troduced into a composition instead of a non-gas In fact, tests of using ADCA in the 10 % range were per-
generating material, such as chlorinated rubber, a relatively formed, but the combustion process became uneven and
large flame envelope can result due to gas flame expansion. pieces of burning composition were shooting out of the
That, in turn, has a higher emissive area, which results in an flaring chimney as it was burning. This phenomenon is like-
increased light emission intensity. Other effects could be ly to occur when the gas generating material, present in the
the change in the flames absorbance and emissivity factors flare composition, starts to decompose in the pre-reaction
that directly determine the Lsp. When gaseous molecules re- zone of the composition column. In this case by “pre-re-
place less volatile or incandescent species, the inner ab- action zone,” we refer to the compacted flare composition
sorbance in the flame envelope is decreased resulting in a just a short distance below the combustion surface. As a re-
more luminous flame. sult, a piece of solid unburnt composition detaches and

Propellants Explos. Pyrotech. 2020, 45, 671–679 © 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.pep.wiley-vch.de 675
87
 Full Paper D. Juknelevicius, P. Alenfelt, Prof. A. Ramanavicius

ejects through the flaring chimney disturbing the combus-

Urea Guanidinium Hexamine Sulfur Chlorinated

tion performance.
Compositions were designed with a higher amount of
rubber

1,53
2,28

16,2
21,3
GN (10–25 % w.t. range). In this series of tests, GN mainly

7,1
29
6

2
0
0
0 replaced some of the chlorinated rubber and Sr(NO3)2, while
1,68
2,92

17,9
33,0
the Mg amount was kept in 40–50 % range (Table 5). The

6,2
33

0
0
0
0
0

experiments proved that GN can be used in rather large

amounts as an additive in Mg-rich compositions for low BR
1,64
2,40

15,3
22,8 and higher color purity and Lsp.
6,4
32

51

40
40
9
0

Sparks or glowing slag were observed to form during

nitrate (GN)

combustion only in certain compositions. In the first experi-

ments, compositions with the highest Mg content had high
burning rates of > 3 mm·s1. These compositions produced
1,67
2,39

20,1 22,7
26,4 34,2

10,1 9,5
31

10

39
46
85
5

fine sparks that were ejected upwards and were not ob-
1,76
1,99

served falling down.

30

20

27
47
73
7

Some GN modified compositions from Table 5 were

sulfonate tetrazole
5-amino

rather a slow-burning (< 1.8 mm s1). These had a tendency

(5-AT)

1,65
2,56

21,0
33,5

8,2

to produce larger sparks or glowing slag pieces that fell

29

14

82
82
4
0

down. None the less, certain slow-burning compositions

Melamine Azodicarbonamide Calcium Shellac Strontium Lactose Dextrin Phenolic Lignin

possessed high color saturation (39, 40, 43), and some oth-
1,64
2,63

22,5
36,6

8,6

ers (37, 38, 41) burned efficiently and clean (Table 5, Fig-
28

45

30

30
5

ure 6)
When compositions from Table 5 were tested, an inter-
resin

1,62
2,15

10,9
14,3

5,1
27

80

14

14

esting phenomenon was observed when evaluating the

6

emission spectra. In Figure 7 emission spectra of composi-

1,63
2,31

19,2
27,3

tions 37–40 are compared. The incandescence emission

8,3
26

44

49

49
6

seemed to have decreased with decreasing burning rate

and increasing GN w.t. %. The lower burning rate can affect
1,66
2,26

17,6
24,9

7,8
25

40

53

53

the proportion of how much solid particles remain de-

7

posited on the flare’s metal chimney and how much that

oxalate

enters the flame. Also, the lower burning rate can affect the
1,71
3,11

20,3
40,5

6,5

kinetic parameters, which determine how big spark par-

24

14

36

36
0

ticles are formed. Perhaps the explanation behind comp 40

1,62
2,17

10,7
14,1

performance (lots of sparks, yet a high purity red flame) is

4,9
Table 4. Performance parameters of compositions 19–33 with different additives.

23

68

23

23
9

that spark particles that formed were relatively big and not
resinate

as hot to give strong incandescence. Fine solid particles

1,63
2,27

12,3
8,8

3,9

that easily can be dispersed and heated up possibly emit

22

75

10

10
9

stronger incandescence emission, which could explain com-

position 37 performance (higher burning rate and lower
color purity). Additionally, a low burning rate can cause
(ADCA)

1,69
2,23

19,7
28,0

8,8
21

21

28
48
76
3

1,65
2,51

18,3
28,6

7,3
20

29

67
67
5
0
2-cyanoguani-
dine

1,62
2,39

17,4
25,4

7,3
19

29

67
67
5
0 O
N
ical consistency H
Additive chem- C

[kcd s2 g1 mm1]

O,N amount [%]
Density [g cm3]

L(sp) [kcd s g1]

Light intensity
(organic ele-

Burning rate
ments, %)

Figure 6. Combustion of flares with composition 11 (A) producing

[mm s1]
Additive

L(sp)/BR
Comp.

some sparks and glowing slag, and 38 (B) producing very few
[kcd]

sparks.

676 www.pep.wiley-vch.de © 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Propellants Explos. Pyrotech. 2020, 45, 671–679
88
The Performance of Red Flare Pyrotechnic Compositions Modified with Gas Generating Additives

Table 5. Formulations with 10–25 % of GN and their burning characteristics.

Compositon # 34 35 36 37 38 39 40 41 42 43
Sr(NO3)2 36 31 26 36 31 26 21 31 26 21
Mg rounded (45–100 μm) 40 40 40 45 45 45 45 50 50 50
GN 15 20 25 10 15 20 25 10 15 20
Chlorinated rubber 6 6 6 6 6 6 6 6 6 6
Linseed oil 3 3 3 3 3 3 3 3 3 3
Density [g cm3] 1,71 1,67 1,64 1,69 1,66 1,62 1,58 1,63 1,60 1,57
Burning time [s] 39,0 47,8 54,4 32,7 41,0 50,4 57,0 33,0 43,5 54,9
Burning rate [mm s1] 1,83 1,53 1,37 2,22 1,80 1,50 1,36 2,27 1,75 1,42
Light intensity [kcd] 17,9 14,5 11,8 25,5 21,0 16,1 13,5 30,9 22,1 15,4
L(sp) [kcd s g1] 15,1 15,0 14,0 18,1 18,7 17,6 16,7 22,1 20,9 18,3
L(sp)/BR [kcd s2 g1·mm1] 8,3 9,8 10,2 8,2 10,4 11,8 12,3 9,7 11,9 12,9
Flame color purity [%] 85 92 94 83 91 97 99 89 95 97
Spark formation* 0 1 1 0 0 1 2 0 1 2
* defines the relative amount of sparks generated outside the flame envelope. 0 – no sparks; 1 – some sparks; 2 lots of sparks.

more solid particles to be deposited on the walls of the

chimney because reaction products travel slower from the
burning surface, which provides more time for solids to set-
tle. That can also cause an increase in flame color purity.
Additionally, not a single composition from Table 4 shot
out burning pieces of solid composition, which was a rare,
but possible case for compositions 1–33. The uniform and
slow-burning is likely achieved due to GN property to melt
prior to decomposition, which seals the pores in the com-
position, before generating gases and combusting.
A final comparison of compositions that possess similar
BR (in groups) is presented in Table 6, together with the fi-
nal chlorine-free composition 44.

Figure 7. Flame emission spectra of compositions 37–40. A de-

crease of incandescence emission results in increasing color purity 3.4 Properties of Chlorine-Free Composition
and a dominant wavelength (DW) shift.
By further tuning compositions with GN and 6 % of chlori-
nated rubber, and inspiration from recent papers [4, 11], a
chlorine-free, Mg-rich composition was evaluated. Even

Table 6. Comparison of two groups of compositions from this work, that possessed similar burning/mass consumption rates and Cl free
composition 44.

Group A B C (Cl-free)
Comp. No. 6 21 31 3 39 44
Sr(NO3)2 34 34 34 31 26 34
Mg rounded (45–100) 48 48 48 48 45 48
Chlorinated rubber 14 9 9 17 6
Linseed oil 4 4 4 4 3 4
ADCA 5
GN 5 20 13
Magnesium stearate 1
Burning rate [mm s1] 2,24 2,23 2,39 1,68 1,50 1,71
Lsp [kcd s g1] 16,7 19,7 22,7 9,8 17,6 16,2
Light intensity [kcd] 21,5 28,0 34,2 9,2 16,1 17,4
Lsp /BR [kcd s2 g1·mm1] 7,4 8,8 9,5 5,8 11,8 9,5
Flame color purity [%] 89 88 87 92 97 80

Propellants Explos. Pyrotech. 2020, 45, 671–679 © 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.pep.wiley-vch.de 677
89
Full Paper D. Juknelevicius, P. Alenfelt, Prof. A. Ramanavicius

Finally, Mg-rich and Cl-free composition were inves-

tigated using GN, which burned with a similar light in-
tensity and efficiency as chlorine-containing composition,
but with decreased flame color purity in the 80 % range.

