Journal of Pyrotechnics - Issue 33 by CarnDu

Journal of

Issue 33, 2014 $100.00

Pyrotechnics

Fireworks - Pyrotechnic Special Effects - Propellants & Rocketry - Civilian Pyrotechnics

Full papers:

Information for Readers and Authors Journal of Pyrotechnics Archive on the web Events, Sponsors and Caution

ks or

Ro ck

ew Fir

Journal of Pyrotechnics

d an lE ia

la n

3 9 16

38 52 64

s ct

ffe

op el

Screening of hexachlorobenzene (HCB) contents in fireworks Silke Schwarz, Annett Knorr and Christian Lohrer Issues with UN explosives classification – a personal perspective Tom Smith Quantification of Visible Aerosols from Pyrotechnics: Metal and Metal Compound Additives Rene Yo Abe, Yoshiaki Akutsu, Akihiro Shimada and Takehiro Matsunaga Green Pyrotechnic Formulations Based on Metal-Free and Nitrogen-Rich Tetrazolylborate Salts Thomas M. Klapötke, Magdalena Rusan and Jörg Stierstorfer Display Fireworks And Stage Pyrotechnics In Use – Which Distances Are ‘Safe’ In Germany And Other Parts Of the EU? Christian Lohrer Comparison of national “safety distances” at professionally fired firework displays and distances derived from ShellCalc© Tom Smith and& Christian Lohrer Quantification of Visible Aerosols from Pyrotechnics: The Effect of Relative Humidity Rene Yo Abe, Yoshiaki Akutsu, Akihiro Shimada and Takehiro Matsunaga

iv il ia C

Dedicated to the Advancement of Pyrotechnics Through the Sharing of Information

Dedicated To The Advancement Of Pyrotechnics Through The Sharing Of Information

Other titles available from the Journal of Pyrotechnics Pyrotechnic Reference Series No 1

The Illustrated Dictionary of Pyrotechnics, 1995 [ISBN 1-889526-01-0]

No 2

Lecture Notes for Pyrotechnic Chemistry, 2004 [ISBN 1-889526-09-6]

No 3

Lecture Notes for Fireworks Display Practices, 2006 [ISBN 978-1-88952617-1]

No 4

Pyrotechnic Chemistry, 2013 [ISBN 1-889526-31-7 – Version 1.1].

No 5

Encyclopedic Dictionary of Pyrotechnics, 2012 [ISBN 978-1-889526-21-8] – CD Encyclopedic Dictionary of Pyrotechnics, 2012 [ISBN 978-1-889526-20-1] – 3-volume set – color Encyclopedic Dictionary of Pyrotechnics, 2012 [ISBN 978-1-889526-19-5] – 3-volume set – B&W

Pyrotechnic Literature Series (Please note new versions which replace previous individual publications) No 13 No 14

Selected Pyrotechnic Publications of K.L. and B.J. Kosanke, (1981 through 2009), 2012 [ISBN 1-889526-30-0]. Selected Pyrotechnic Publications of Dr. Takeo Shimizu, 2013 [ISBN 978-1-889526-29-4].

Shellcalc© Pro It is somewhat over 10 years since the first paper on Shellcalc© was published in the Journal of Pyrotechnics, although the program itself had been developed some time before by John Harradine in Australia following 2 firework incidents. This paper looks at the development and use of the Shellcalc© Pro program – a useful tool in the planning stage for any fireworks display company. Shellcalc© is a Microsoft Excel© based modelling tool useful to predict debris patterns from a variety of fireworks and special effects and is used by many of the World’s most accomplished firework display companies. Shellcalc© has undergone significant recent development and is now available in TWO versions – a free Standard version and an enhanced Pro version that allows multiple plots and specific tailoring of display output

For more information and pricing options please see www.shellcalc.co.uk or www.facebook.com/shellcalc or contact: CarnDu Limited 8 Aragon Place Kimbolton Huntingdon Cambs UK. PE28 0JD. Tel: +44 1480 878620 email: tom@carndu.com

Table of Contents - Issue 33, 2014 Full papers: Screening of hexachlorobenzene (HCB) contents in fireworks Silke Schwarz, Annett Knorr and Christian Lohrer Issues with UN explosives classification – a personal perspective Tom Smith Quantification of Visible Aerosols from Pyrotechnics: Metal and Metal Compound Additives Rene Yo Abe, Yoshiaki Akutsu, Akihiro Shimada and Takehiro Matsunaga Green Pyrotechnic Formulations Based on Metal-Free and Nitrogen-Rich Tetrazolylborate Salts Thomas M. Klapötke, Magdalena Rusan and Jörg Stierstorfer Display Fireworks And Stage Pyrotechnics In Use – Which Distances Are ‘Safe’ In Germany And Other Parts Of the EU? Christian Lohrer Comparison of national “safety distances” at professionally fired firework displays and distances derived from ShellCalc© Tom Smith and Christian Lohrer Quantification of Visible Aerosols from Pyrotechnics: The Effect of Relative Humidity Rene Yo Abe, Yoshiaki Akutsu, Akihiro Shimada and Takehiro Matsunaga

Information for Readers and Authors Journal of Pyrotechnics Archive on the web Events, Sponsors and Caution

Journal of Pyrotechnics, Issue 33, 20143

3 9 16

38 52 64

Page 1

Journal of Pyrotechnics

Policy Board Members

Ettore Contestabile Canadian Explosive Research Lab 555 Booth Street Ottawa, Ontario KA1 0G1 Canada

Izaskun Astondoa Pirotecnia Astondoa, S.A. Barrio Irupago s/n 48143 Areatza (Bizkaia) Spain

Andrew Tang Tian Cheng Pyrotechnics Laboratory Lihua Village, Yanxi Town, Liuyang City, Hunan, China 410304

Alexander van Oertzen BAM Federal Institute for Materials Research and Testing Division 2.3 “Explosives” Unter den Eichen 87 12205 Berlin, Germany

Tadao Yoshida Ashikaga Institute of Technology 268-1 Omae-cho, Ashikaga-shi, Tochigi 326-8558, Japan

Pierre Thebault Etienne LACROIX Tous Artifices S.A. Route de Gaudies 09270 MAZERES France

Bonnie Kosanke PyroLabs Inc 1775 Blair Road Whitewater CO 81527, USA

Tom Smith CarnDu Ltd 8 Aragon Place, Kimbolton Huntingdon, Cambs. PE28 0JD, UK

Christian Lohrer BAM Federal Institute for Materials Research and Testing Division 2.3 “Explosives” Unter den Eichen 87 12205 Berlin, Germany

Production Team Publisher: Tom Smith - CarnDu Ltd Production Editors: Helen Saxton - CarnDu Ltd Avril DiPalma - CarnDu Ltd Publishing Consultant: Bonnie Kosanke

Journal of Pyrotechnics. 8 Aragon Place, Kimbolton, Huntingdon, Cambridgeshire UK tel: +44 (1480) 878620 email: jpyro@carndu.com Page 2

Journal of Pyrotechnics, Issue 32, 2013

Screening of hexachlorobenzene (HCB) contents in fireworks Silke Schwarz, Annett Knorr and Christian Lohrer

Federal Institute for Materials Research and Testing (BAM) - Unter den Eichen 87, 12205 Berlin, Germany Email: christian.lohrer@bam.de

Abstract: This paper gives a brief overview of the findings of hexachlorobenzene (HCB) in fireworks compositions observed within EC type-examinations according to Module B (as set out in the annex II of the Directive 2007/23/EC) of the notified body BAM. In this work, roughly 220 samples were analysed, originating from 49 consumer fireworks articles of the category F2, such as batteries, combinations, fountains and rockets. It was found that the vast majority of samples showed concentrations below 5 mg HCB/kg. However, in three cases concentration values between 5 mg HCB/kg and 50 mg HCB/ kg were observed, and in a further four cases extreme values of up to 8046 mg HCB/kg were detected. The results of this study are compared with published HCB findings originating from market surveillance activities in Europe.

Introduction Hexachlorobenzene (HCB, sum formula C6Cl6) is a toxic, carcinogenic, persistent and bio-accumulative compound. Due to its highly persistent and lipophilic properties HCB degrades slowly and accumulates preferentially in fatty tissue. Therefore, HCB is categorized as a so-called persistent organic pollutant (POP substance). In Europe HCB has not been used or intentionally produced for many years. However, the use of HCB in the past was manifold. The widest use was probably within the agriculture and farming industry, as a fungicide or pesticide, as well as for seed treatments. In addition, it was widely used as a softener for polyvinylchloride (PVC) and fire protection agents in plastic materials. To a minor extent HCB was used as a chlorine donor, color enhancer and smoke intensifier in pyrotechnic compositions. Currently, HCB is an unintentional by-product of several industrial sectors where both chlorine and carbon are present. The concentration in the environment is mainly due to historical pollution and accumulation.1 Further continuative information on the historic use, properties, and typical reaction schemes of HCB was recently published by Smith and Guest.2 HCB is controlled by a number of European regulations. The placing on the market and use of HCB as a plant protection product has been forbidden in the European Union since 1981 by the Council Directive 79/117/EEC prohibiting the placing on the market and use of plant protection products containing certain active substances.3 The import and export of HCB was restricted by the Council Regulation (EEC) No 2455/92 of 23 July 1992 4 and by the Regulation (EC) No 304/2003 of the European Parliament and of the Council of 28 January 2003.5,1 Advanced legislation on HCB was achieved, within the framework of the United Nations Environment Programme (UNEP), by the Aarhus Protocol (1998) to the 1979

Convention on long-range transboundary air pollution on POPs and the Stockholm Convention (2001) on POPs.6 These international agreements establish a global regime for controlling POPs, including HCB, and aim at eliminating or reducing their use. They were implemented at the European level in 2004 by Regulation 850/2004,7 which prohibited, among others, the production, placing on the market and use of HCB. The Stockholm Convention sets out measures for the minimization and elimination of unintentional HCB emissions and was implemented at the European level by the Regulation (EC) 850/2004.7,1 At last HCB is totally banned (except for research) in Europe but can still be found as unintentional release or unintentional trace contaminant in products. Within the scope of CLEEN – Chemicals Legislation European Enforcement Network – the project EUROPOP was launched in 2010 with the aim of investigating the content of HCB in fireworks (mostly imported from China). Within the EUROPOP project a threshold limit value of 50 mg HCB/kg was stated. This value is based on the waste provisions following Article 7 of the POPs regulation.8 Some further information about a reasonable measurement threshold value of HCB in fireworks was recently published by Smith and Guest.2 They also proposed a cut-off limit of 50 ppm (which corresponds to 50 mg kg−1) in a colored star composition at which level analysis can conclude that the HCB has not been added deliberately. This value seems to be broadly accepted and applied by the market surveillance authorities and testing institutes in Europe. The latest standardization developments in Europe have taken this issue into account. According to the Directive 2007/23/EC of the European Parliament and of the Council on the placing on the market of pyrotechnic articles9 (article 8), European standardization bodies are requested to draw

Article Details

Article No:- 0103

Manuscript Received:- 5/2/2014

Final Revisions:-25/2/2014

Publication Date:-26/5/2014

Archive Reference:- 1665

Journal of Pyrotechnics, Issue 33, 2014

Page 3

up or revise European standards in support of this Directive. Therefore, the European Committee for Standardization (CEN) set up the Technical Committee CEN/TC 212. This technical committee consists of five Working Groups (WGs), each corresponding to the different pyrotechnic categories. With regard to fireworks articles, working groups 1 (consumer fireworks of the categories 1–3) and 2 (display fireworks of category 4 for persons with specialist knowledge only) were established. Up to this date, the following standard series were published: •

“EN 15947 Pyrotechnic articles – Fireworks, Categories 1, 2, and 310” consisting of 5 parts: 1—Terminology, 2— Categories and types of firework, 3— Minimum labelling requirements, 4—Test methods and 5— Requirements for construction and performance.

•

“EN 16261 Pyrotechnic articles – Fireworks, category 411” consisting of 4 parts: 1—Terminology, 2— Requirements, 3—Test methods and 4—Minimum labelling requirements and instructions for use.

In both relevant standard parts – EN 15947-510 and EN 16261-211 – the scopes are defined and it is stated that both standards do “not apply for articles containing pyrotechnic compositions that include any of the following substances”: hexachlorobenzene and polychlorobenzene, respectively (amongst others). These lists of substances are not exhaustive and these substances are often referred to as “forbidden substances” in fireworks. Due to the fact that the absence of HCB is not explicitly given as a requirement in the subsequent chapters of these standard parts, connected with a specific test method or procedure, the experimental proof of the absence within conformity assessment procedures

Figure 2. Dismantled fireworks rockets of category F2 (photo: BAM). applying these standards is not mandatory. However, some notified bodies under the Directive 2007/23/EC9 carry out tests regarding HCB within EC type-examination procedures of fireworks. BAM as one of the notified bodies saw the need to implement a screening procedure for HCB within the type tests of fireworks. An analytical method for the quantification of HCB was developed and different pyrotechnic compositions of 49 firework articles have been tested to date. Figure 1 and Figure 2 show examples of dismantled fireworks articles during the type examination procedures at BAM. This paper presents the main relevant results of the HCB screenings and gives comparison with results of other published studies.

Experimental set up Fireworks articles During a period of two years about 220 pyrotechnic compositions were analysed for HCB in the BAM laboratory. The samples were taken randomly from 49 firework articles. The fireworks articles belonged to the category F2 potentially presenting a low hazard and low noise level and which are intended for outdoor use in confined areas. Within this category the following fireworks types according to EN 15947-210 were part of this investigation: •

Batteries of shot tubes (various numbers of tubes),

•

Batteries of fountains,

•

Combinations of various types (e.g. shot tubes and mines),

•

Fountains, and

•

Rockets.

Analytical procedure Figure 1. Dismantled fireworks combination of category F2 (photo: BAM). Page 4

An analytical method was developed for the determination of HCB in fireworks on the basis of the procedures proposed by the CLEEN project EUROPOP8 and the Austrian Journal of Pyrotechnics, Issue 33, 2014

Environmental Agency.12 The procedure is described as follows. Sample preparation Depending on availability of the substance, masses of the pyrotechnic compositions of approximately 2 g to 7 g are weighed into a beaker. Distilled water is added to dissolve water-soluble compounds or to soften any starch based binder. Thereby, the subsequent crushing of larger composition particles is easier and safer. The water phase is separated by filtration and then discarded. During the developing of the analytical procedure the HCB content in the water phase was checked several times; HCB was not detected. After drying of the composition residue, an extraction with a mixture of hexane and acetone (1 : 1 v/v; approx. 50 mL) follows. The solvent is separated again by filtration. Analytical measurement For analytical measurement gas chromatography combined with mass spectrometry (GC-MS) was applied. The prepared sample solution, as described above, is used for a first screening of the HCB amount. 50 µL of it are dissolved in 1 mL of a solvent mixture of hexane and acetone (1 : 1 v/v) and measured by GC-MS in SCAN-mode. If an HCB peak is detected in this first test, the assumption can be made that the HCB content of the pyrotechnic composition exceeds the threshold value of 50 mg kg−1 (considering the above mentioned sample mass and the volume of the prepared sample solution). For quantification in SIM-mode the previously prepared sample solution must be diluted in an appropriate way to fit the concentration range, which is described by an HCBcalibration. HCB concentration lay between 0.1 µg HCB/mL and 2.5 µg HCB/mL for calibration. Six calibration levels were set. HCB with a purity of 99.8 area percentage GC

(analytical standard) was used. As internal standard lindane (C6H6Cl6, γ-hexachlorocyclohexane, analytical standard, 99.8 area percentage GC) was selected. It was dissolved in hexane to get a concentration of about 20 µg lindane/mL. An amount of 50 µL of the internal standard solution was added to the individual HCB solution. The same amount of internal standard solution was added to the eventually diluted sample solution for the final quantification measurements. The measuring process was run in cycles, where each cycle consists of analysis of the pure solvent mixture (hexane : acetone, 1 : 1 v/v), threefold measurement of the sample solution, followed by analysis of the pure solvent mixture again. After 20 measuring cycles an already used calibration solution had to be analysed again. When a deviation of 10% or more was observed a new calibration had to be done. Method parameters are given below. The SCAN- and SIM-modes differed only in detection of selected ions. Whereas in SCAN-mode ions in the range of 20 amu to 550 amu were detected, in SIM-mode only the target and qualifier ions of HCB and lindane were of interest. A typical GC-chromatogram with MS spectrum in SIMmode is displayed in Figure 3. According to the selected method parameters HCB was detected at a retention time at about 11.8 min, and the internal standard at about 12.7 min. The limit of detection is lower than 5 mg HCB/kg pyrotechnic composition. The limit of quantification is generally lower than the limit value of 50 mg kg−1. For the determination of the recovery rate several pyrotechnic compositions, which were probably clear of HCB, were spiked with different HCB concentrations. The spiked samples were run through all steps of the procedure. A recovery rate of around 70 % was determined.

Table 1. Method parameters of GC-MS Gas chromatograph (HP 6890 GC) Inlet conditions

230 °C, split-mode 100 : 1, sample volume 1 µL

Oven program

100 °C (hold time 2 min)

1st heating rate

15 K min−1 to 160 °C (holding time 0 min)

2nd heating rate

5 K min−1 to 280 °C (holding time 10 min)

Flow

Constant, 1 mL min−1 (carrier gas He)

Column

Nonpolar ((5%-phenyl)-methylsiloxane), length 30 m, ID 0.25 mm, film thickness 0.25 µm

Transfer line

300 °C

Mass spectrometer (HP 5973 MSD) Ion source

Electron impact ionization (70 eV), 230 °C

MS quadrupol

150 °C

SIM-mode

HCB target ion 284, qualifier 249, 282, 286 Lindane (int. standard), target ion 181, qualifier 183, 217, 219

Journal of Pyrotechnics, Issue 33, 2014

Page 5

the effect charges also other parts of the firework articles were investigated, e. g. priming composition, bursting and lift charges. The main results are presented in Table 2, the detailed single values are illustrated in Figure 4. As shown in Table 2 and Figure 4, HCB could only be detected for the effect charges. HCB was not found in primer, bursting or lift charges, as expected. Detectable HCB concentrations could be analysed in less than 4% of the effect charges. Only four samples had HCB contents above the threshold value of 50 mg HCB/kg. Although the overall detection rate of this value of 1.5% is low, the observed overstepping of the threshold value was found to be very dramatic in these cases (2000 mg HCB/kg up to 8046 mg HCB/kg). These high concentrations indicate an intentional addition of HCB to the effect charge. As a consequence of this the respective EC type-examinations were refused for the articles concerned. The newly requested types have not shown any relevant HCB contamination. In Table 3 the results for HCB in fireworks analysed within the scope of the EUROPOP project8 are shown. The samples for this investigation were taken within the scope of market surveillance activities from fireworks articles which were already placed on the market, i.e. having been already type and batch tested by notified bodies and manufacturers/ importers in Europe.

Figure 3. Gas chromatogram of HCB and lindane (SIMmode) and MS spectrum of HCB (SCAN-mode).

Results and discussion Within the EC type-examination procedures of fireworks according to the Directive 2007/23/EC9 a screening for the so called “forbidden substances” was established at BAM. As well as some heavy metals or certain substance combinations with chlorates, HCB is among this group. For the EC type-examination the manufacturer provides a sample representative of the production envisaged. Mainly the effect charges (e.g. stars) were analysed in each article due to the enhanced probability of finding HCB, potentially added as a chlorine donor and color intensifier. Besides

The percentage of samples in which the limit value of 50 mg HCB/kg was exceeded was found to be 10% on average. The observed violations ranged from 62 mg HCB/kg up to 27 000 mg HCB/kg. The EUROPOP report8 remarks that in only 13 cases was a legal action (e. g. product warning within the RAPEX notification system) taken. The findings of the EUROPOP project differ strongly in comparison with the results of this work. The violation rate observed in the BAM survey (1.5%) is much lower than the violation rate observed in the EUROPOP project (10%). One reason for this is maybe the different sources of fireworks which were investigated. For the BAM screening fireworks articles were investigated which were subjected to EC type-examination procedures, and which might have been specially manufactured or selected for this examination and the addition of HCB was avoided. In contrast to this, the EUROPOP project investigated only firework articles which were already placed on the European market after being type and batch tested by either notified bodies or

Table 2. Main HCB concentration results of the BAM screening Pyrotechnic composition

HCB concentration ranges <5 mg HCB/kg

5 to 50 mg HCB/kg

>50 mg HCB/kg

Effect charges (total number of samples = 201) Number of samples matching the HCB ranges

194

Percentage

96.5 %

2.0 %

1.5 %

Priming composition, bursting charges, lift charges (total number of samples = 20) Number of samples matching the HCB ranges

Page 6

Journal of Pyrotechnics, Issue 33, 2014

Figure 4. HCB concentrations measured during the BAM screening. manufacturers/importers. One has to consider the fact that within batch test procedures no dismantling or chemical analysis is mandatory. A deliberate addition of HCB to the pyrotechnic compositions is likely to remain undetected during batch testing according to the current requirements as set out in the respective standard part EN 15947-5. Taking the findings of this work and the referenced market surveillance activities into account, it might be worthwhile for the experts in the CEN/TC 212 (such as manufacturers, enforcement and notified bodies, market surveillance institutions), to reconsider if the experimental proof of the absence of some “forbidden” chemicals listed in the scope of the European EN 15947-5 like HCB should be included as

a specific requirement for EC type-examinations and, more importantly, for batch tests performed within quality systems of the manufacturers and importers in Europe. Especially the inclusion within batch testing procedures would require significantly more efforts compared to the current situation, as dismantling of an article would be a necessary step before the actual screening could be carried out. However, these additional measures for consumer safety must be carefully weighed against the possible unnecessary burden for all included parties (economic operators and notified bodies). Future developments and market surveillance results in this field should therefore be carefully observed and assessed.

Table 3. HCB contents in fireworks tested within the EUROPOP project8 Total number of samples

<5 mg HCB/kg Number

5 to 50 mg HCB/kg Number

>50 mg HCB/kg Number

Austria

Belgium

Denmark

Estonia

Finland

182

126

Sweden

Switzerland

439

317

Germany Iceland Netherlands

UK ∑ Percentage

Journal of Pyrotechnics, Issue 33, 2014

72.2%

— —

2 9

—

— 10 — —

17.5%

10.2%

Page 7

Acknowledgments The authors would like to thank Cordelia Schmidt and Ralf Jungnickel for analyzing the large quantity of samples.

References 1

A. Genty, “An Inventory and Assessment of Options for Reducing Emissions: Hexachlorobenzene (HCB)”, SOCOPSE Source Control of Priority Substances in Europe, Project contract no. 037038, 2009.

T. Smith and M. Guest, “A proposal to quantify trace levels of hexachlorobenzene in fireworks”, Journal of Pyrotechnics, Issue 30, 2011, pp. 28–35.

Council Directive 79/117/EEC of 21 December 1978 prohibiting the placing on the market and use of plant protection products containing certain active substances, Official Journal of the European Union, 08.02.1979.

Council Regulation (EEC) No 2455/92 of 23 July 1992 concerning the export and import of certain dangerous chemicals, Official Journal of the European Union, 29.08.1992.

Regulation (EC) No 304/2003 of the European Parliament and of the Council of 28 January 2003 concerning the export and import of dangerous chemicals, Official Journal of the European Union, 06.03.2013.

Stockholm Convention on Persistent Organic Pollutants; http://chm.pops.int/default.aspx [accessed on January 15th, 2014].

Regulation (EC) No 850/2004 of the European Parliament and of the Council of 29 April 2004 on persistent organic pollutants and amending Directive 79/117/EEC, 30.04.2004.

CLEEN - Chemical Legislation European Enforcement Network, “EUROPOP Final Report. Published by the EUROPOP Working Group”, http:// www.cleen-europe.eu/projects/europop.html, 2012.

Directive 2007/23/EC of the European Parliament and of the Council of 23 May 2007 on the placing on the market of pyrotechnic articles, Official Journal of the European Union, 14.6.2007.