Acknowledgements
The company Hansson PyroTech AB, owned by Nammo AS, is
gratefully acknowledged for providing test facilities and equipment
for the experimental work. The Erasmus + program and Vilnius Uni-
versity doctoral studies program is acknowledged for a scholarship
(D.J.). The author is very grateful to Rutger Webb from Clearspark
LLC, Prof. Thomas M. Klapötke from LMU Munich and Dr. Rytis Ku-
bilius from Blikas LTD, for various inspiring discussions related to
the experiments presented herein. Also Dr. Ernst-Christian Koch for
consultations and useful comments for improvements.
Figure 8. Flame emission spectrum of chlorine-free, Mg-rich com-
position 44. Na(g), K(g) peaks are present due to impurities.

References
though it was similar to compositions 34–43, it possessed [1] Military Explosives. Washington, D. C.: Headquarters Depart-
slightly different spectral characteristics. Since SrCl(g) emis- ment Of The Army, 1984.
sion was absent, the emission spectrum was solely com- [2] B. E. Douda, Theory of Colored Flame Production,” U. S. Naval
posed of strong SrOH(g) emission together with atomic Na(g), ammunition depot, Crane, Indiana, 1964.
K(g) emissions (Figure 8). Chlorine is known to quench in- [3] D. L. Naud, M. A. Hiskey, D. E. Chavez, Perchlorate-Free Py-
candescence emission of solid particles such as MgO(s/l), rotechnic Formulations Utilizing Energetic Chlorine Donors: 1-
SrO(s/l), by reacting with them, and forming more volatile CDN, 11-CDN, 13-CDN, Z. Anorg. Allg. Chem. 2013, 639, 702–
706.
chlorides MgCl(g), SrCl(g). Therefore, with the absence of Cl,
[4] J. J. Sabatini, E. C. Koch, J. C. Poret, J. D. Moretti, S. M. Harbol,
the incandescence emission of the chlorine-free composi- Chlorine-Free Red-Burning Pyrotechnics, Angew. Chem. Int. Ed.
tion was stronger and resulted in a decreased color purity 2015, 54, 10968–10970; Angew. Chem. 2015, 127, 11118–
of around 79–82 %. 11120.
[5] G. Steinhauser, T. M. Klapötke, Green’ Pyrotechnics: A Chem-
ists’ Challenge, Angew. Chem. Int. Ed. 2008, 47, 3330–3347; An-
4 Conclusions gew. Chem. 2008, 120, 3376–3394.
[6] D. Juknelevicius, E. Karvinen, T. M. Klapötke, R. Kubilius, A.
In this work, an effective area of red flare formulations spe- Ramanavicius, M. Rusan, Copper(I) Bromide: An Alternative
Emitter for Blue-Colored Flame Pyrotechnics, Chem. Eur. J.
cific for their high Mg content (40–60 % w.t.), low burning
2015, 21, 15354–15359.
rate (BR), and presence of non-exotic gas generating mate- [7] J. Glück, T. M. Klapötke, M. Rusan, J. J. Sabatini, J. Stierstorfer, A
rials was investigated. Strontium- and Chlorine-Free Pyrotechnic Illuminant of High
A series of additives were tested in order to increase the Color Purity, Angew. Chem. Int. Ed. 2017, 56, 16507–16509; An-
performance of an illuminating flare composition. Certain gew. Chem. 2017, 129, 16733–16736.
oxygen and nitrogen-rich additives were noted for their [8] T. M. Klapotke, B. Krumm, M. Rusan, J. J. Sabatini, Improved
ability to increase the specific luminosity (Lsp). Finally, a Green-Light-Emitting Pyrotechnic Formulations Based on Tris
trend was found indicating that Lsp value is proportional to (2,2,2-Trinitroethyl)borate and Boron Carbide, Chem. Commun.
2014, 50, 9581–9583.
the total amount of oxygen and nitrogen present in the gas
[9] J. Gluck, T. M. Klapotke, M. Rusan, J. Stierstorfer, Green Colo-
generating additive. Guanidinium nitrate showed the best rants Based on Energetic Azole Borates, Chem. Eur. J. 2014, 20,
performance as an additive in the tested system. While 15947–15960.
maintaining low BR, it produced 40 % higher Lsp and 60 % [10] J. C. Poret, J. J. Sabatini, Comparison of Barium and Amor-
higher light intensity in comparison with a similar GN-free phous Boron Pyrotechnics for Green Light Emission, J. Energ.
composition. Changes in flame’s emissivity and absorbance Mater. 2013, 31, 27–34.
factors, as well as the total emissive area, have probably [11] J. J. Sabatini, A Review of Illuminating Pyrotechnics, Propellants
had the greatest influence on the increase of flare’s illumi- Explos. Pyrotech. 2017, 43, 28–37.
nating performance. [12] E.-C. Koch, Special Materials in Pyrotechnics: III. Application of
Lithium and its Compounds in Energetic Systems, Propellants
Different from conventional compositions, a rather small
Explos. Pyrotech. 2004, 29, 67–80.
amount of chlorine donor (4–6 % w.t.) was found to be suf- [13] P. Alenfelt, Corrosion Protection of Magnesium Without the
ficient to achieve a deep red flame when used in combina- Use of Chromates, Pyrotechnica, vol. XVI, 1995.
tion with gas-generating additives and high amount of Mg.

678 www.pep.wiley-vch.de © 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Propellants Explos. Pyrotech. 2020, 45, 671–679
90
The Performance of Red Flare Pyrotechnic Compositions Modified with Gas Generating Additives

[14] K. Kosanke, Pyrotechnic Chemistry. Journal of Pyrotechnics, In- Flames, US8486207B2, Clearspark, LLC, Glendale, CA (US),
corporated, 2004. 2013.
[15] I. Glassman, R. A. Yetter, Combustion. Elsevier Science, 2008. [23] M. A. Hiskey, D. E. Chavez, D. L. Naud, Low-Smoke Pyrotechnic
[16] H. Ellern, Military and Civilian Pyrotechnics. Chemical Publishing Compositions, US6312537B1, The Regents of the University of
Company, 1968. California, Los Alamos, NM (US), 1999.
[17] A. A. Shidlovsky, Fundamentals of Pyrotechnics. CFSTI, 1965. [24] B. J. McBride, S. Gordon, NASA CEA. Cleveland, Ohio: Glenn Re-
[18] J. D. Moretti, J. J. Sabatini, J. C. Poret, High-Performing Red- search Center, 2015.
Light-Emitting Pyrotechnic Illuminants through the Use of Per- [25] T. Shimizu, Selected Pyrotechnic Publications of Dr. Takeo Shimi-
chlorate-Free Materials, Chem. Eur. J. 2014, 20, 8800–8804. zu Part 3: Studies on Fireworks Colored Flame Compositions.
[19] J. D. Moretti, S. M. Harbol, D. E. Riegner, N. Carlucci, J. J. Sabati- Journal of Pyrotechnics Inc., 1999.
ni, J. C. Poret, Single-Point-of-Failure Mitigation: Prove-Out of a [26] T. M. Klapötke, M. Rusan, J. J. Sabatini, Chlorine-Free Py-
Green Light-Emitting Illuminant Composition in the Handheld rotechnics: Copper(I) Iodide as a ‘Green’ Blue-Light Emitter, An-
Signal Cluster and 40-mm Parachute Configurations, J. Energ. gew. Chem. Int. Ed. 2014, 53, 9665–9668; Angew. Chem. 2014,
Mater. 2014, 32, 293–299. 126, 9820–9823.
[20] Hansson PyroTech AB, Historic product portfolio, Sweden, [27] C. Jennings-White, Nitrate colors, Pyrotechnica, vol. XV, 1993.
1990.
[21] C. Bernardy, Illuminating Pyrotechnic Composition which Gen-
erates Gases, US patent 4078954, Société Nationale des Pou-
dres et Explosifs, France, 1978. Manuscript received: September 14, 2019
[22] R. Webb, M. Zebregs, J. F. Zevenbergen, M. P. Van Rooijen, Revised manuscript received: October 11, 2019
Low-Smoke Pyrotechnic Composition for Producing Colored Version of record online: February 12, 2020

Propellants Explos. Pyrotech. 2020, 45, 671–679 © 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.pep.wiley-vch.de 679
91
 Paper 3

Blue Strobe Pyrotechnic composition based on Aminoguanidinium

Nitrate

D. Juknelevicius, T. M. Klapötke, A. Ramanavicius

Propellants Explos. Pyrotechn., 2019, 44: 1466-1471

Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with

permission.