10 EN 15947 Standard series for Pyrotechnic articles – Fireworks of Categories 1, 2, and 3, consisting of five parts, CEN/TC 212 WG1, 2010. 11 EN 16261 Standard series for Pyrotechnic articles – Fireworks of Categories 4, consisting of five parts, CEN/TC 212 WG2, 2012. 12 Umweltbundesamt Österreich, “Bestimmung von Hexachlorbenzol HCB in Feuerwerkskörpern mittels GC-MS. MKB”, Prüfart 07.2, Nummer 4, Version 01, Abteilung O [in German].

Page 8

Journal of Pyrotechnics, Issue 33, 2014

OPINIONS This article is the first, we hope, of many written by experts in the Pyrotechnic and Explosives sectors to highlight issues of the day and to stimulate discussion both within the pages of the Journal and elsewhere. Opinions are just that â&#x20AC;&#x201C; they represent views by their respective authors which are not reporting of scientific findings, but nonetheless have a place in the pyrotechnic literature. Opinions will be peer reviewed, as are all articles in the Journal, and authors will, we hope, take account of comments from the Board and reviewers before publication. However the subjects discussed and the opinions expressed remain with the respective authors and may not reflect the positions of the Journal of Pyrotechnics, its Board or other contributors. If you have a topic for consideration in future issues, please contact the Publisher directly.

Issues with UN explosives classification â&#x20AC;&#x201C; a personal perspective Dr Tom Smith

CarnDu Ltd, 8 Aragon Place, Kimbolton, Huntingdon, Cambs UK PE28 0JD Email: tom@carndu.com

Abstract: The United Nations classification regime for explosives is well established and has undoubtedly led to improvements in transport safety. However there are challenges that the hazard-based approach faces and this paper attempts to highlight some of the more pressing issues.

Introduction The United Nations (UN) classification regime1 for determination of the hazard as presented in transport of explosives (Class 1 dangerous goods) is well established and well recognised throughout most of the world. Even those countries that do not conform entirely to the UN requirements recognise the merit in such a system and are exposed to it through, for example, imports and exports. Table 1. UN Hazard Divisions. Hazard Division HD 1.1 HD 1.2 HD 1.3

However it is apparent to many that the system is being used for purposes that it was never designed for, and that it copes poorly with modern risk reduction approaches to explosive safety. This paper is not intended to propose a sudden change from the hazard-based approach that exists now, but instead to highlight areas where developments may be needed to maintain the credibility and practicability of the system in the future.

Definition Substances and articles which have a mass explosion hazard (a mass explosion is one which affects almost the entire load virtually instantaneously) Substances and articles which have a projection hazard but not a mass explosion hazard Substances and articles which have a fire hazard and either a minor blast hazard or a minor projection hazard or both, but not a mass explosion hazard: (i) combustion of which gives rise to considerable radiant heat; or

HD 1.4

HD 1.5 HD 1.6

(ii) which burn one after another, producing minor blast or projection effects or both Substances and articles which present no significant hazard: substances and articles which present only a small hazard in the event of ignition or initiation. The effects are largely confined to the package and no projection of fragments of appreciable size or range is to be expected. An external fire shall not cause virtually instantaneous explosion of almost the entire contents of the package Very insensitive substances which have a mass explosion hazard Extremely insensitive articles which do not have a mass explosion hazard: articles which contain only extremely insensitive detonating substances and which demonstrate a negligible probability of accidental initiation or propagation

Article Details

Article No:- 105

Manuscript Received:- 01/06/2014

Final Revisions:- 14/06/2014

Publication Date:- 15/06/2014

Archive Reference:- 1672

Journal of Pyrotechnics, Issue 33, 2014

Page 9

What is classification for and why do we do it? In essence the UN classification regime determines the hazard of an explosive (as a substance or an article) as packaged for transport and assigns three parameters – a Hazard Division and a Compatibility Group as well as a four-digit “UN Number” – to that explosive in the specified packaging as shown (in abbreviated form) in Table 1. Fundamentally, the Hazard Division determines the quantities that may be transported in vehicles of different types, the Compatibility Group determines which explosives may be transported together, and the UN Number allows the correct Shipping Name to be presented, the correct types of packaging to be used and leads to any special provisions which should be observed. In addition to providing information to the consignor of the explosives, and information to the transporters, critically the correct assignment of the three parameters provides emergency information to the transporter and to the emergency services in the case of accident during transport.

likelihood of an initiation taking place should be effectively redundant. For example, in the case of Rocket Motors (e.g. UN 0181) the special packaging provisions (L1) require that “their ignition systems shall be protected against stimuli encountered during normal conditions of carriage” – but this is plainly a provision to reduce the likelihood of accidental ignition and not one to contain or reduce the effect once ignition has taken place. In other words, this provision affects risk and not hazard. Another example is the insistence by some that items are transported without electric igniters attached, or that items with or without electrical ignition require different formal classifications. While the presence or absence of a few milligrams of an igniter composition will not affect the hazard resulting from an ignition of the item, it is fair to say that it may affect the likelihood of an ignition taking place (although this is overemphasised by many) but this is an issue of risk not of hazard. Modern regulations tend to consider risk not hazard and a sensible approach to risk reduction in the carriage of explosives would be to:

What are the issues?

•

Reduce the likelihood of initiation

The main issues that have been identified by the author and in discussions with other industry parties are:

•

Reduce the likelihood of communication within the load

•

The difference between risk and hazard

•

Increase the time an event takes

•

The consequential anomalies with the HD 1.5 and HD 1.6 classifications

•

Reduce the consequences of the event presuming all of the above have failed

•

Initiation issues

•

The importance of getting classification right

•

The absence of time factors in providing information to emergency services

It is only the last of these which is truly related to hazard – the first three reduce the likelihood of, for instance, the whole of a transport load exploding simultaneously (or even in quick succession) and hence not only reduce the risk but allow a sensible and effective response by the emergency services.

•

The relationship between classification and packaging

•

The improper extension of the UN scheme to nontransport situations

Hazard divisions 1.5 and 1.6

•

UN Hazard Division and security

•

Effects of bulk

•

The default classification system for fireworks

•

Time/pressure testing for determination of flash powder

•

Mutual recognition of classifications

•

GHS

Each will be examined in turn.

Hazard and risk The UN system essentially only assesses the hazard of an explosive event – it considers the worst case effect of an explosion of the packaged explosives, having made the assumption that the explosion occurs (by whatever initiation means). As such any special provisions which attempt to reduce the Page 10

The distinction could not be clearer than for explosives classified as 1.5 and 1.6. In both cases the definition refers to “insensitive” – but insensitivity is a measure of the likelihood of initiation not the consequences of it. A large amount of expense and time has been devoted in developing, for example, insensitive munitions – the reasons are obvious and desirable – but this insensitivity should have no bearing on the determination of hazard during transport. Indeed the definition of HD 1.5 recognises this – it acknowledges that such explosives have a mass explosion hazard and thus, we conclude, should be classified as HD 1.1. HD 1.6 states that such explosives have “a negligible probability of accidental initiation or propagation” – but probability is a determinant of risk and not hazard. These inconsistencies in approach devalue the UN process and provide little valuable information to the emergency services in the case of incident.

Journal of Pyrotechnics, Issue 33, 2014

Initiation issues The issue of initiation is wider than just 1.5/1.6 issues. Many explosive substances will not behave as explosives unless they are properly initiated – for instance by a detonator. In the absence of such initiation, most modern plastic explosives will simply burn – yet the UN 6(a) and 6(b) tests require a detonator to be deliberately inserted into such explosives solely for the purposes of the test, with inevitable consequences which are completely unrelated to any realistic transport situation. There seems little point in striving to develop and manufacture such explosives if the tests required for classification will rate them equal in hazard to the most sensitive explosives.

The importance of getting classification right It is obvious that it is important to get the hazard classification “right”. Although there have long been arguments that in the absence of test data all classifications should default to the “worst case”, HD 1.1, this is neither practical nor sensible. There is a danger that emergency services would over-react in case of incident with the potential for future incidents to be treated complacently because the last one was seen to have been over-rated. Of course it is also vital that the incident is not under-rated – and hence it is obvious that the only proper solution is to rate hazards properly! A better understanding (by operators, enforcers, the emergency services and the public) of the difference between risk and hazard would be beneficial but, sadly, unlikely. It is important, however, that the classification process is done in a timely and practical way (mutual recognition would help this – see later) in which means are taken to ensure that single items and mixed packages have an accurate classification assigned to them. It does appear anomalous that explosives require classification by Competent Authorities when all other classes of dangerous goods allow determination of hazard by the producer and/ or consignee. The undesirable effect of Competent Authority involvement is often that the process becomes bureaucratic and costly

– and hence there is a temptation by some to “cheat”. In principle a “simple process/strong enforcement” approach, where the enforcement activity is centred around ensuring compliance by all, would satisfy these requirements. Indeed, it does appear that in many countries the complications and costs of the process of classification have meant that items have been produced (often imported), transported, stored, used and have ceased to exist before the bureaucratic process of assigning classification has been completed. This is not an acceptable position and although no one should condone it, it is understandable.

Time factors One of the failures of the UN system is the lack of timecritical information about the possible explosive event. For instance HD 1.3C propellants and HD 1.3G fireworks seemingly present similar hazards to the emergency services. The former might produce a very high thermal flux on ignition for a period of a few seconds whereas the latter might eject an occasional star from a package over a period of minutes or even hours. In dealing with an incident both the transporters and the attending emergency services would like to have an appreciation of whether the incident is likely to take •

Milliseconds

•

Seconds

•

Minutes

•

Hours

•

Days

Examples are shown in Table 2. It would be of great benefit if there was some method of providing such information to the emergency services.

The relationship between classification and packaging It is generally not appropriate to attempt to give an explosive an inherent classification – the classification is ultimately dependent on the packaging in which the explosive is transported. In some cases (e.g. the default classification of fireworks – see later) where the packaging is closely defined

Table 2. Effects of time on explosion effects and possible emergency reponses. Timescales Milliseconds to seconds Seconds to minutes Hours Days

Example and comments HD 1.1 event – little possibility of subsequent explosions – in all likelihood the event will be over by the time the emergency services have been notified, let alone arrived at the scene Possibly an HD 1.2 event when pieces of the explosive article itself or of its packaging are ejected or an HD 1.3C propellant fire. Again in all likelihood the event will be over by the time the emergency services have been notified let alone arrived at the scene A firework fire of predominantly HD 1.3G fireworks For example an HD 1.4 event where atmospheric oxygen is required to maintain fire between the packaged items (burning fibreboard boxes). If the oxygen is depleted (e.g. in a closed vehicle or container) only when the oxygen is re-admitted (on opening the doors) will the fire and subsequent explosive events restart.

Journal of Pyrotechnics, Issue 33, 2014

Page 11

this may be possible but as a hypothetical example: •

Packing 72 theatrical maroons in a heavy-walled fibreboard box – HD 1.1G (UN 0428)

•

Packing 12 of the same maroons in a metal “ammunition” style box – HD 1.2G (UN 0429) from fragments of the box

items packaged for transport and classified by the UN properly as simply HD 1.2. These terms, intended to discriminate between the different behaviour of HD 1.2 explosives in bulk storage, lead to confusion and are not used within the civil sector. Again it is attempting to use the UN system for transport to reflect hazards in storage – a purpose for which it was not designed.

•

Packing 12 of the same maroons in a weak-walled fibreboard box – HD 1.3G (UN 0430) (as individual maroons are thrown and then function to throw fiery projections outside the 1.3 test distance)

In the UK a separate (but related) regime has been introduced to attempt to separate the determination of hazard as packaged for transport (the UN regime) from the hazard presented in manufacture, storage and even use.

•

Packing a single maroon in a light-walled fibreboard box – HD 1.4G (UN 0431); the effect is limited to the immediate area around the box if it ruptures

•

Packing a single maroon in a very large fibreboard box – HD 1.4S (UN 0432) as the effects are confined within the box

The UK system is called “Hazard Type” and there are four divisions HT1, HT2, HT3 and HT4 which are related to HD 1.1 etc provided that the explosives are stored in their transport packages. Again it is notable that there is no HT5 or HT6.

Improper use for use, storage and major hazard legislation The UN system properly determines the hazard of explosives as presented for transport. We make no apology for repeating that phrase often – it is critical to the understanding of the UN system and its proper application. Unfortunately the use of Hazard Divisions of packaged explosives as a determinant has been extended into other areas: •

•

The suitability for sale – for instance it appears some countries regard only 1.4G (UN 0336) fireworks as suitable for sale to the public rather than striving to ensure the fireworks suitable for sale obtain a classification of HD 1.4G so that they may be transported by members of the public in their own vehicles more safely. As seen above, the classification is often dependent on packaging. Attempting to equate hazard in transport and hazard in use is unrealistic and makes no logical sense. The suitability for transport by air – in some countries articles have been awarded an HD 1.4S classification so they may be flown, rather than being allowed to be flown because they have been awarded an HD 1.4S classification! Recent attempts have been made to address this anomaly but they have been only partly successful.

•

The thresholds for explosives in the Seveso Directives2 are based on UN classifications – irrespective of the fact that the explosives may be stored in bulk, may not be in their transport packagings or may be in manufacture. The Seveso directive attempts to address this but fails because of the convenience of using UN Hazard Divisions. (It is notable that the Seveso Directives ignore HD 1.5 and HD 1.6.)

•

The NATO use of mixed transport/storage classifications such as 1.2.1 and 1.2.2 to indicate hazard in storage of

Page 12

However in bulk and in situations where the explosives are not in their transport packages the Hazard Type may differ (up or down) from the equivalent Hazard Division. In essence the Hazard Type may be related to (all related to an explosive storage situation for simplicity): •

The packaging (or not) of the explosives

•

The quantity (or the explosive loading density within the store)

•

The orientation within a store

•

The orientation of the store itself (e.g. a store could present an HT3 hazard through the doors but an HT4 hazard in all other directions

•

The construction of the store

UN Hazard Divisions and security ADR has, in recent editions, extended its scope to embrace transport security as well as transport safety. The rationale for combining the two is sound, but the implementation is not. As noted above, the UN Hazard Division of an explosive substance or particularly an explosive article is intimately related to its packaging. The possible security threat of a particular article is not and cannot be related to its packaging – after all it is most likely that the article would be removed from its outer packaging before use! There are many real examples of where an explosive article has had specific packaging designed to reduce the hazard in transport – for instance 1.2G pyrotechnic articles can be packaged to present a 1.3G or 1.4G hazard. The hypothetical example given above could even make items classified as 1.1G in one configuration present a 1.4S hazard in another. This is a sensible approach for the manufacturer/consignor to take – it reduces the hazard in transport and potentially reduces costs and allows more economical shipments. However, the article remains the same – take the article out of the box and it must necessarily present the same security risk. To use the crude criteria in ADR section 1.10, based almost Journal of Pyrotechnics, Issue 33, 2014

solely on the Hazard Division, misses the point that the safety hazard posed by packaged explosives is intimately linked to its packaging! The only simple way for ADR to proceed to address the security provisions for explosives would be to make decisions about the “High Consequence Dangerous Goods” status by reference to individual UN numbers, and present this information in the existing Table A – The Dangerous Goods list. In this respect it is vital that regulators understand that the hazards posed by packaged explosives are dependent on the packaging. This will be explored further in the section dealing with the provisions of GHS.

the complexity of testing. It was also in part a recognition that it was better to have a scheme to allow users to obtain a meaningful classification easily, than for products to be wrongly classified or for producers and consignors to classify for convenience! Any default scheme must necessarily be rather pessimistic, and the option is always available (and indeed encouraged) to applicants to demonstrate (by test or by analogy) a classification related to a lower hazard, perhaps by specifically designed packagings. However there are problems with the UN default regime: •

There are anomalies within it – for instance between Roman candles and shot tubes (which are in essence single shot Roman candles) or between fountains and waterfalls. The CEN Standard EN 16261 recognises these differences and so should the UN

•

There are default classifications missing. It would be sensible to remove these omissions even if the items are not in common circulation – they may be in the future. Again the definitions in EN 16261 could assist here

•

It should be extended to identical items which are not fireworks. For instance, following the adoption of the EU Directive on pyrotechnic articles many items that were previously regarded as fireworks (and were classified as such) are now properly termed “Articles Pyrotechnic for technical purposes” (UN 0428-0431). The default classification scheme could usefully be extended to such articles and a formal proposal is being submitted to the UN for consideration

•

It may well be appropriate to introduce a new UN series for “Articles Pyrotechnic – for special effects use” to go alongside “Articles Pyrotechnic for technical purposes” (UN 0428-0431) and relate these to European Standard (CEN) T1 and T2 or similar articles and hence restrict UN 0428-0431 to CEN P1 and P2 articles or similar

•

It calls into question whether an item in a specific packaging should in theory have more than one classification at all – but this can only happen when the erroneous link between classification for transport and other use of UN Hazard Divisions in non-transport regulations is broken.

Effects of bulk The hazard identified by application of the UN tests may be realistic for the quantities involved but may not represent the hazard in bulk storage (for which classification is not appropriate and should not be used anyway) or bulk transport (for instance within an ISO container). Following well publicised incidents at Enschede3 and Kolding4 much work has been done to investigate the true “hazard classification” of bulked up fireworks. Many authorities within Europe collaborated to address the “Quantification and control of hazards associated with the transport and bulk storage of fireworks” – the CHAF trials5 – and although most of the trials gave results analogous to the classifications awarded from UN test data and the derived default classification table, there were anomalous results where a higher explosive event was observed (although explanations are somewhat contentious). The failure is not with the classification of fireworks per se, it rests with the inappropriate extension of trials data on small samples to the bulking up in ISO containers for transport. We believe however that this is not a unique feature of fireworks – most explosives (especially those containing sensitive explosive substances or significant proportions of flash powder) would react in the same way. However to date no large scale trials have been carried out on items other than fireworks. The CHAF project and subsequent small scale investigations take a fundamentally different approach to that of the existing UN regime – in time (but not yet) it may be appropriate to re-examine the whole basis of the UN regime to adequately reflect the results across the whole of the explosives sector.

Furthermore similar approaches could be adopted for other explosives articles, for instance small arms ammunition. Parameters could include:

The default classification system for fireworks

•

Calibre

•

Net explosive content

The UN developed (following the adoption of a similar scheme in the UK many years before) a “default” classification regime for assignment of Hazard Divisions to fireworks.

•

Type of package

•

Number in package (or explosive loading density of packages)

In part at least, the need for such a scheme was driven by the diversity of products available, their individual value, and Journal of Pyrotechnics, Issue 33, 2014

The advantages are many – simplification, better understanding of the issues of changing packaging etc and such a scheme could be developed with no reduction in Page 13

safety.

Time/pressure testing for determination of flash powder Following some extensive arguments, the UN adopted an approach of assigning pyrotechnic composition to some fireworks classifications (essentially UN 0333 to 0336) on the basis of amount of “flash composition” contained within a particular article and the results of the “time/pressure” test carried out on such compositions. This decision followed the seeming assumption by members of the UN committee charged with investigating this area that the existing definition (based on composition) was inadequate and had led manufacturers to change their compositions to avoid being deemed “flash”. Unfortunately there were severe problems in this approach, highlighted but ignored at the time, and which still remain unresolved. Firstly the initial results5 showed extreme variations in the time criterion applied. This highlighted potential practical issues with the test and which indicated the need for further work to be carried out before the tests were adopted into the UN regime. Secondly the body which carried out such tests recommended a value of 4 ms as the cut off point for assigning a composition as flash powder – this took into account the desire to exclude blackpowder (UN0066) from the definition of flash powder – and as a consequence of the deviations highlighted above. Unfortunately the UN decided (without further tests) to adopt a figure of 8 ms which meant that many non-flash compositions potentially required testing. Thirdly, the test required compositions to be tested “in the form” in which they were found in the article – but necessarily removing them from the article means that they may no longer behave in a similar manner to how they do in the article itself. Although the time/pressure tests were an attempt to address a perceived problem of the potential of flash powder in fireworks to mean a packaged article produced an HD 1.1 hazard, there are several fundamental problems with this approach: •

Testing compositions in isolation is not a determinant of the behaviour of those compositions within articles. For example, the flash powder used as the bursting charge of a firework shell does not spontaneously ignite to detonation when the shell is initially functioned (i.e. is fired from a mortar) – it is protected by the casing and physical arrangement of the shell itself

•

The tests takes no account of any special packaging features designed to prevent any mass explosion taking place

•

The approach is limited to fireworks, but if it is a valid concern that items containing flash powder could

Page 14

present an HD 1.1 hazard then the principles should apply to any article containing flash powder. The time/pressure saga was a particularly low point in industry’s relationship with the UN and, some 10 years on from the adoption of the principles, industry’s view that it was hurried and unscientific have been vindicated.

Mutual recognition of classifications ADR,7 the European agreement concerning International Carriage of Dangerous Goods by road (and the equivalents for transport by other modes), which are derived from the UN Recommendations and Model Regulations, require explosives to be classified. They do not require that explosives are reclassified by different Competent Authorities depending on transport arrangements and eventual use. In the UK, after long discussions involving the author and others with the Competent Authority (DfT and HSE), this position was accepted and now classifications of explosives that have been carried out by a contracting party to ADR are accepted without further bureaucratic involvement of the UK Competent Authority. Consignors are still required to have evidence of proper classification, but they are not required to carry out their own tests or to formally demonstrate classification “by analogy”. Explosives that have been classified by non-ADR contracting parties do not, understandably, benefit from this position, although it is hoped in time that the scheme will be extended to all contracting parties to the UN Recommendations (ADR technically is a European agreement only). Freedom of Information requests of the UK Government asking how many classifications had been changed in the UK from those awarded by other ADR signatories was the final decider – in the period sought not a single classification had been changed from that awarded elsewhere, although UK industry was charged for generating new UK classifications and the process took, in many cases, several months to complete. We would encourage the UN (and ADR) to formally adopt mutual classification recognition for all dangerous goods, but specifically for explosives where at present classifications are already subject to a Competent Authority approval. The logic of this approach is unchallengeable – the purpose of the UN Recommendations (and ADR) exist to ensure a commonality of approach and to make domestic and international journeys safer as a result. There is nothing to be gained by duplication of process. Again, where there are obvious or systemic failings then enforcement action should be taken to eliminate such issues.

GHS The adoption of the Globally Harmonised System for classification of dangerous goods (GHS) and its convergence with the UN recommendations has inherit merit, but it fails to appreciate the role of packaging in explosive classification. Journal of Pyrotechnics, Issue 33, 2014

In contrast, for instance, the corrosive nature of a particular strength solution of sodium hypochlorite is the same regardless of whether the container is glass, plastic or even metal. As outlined above the classification of an explosive (and particularly an explosive article) is dependent on the nature and construction of the article, and the nature of the packaging. Explosive articles cannot and should not be classified outside their packaging. There are a limited number of exemptions within the UN system itself (for items over 400 kg, and nationally8 where this is permitted, but in essence the packaging is integral to the classification.

from industry. I hope that the paper will stimulate discussion around the explosives community and help inform debate and revisions to the UN system, and hence lead to a greater understanding of the unique properties of explosives and hence ultimately to safer transport and to safer storage.

References 1

UN Recommendations on the Transport of Dangerous Goods – Model Regulations and UN Manual of Tests and Criteria, accessed via http://www.unece.org/trans/ danger/danger.html

There are now three Seveso directives. For more information see http://ec.europa.eu/environment/ seveso/index.htm

See the final report on the Enschede disaster for more information – http://www.nbdc.nl/cms/servlet/nl.gx. nibra.client.http.GetFile?id=498631&file=Final_ consideration_(Slotbeschouwing_Engels).pdf

For more information see the Kolding website – http://www.kolding.dk/brand/0032135.asp (in Danish)

The extension of UN classification of explosives as packaged for transport to non-transport situations (and especially to GHS labelling of articles) highlights the problems that have been outlined above and demonstrates a lack of understanding of the principles of explosive classification. We can only hope that efforts to divorce GHS from UN classification of explosives is successful.