77
Full Paper
DOI: 10.1002/prep.201900184

Blue Strobe Pyrotechnic Composition Based on

Aminoguanidinium Nitrate
Dominykas Juknelevicius,[a] Thomas M. Klapötke,[b] and Arunas Ramanavicius*[a]

Abstract: A new blue strobe pyrotechnic composition rods. Strobe frequency, linear burning rate, chromaticity co-
based on aminoguanidinium nitrate (AGN) is described. Al- ordinates and color purity (of the flash reaction), sensitivity
ternative compositions from the literature that contain ei- to mechanical stimulii, DTA curves, humidity tests, high
ther tetramethylammonium nitrate or guanidinium nitrate speed camera footage are reported and discussed. Finally a
are compared to the new AGN composition. Pyrotechnic discussion of the strobe reaction mechanism has been in-
compositions were processed into 13 mm pellets and 5 mm cluded.
Keywords: Blue strobe · Copper chloride emitter · Aminoguanidinium nitrate · Fireworks · Pyrotechnics

1 Introduction 2 Experimental Section

A strobe composition burns in an oscillatory manner with a CAUTION! The mixtures described herein are potential ex-
smolder reaction occurring at all time and flash reaction oc- plosives, which are sensitive to mechanical stimuli, such as
curring periodically. The smolder reaction produces rela- impact, friction, heat, and electrostatic discharge. Although
tively a low amount of heat and forms a slag at the burning we encountered no problems in the handling of these ma-
front. When sufficient heat is generated, the semi-reacted terials, appropriate precautions and proper protective
slag is ignited, and a flash occurs. Then the smolder re- measures (safety glasses, face shields, leather coats, Kevlar
action continues further to the depth of the pyrotechnic gloves, and ear protectors) should be taken when preparing
column [1]. and manipulating them.
While colored strobes are created without significant N.B., in previous work AGN was found to be extremely
difficulties using a nitrate oxidizer or ammonium perchlo- dangerous when mixed with copper bromate, Cu(BrO3)2
rate (AP) combined with Mg, or magnalium (MgAl) powder and copper iodate, Cu(IO3)2. The sensitivity to friction was
[2, 3], a blue strobe is more difficult to make, since the high extreme and small amounts of experimental mixtures self-
temperature flash reaction involving oxidation of Mg or ignited during storage. Also a side product of AGN reaction
MgAl can destroy the temperature sensitive CuCl blue with BCC (black in color, obtain under acidic conditions)
flame emitter. had self-ignited while being wet on a filter paper.
Therefore, a blue strobe composition calls for a some- AP and AGN were synthesized in Prof. Klapötke’s en-
what different chemical composition. A few of such compo- ergetic materials research group (LMU, München). AP by
sitions have been described by Jennings-White, that uses neutralizing perchloric acid with ammonia solution. AGN by
the AP/tetramethylammonium nitrate (TMAN)/Cu system [4] reacting nitric acid with aminoguanidinium bicarbonate.
and by McCaskie who described guanidinium nitrate (GN) The final product was dried in a desiccator before use and
based blue strobe compositions [5]. Also, similar composi- water solution of AGN had pH 7–8. Copper powder (<
tions as mentioned above have been analyzed in our recent 150 μm) was from Grüssing. Basic copper carbonate, guani-
work [6]. However, TMAN is somewhat difficult to obtain dinium nitrate and PVC powder were from Sigma Aldrich.
and possibly an expensive material. In our experience, such
TMAN based compositions have difficulties of sustaining
combustion, which can lead to a low wind resistance in [a] D. Juknelevicius, A. Ramanavicius
practical applications. Faculty of Chemistry and Geosciences
In this work GN and aminoguanidinium nitrate (AGN) Vilnius University
were employed as replacements for TMAN in blue strobe Naugarduko st. 24, 03225, Vilnius, Lithuania
compositions containing AP, polyvinyl chloride (PVC) and *e-mail: arunas.ramanavicius@chf.vu.lt
[b] T. M. Klapötke
basic copper carbonate, CuCO3 · Cu(OH)2 (BCC). Due to bet- Department of Chemistry, Ludwig-Maximilian University of Munich
ter performance, AGN composition was studied in more de- Butenandtstr. 5–13 (D), 81377 Munich, Germany
tail. Fax: + 49 (0)89-2180 77492
*e-mail: tmk@cup.uni-muenchen.de

1466
93
© 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Propellants Explos. Pyrotech. 2019, 44, 1466–1471
Blue Strobe Pyrotechnic Composition Based on Aminoguanidinium Nitrate

Nitrocellulose (NC) with a nitrogen content of 13.25 % was option. A series of strobe compositions containing AP, GN,
from Nitrochemie Aschau GMBH. All chemicals used were AGN, different copper sources (Cu, CuO, Cu2O, BCC) were
ground with a mortar and pestle and passed through tested.
30 mesh screen before conducting experiments. BCC was chosen over copper and copper oxides as it
The strobe compositions were mixed using a mortar was capable of producing a blue flame flash of higher color
and pestle. Pellets of 2 g (13 mm in diameter, ~ 10 mm in purity and a better flash separation. However, CuO can also
height, 1 ~ 1.57 g cm3) were pressed in one increment by a work. PVC was also employed as recommended by E.
consolidation dead load of 2 tons. For extruding of 5 mm McCaskie. PVC seems to serve well as a low-energetic fuel
diameter rods the following procedure was followed: 2–4 g that possibly assists the smolder reaction [7]. After several
of composition was moistened with either MEK or Acetone experiments of testing different additives and varying the
to swell or dissolve the binder material (PVC or NC). Then a oxygen balance of such compositions led to a narrow area
5 mm ID plastic syringe with a cut end was used as a simple of a working blue strobe compositions. Such working com-
pump to press 10–25 mm long rods (~ 0.4 g each, 1 ~ positions possessed a distinct strobing behavior, sustain-
1,15 g cm3). The syringe was filled with the moist strobe able combustion and a seemingly good blue color of the
composition and while facing a hard surface the plunger strobe flashes. In Table 1 three different blue strobe compo-
was pressed down by hand to eject entrapped air and to sitions (GN, AGN, TMAN based) are compared. In Figure 1
consolidate the compositions as much as possible. Finally their strobing patterns are presented, that were obtained
the compressed rod was extruded and left to dry overnight by monitoring the light intensity at 450 nm in time. Two
at room temperature. types of samples were tested: 5 mm rods (0.4 g sample) and
Spectrometric measurements were carried out using a 13 mm pellets (2 g). The 5 mm rod samples were made with
HR2000 + ES spectrometer with an ILX511B linear silicon an intention of having a minimal energetic feedback, that
CCD-array detector controlled by software from OCEAN OP- could help lowering the strobe frequency.
TICS. The integration time for recording the emission spec- The composition A with GN produced the strobing ef-
tra was set to 20 ms whereas it was set to 5–10 ms for fre- fect (13 mm pellet), even though the light intensity of the
quency measurements. The detector-sample distance was flashes was rather weak. The 5 mm rod of A burned con-
1 m for 13 mm pellets and 0.5 m for 5 mm pellets. The DTA stantly with a very weak re-appearing blue flame that was

alyzer from OZM at heating rates of 5 °C min1. Visario G2

curves were measured with a 552-Ex differential thermal an- not registered by spectrometer.
The reference composition with TMAN burned with an
1500 Weinberger speed camera was used for filming at overlapping blue flashes as seen in Figure 1.
1000fps and SONY RX10 III for 100 fps and 500 fps re- The B composition performed quite well producing sta-
spectively. The impact and friction sensitivities were de- ble and distinct (Figure 2) blue flashes when tested as both
termined using a BAM drophammer and a BAM friction 5 rods and 13 mm pellets. Due to good performance as a
tester. The ESD test was performed with Xspark 10 instru- blue strobe the composition B was analyzed further to eval-
ment from OZM. The sensitivities of the compounds are in- uate its chemical stability and strobe mechanism. The chro-
dicated according to the U.N. Recommendations on the maticity diagram (Figure 3) shows the color purities of the
Transport of Dangerous Goods (+): impact: insensitive flash reactions of A, B and reference composition.
> 40 J, less sensitive > 35 J, sensitive > 4 J, very sensitive
4 J; friction: insensitive > 360 N, less sensitive = 360 N, sensi-
tive < 360 N > 80 N, very sensitive < 80 N, extreme sensitive 3.2 Humidity Test
< 10 N.
Blue strobes are generally known to degrade over time due
to possible ageing reactions that can occur. For example,
3 Results and Discussion AP can react with metallic copper, salts can undergo double
exchange reactions, acid-base as well as complex formation
3.1 Aminoguanidinium Nitrate Blue Strobe reactions are also possible.
Therefore, a humidity test was performed in order to es-
The idea of using a AP/GN base for blue strobes has been timate how moisture sensitive is the composition B at room
already investigated by McCaskie [5]. The high percentage temperature (Table 2). Each test was performed in a desic-
of GN requires a lot of energy to melt it and even with cata- cator with a saturated salt solution over a 3-day period.
lytic help of copper compounds, it’s decomposes slowly. Three samples of ~ 0.6 g were weight with an analytical
Therefore, GN based strobe compositions are possible, scale before and after exposure to enhanced relative hu-
however their performance does not surpass TMAN based midity.
compositions. After exposure to RH 75 % the first significant change in
Hence AGN was suggested as a possible replacement sample’s weight was registered. Moisture induces a reaction
that could surpass GN in its reactivity and have an advant- that releases gases, which causes the weight loss and the
age over TMAN being more easily available and cheaper color change.

1467
Propellants Explos. Pyrotech. 2019, 44, 1466–1471
94
© 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.pep.wiley-vch.de
Full Paper D. Juknelevicius, T. M. Klapötke, A. Ramanavicius

Table 1. Blue strobe chemical compositions of a GN base (A), AGN base (B), TMAN base (reference). Strobe frequency, linear burn rate,
chemical stability and sensitivity parameters are present.