An Introduction to the European CHAF Project, D. Chapman – see http://www.jpyro.com/wp/?p=116

Defining Flash Compositions: Modifications to UN Time/Pressure Test, D. Chapman and K. Howard . See http://www.jpyro.com/wp/?p=1176

European Agreement concerning the International Carriage of Dangerous Goods by Road (ADR). See http://www.unece.org/trans/danger/publi/adr/ adr2013/13contentse.html

Conclusions

For instance see UK Authorisation 151 – https:// www.gov.uk/government/uploads/system/uploads/ attachment_data/file/3283/authorisation-151.pdf

For instance, some have suggested that any article containing blackpowder should be labelled under GHS as HD 1.1D – the classification of “raw” blackpowder itself, irrespective of the fact the article does not behave like “raw” blackpowder. There are many examples of, for instance, fireworks and flares where the correct classification of the article as packaged for transport is HD 1.4G and that even articles outside of their transport packaging present only a very localised effect in case of incident or in proper functioning. It would be meaningless to label such items as HD 1.1D because they happen to contain a small quantity of blackpowder.

The United Nations Recommendations and Model Regulations regarding the classification of explosives have undoubtedly improved the safety of transport of such items through a consistent approach and education of all parties involved in obtaining proper classifications for their shipments which truly reflect the hazards as presented for transport. However it has become apparent that there are problems associated with the process itself, and the fundamental principles which lie behind it. Most of the issues relate to the attempt to extend the principles of classification into areas that it was never designed for – and we would urge those developing future editions of the UN Recommendations and ADR to revert to the sole and proper purpose: …to determine the hazard as presented for transport…. This opinion paper addresses some of these issues and is intended to stimulate future discussions.

Acknowledgements This paper represents the views of the author alone; however I am grateful for the support of many people who have commented upon and expressed support for the challenges raised, be they from enforcing authorities, the military or Journal of Pyrotechnics, Issue 33, 2014

Page 15

Quantification of Visible Aerosols from Pyrotechnics: Metal and Metal Compound Additives Rene Yo Abe,a Yoshiaki Akutsu,a Akihiro Shimadab and Takehiro Matsunagab

Department of Environment Systems, Graduate School of Frontier Sciences, The University of Tokyo, 5-1-5 Kashiwanoha Kashiwa, Chiba 277-8563, Japan. Fax: (+) 81-4-7136-4729, email: abe@geel.k.u-tokyo.ac.jp b Research Institute of Science for Safety and Sustainability, National Institute of Advanced Industrial Science and Technology, 1-1-1 Higashi, Tsukuba 305-8565, Japan a

Abstract: The effect of metal and metal compounds commonly used in pyrotechnics on visible aerosol development at high relative humidity has been investigated in combustion experiments using a combustion chamber. Ammonium perchlorate/ hydroxyl-terminated polybutadiene as oxidant/fuel system allowed aerosols generated from the additives to be observed in the absence of particles generated from the base composite. For magnesium and magnalium and all flame coloring agents except barium nitrate, light extinction measurements at 80% relative humidity were found to be proportional to the mass concentration of hygroscopic metal compound particles which are formed at high temperatures from metal chloride or metal vapors during combustion. Low visible aerosol development under humid conditions was observed for aluminium and titanium which have higher boiling points than magnesium and do not readily vaporize during combustion, as well as for barium nitrate which forms too small hygroscopic barium chloride particles and iron(iii) oxide which, because of its low boiling point, forms coarser iron(iii) chloride particles at lower temperatures.

1. Introduction The use of metal compounds and metal powders is essential for flame coloration and the generation of spark effects in pyrotechnics. Together with combustion products from the pyrotechnic composition base components their reaction products are found in emissions of very fine aerosol (smoke) particles in the range of a few hundred nm.1–3 The particles consist of solid metal chlorides or oxides which can act as condensation nuclei for HCl and H2O vapor4–6 emitted by the pyrotechnic composition. Though such particles scatter light only weakly at their initial size, they produce visually opaque aerosols7,8 when they grow in size, reaching a maximum in light scattering efficiency at around 1 µm for visible light (~500 nm), if number concentrations are conserved.9,10 Metal powders of various particle size and type find use in gold to silvery-white light, flash, sparkle and tail effects. These light effects are based on continuous spectrum emissions of high temperature particles.11 Practical safety and combustion performance limit the common metal types to aluminium (Al), magnesium (Mg), magnalium (MgAl, i.e., ~50% alloy of Mg and Al) and titanium (Ti). For colored flames of green, blue, yellow and red, respectively, as well as their mixtures, only compounds of barium, copper, sodium and strontium are commonly used in pyrotechnic compositions. These metals generally need to form volatile

chlorides during combustion to exhibit colored emission lines.11,12 Because chlorides of these metals are hygroscopic, they cannot be directly used as additives in pyrotechnic compositions. Therefore the necessary metals are added as compounds which are decomposed and converted to metal chlorides during combustion. The necessary amount of chloride is generally provided by a dedicated chlorine donor compound or hydrochloric acid produced by the oxidant such as ammonium perchlorate (AP). These metal chloride vapors produce fine aerosol particles at high number concentrations when the exhaust gas is rapidly diluted and cooled to ambient temperatures. AP, despite its good performance in producing colored flames, is less common in firework compositions. It also corresponds to oxidant/fuel (binder) systems in pyrotechnics due to the recent desire to reduce solid residue fallout. It finds heavy use in propellants and produces virtually no solid ashes or aerosols under dry conditions in contrast to oxidants based on the more common potassium salts.11 Though its relatively high excess of HCl emissions is known to produce visible aerosols under high humidity, combustion chamber experiments have shown in previous work that the effect is limited to relative humidity (RH) higher than 85%.13 The effect of additive metals and metal compounds on visible aerosol generation, particularly at high relative humidity conditions, was studied in this work by analyzing aerosol development trends in that combustion chamber.

Article Details

Article No:- 0106

Manuscript Received:- 07/07/2014

Final Revisions:- 24/10/2014

Publication Date:-01/11/2014

Archive Reference:- 1694

Page 16

Journal of Pyrotechnics, Issue 33, 2014

Table 1. Formulations of composites based on ammonium perchlorate/polybutadiene Main formulation

Flame coloring and sparkling effect additives

HTPB

Curing agent

14.5%

81.9%

3.6%

Compound Ba(NO3)2 CuO Na2C2O4 SrCO3 Fe2O3 Al Mg Mg–Al

Effect Green Blue Yellow Red Catalyst Sparks, silver flames Flame modifier Sparks, white flames

Sparks

2. Experimental 2.1

Pyrotechnic compositions

As oxidant/fuel system with only negligible visible aerosol over a wide relative humidity (RH) range up to 85% allowing quick laboratory scale production, ammonium perchlorate (AP)/hydroxyl-terminated polybutadiene (HTPB) was chosen. Potassium perchlorate is much more common as an oxidant in pyrotechnics but is known for producing large amounts of solid particles including large fallout particles and fine visible aerosols. Another alternative is potassium nitrate which, as a chlorine-free system, has little visible aerosol output but generates less intense light and large amounts of large particle residue/fallout. The AP/HTPB

Figure 1. Combustion chamber features (front view). (a) Laser emitter–transmittance sensor assemblies; (b) combustion platform; (c) sample suction port; (d) circulation fan and psychrometer; (e) ducts to humidity control unit.

Journal of Pyrotechnics, Issue 33, 2014

Additive ratio (%) 5–20 0.5–20 1–20 1–10 2–10 10–30 5–20 5–30 5–30

system, on the other hand, allows aerosol generation induced and amplified by additive compounds to be observed with little interference from fallout losses or particles generated by the oxidant/fuel system of the base composite. HTPB was treated with a curing agent (polymeric methylene diphenyl diisocyanate) mixed with AP and varying amounts of additives as shown in Table 1. The composites were then rolled out as sheets that could easily be cut to appropriate sample sizes after curing overnight. 2.2

Combustion chamber

A chamber already described in previous work13 consisting of a cubic steel frame covered with clear acrylic sheets at the top and three sides, a stainless steel floor and a detachable soft PVC sheet at the front side was constructed as shown in Figure 1. Light extinction measurements were provided by laser sensors (a) (Keyence LX-100, λ = 670 nm) installed in the chamber at distances of 77.5 and 10 cm and a flow of aerosol was transferred to a modified particle size distribution (PSD) analyzer (HORIBA LA-920) through a pipe (12 mm i.d.) of stainless steel serving as sample suction port (c). After initial temperature and humidity conditions were applied with a humidity controllable type air-conditioning unit (Apiste PAU300S-HC) connected through ducts (e), the chamber was sealed. Room temperature was also controlled by air conditioning to match the temperature setting inside the chamber. Aerosol opacity which was previously found to be virtually independent of temperature was set at 20– 30 °C. Samples were placed on a combustion platform (b) and ignited electrically using a nichrome wire embedded in a ceramic tube and air in the chamber was mixed using an electric fan (d) so that a sufficiently homogeneous aerosol was yielded within about ten seconds after combustion. A psychrometer composed of two thermocouples was used as robust humidity sensor. 13 2.3

Theoretical considerations

2.3.1

Chemical equilibrium analysis

Combustion of AP/HTPB composite pyrotechnics can be interpreted as a two-step process. A premixed flame of Page 17

ammonia with perchloric acid closely coupled with the sublimation and decomposition of AP and HTPB binder provides excess oxidizer to form diffusion flames with fuel fragments released from the binder.14 Though the kinetics of this process are complex near the surface, especially if additives have to be taken into account, combustion products and flame temperatures can roughly be estimated using chemical equilibrium calculations with the NASA Chemical Equilibrium with Applications (CEA) program package15 which was extended with thermodynamic data available online.16 Equilibrium states were analyzed for a dilution series with air for the slightly oxidant-deficient composite. The results were visualized as mass fractions excluding the mass of air additions to track conversion of all species related to the corresponding metals. 2.3.2

Nucleation, coagulation and growth kinetics

Homogeneous and heterogeneous nucleation mechanisms both play important roles in aerosol generation from pyrotechnics. Previous investigations on aerosol size distributions in the range of some hundred nm1–3 strongly support primary particle generation by homogeneous nucleation. Particle nucleation and growth kinetics have however not been extensively studied theoretically for pyrotechnics. Studies on such processes include both experimental measurements and numerical simulations in the vicinity of burning coal particles17 or in turbulent jets18,19 of hot gas. A program for solving particle nucleation and growth problems (Nodal General Dynamic Equation solver; NGDE)20 was extended for calculations of PSDs resulting from condensation of gaseous metal chlorides when they are cooled down by dilution with ambient air. Vaporized metal chloride in hot exhaust from the composite corresponding to a volume of around 2.5 L at standard conditions is diluted to a volume of 1 m3 following an exponential dilution function to imitate the dilution process occurring in the combustion chamber. Particle nucleation, condensation and coagulation are simulated according to the physical properties vapor pressure, surface tension and molecular mass of each of the metal chlorides in liquid state. Trends of mean particle sizes and number concentrations for different concentrations and dilution speeds were analyzed.

Results and discussion 3.1

Analysis of humidity characteristics

Fine aerosol particles are generated at high number concentrations from metal chloride or metal vapors during combustion and act as condensation nuclei for H2O and HCl vapors when cooled to ambient temperatures. At high relative humidity conditions, they grow to particle sizes up to diameters of around 1 µm within minutes as can be seen in Figure 2a for an aerosol generated from 2 g of an AP/ HTPB composite with 10% of CuO additive at 89% RH. Light scattering and absorption processes on such spherical particles of sizes near the wavelength of light are described in detail by the Mie theory21 where the corresponding solutions are obtained from Maxwell’s equations. Light extinction has been shown previously13 to be proportional to samplemass and can be normalized as a sample-mass specific scattering or extinction coefficient b*. In the initial growth phase (Phase I) driven by condensation of H2O and HCl vapors, particles gain in light scattering efficiency until light extinction reaches its maximum at the point indicated by the dashed line. Although maximum light scattering efficiency for a single aqueous particle should lie at a diameter of 1.3 µm (1.0 µm for λ = 500 nm10) according to Mie theory, mode diameters of PSDs have only been observed at around 0.6 µm at the maximum of light extinction in 670 nm laser measurements (Figure 2b). Thereafter growth by coagulation and, if particles grow too large, deposition processes on the chamber surfaces take over and light extinction starts to decrease due to loss in particle number-concentration (Phase II). It was not possible to obtain quantitative data as number concentrations from PSD measurements in the current setup, but the continuing growth in size suggests coagulation as the growth mechanism during Phase II. Compared to the rising slope b* displays during Phase I, Phase II can be considered as having only a small influence on the maximum value of b*. At very dry conditions the extinction coefficient even reaches a constant value almost instantly and does not decrease notably within the observed time span. The maximum value of sample-mass specific light extinction b*max can be extracted for each combustion experiment as a measure of the aerosol’s obscuring potential.