Compositions A B Ref.
NH4ClO4 30 25 55
GN 50
AGN 55
TMAN 30
PVC 5 5
CuCO3 · Cu(OH)2 (BCC) 15 15
Cu powder (40–100 mesh) 15
NC powder +1%
Frequency, Hz 13 mm 3,5 � 2 8�2 > 10
5 mm rods const. 3,8 � 2 > 10
Linear burning rate 1,4 1,1 5,2*
(5 mm rod), mm/s
CIE coordinates** X 0,262 0,249 0,236
Y 0,296 0,275 0,289

Chemical stability (DTA), °C

Color purity,**% 27 32 35
247 179 245
Sensitivity tests Impact (J) 15 10 6
Friction (N) > 360 > 360 240
ESD (mJ) 42 33 51
* unusually high combustion rate resulted from the enhanced surface flame propagation of this particular composition. The linear burning
rate is expected to be lower. ** Light produced by the flash reaction.

Figure 1. Light intensity at 450 nm vs time.

The color change was also registered in several sample It must be noted that dry and acid-free AGN was used in
compositions that were stored before pressing pellets. our experiments (water solution shows pH of 7–8 on uni-
However, well dried compositions did not cause such re- versal indicator paper). In our experience, compositions
action for up to 2 months of storage at room temperature. with non-acidic AGN have significantly greater shelf life.

1468 www.pep.wiley-vch.de © 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Propellants Explos. Pyrotech. 2019, 44, 1466–1471
95
Blue Strobe Pyrotechnic Composition Based on Aminoguanidinium Nitrate

Figure 2. Consistent over time strobing pattern of composition B

Figure 4. DTA curves of pure AGN, AGN/BCC (55/15 w.t. ratio) and
tested as a 5 mm rod and 13 mm pellet. From the ignition point
composition B.
the rod was consumed in 25 s at a linear burn rate of 1,1 mm/s.
Light intensity was registered at 450 nm.

formed, which is likely to be a Cu aminoguanidinium com-

plex. When dried and ignited as loose powder, the complex
material burns fiercely producing a green flame that comes
from CuOH(g) emissions. The chemical stability this complex
species is quite unclear, and as mentioned before, AGN and
BCC reaction’s side product had self-ignited one time dur-
ing drying. Therefore, long term stability tests at elevated
temperatures are suggested before using in practice.

3.3 Differential Thermal Analysis (DTA)

DTA measurement was performed for composition B and

certain compositions with AGN. It was found that BCC cata-
lyzes the decomposition of AGN and significantly changes

to AGN/BCC with the same decomposition point at 179 °C,

the DTA curve of AGN (Figure 4). The B curve is very similar
Figure 3. Chromaticity diagram indicating color points of the flash
reactions of A, B and reference compositions. meaning that AP and PVC do not participate in this process.
Sometimes during the incremental heating of the DTA

showed a high-order deflagration (likely at 238 °C) that

sample, the test composition B (and very similar ones)
When 3 g of acid-free AGN is dissolved in water and 1 g
of finely ground BCC is added to the solution, the color of were powerful enough to rupture the mini glass test tube
undissolved BCC changed first to light violet and then to of the DTA instrument. The sample weight was ~ 40 mg.
purple. Bubbling is observed. After stirring the suspension
and letting react overnight, 2,8 g of purple precipitate is

Table 2. Exposure of composition B to different relative humidity generated by saturated salt solutions at room temperature.

Relative humidity at 25 °C, %

Salt K2CO3 NaBr NaCl KNO3
43 57,6 75,3 93,6
Pellets Weight change 0 0,0004 -0,0005 0,0246
Color change/cracks no no no yes*
Powder Weight change 0,001 0,002 0,0115 0,0375
Color change no greyish violet (lite) greyish violet deep purple**
* In 6 h purple spots appear. ** In 3 h turns grey-violet, in 6 h purple, after three days – deep purple.

Propellants Explos. Pyrotech. 2019, 44, 1466–1471

96
© 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.pep.wiley-vch.de 1469
Full Paper D. Juknelevicius, T. M. Klapötke, A. Ramanavicius

Figure 5. Smoldering surface of composition B 5 mm rod (A);

13 mm pellet (B) (capture from the high-speed camera footage at
1000 fps). Heavy boiling and gas generation is observed on the
molten reaction surface. Whitish boiling reactants can be observed
in the center of both samples.

3.4 Origin of the Bright Blue Flash

The high-speed camera videos give interesting information

on the strobe reactions observed in this work. There have
been a few curious phenomena registered that are de-
scribed, however it is somewhat difficult to draw a clear Figure 6. High speed camera footage of the oscillatory burning B
conclusion from the latter that would help to explain the composition. A) glowing gases appear at the reaction surface at all
strobe reactions that occur in AGN/AP/BCC system. How- time (500 fps); bright flame spot appearing prior to the flash re-
ever, they give some clues. action can be seen in the fifth frame; in the last frame flash reaction
starts. B,C) capture series of the occuring blue flash at 100 fps and
During the smolder phase gases are generated from the
500 fps respectively. The blue color is associated to the emission of
reaction surface (Figure 5). Around 50 % of the smolder re- CuCl and redish tip to CuO.
action time small blue-glowing lines of gases appear at the
reaction surface as well as micro blue flashes (Figure 6A).
As the glowing gases eject from the burning surface,
sometimes, especially few miliseconds before the flash re-
action, a significantly brighter flame spot appear (Figure 6A,
fifth frame) that travels upwards with the flow of gases. Per-
haps this is a region of higher temperature or a piece of
molten reactive composition being ejected from the pellet‘s
surface, that has a delayed ignition and burns to produce a
significantly bright blue flash/spot, that is associated to the
same brightness and color of the flash reaction (Figure 6A).
The flash reaction seems to be a result of (i) rapid exo-
thermic reaction that causes the ejection of flammable gas-
Figure 7. The ash scaffold is formed due to a strong heat loss to the
es from the boiling surface or (ii) an ignition of gases that
surounding atmosphere during the combustion of 5 mm rod
have already been formed above the pellet (Figure 6B). Pos- (comp. B). Partially red glowing ash scaffold in the last frame is ob-
sibly it can be a combination of the two as well. served.
The first assumption is made from our previous work

measurement, as the exothermic decomposition at ~ 240 °C

with TMAN strobes and the evidence from the current DTA

900 °C.
combustion, indicating the surface temperature of 600–
tends to be quite energetic often breaking the mini test
tube of the sample. Interestingly, being thin and fragile as the scaffold is, it
The second assumption followed observing the ash did not fall apart due the strobe reaction that would make
scaffold formation (Figure 7) and analyzing the high-speed one expect to have a certain pressure fluctuation at the sur-
camera footage. Most test samples burned rather clean face that is accompanied by the popping sound observed
without significant ash formation, however, few samples of every time when testing compressed and uncompressed B
a 5 mm rods burned leaving a thin scaffold of ash that was composition. For example, a classical white strobe composi-
almost as tall as the 20 mm long test sample (Figure 7). Also tion based on AP/MgAl/BaSO4 [8] burns to produce flashes
the ash scaffold was observed to be glowing red during that are more similar to a small portions of flash powder
deflagrating. This produces pressure fluctuations, that can

1470 www.pep.wiley-vch.de
97
© 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Propellants Explos. Pyrotech. 2019, 44, 1466–1471
Blue Strobe Pyrotechnic Composition Based on Aminoguanidinium Nitrate

be seen when such composition is compact into a card- GN guanidinium nitrate

board casing, that eventually burns off and the remaining’s BCC basic copper carbonate (malachite)
of the paper casing are blown away by the pressure waves Cu copper powder (electrolytic)
of the flashes. However, in the case of rod, the ash scaffold CuCl copper (I) chloride
remained still, indicating that there are no significant pres- CuO copper (II) oxide (black)
sure fluctuations at the burning surface, what supports the MgAl magnalium powder (Mg and Al alloy 50 : 50)
second assumption. NC nitrocellulose powder
Moreover, in Figure 6B it can be seen, the erupted blue PVC polyvinyl chloride
flash did not create pressure that would blow away the RH relative humidity
smoke cloud that had been formed from the beginning.
This would indicate that the flame had spread through the Acknowledgements
flammable semi-reacted smoke above the pellet, consum-
ing it, however not creating any significant pressure in the The DAAD (one-year grant) program is acknowledged for a scholar-
flame envelope. This also supports our personal ob- ship (D.J.). The authors are very grateful to Per Alenfelt from Hans-
servations during measurements. Visually, the combustion son Pyrotech AB, Kaj Fredriksson from Sweden and Rutger Webb
of the B comp. rod produced a rather steady flow of gases from Clearspark B.V. for many inspiring discussions related to the
experimental work presented herein. Also Marcel Holler and Mar-
from the rod’s surface that only flashed rapidly, and the cus Lommel for helping with the high speed camera experiments,
flashes did not disturb the uniform flow of the gases. Also Stefan Huber for the sensitivity measurements and Maximilian Wur-
the popping sound was very distinct and observed all time zenberger for various discussions and the help with the DTA meas-
for tests of Comp. B. urements.