Figure 2. a. Change of sample-mass specific extinction coefficient b* and mode diameter of aerosol from AP/HTPB composites with 10% CuO additive over time; b. PSD measured at maximum light extinction, 4.7 minutes after ignition. Page 18

Journal of Pyrotechnics, Issue 33, 2014

3.2

Metal powders

Sparkling effects in pyrotechnics are produced by burning metal particles which burn at lower combustion rates than the composite itself. The metal particles are ignited but not evaporated in the pyrotechnic flame and continue burning after being ejected into ambient air. The particles’ incandescent light emissions depend on particle temperature and combustion times which are influenced by metal type and particle properties such as size and shape. If particles burn long enough, tail or willow-like effects can be produced. For some metals, particles can also burst during combustion (e.g., Ti) and add a crackling effect. Most of the material, however, remains in the condensed phase during the whole process and very low levels of smoke have been observed for composites containing Al and Ti powders compared to samples containing Mg, as a plot of b*max at 80% RH against amount of added metal shows in Figure 4.

Figure 3. Humidity dependence of maximum sample-mass specific extinction coefficient of KP/HTPB base composite and AP/HTPB composites with 5% of SrCO3, Ba(NO3)2 and without additives. A series of b*max measurements can efficiently describe the humidity characteristics of a test composite. The base composition of AP/HTPB composites used in this work produces only negligible amounts of visible aerosol up to about 80% RH. In Figure 3, b*max is plotted against initial relative humidity. Usage of potassium perchlorate (KP) would result in a high baseline throughout the whole humidity range with a sharp increase above 85% RH. Metal compound additives like SrCO3 induce visible aerosol development from AP/HTPB composites even below 80% RH. b*max increases moderately with rising relative humidity until it shows a steep rise above 80% RH. Hygroscopic aerosol particles already form aqueous solutions by absorbing H2O vapors at lower RH and efficiently scavenge excess HCl vapors emitted by the composite. The growth of such particles can be considered as being controlled mainly by the HCl/H2O system and causes the steep rise in light scattering efficiency above 80% RH. Visible aerosol increase can therefore be best extracted at 80% RH where the effect of metal compound additives is high while aerosol which may originate from the base composition itself is negligible. Also, losses due to coagulation and deposition processes, which would increase with growing particle sizes at higher RH, can be kept within acceptable limits. Results for metal powder additives and flame coloring additives are analyzed separately, because of differences in their combustion behavior and combustion products involved. Further Fe2O3, which is often used as catalyst in pyrotechnics to increase the combustion speed, is discussed together with the coloring agents. Journal of Pyrotechnics, Issue 33, 2014

Mg represents an exception to the combustion mechanism observed for other metals. Due to its low boiling point of 1363 K, compared to the other metals (Al: 2743 K; Ti: 3560 K), it evaporates within the flame, unless the metal particles are very large, and can contribute to higher flame temperatures there. Thus AP/HTPB composites containing only Mg powder additive do not emit incandescent light at all, despite the very bright flame burning Mg produces as bulk material. If solid or liquid particles are present in the flame, however, a bright high temperature incandescent flame can be produced (e.g., white stars with Ba(NO3)2 or Mg flash-compositions). As a result Mg produces fine aerosol particles of MgO at high number concentrations by homogeneous nucleation from the gas phase. These hygroscopic particles absorb HCl and H2O vapor to form aqueous solutions of MgCl2 and HCl which grow in size, particularly under high RH. The slope for AP/HTPB with Mg flattens at high ratios of Mg in Figure 4, due to reduced emissions of free HCl from a lower proportion of AP/HTPB

Figure 4. Maximum sample-mass specific extinction coefficient of AP/HTPB composites with Mg, Mg-Al, Al and Ti metal powder aditives at 80% RH. Boiling points22 denoted as “b.p.”. Page 19

base composite (in a 2 g sample) and accelerated coagulation of particles at high aerosol concentrations. Mg-Al behaves very similar to Mg in respect to visible aerosol formation, but produces bright white sparks. Clearly Al is responsible for the metal particles not evaporating completely during combustion in the composite flame due to its higher boiling point at 2743 K. Some of the Al seems to be evaporated, though, as light extinction measurements suggest. Values of b*max are significantly higher than would account for half the amount of Mg as would be expected from the proportion of Mg in the alloy. Higher temperatures of metal-particles resulting from better combustion performance of Mg-Al compared to Al also support this. Elemental analysis of aerosol particles should provide insight into the mechanism of this phenomenon. Particle emissions of aluminium oxides have been reported for combustion plumes of solid rocket motors to have maxima in their PSDs at 0.1 and 2 µm. 23 These particles act as condensation nuclei for H2O vapor and HCl emissions, which can be observed as thick white rocket trails. Combustion in these propellants, however, follows a different temperature and pressure profile during combustion and uses atomized aluminium, which has combustion characteristics different from the aluminium flakes common in firework pyrotechnics. Aluminium is therefore much more readily vaporized in those propellants than Al and Ti used as spark effect metals in pyrotechnics. In the case of magnalium, Al vapors may be present more abundantly during combustion due to Mg supporting the combustion and evaporation process of the metal particles. 3.3

Flame coloring agents

Flame coloring compounds form volatile metal chlorides

Figure 5. Maximum sample-mass specific extinction coefficient of AP/HTPB composites with CuO, SrCO3, Na2C2O4, Ba(NO3)2 calculated as chlorides formed during combustion. Deliquescent relative humidity25 annotated as “DRH”. *DRH of CuCl2.26 Page 20

(BaCl2, CuCl, NaCl, SrCl2) during combustion, which produce fine aerosol particles by homogeneous nucleation during rapid cooling of the exhaust gas. Figure 5 shows linear plots of b*max against metal chloride mass formed per gram of composite during combustion, which flatten out at high additive ratios because combustion performance cannot be sustained by the base composite. Furthermore, all metal chlorides except BaCl2 show the same slope, which implies that aerosol nucleation and growth processes at high temperatures (boiling points are 1833, 1763, 1686 and 1523 K, respectively24) immediately after combustion result in particle number-concentrations proportional to mass concentrations, even for metal chlorides. Mg from the previous section would also fit the common slope of the colorants if it is considered an analogous nucleation process of MgO from Mg vapor. This suggests that particles produced by rapid condensation at high temperatures have the same average masses rather than diameters or volumes for all compounds. When cooled to ambient temperatures at high RH, these hygroscopic particles grow by uptake of HCl and H2O and become highly efficient in light scattering. BaCl2, which showed low visible aerosol response, has the highest deliquescence relative humidity value (91% RH at 20 °C)27 among the tested metal compounds (68%, 73%, 75% RH for CuCl2, SrCl2 and NaCl respectively – CuCl is oxidized and converted to CuCl2 by O2 and HCl over 50% RH28). HCl vapor therefore does not condense on BaCl2 particles. Instead, HCl and H2O vapors are absorbed by heterogeneous nucleation on other hygroscopic particles present in ambient air. This results in the formation of large particles at low number concentration analogous to the base composite without additives. This is observed over the whole humidity range and is also seen in Figure 3, where the values of b*max of AP/HTPB with 5% of Ba(NO3)2 additive only slightly surpass those of the base composite. For Ba(NO3)2 additives, PSDs could not be measured in the current setup, because overall scattering efficiencies were too low. Weak forward-scattered light of the transmission lasers has however been observed with the naked eye through the transparent chamber walls at relatively large scattering angles up to around 45°, suggesting particles sized in the range of a few hundred nm. Fe2O3, which catalyzes decomposition of the HTPB binder during combustion, is converted to gaseous FeCl3, and has a much lower boiling point of 553–589 K because of its molecular structure and ability to form dimers. It will therefore stay gaseous longer than chlorides from coloring agents. Although FeCl3 is very hygroscopic, its presence does not increase light extinction noticeably. In Figure 5, its plot rises only marginally to 0.07 at its maximum addition ratio of 10%. It can be therefore said that FeCl3 will only form aerosol particles at much lower concentrations than compounds condensing at high temperatures if at all. A plot of b*max over humidity is practically indistinguishable from plots of aerosol produced by the base composite. Under very high RH of over 90%, aerosol particles grow large enough Journal of Pyrotechnics, Issue 33, 2014

for a faint brownish coloration from dissolved FeCl3 to be observed. When excessive amounts of additive are used, combustion performance (combustion speed) of the composite is degraded and aerosol opacity does not further increase linearly, because reduction of particle number-concentrations by coagulation effects is accelerated and energy output of the base composite is not sufficient to support complete reaction and vaporization of the metal compounds or even stable combustion. For Na2C2O4 the negative effect of its endothermic decomposition reaction appears most distinctly at additions as low as 10% (0.08 as NaCl in Figure 5), while for the same amounts of SrCO3 (0.09 as SrCl2) and CuO (0.11 as CuCl) b*max at 80% RH deviates only slightly from the linear trend. Also, color quality can be reduced by continuous spectrum emissions (condensed phase) or emission spectra (vapors) of interfering by-products, if too much additive is used. Additive ratios for a pyrotechnic composite therefore have to be optimized for each additive. 3.4 Chemical equilibrium and particle nucleation calculations All metal compounds used as coloring agents are transformed to their metal chlorides according to chemical equilibrium calculations and exist only in the gas phase at adiabatic combustion temperatures. Figure 6 shows the main species containing the respective metal in the case of SrCO3 as additive. Dilution with ambient air first causes further increase in temperature, because combustion of the oxygen deficient composite can progress. Further dilution, however, cools the mixture and at around 1100 °C after 5-fold dilution the first condensed-phase species of liquid SrCl2(L) appears, which subsequently transforms to solid phases of SrCl2(b) and SrCl2(a) representing high-temperature superionic and low-temperature phases29 of a cubic fluorite crystalstructure, respectively. Analogous formation of condensed phase metal chlorides from their vapors has been obtained as a result for all coloring agents. Only CuO shows the peculiarity of forming Cu metal vapor (which in experiments

can be condensed as metal on cold surfaces held into the flame) before condensation as CuCl. Rapid cooling of those gaseous species by turbulent mixing lead to supersaturated conditions at high temperatures of over 1000 °C under which particles are formed by a homogeneous nucleation process. Except Mg which at least partly forms metal Mg and MgCl2 vapors before condensing as MgO, metals used for spark effects only form their oxides in liquid state. Mg follows the same scheme as the coloring agents with the difference of MgO forming from Mg or MgCl2 vapor. When plotted against MgO mass the slope of b*max also concords with the metal chlorides. Other metals’ actual combustion processes may not be sufficiently reproduced with chemical equilibrium calculations, but results showing only little aerosol formation support a combustion mechanism with minimal nucleation from the gas phase. Particle nucleation, growth by condensation of monomers and coagulation processes of vaporized metal chlorides could be described by solving the General Dynamic Equation with the available software.20 The model results in smaller particles and higher number concentrations at higher dilution (and consequently cooling) rates. Dilution rates were chosen so that exponential dilution of exhaust gas from the combustion to the chamber volume of 1 m3 would be complete in 0.0008 to 8 seconds. For all calculations an increase in metal chloride concentration formed from the additive compound would form significantly larger particles as Figure 7 shows for the case of SrCO3. Number concentrations show therefore a decreasing trend with mass concentration of introduced metal chloride. This contradicts the trends observed for light extinction properties of the aerosol, which would suggest equally sized particles regardless of concentration. It is an indication of too many assumptions having been made in this application of the simple model beginning at the dilution rates and also including the particle growth and coagulation model itself, which assumes liquid particles, although temperatures approach or even fall below the melting points during the particle growth processes.

Figure 6. Equilibrium calculation results for AP/HTPB with 5% SrCO3 additive and dilution with air. Mass fractions exclude mass added by dilution air. Journal of Pyrotechnics, Issue 33, 2014

Page 21

Figure 7. Simulation results modeling particle nucleation, growth and coagulation during dilution of vaporized SrCl2 with ambient air within about 0.08 seconds.