4 Conclusions References
[1] J. M. Corbel, J. N. van Lingen, J. F. Zevenbergen, O. L. Gijzeman,
AGN was proved to be a suitable material in combination
A. Meijerink, Strobes: pyrotechnic compositions that show a cu-
with AP and BCC for producing a blue strobe pyrotechnic rious oscillatory combustion, Angew. Chem. Int. Ed. 2013, 52,
pellet. The composition B burned producing sharp and well 290–303.
defined flashes accompanied by a popping sound. [2] C. Jennings-White, Strobe Chemistry in: Pyrotechnic Chemistry,
The smolder reaction seems to be caused by the de- J. Pyrotech., Whitewater, 2004.
composition of AGN with a copper catalyst and the flash re- [3] R. Cardwell, Review on strobe light pyrotechnic compositions,
action is possibly a combination of an exothermic reaction Pyrotechnica 1982, 5, 6–24.
[4] C. Jennings-White, Blue strobe light pyrotechnic compositions,
at the surface and ignition of flammable gases above. How-
Pyrotechnica 1992, 14, 33–45.
ever, the origin of the cycling burning is unclear. [5] E. McCaskie, A new method for the production of blue strobe
Composition B powder showed sensitivity to moisture stars, Pyrotechnica 1993, 15, 35–45.
at RH = 58–75 %, and better resistance with RH = 75–94 % [6] D. Juknelevicius, A. Dufter, M. Rusan, T. M. Klapötke, A. Ram-
when tested in as a compressed pellet bind with PVC/MEK. anavicius, Study of Pyrotechnic Blue Strobe Compositions Based
In both cases the composition turned purple with a regis- on Ammonium Perchlorate and Tetramethylammonium Nitrate,
tered weight loss in the sample. Eur. J. Inorg. Chem. 2017, 1113–1119.
[7] T. Shimizu, Studies on strobe light pyrotechnic compositions,
Pyrotechnica 1982, 8, 5–28.
[8] T. Shimizu, Fireworks: The art, science, and technique, Py-
5 Abbreviations rotechnica publications, Austin TX, 1981.

AP ammonium perchlorate
Manuscript received: June 6, 2019
TMAN tetramethylammonium nitrate Revised manuscript received: August 1, 2019
AGN aminoguanidinium nitrate Version of record online: September 25, 2019

1471
Propellants Explos. Pyrotech. 2019, 44, 1466–1471
98
© 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.pep.wiley-vch.de
 Paper 4

Study of Pyrotechnic Blue Strobe Compositions Based on

Ammonium Perchlorate and Tetramethylammonium Nitrate

D. Juknelevicius, A. Dufter, M. Rusan, T. M. Klapötke, A. Ramanavicius

Eur. J. Inorg. Chem., 2017, 7: 1113-1119

Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with

permission.

78
 DOI: 10.1002/ejic.201601486 Full Paper

Energetic Materials
Study of Pyrotechnic Blue Strobe Compositions Based on
Ammonium Perchlorate and Tetramethylammonium Nitrate
Dominykas Juknelevicius,[a] Alicia Dufter,[b] Magdalena Rusan,[b] Thomas M. Klapötke,*[b]
and Arunas Ramanavicius[a,c]
Dedicated to Professor Dr. Wolfgang Beck on the occasion of his 85th birthday

Abstract: Pyrotechnic strobes, both colorless and colorful, are and a whitish-blue flame is observed. Therefore, it was interest-
compositions that burn in an oscillatory manner. They are used ing to see whether blue pyrotechnic strobes can be made with-
in military illumination devices and in fireworks as one of the out using conventional high-temperature fuels, such as Mg or
most attractive pyrotechnic effects. Most colored strobes con- Al, in order to conserve a deep flame color. The formulation
tain magnesium (Mg) or magnesium–aluminum (MgAl) alloy as called the Jennings-White blue strobe No. 5 [55 % ammonium
a fuel. However, this is an issue for blue strobes since the choice perchlorate (AP), 30 % tetramethylammonium nitrate (TMAN),
of stable blue-flame emitters is extremely limited. In fact, cop- and 15 % copper] is considered the top-of-the-line formulation,
per halides have been observed to produce blue emissions in which is described as being capable of producing an intense
pyrotechnic flames exclusively at temperatures ranging from blue strobe. However, its mechanism is not well-understood,
1200 to 1500 °C. At higher temperatures, the emitter is de- and the formulation is not chemically compatible. This paper
stroyed, or more specifically, a preferred trimer Cu3Cl3 species presents research focused on improving the blue strobe formu-
is converted into monomers. As a result, the flame color fades lation by Jennings-White.

Abbreviations Introduction
AP: ammonium perchlorate The pyrotechnic-strobe effect is defined as a cyclic variation of
TMAN: tetramethylammonium nitrate emitted light intensity as a function of time produced by a
TMAP: tetramethylammonium perchlorate light-emitting pyrobody.[1,2] The two conditions expressed by
this definition are the production of at least two flashes sepa-
GN: guanidinium nitrate
rated by a “dark” phase and a strobe frequency sufficiently low
Cu: copper powder (electrolytic)
for the human eye to distinguish between flashes.
CuO: copper(II) oxide (black)
Although the first pyrotechnic-strobe compositions were
MgAl: magnalium powder (Mg and Al alloy 50:50) published under the name “Orion Flashing Guns” as early as
EDN: ethylenediammonium nitrate 1898,[3] there has been little knowledge of how the strobing
TMEDN: tetramethylethylenediammonium nitrate behavior of such compositions comes about until Krone's[4,5]
PR: phenolic resin and Wasmann's[6,7] studies during the 1970s. Krone's experi-
NC: nitrocellulose powder ments, with compositions that comprised 20 % magnalium
PVC: poly(vinyl chloride) serving as fuel, and 65–75 % nitrates or perchlorates of the al-
K4–K112: Tested formulation numbers (full formulations are kali-earth metals serving as oxidizers and additives, gave a sig-
present in the Supporting Information) nificant hint about the strobing mechanism: during a reaction
phase, he observed the formation of a partly solid, partly mol-
[a] Faculty of Chemistry, Vilnius University, ten foam and he called it the semireacted slag, which contained
Naugarduko St. 24, 03225 Vilnius, Lithuania only small amounts of magnesium.[4,5] This result led Krone to
E-mail: arunas.ramanavicius@chf.vu.lt the conclusion that the oxidation of magnesium mainly takes
[b] Department of Chemistry, Ludwig-Maximilian University of Munich,
place in the previous dark phase, whereas aluminum is only
Butenandtstr. 5-13 (D), 81377 Munich, Germany
E-mail: tmk@cup.uni-muenchen.de oxidized in the presence of active gases in the semireacted slag
march@cup.uni-muenchen.de zone when the available surface area is higher. The oxidation
www.hedm.cup.uni-muenchen.de of aluminum results in a light flash. However, the occurrence of
[c] Laboratory of NanoBioTechnology, Institute of Semiconductor Physics, competing chemical reactions with opposite reaction pathways
State Research Institute, Centre for Physical and Technological Sciences,
A. Goštauto st. 11, 01108 Vilnius, Lithuania
(in this case, the slow oxidation of magnesium and the fast
Supporting information for this article is available on the WWW under oxidation of aluminum)[2] is only one way to explain strobing
http://dx.doi.org/10.1002/ejic.201601486. 100 behavior.
Eur. J. Inorg. Chem. 2017, 1113–1119 1113 © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Full Paper

Wasmann suggested that physical parameters essentially melted. When the layer remained partly solid, sometimes the
contribute to strobe reactions.[6,7] This hypothesis emerged gas escaped, breaking or even ejecting the layer due to the
from studies involving copolymers of unsaturated monomers, inflation underneath. When too much gas escaped, the flashes
which contained nitrate groups with conventional unsaturated were irregular and less sharp. The concentration of the gaseous
monomers and metal perchlorates that were soluble in the intermediate species oscillated due to their synthesis and con-
monomers. Due to the low thermal conductivity of the micro- sumption by competing chemical pathways.
porous residual mixture of the former reaction cycle, the heat The thermokinetic model developed by Davies,[10] which
inside the system built up in some areas until it reached the adapts a scheme created to model homogeneous oscillators
ignition temperature and light was emitted. Here, the periodic and uses the Salnikov model,[11] focuses on a binary composi-
strobing behavior is controlled by the heat transfer from the tion of Mg and AP. Herein, a set of compounds consisting of
burning surface into the interior of the system. solid and liquid magnesium react in a slow reaction, thereby
Another theory concerning the strobing mechanism comes producing gaseous magnesium. The fast and exothermic flash
from Shimizu.[8] Here, the ignition of a strobe composition on reaction occurs only if the heat generated is sufficient and if
the surface triggers a reaction in the dark zone, which is at a the concentration of intermediate species has reached a critical
different temperature, with a small activation energy and a level. The model claims not only thermal oscillations responsi-
small heat output,[1] and this leads to the formation of bright ble for the regulation of the progress of strobe reactions, but
spots, which are glowing and moving on the surface.[9] The hot also periodic variations in the concentration of gaseous inter-
spots magnify with increasing temperature in these areas. mediate species, which is more likely to explain strobe mecha-
When the temperature of the dark zone meets the ignition tem- nisms.[2]
perature, the flash reaction takes place. A small layer of the dark A common approach towards white, yellow, orange, red, and
zone on the burning surface, in which the temperature remains green strobe compositions is combining AP and magnesium,
below the ignition point, is left behind and is the starting point with a metal sulfate generating the desired color.[12] Blue strobes
for the next cycle.[8] Obviously, in this case, the temperature are not accessible in this system, since the high flame tempera-
controls the progress of the reaction. Furthermore, this mecha- ture during the combustion of such formulations keeps the con-
nism suggests that strobe compositions consist of one dark centration of the preferred copper(I) chloride emitter low.
composition, which gives rise to the dark reaction, and one The first blue strobes evolved from Jennings-White's studies
flash composition, which gives rise to a flash reaction. Shimizu's in 1990 were inspired by the existence of white-strobe composi-
hypothesis might be a useful criterion to decide whether a tions containing AP, magnalium, and some nitrogenous strobe
chemical composition is a possible strobe candidate or not.[1] material, as well as Shimizu's observation of a good dark reaction
However, theoretical and experimental analysis of diverse po- produced by a mixture of AP and 2 % of copper powder.[8,13] In
tential strobe compositions revealed that Shimizu's hypothesis these experiments, among others, TMAN was tested in combina-
does not always prove to be correct. The black strobe consisting tion with AP and copper-substituting magnalium.[13] One com-
of 30 % AP, 40 % TMAN, and 30 % Cu can also be referred to as position [the Jennings-White blue strobe No. 5, (Table 1)]
a blue strobe, because the combustion gases produced during showed outstanding burning characteristics: it produced flashes
the “flash” phase may be ignited to generate a blue flame and of a high-quality color that were well-separated. Additionally, the
increasing the amount of AP up to 40 % results in the ignition ignition sensitivity of this blue strobe was sufficient. The combi-
of these gases in air, of their own accord, to produce the blue nation of these two properties makes the Jennings-White blue
flame. However, neither a dark reaction nor a flash reaction can strobe No. 5 a promising candidate for practical use.
be postulated that complies with the constant change of a slow
Table 1. The Jennings-White blue-strobe formulation No. 5.
and a fast smolder without the emission of visible light at any
stage of the combustion. Vice versa, mixtures of a smolder com- Ingredient (wt.-%) No. 5
position and a flash composition do not generally show a strob- AP 55
ing behavior. The breakdown of the Shimizu hypothesis to ex- TMAN 30
Cu (atomized, 40–100 mesh) 15
plain the strobe effect of some compositions can be justified by
the fact that Shimizu did not claim the presence of a smolder
composition and a flash composition as a sufficient condition The aim of this work is to compose a blue-strobe formulation
for a strobe, but only as a necessary one. This implies that the that possesses lower strobe frequency and improved flash sepa-
description of the strobe mechanism by Shimizu might not be ration than the Jennings-White blue strobe No. 5, without a sig-
fully satisfactory. It also means that the mechanism requires nificant loss in flame color quality or combustion properties, and
experiments to detect the chemical changes occurring in differ- also to tune the chemical stability of this strobe composition.
ent phases of the cycle as well as studies of the thermodynam-
ics and kinetics involved.
Results and Discussion
Recordings of the combustion of several strobe compositions
by a high-speed camera[2] not only confirmed the previous find- The flashes produced by the Jennings-White blue strobe No. 5
ings, but revealed the formation of gaseous species under the (Table 1) are sharp and of an exceptional blue color, but the
surface layer at the stage of the dark reaction. The latter es- interval between the flashes is too small.[13] Although the igni-
caped through the foam forming bubbles when the layer101 tion sensitivity of the Jennings-White blue strobe No. 5 is rela-

Eur. J. Inorg. Chem. 2017, 1113–1119 www.eurjic.org 1114 © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Full Paper

Table 2. Selected strobe formulations for further analysis. K4: Original formulation (ref.[13]); K25: copper(II) oxide variation; K66: copper(II) benzoate variation;
K97: basic copper carbonate variation; K102: Nonstrobing, fast-burning; K105: sulfur-containing; K108: NC-containing. Reference: AP white strobe.

Ingredient (wt.-%) K4 K25 K66 K97 K102 K105 K108 Reference values
AP 55 55 55 55 55 55 55 60
TMAN 30 30 30 30 30 30
CuO 15 7.5 15 15 15
Cu 15
Copper(II) benzoate 7.5
Basic copper carbonate 15
TMEDN 30
MgAl (40–80 mesh) 25
BaSO4 15
Additive (add. %) +2 % sulfur +2 % NC +5 % K2Cr2O7
Sustainable combustion[a] S D D S S S S S
Relative strobe frequency[a] H M M M n/a M M L
Chemical stability (DSC) (°C) 245 270 305 245 165 300 270 > 330
Sensitivity tests
Impact (J) 6 7 5 15 7.5 4 7 7.5
Friction (N) 240 288 240 360 120 120 288 144
[a] Abbreviations are described in the Supporting Information.

tively high, it does not stay reliably lit. Furthermore, its chemical tion in the strobe composition, there are two different emission
stability, which determines the susceptibility to degradation spectra acquired during the measurement. Ideally, any strobe
during storage, and thus, the potential of this composition for should produce minimum light during the smolder stage and
industrial use, leaves much to be desired. Therefore, the aim of maximum light output during the flash stage. Therefore, during
this research project was to lower the strobe frequency, to im- the smolder reaction, the acquired emission spectra should rep-
prove the flash separation of the Jennings-White blue strobe resent the background level. Also, these stages must be well-
No. 5 without a loss in color quality, to make the combustion separated. This makes the strobe process more easily distin-
sustainable, and to tune the chemical compatibility of this guishable to the human eye. In the case of colored strobes,
strobe composition. Also, an understanding of the strobe mech- another criterion is added – a good colored strobe must pro-
anism would greatly benefit the empirical evaluation method duce a preferred emission during the flash reaction so that the
that was used for this experiment as the strobe mechanism of flash is colorful in and by itself. For blue strobes, copper halides
this particular system is not well-understood.[1,14] Parallel to the must be excited during the flash reaction in the flame to pro-
empirical optimization experiments, DSC measurements were duce dominating emissions in the blue region of the visible
performed and an ammonium perchlorate/tetramethylammo- spectrum. For this reason, the system of AP and TMAN is one
nium nitrate binary system was analyzed to get an insight into of the very few that have the right parameters for both the
the strobing mechanism. The empirical evaluation method was strobe reaction and blue-flame production.
used in the experimental section, which led to some general As all pyrotechnic compositions presented herein contain AP
conclusions on various strobe formulations that contained as an oxidizer, as well as some form of copper as a colorant, it
mostly AP and TMAN. During the experiment, a total of 112 can be assumed that copper(I) chloride is the only blue emitter
formulations were tested (Supporting Information). From these, formed during the burning processes of these formulations. For
eight formulations were selected for discussion for their con- this reason, the emission spectra are expected to be very simi-
spicuous performance (Table 2). These formulations were ana- lar, which is indeed the fact. The emission spectrum of K66 is
lyzed further by collecting additional physical and chemical pa-
rameters. Physical measurements include emission spectra and
strobe frequencies, and chemical analysis includes DSC meas-
urements and chemical stability tests of pressed formulations.
The sensitivity tests revealed that most formulations are sta-
ble up to 245 °C, except formulation K102, which contains
TMEDN. The impact sensitivity is in the range of 4–15 J, and
the friction sensitivity ranges from 120 to 360 N. According to
U.N. recommendations on the transport of dangerous goods,[15]
such formulations are classified as sensitive to very sensitive to
shock and sensitive to friction.

Emission Spectra
Emission spectra were acquired during the combustion process
of each strobe formulation. As there are two stages of combus-102Figure 1. Emission spectrum of the blue strobe formulation K66.

Eur. J. Inorg. Chem. 2017, 1113–1119 www.eurjic.org 1115 © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Full Paper