Conclusion Promotion of visible aerosol development by common additive metals (Mg, Al, magnalium, Ti) and flame coloring metal compounds (Na2C2O4, CuO, SrCO3, Ba(NO3)2) after combustion of pyrotechnic compositions based on the low-smoke oxidant/fuel system of ammonium perchlorate/ hydroxyl-terminated polybutadiene has been confirmed at high relative humidity above around 80%. With the exception of Al and Ti metal powders for sparkling effects and Ba(NO3)2 for green flames, which produce only minimal visible aerosols, light extinction has been found to increase proportionally with the mass concentration of the corresponding metal chloride (CuCl, NaCl, SrCl2) or metal oxide (MgO) emissions. Fine aerosol particles are produced through homogeneous nucleation during the combustion process and act as condensation nuclei for HCl produced during combustion and H2O vapors present in ambient air after dilution. Growth of these particles to highly opaque aerosols occurs distinctly at relative humidity above 80%. For low smoke pyrotechnics producing only little visible smoke even under high relative humidity, Al and Ti are found to be effective spark-generating metals, while from the available coloring agents only Ba (green) is found to be unaffected by humidity. Under RH above 80%, use of Mg and magnalium should be avoided and replaced by Al and Ti as much as possible. Use of flame-coloring compounds of Cu (blue), Na (yellow) and Sr (red) produce thickening smoke at high humidity and needs to be minimized contrary to Ba (green) which does not increase smoke production.

References 1

B. Wehner, A. Wiedensohler and J. Heintzenberg, “Submicrometer aerosol size distributions and mass concentration of the millennium fireworks 2000 in Leipzig, Germany,” Journal of Aerosol Science, vol. 31, no. 12, pp. 1489–1493, 2000.

A. Dutschke, C. Lohrer, L. Kurth, S. Seeger,

Page 22

M. Barthel and U. Panne, “Aerosol Emissions from Outdoor Firework Displays,” Chemical Engineering & Technology, vol. 34, no. 12, pp. 2044–2050, 2011. 3

A. Dutschke, C. Lohrer, S. Seeger and L. Kurth, “Gasförmige und feste Reaktionsprodukte beim Abbrand von Indoor-Feuerwerk,” Chemie Ingenieur Technik, vol. 81, no. 1–2, pp. 167–176, 2009.

A. T. Cocks and R. P. Fernando, “The rates of formation of hydrogen chloride/water aerosols by homogeneous nucleation,” Atmospheric Environment, vol. 15, no. 7, pp. 1293–1299, 1981.

N. Kubota, “Propellant Chemistry,” in Pyrotechnic Chemistry, Whitewater, Journal of Pyrotechnics Inc., 2004, pp. 7–11.

A. Schenkel and K. Schaber, “Growth of salt and acid aerosol particles in humid air,” Journal of Aerosol Science , vol. 26, no. 7, pp. 1029–1039, 1995.

J. A. Conkling, “Smoke and Sound,” in Chemistry of Pyrotechnics, New York, Marcel Dekker Inc., 1985, pp. 167–179.

J. T. Hanley and E. J. Mack, “A laboratory investigation of aerosol and extinction characteristics for Salty Dog, NWC 29 and NWC 78 pyrotechnics,” Calspan Report No 6665-M-1, Department of the Navy Naval Air Systems Command, Washington DC, 1980. http://oai.dtic.mil/oai/ oai?verb=getRecord&metadatPrefix=html&identifier =ADA093098

C. Bohren and D. R. Huffman, Absorption and scattering of light by small particles, Weinheim, Wiley-VCH Verlag GmbH & Co. KGaA, 2004.

10 M. Z. Jacobson, “Absorption and scattering by gases and particles,” in Fundamentals of Atmospheric Modelling, Second Edition, New York, Cambridge University Press, 2005, pp. 301–312. 11 T. Shimizu, Fireworks: The Art Science and Technique, third edn, Austin, Pyrotechnica Journal of Pyrotechnics, Issue 33, 2014

Publications, 1981. 12 K. L. Kosanke and B. J. Kosanke, “The Chemistry of Colored Flame,” in Pyrotechnic Chemistry, Whitewater, Journal of Pyrotechnics Inc., 2004, pp. 25–49. 13 R. Y. Abe, Y. Akutsu, A. Shimada and T. Matsunaga, “Quantification of Visible Aerosols from Pyrotechnics: The Effect of Relative Humidity,” Journal of Pyrotechnics, in press, 2014, preceding paper. 14 N. Kubota, “Combustion of Composite Propellants,” in Propellants and Explosives, second edn, Weinheim, Wiley-VCH, 2007, pp. 181–233. 15 M. J. Zehe, “Chemical Equilibrium with Applications,” NASA, 2010. http://www.grc.nasa. gov/WWW/CEAWeb/

26 L. B. Rockland, “Saturated Salt Solutions for Static Control of Relative Humidity between 5° and 40° C,” Analytical Chemistry, vol. 32, no. 10, pp. 1375–1376, 1960. 27 World Meteorological Organization, “Measurement of Humidity,” in Guide to Meteorological Instruments and Methods of Observation, 2008, p. I.4. 28 H. W. Richardson, “The Manufacture of Copper Compounds,” in Handbook of copper compounds and applications, New York, Marcel Dekker, Inc., 1997, pp. 53-92. 29 S. Hull, S. T. Norberg, I. Ahmed, S. G. Eriksson and C. E. Mohn, “High temperature crystal structures and superionic properties of SrCl2, SrBr2, BaCl2 and BaBr2,” Journal of Solid State Chemistry, vol. 184, pp. 2925–2935, 2011.

16 “NIST-JANAF Thermochemical Tables,” National Institute of Science and Technology, 2013. http:// kinetics.nist.gov/janaf/ 17 C. L. Senior and R. C. Flagan, “Ash Vaporization and Condensation During Combustion of a Suspended Coal Particle,” Aerosol Science and Technology, vol. 1, no. 4, pp. 371–383, 1982. 18 K. Zhou and T. L. Chan, “Simulation of Homogeneous Particle Nucleation in a Free Turbulent Jet,” Aerosol Science and Technology, vol. 45, pp. 973–987, 2011. 19 T. K. Lesniewski and S. K. Friedlander, “Particle nucleation and growth in a free turbulent jet,” Proceedings of the Royal Society of London A, vol. 454, pp. 2477–2504, 1998. 20 A. Prakash, A. P. Bapat and M. R. Zachariah, “NGDE: Software for Solution of Nucleation, Surface Growth and Coagulation Problems,” Departments of Mechanical Engineering and Chemistry, University of Minnesota, 2003. http://www.me.umn.edu/~mrz/ software.htm 21 G. Mie, “Beiträge zur Optik trüber Medien, speziell kolloidaler Metallösungen,” Annalen der Physik, ser. IV, vol. 25, no. 3, pp. 377–445, 1908. 22 Y. Zhang, “Corrected Values for Boiling Points and Enthalpies of Vaporization of Elements in Handbooks,” Journal of Chemical Engineering, vol. 56, pp. 328–337, 2011. 23 M. N. Ross, P. D. Whitefield, D. G. Hagen and A. R. Hopkins, “In Situ Measurement of the Aerosol Size Distribution in Stratospheric Solid Rocket Motor Exhaust Plumes,” Geophysical Research Letters, vol. 26, no. 7, pp. 819–822, 1999. 24 D. L. Perry, Handbook of Inorganic Compounds, Second Edition, Boca Raton, Taylor & Francis Group, 2011. 25 L. Greenspan, “Humidity Fixed Points of Binary Saturated Aqueous Solutions,” Journal of Research for the National Bureau of Standards – A. Physics and Chemistry, vol. 81A, no. 1, pp. 89–96, 1977. Journal of Pyrotechnics, Issue 33, 2014

Page 23

Green Pyrotechnic Formulations Based on Metal-Free and NitrogenRich Tetrazolylborate Salts Thomas M. Klapötke,* Magdalena Rusan and Jörg Stierstorfer

Department of Chemistry, Energetic Materials Research, University of Munich (LMU), Butenandtstr. 5-13, D-81377 Munich, Germany email: tmk@cup.uni-muenchen.de

Abstract: The investigation of green-burning boron-based compounds as colorants in pyrotechnic formulations as an alternative for environmentally and health hazardous barium nitrate is reported here. Metal-free and nitrogenrich dihydrobis(5-aminotetrazolyl)borate salts and dihydrobis(1,3,4-triazolyl)borate salts have been synthesized and characterized by NMR spectroscopy, elemental analysis, mass spectrometry and vibrational spectroscopy. The energetic and thermal properties have been determined as well. Crystal structures of compounds 5b, 7 and 13 were obtained. Pyrotechnic compositions have been prepared using selected dihydrobis(azolyl)borate salts as green colorants. In these compositions ammonium dinitramide and ammonium nitrate haven been used as oxidizers, and boron and magnesium as fuels. The burn time, dominant wavelength, spectral purity, luminous intensity and luminous efficiency as well as the thermal and energetic properties of these compositions were measured.

Introduction In the course of the elimination of hazardous materials from the wide field of pyrotechnics, the search for alternatives has become important.1 In green-light emitting pyrotechnic formulations barium nitrate is mostly used as both the flame colorant and the oxidant.2 Despite producing intense green colored flames in the presence of a chlorine source, barium nitrate is known to be toxic to humans and the environment and therefore it is desirable to replace it by non-toxic green colorants. As alternatives boron and boron compounds are taken into consideration. In particular boron compounds with nitrogen-rich moieties are considered to be interesting green colorants. The use of nitrogen-rich compounds in pyrotechnics was reported by Douda and co-workers in the 1960s.3 Later publications describe further investigations of nitrogenrich compounds like dihydrazinotetrazine in pyrotechnic formulations.4 A boron-based pyrotechnic formulation has been investigated by Sabatini using boron carbide as green colorant.2b The combination of boron carbide and potassium nitrate resulted in an intense green-burning and insensitive formulation revealing good color properties. The synthesis of nitrogen-rich and therefore energetic boron-containing compounds like sodium and potassium dihydrobis(5aminotetrazolyl)borates has been described in the literature.5 The reaction of sodium and potassium borohydrides and two equivalents of 5-aminotetrazole led to the formation of the corresponding metal dihydrobis(5-aminotetrazolyl) borates. In addition, a series of metal complexes containing dihydrobis(tetrazolyl)borate as ligand have been described in the literature.6 Higher coordinated compounds like hydrotris(tetrazolyl)borate moieties have been synthesized

using the solvent diglyme, high temperatures and a crown ether revealing unique properties in metal complexes and representing nitrogen-rich and energetic metal complexes.7 Shreeve and co-workers investigating ionic liquids for solvent applications reported the synthesis of triazole-based borates, which fit the ionic liquid criteria. The synthesis of potassium and barium dihydrobis(1,2,4-triazolyl)borates as well as metal and metal-free hydrotris(triazolyl)borates has been described.8 Also, the preparation and investigation of silver(i) poly(1,2,4-triazolyl)borate complexes containing phosphane ligands have been reported.9 In this work the synthesis and characterization of nitrogenrich and metal-free dihydrobis(5-aminotetrazolyl)borate and dihydrobis(1,2,4-triazolyl)borate salts are reported. Their thermal and energetic properties were determined as well. Several selected metal-free dihydrobis(azolyl)borate salts are employed as colorants in pyrotechnic formulations. Mixtures of those salts and ammonium dinitramide (ADN), ammonium nitrate, magnesium, boron and vinyl alcohol acetate resin binder (VAAR) were made and their color performance and thermal and energetic properties investigated.