shown as an example (Figure 1). Dominating copper(I) chloride ment of a good strobe that possesses a sustainable reaction, a
emissions are present in the 400 to 500 nm region, and copper distinguishable strobing effect, and good chemical stability.
hydroxide produces a wide band in the 550 nm region, which Jennings-White has made a huge effort in the quest to eluci-
overlaps with the CuCl emissions. Copper oxide has two bands date certain strobe reaction mechanisms.[1] However, the AP/
in the red region. Sodium and potassium impurities give sharp TMAN system was the most difficult to explain, since the occur-
atomic lines at 589 and 767 nm, respectively. This pattern is ring combustion reactions were uncertain. It is usually referred
visible in the emission spectra of all formulations that contain to as a “black strobe”, because the base of this formulation does
AP, TMAN, and a copper source. The only observed fluctuations not contain any strong pyrotechnic flame emitters. Therefore,
are within the error limits and do not essentially influence the the flame is pale white, due to weak incandescent emission.
overall quality of the colored flame.
The formulation of AP and TMAN in the ratio of 2:1 is the
basis for this type of strobe formulation. However, the mixture
only stays alight with the help of an external heat source. The
Strobe Frequencies
third component added to this mixture, in our case, a mixture
Unsynchronized reaction zones progressing simultaneously of copper and its salts and oxides, acts as catalysts in AP decom-
through the pellet caused overlapping flashes that complicated position. This is what makes the composition sustain combus-
these measurements. The strobe frequencies of the selected tion. Copper, in this case, acts as a double-duty material, as it
formulations are estimated to be between 5 and 15 Hz (Fig- not only has a catalytic effect on AP decomposition, but also
ure 2), with strobe frequencies of 1–2 Hz for the reference. Ac- helps to sustain the reaction quite well (composition K4). This
cording to Jennings-White,[13] strobe frequencies between 1 could be related to the fact that copper acts as a fuel in either
and 10 Hz are applicable in pyrotechnics; thus, the determined flash or smolder reactions. In any case, the net heat produced
frequencies are within a good regime. during combustion is higher, thus increasing the energy feed-
back to the unburnt composition.
For the following experiment, we prepared formulation K98,
in which the ratio of AP to TMAN was the most original of
our tested formulations 11:6 (K98 contained 64.7 % ammonium
perchlorate and 35.3 % TMAN). The external heat source was a
butane torch. When the flame of the torch is pointed towards
a 13 mm, 2 g pellet of the pressed K98 formulation, it starts to
produce gases due to a low-temperature decomposition proc-
ess that occurs for AP below 300 °C. At the same time, the
pellet's surface starts to visibly melt and turns dark yellow, and
upon further heating, brown. At the moment it turns brown, a
certain temperature is reached and the molten residue reacts
very rapidly, producing a flash with a big flame envelope. At
this point, the molten layer is consumed in the flash reaction,
the torch starts to heat up the next layer of the unreacted for-
mulation, and the cycle repeats several times until the pellet is
consumed. Finally, the pellet is fully combusted till no residue
remains on the test plate. From this experiment, it is clear that,
the smolder reaction does not occur without the presence of a
third component, which acts as a catalyst or a low-temperature
fuel. Also it was important to understand the chemical consist-
ency of the slag, as it gave an insight into the occurring reac-
tions. So, in one experiment when the slag was already formed
and brown, just before the flash, the butane torch was removed
and the slag was cooled. Then it was collected, with possibly a
small amount of the remaining unburnt formulation, and ana-
lyzed.
Theoretically, during slag formation, volatile, low-molecular-
Figure 2. Light intensities at 450 nm versus time. (A) Strobing pattern of
mass species evaporate and high-molecular mass species,
formulation K25; (B) strobing pattern of reference (see Table 2).
oligomers, and polymers, which are not volatile, remain within
the slag. The formed slag is visibly yellow or brown in color.
This indicates the presence of unsaturated hydrocarbons, which
Strobe Mechanism
are essentially formed from the partial decomposition of TMAN.
The effort to explain the AP/TMAN strobe reaction mechanism To get more information of the slag's composition, NMR meas-
is an important part of this study. A more thorough understand- urements were performed. The results showed that the slag
ing would lead to a more purposeful approach to the develop-103 contains: (a) monosubstituted six-membered ring aromatic sys-

Eur. J. Inorg. Chem. 2017, 1113–1119 www.eurjic.org 1116 © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Full Paper
tems, which can be either benzene or pyridine moieties in an Conclusion
oligomer chain; otherwise, being volatile, they would have
evaporated at temperatures close to 300 °C; (b) alkyl chains; The substitution of TMAN in the Jennings-White blue strobe No.
and (c) carbonyl species. 5 (55 % AP, 30 % TMAN and 15 % Cu) by the less reactive fuels,
such as hexamine, TMAP, EDN, or TMEDN, causes the flash fre-
DSC Measurement quencies to rise so dramatically that the resulting compositions
Furthermore, differential scanning calorimetry was applied to are either very fast-strobing or they burn constantly. As a conse-
estimate the decomposition temperature under ambient pres- quence, formulations using the AP/TMAN/Cu system and/or
sure. The first observed exothermic reaction peak was the indi- CuO are still favorable over the analogous ones with the tested
cation of a decomposition reaction. The exact decomposition fuels, with regard to the strobing behavior. Nevertheless, it
temperature of a formulation was sometimes hard to predict, should be mentioned that the use of TMEDN as a fuel in pyro-
because multiple decomposition points and phase transitions technics is suggested due to its outstandingly high burning rate
took place. K66, for example, started to decompose at 305 °C, with ammonium perchlorate.
but the final ignition point was at 365 °C (strongly exothermic). Replacing Cu in the Jennings-White blue strobe No. 5 by CuO
For K102, the first decomposition could be noticed at 165 °C, results in a less consistent strobing behavior, but it improves
whereas the second one began at 350 °C. In the case of K97, the flash separation. A complete substitution of Cu by CuO ad-
an AP phase transition took place at 240 °C, before the actual ditionally lowers the strobe frequency. Substituting Cu by basic
decomposition started at 245 °C. In the DSC curve of K105, a copper carbonate yields a composition that produces very reg-
slight endothermic reaction, which could be attributed to the ularly occurring flashes and even sustains combustion. A re-
oxidation of sulfur, occurred at 185 °C. An exothermic reaction placement of half the amount of Cu by cupric benzoate in-
at 175 °C in the DSC curve of K108 indicated the start of the creases the ignition sensitivity and results in a burning behavior
decomposition of NC, which proceeded at 193 °C and reached superior to that of a formulation using CuO as the only colorant.
its maximum at 200 °C. Generally, such formulations are stable In that sense, the substitution of copper in the Jennings-White
up to 245 °C, except K102, which is not a strobe, but rather a blue strobe No. 5 by the less sensitive and cheaper colorants,
rapidly burning blue-flare formulation, with TMEDN as the fuel. cupric oxide, cupric benzoate, and basic copper carbonate, is
In addition to the DSC test, a chemical stability test was per- certainly possible, and it even improves the burning properties
formed. Pressed pellets were left to stand in humid air at tem- and physical characteristics of the pyrotechnic strobe.
peratures ranging from 0 to 8 °C. After 30 days, pellets were When the percentages of AP and TMAN deviate from the
inspected. All formulations had degraded over time to some 11:6 ratio used in the Jennings-White blue strobe No. 5, the
degree. Pellet K4, containing copper, completely changed its strobing behavior of the compositions gradually transitions to
color as a result of the metallic copper reaction forming Cu2+ a constant burning. Therefore, the ratio 11:6 between AP and
salts. Small blue spots appeared on formulations containing TMAN is suggested as the optimum ratio for all investigated
copper oxide and the reference formulation showed the maxi- blue strobe systems.
mum resistance to degradation. It must be noted that even The strobe frequencies of mixtures containing AP, TMAN, and
after the chemical stability test, all formulations still possessed CuO increase with an increasing percentage of this color im-
the strobing effect. parter.
Finally, with all data combined, the following explanation of Sulfur proved to be a dark-reaction sustainer for the system
the reaction mechanism in AP/TAMN strobe formulations is sug- AP/TMAN/Cu and/or CuO. The addition of 2 % sulfur to the sys-
gested: tem AP/TMAN/CuO even improves its strobing behavior. Higher
(1) AP and TAMN decompose at temperatures close to 300 °C amounts of this additive are not beneficial, since they lead to
in a complex way.[16,17] One of the dominant reactions is exo- increased strobe frequencies. The same effect was observed
thermic (smolder reaction), which occurs at around 300 °C and upon the addition of lactose to the same systems. The use of
increases the temperature of the reaction front further. Slag NC as an additive for compositions containing AP, TMAN, and
starts to form. cupric oxide did not only improve the performance and pro-
(2) TMAN partly decomposes to form a slag, which is stable mote a sustainable combustion for these formulations, but also
at temperatures above 400 °C. Slag is composed of highly react- proved to be a binder for the formulation. The use of dextrin
ive resins, oligomers with phenyl or pyridyl moieties, and aro- and phenolic resin in the system AP/TMAN/copper and/or
matics that possess low volatility. Otherwise, they would vapor- cupric oxide results in hardly ignitable or continuously burning
ize and leave the slag. compositions. Apart from sulfur, which causes the mixtures to
(3) When the temperature reaches 400 °C, due to an exother- stay alight, thereby decreasing their ignition sensitivity, none of
mic smolder reaction, the flash reaction's Ea is reached, the slag the tested additives are beneficial for the AP/hexamine/copper
gets oxidized, and a flash occurs. The oxidizer for the flash reac- system.
tion is possibly: (a) the partly decomposed oxidizers present in The strobe reaction has been explained in the last section of
the slag; and (b) AP and TMAN from the pre-ignition layer. this work. The mechanism, by itself, is closely related to the
(4) After the flash reaction, the smolder reaction progresses strobe theory by Shimizu,[1] which emphasizes that any strobe
to the depth of the pellet and the cycle repeats until the pellet formulation can be deconstructed to smolder and flash formu-
is completely combusted. 104lations.
Eur. J. Inorg. Chem. 2017, 1113–1119 www.eurjic.org 1117 © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Full Paper

Experimental Section with an ILX511B linear silicon CCD-array detector controlled by soft-
ware from Ocean Optics. The integration time for recording the
CAUTION! The mixtures described here are potential explosives, which emission spectra was set to 20 ms, whereas it was 10 ms for the
are sensitive to environmental stimuli such as impact, friction, heat, frequency measurements. The detector–sample distance was 1 m.
and electrostatic discharge. While we encountered no problems in Three samples were taken for each of the blue strobes K25, K66,
handling these materials, appropriate precautions and proper protec- K97, K102, K105, and K108 and all measured values were averaged.
tive measures (safety glasses, face shields, leather coats, Kevlar gloves, Averaging was based on the full burn of the mixture. Strobe compo-
and ear protectors) should be taken when preparing and manipulat- sitions that were not pressed, primed, and burned on the day of
ing them. manufacture were stored under dry conditions. For studying the
AP was employed as an oxidizer for most of the pyrotechnic compo- strobing mechanism, formulation K98 was chosen, because it was
sitions described in this report and was synthesized in Prof. Kla- the base mixture for the majority of nonmetallic strobe composi-
pötke's energetic materials research group (LMU, München) by neu- tions studied, and it did not sustain combustion. Each time the
tralizing perchloric acid with ammonia solution. A fine, 50 mesh formulation extinguished, a yellow-brownish combustion product
powder of the fuel TMAN was used as received. Hexamine was pur- containing a mixture of intermediate reaction products was
chased from Sigma–Aldrich. Tetramethylammonium perchlorate scratched off the white pellet. The latter was then investigated by
(TMAP) was synthesized according to the following protocol: per- means of 1H NMR measurements in [D6]DMSO and [D2]H2O at
chloric acid (5.79 g, 0.0576 mol) was diluted with distilled water, 400 MHz, 25 °C with a JEOL Eclipse: 400 NMR instrument. The de-
and tetramethylammonium hydroxide (5.25 g, 0.0576 mol) was composition points of K25, K66, K97, K102, K105, and K108 were
cooled down, separately, in an ice bath while stirring. Merging the measured with a LINSEIS PT10 DSC apparatus at heating rates of
two solutions resulted in the flocculation of the white metathesis 5 °C min–1. The impact and friction sensitivities were determined
product. Water was evaporated at 150 °C to form a crystalline prod- using a BAM drop hammer and a BAM friction tester. The sensitivi-
uct which was then placed in a drying cabinet. The fuels ethylenedi- ties of the compounds are indicated according to the definitions in
ammonium nitrate (EDN) and tetramethylethylenediammonium the U.N. Recommendations on the Transport of Dangerous Goods.
nitrate (TMEDN) were prepared by reacting ethylenediamine or The degrees of impact sensitivity are defined as: insensitive for val-
tetramethylenediamine with an excess of nitric acid in ethanol. The ues greater than 40 J, less sensitive for values between 40 and 35 J,
precipitate products were filtered and washed with diethyl ether. sensitive for values between 35 and 4 J, and very sensitive for values
As flame colorants, electrolytic 200 mesh copper powder, cupric below 4 J. The degrees of friction sensitivity are defined as: insensi-
oxide with a purity of 99+ % (Acros Organics), basic copper carb- tive for values greater than 360 N, less sensitive for a value of 360 N,
onate [Sigma–Aldrich, CuCO3·Cu(OH)2], and cupric benzoate were sensitive for values between 360 and 80 N, very sensitive for values
tested. The last one was synthesized by reacting potassium benzo- between 80 and 10 N, and extremely sensitive for values below
ate with CuII chloride dihydrate in aqueous media. Sulfur flower, 10 N.
lactose, nitrocellulose (NC) with a nitrogen content of 13.25 % A summary of all experimental formulations is given in Table S1.
(Nitrochemie Aschau GMBH), dextrin, phenolic resin (Blikas Ltd.,
Lithuania), poly(vinyl chloride) (PVC), and red gum were used as
received. All chemicals used were ground and sieved through a Acknowledgments
50 mesh screen before conducting experiments.
Financial support of this work by the Ludwig-Maximilian Univer-
After mixing the components for each strobe composition in a mor- sity of Munich (LMU), the Office of Naval Research (ONR) under
tar and pestle, pellets of 1 g (13 mm in diameter, 4.0 mm to 4.5 mm grant no. ONR.N00014-16-1-2062, and the Bundeswehr - Wehr-
in height) were pressed in one increment by a consolidation dead technische Dienststelle für Waffen und Munition (WTD 91) un-
load of 2000 kg. For initial assessment of strobe formulations K75,
der grant no. E/E91S/FC015/CF049 is gratefully acknowledged.
K92 to K96, K98 to K101, K103 and K104, K106 and K107, and K109
The authors acknowledge collaborations with Dr. Mila Krupka
to K112, no priming composition was employed. To ignite strobe
compositions K26 and K27, a droplet of a priming mixture consist- (OZM Research, Czech Republic) in the development of new
ing of 75 % potassium perchlorate, 20 % birch charcoal, and 5 % testing and evaluation methods for energetic materials and
shellac was pasted onto the surface of the pellets. However, the with Dr. Muhamed Suceska (Brodarski Institute, Croatia) in the
poor ignitability of these formulations gave reason to improve the development of new computational codes to predict the deto-
priming system. For all other pyrotechnic compositions presented nation and propulsion parameters of novel explosives. We are
herein, a thin layer (0.1 g) of a priming formulation, which was indebted to and thank Drs. Betsy M. Rice, Jesse J. Sabatini, and
composed of 65 % cupric oxide and 35 % magnesium, was applied Brad Forch (ARL, Aberdeen, Proving Ground, MD) for many in-
to the surface of the pellets, and a droplet of a mixture consisting spiring discussions. The authors greatly acknowledge Dr. Wag-
of 75 % potassium perchlorate and 25 % birch charcoal was placed
ner from Nitrochemie Aschau GmbH for providing nitrocellulose
on top. Both priming systems were wetted down with a 4 % solu-
for the experiments and Dr. Jesse J. Sabatini from ARL for many
tion of collodion (Merck), and homogeneous slurry was formed be-
fore pasting the primers onto the top of the pellets. Finished test inspiring discussions and support. The Erasmus plus program
units were dried for at least 1 h at room temperature. Ignition was and DAAD (German academic exchange service) is gratefully
imparted by touching the tip of the primer with a butane torch acknowledged for the scholarship (D. Juknelevicius).
flame. The combustion process was filmed with a digital video cam-
era recorder (Sony, DCR-HC37E) and the burning processes of the Keywords: Strobes · Oscillatory combustion · Copper ·
most promising strobe formulations K25, K66, K97, K102, K105, and
Magnesium · Oxidation
K108 were recorded with a high-speed camera. The performance of
each composition was evaluated with respect to flame color, smoke
generation, and the amount of solid residues. Spectrometric meas- [1] C. Jennings-White in Pyrotechnic Chemistry, Journal of Pyrotechnics Inc.,
urements were carried out by using a HR2000+ES spectrometer105 Whitewater, 2004.

Eur. J. Inorg. Chem. 2017, 1113–1119 www.eurjic.org 1118 © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Full Paper
[2] J. M. Corbel, J. N. van Lingen, J. F. Zevenbergen, O. L. Gijzeman, A. Meijer- Perchlorate, Proceedings of the 16th International Pyrotechnics Seminar,
ink, Angew. Chem. Int. Ed. 2013, 52, 290–303; Angew. Chem. 2013, 125, 1991, 16, 373–383.
306. [10] M. L. Davies, J. Pyrotech. 2008, 27, 42–49.
[3] A. S. H. Brock, Pyrotechnics: The History and Art of Firework Making, Daniel [11] M. J. Roberts, B. F. Gray, Proc. R. Soc. London Ser. A 1988, 416, 391–402.
O'Connor, London, 1922. [12] T. Shimizu, Fireworks: The Art, Science, and Technique, Pyrotechnica publi-
[4] U. Krone, Pyrotechnik: Grundlagen, Technologie und Anwendung, Das Insti- cations, Austin, 1981.
tut, Karlsruhe, 1975, pp. 225–237. [13] C. Jennings-White, Pyrotechnica 1992, 14, 33–45.
[5] U. Krone (Nico Pyrotechnik), DE-2164437, 1973. [14] E. McCaskie, Pyrotechnica 1993, 15, 35–45.
[6] F. Wasmann, The Phenomenon of Pulsating Burning in Pyrotechnics, Pro- [15] U.N. Recommendations on the Transport of Dangerous Goods, Model Regu-
ceedings of the 5th International Pyrotechnics Seminar, 1976, 5, 643– lations, 15th ed., United Nations, New York and Geneva, 2007.
651. [16] M. Rajić, M. Sućeska, J. Therm. Anal. Calorim. 2000, 63, 375–386.
[7] F. Wasmann, Pyrotechnik: Grundlagen, Technologie und Anwendung, Das [17] S. R. Jain, M. V. Rao, V. R. Pai Verneker, Proc. Indian Acad. Sci. Chem. Sci.
Institut, Karlsruhe 1975, pp. 239–250. 1978, 87, 31–36.
[8] T. Shimizu, Pyrotechnica 1982, 8, 5–28.
[9] T. Shimizu, The Effect of Hot Spots on Burning Surface and Its Application
to Strobe Light Formation with Mixtures Which Contain No Ammonium Received: December 13, 2016

106
Eur. J. Inorg. Chem. 2017, 1113–1119 www.eurjic.org 1119 © 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
NOTES

107
 Vilniaus universiteto leidykla
Saulėtekio al. 9, III rūmai, LT-10222 Vilnius
El. p.: info@leidykla.vu.lt, www.leidykla.vu.lt
Tiražas 20 egz.
108