Results and discussion Syntheses The starting materials potassium dihydrobis(5aminotetrazolyl)borate (1), (5-aminotetrazole)dihydro(5aminotetrazolyl)borane hemihydrate (2), potassium dihydrobis(1,2,4-triazolyl)borate (11) and barium dihydro bis(1,2,4-triazolyl)borate trihydrate (13) were described in the literature and synthesized according to those literature procedures.5a,8 The reaction of potassium borohydride and

Article Details

Article No:- 0104

Manuscript Received:- 25/05/2014

Final Revisions:-2/11/2014

Publication Date:-4/11/2014

Archive Reference:- 1698

Page 24

Journal of Pyrotechnics, Issue 33, 2014

NH2

KBH4 + 2 N

MeCN

NH2 H2N

N N

NH2 H2N

CH3COOH

H2O

NH2 H2N

HN N

B H

NH2 H2N

Ba(OH)2

N N 0.5 H2O

H2O

N N

3 H2O

Ba2

Scheme 1. Synthesis of potassium and barium dihydrobis(5-aminotetrazolyl)borate salts. NH2 H2N

N N

B H

3 H2O 2

NH2 H2N

N H

H2N

NH2

N H

NH2

H2N

Cat

z H2O

NH2

Cat

NH2

H2N

H2O

Ba2

(Cat)2SO4

NH2 H2N

N H

NH2 HN

N H

N N

z=0

z=2

NH4

NH3OH3

z=1

z=0

Scheme 2. Synthesis of metal-free dihydrobis(5-aminotetrazolyl)borate salts.

KBH4 + 2 N

MeCN

B H

CH3COOH

H2O

N N

B H

Ba(OH)2 H2O

11 HN

N N

B H

Ba2

3 H2O

H 2

Scheme 3. Synthesis of potassium and barium dihydrosbis(1,2,4-triazolyl)borate salts. Journal of Pyrotechnics, Issue 33, 2014

Page 25

N N

B H

Ba2

3 H2O

H2N

NH2

N H

NH2

H2N

N H

NH2

H2N

Cat

z H2O

NH2

Cat

NH2 H2N

N H

13 NH2

H2O

(Cat)2SO4

N H

NH2

NH4

z=0

z=1

z=3

z=0

Scheme 4. Synthesis of metal-free dihydrobis(1,2,4-triazolyl)borate salts. two equivalents of 5-aminotetrazole or 1,2,4-triazole, respectively, led to the formation of the corresponding potassium dihydrobis(azolyl)borate salts. By the acidification of the potassium dihydrobis(azolyl)borate salts the corresponding (azole)dihydrobis(azolyl) boranes were obtained. The following reaction with barium hydroxide resulted in the formation of the barium dihydrobis(azolyl)borate trihydrate salts (Schemes 1 and 3). Both barium compounds 3 and 13 served as starting compounds in the synthesis of the metal-free dihydrobis(azolyl)borates (Schemes 2 and 4). Metal-free dihydrobis(5-aminotetrazolyl)borates were obtained by mixing 3 with nitrogen-rich sulfates in water at room temperature (Scheme 2). The compounds 4–10 could be obtained as colorless solids in high yields and high purity. Recrystallization of the aminoguanidinium salt 5 from water/acetone yielded single crystals of hydrazone 5b. According to the literature8 the dihydrobis(1,2,4triazolyl)borate compounds have been synthesized in the same manner as the dihydrobis(5-aminotetrazolyl)borate salts (Scheme 3). First the potassium salt 11 and then the barium salt 13 were synthesized and served as starting materials for the preparation of metal-free compounds. Metal-free dihydrobis(1,2,4-triazolyl)borates were prepared like the dihydrobis(5-aminotetrazolyl)borates by combining compound 13 and nitrogen-rich sulfates in water and at room temperature, obtaining colorless solids in high yields and high purities as well (Scheme 4). Characterization All salts have been identified and characterized by 1H, 13C, 14 N, and 11B NMR spectroscopy, elemental analysis, mass spectrometry and IR spectroscopy. The 1H and 13C NMR shifts of the 5-aminotetrazolyl and 1,2,4-triazolyl moieties as well as those of the cations are in accordance with the literature values.5a,8 Dihydrobis(5-aminotetrazolyl) borates show 11B NMR shifts at around −15 ppm. The Page 26

B NMR signals of dihydrobis(1,2,4-triazolyl)borate compounds have been observed at around −10 ppm. The 1H NMR spectra of the borate anions show broadened shifts between 3.2 and 3.7 ppm attributed to the BH2 group. The IR spectra of all borate anions show B–H stretching vibrations between 2410 and 2450 cm−1. Additionally B–N stretching vibrations between 1548 and 1540 cm−1 could be observed.10 11

Molecular structures Crystal structures of compounds 5b, 6 and 13 were measured by low temperature X-ray diffraction. Suitable single crystals were picked from the crystallization mixture and mounted in Kel-F oil, transferred to the N2 stream of an Oxford Xcalibur3 diffractometer with a Spellman generator (voltage 50 kV, current 40 mA) and a KappaCCD detector using a λMoKα radiation wavelength of 0.71073 Å. All structures were measured at −173 °C. The data collection and data reduction were carried out with the CrysAlisPro software.11 The structures were solved with Sir-92,11 refined with Shelxl 9712 and finally checked using the Platon software13 integrated in the WINGX software suite.14 The non-hydrogen atoms were refined anisotropically and the hydrogen atoms were located and freely refined. The absorptions were corrected by a SCALE3 ABSPACK multi-scan method.15 Detailed data and parameters of the X-ray measurements and refinements are included in Table 1. Further information regarding the crystal-structure determination has been deposited as cif files with the Cambridge Crystallographic Data Centre16 as supplementary publications with the Nos. 1001410 (5b), 1001412 (6), 1001410 (13). Both guanidinium bis(azolyl)borates 5b and 6 crystallize in the monoclinic space group P21/c. The molecular moieties are depicted in Figures 1 and 2. The density of the hydrazone 5b (1.357 g cm−3) is significantly lower than that of the diaminoguanidinium salt 6 (1.551 g cm−3). This might be a consequence of missing hydrogen bonds of the missing hydrazine NH2 group. Nevertheless both structures are dominated by a large variety of hydrogen bonds. The B–H protons in both structures do not participate in any nonJournal of Pyrotechnics, Issue 33, 2014

Table 1. Crystallographic data of 5b, 6 and 13. Compound

Formula

C6H17BN14

C3H14BN15

C8H18B2BaN12O3

FW [g mol ]

296.15

271.10

489.30

Crystal system

Monoclinic

Orthorhombic

Space Group

P21/c (No. 14)

C2221 (No. 20)

Color/Habit

Colorless plate

Colorless block

Colorless rod

Size [mm]

0.09 × 0.21 × 0.40

0.12 × 0.18 × 0.28

0.13 × 0.14 × 0.30

a [Å]

11.8627(6)

9.2353(3)

7.0683(2)

b [Å]

8.5858(5)

13.3916(5)

13.6825(3)

c [Å]

14.8565(9)

9.3994(3)

18.0720(4)

α [°]

β [°]

106.642(6)

92.738(3)

γ [°]

V [Å3]

1449.77(15)

1161.15(7)

1747.78(7)

ρcalc [g cm ]

1.357

1.551

1.860

µ [mm−1]

0.099

0.118

2.311

F(000)

624

568

960

λMoKα[Å]

0.71073

T [K]

173

θ min-max [°]

4.2, 26.0

4.3, 26.5

4.5, 26.0

Dataset h; k; l

−14:14; −9:10; −18:18

−11:11; −16:16; −11:11

−8:8; −16:16; −22:22

Reflect. coll.

7298

18224

8824

Independ. refl.

2833

2400

1716

Rint

0.021

0.037

0.030

Reflections obs

2371

1809

1676

No. parameters

258

228

155

R1 (obs)

0.0325

0.0296

0.0138

wR2 (all data)

0.0862

0.0743

0.0303

1.02

0.96

1.07

Resd. Dens. [e Å ]

−0.17, 0.19

−0.17, 0.17

−0.32, 0.71

Device type

Oxford Xcalibur3 CCD

Solution

SIR-92

Refinement

SHELXL-97

Absorpt. corr.

multi-scan

CCDC

1001410

1001412

1001411

−1

Z −3

S −3

classical hydrogen bond. The 5-aminotetrazole moieties are comparable to similar structures of alkali aminotetrazolates17 in which the N–H protons are bent out of the ring plane. The cations are almost planar. The B–N distances in both structures are observed similarly (5b: B–N1 1.5587(18), B– N6 1.5659(18); 6: B–N1 1.5594(18), B–N6 1.5540(18) Å). In the structure of 5b (and in contrast to 6) the amino groups of the tetrazole ligand show a cis-like conformation.

Journal of Pyrotechnics, Issue 33, 2014

The barium bis(dihydrobis(triazolyl)borate salt 13 as its trihydrate crystallizes in the non-centrosymmetric orthorhombic space group C2221 with four complexes in the unit cell. Its density of 1.860 g cm−3 is relatively low in comparison to other barium azole salts.18 The barium cation and oxygen atom O1 lie on the C2 axis. Taking into account coordination distances up to 3.6 Å, the barium cations are ninefold coordinated. The bistriazolyl anions act as a bidentate ligand coordinating with their nitrogen atoms N4 Page 27

Figure 1. Representation of the molecular unit of 5b, showing the atom-labeling scheme. Thermal ellipsoids represent the 50% probability level and hydrogen atoms are shown as small spheres of arbitrary radius. Selected anion bond lengths [Å]: N1–B 1.5587(17), N6–B 1.5659(17), N1–C1 1.3477(16), N1–N2 1.3676(14), N2–N3 1.2936(15), N3–N4 1.3621(16), N4–C1 1.3335(16), N5–C1 1.3553(16); selected anion bond angles [°]: N1–B–N6 108.50(10), C1–N1–B 130.80(10), N2– N1–B 122.33(10).

Figure 2. Molecular unit of 6, showing the atom-labeling scheme. Thermal ellipsoids represent the 50% probability level and hydrogen atoms are shown as small spheres of arbitrary radius. Selected anion bond lengths [Å]: N1–B 1.5593(17), N6–B 1.5541(17), C1–N1 1.3395(14) N1–N2 1.3630(13), N2–N3 1.2937(14), N4–N3 1.3572(14), N4–C1 1.3273(15), C1– N5 1.3499(16), selected anion bond angles [°]: N6–B–N1 106.51(9), C1–N1–B 130.44(10), N2–N1–B 122.29(9).

Figure 3. Advanced molecular unit of barium salt 13, showing the atom-labelling scheme. Thermal ellipsoids represent the 50% probability level and hydrogen atoms are shown as small spheres of arbitrary radius. B–N and barium coordination distances [Å]: B1–N5 1.559(3), B1–N2 1.543(3) Ba–N4 3.0058(19), Ba–N1 3.0001(18), Ba–O1 2.723(2), Ba–O2 2.7863(18), Ba–N6ii 3.070(2); Symmetry codes: (i) x, −y, −z; (ii) 0.5+x, 0.5−y, −z; (iii) 0.5+x, −0.5+y, z. Page 28

Journal of Pyrotechnics, Issue 33, 2014

and N1. A 3-D structure is formed by the coordination of the triazolyl nitrogen atoms N6 of two other anions.

were performed using an HR2000+ES spectrometer with an ILX511B linear silicon CCD-array detector and included software from Ocean Optics with a detector–sample distance of 1 m. The dominant wavelength and spectral purity were measured based on the 1931 CIE method using illuminant C as the white reference point. Five samples were measured for each formulation and all given values are averaged based on the full burn of the mixture. Amorphous boron was used as fuel and green coloring agent. Ammonium dinitramide (ADN) and ammonium nitrate were used as oxidizer and magnesium and amorphous boron as fuel, respectively.

Energetic properties The sensitivities towards ignition stimuli and the decomposition temperatures of the products have been determined. Table 2 shows the impact (IS) and friction (FS) sensitivities. Also the sensitivities towards electrostatic discharge (ESD), the dehydration (Tdehydr), melting (Tmelt) and decomposition temperatures (Tdec) are summarized in Table 2. All compounds are barely or not at all impact and friction sensitive. The decomposition temperatures vary between 140 °C and 320 °C. Compound 17 shows a decomposition temperature that is higher than 400 °C. Additionally, the DSC curves of compounds 8, 9, 15 and 16 show peaks between 81 °C and 101 °C corresponding to the loss of crystal water.

Two types of metal-free dihydrobis(azolyl)borates – the di hydrobis(5-aminotetrazolyl)borate and the dihydrobis(1,2,4triazolyl)borate salts – have been investigated as green colorants in several pyrotechnic mixtures. Formulations containing different borate salts but the same content and ratio have been tested, investigating the influence of the borate salts on combustion behavior, color properties and sensitivities. The guanidinium, aminoguanidinium, diaminoguanidinium and N-guanylurea salts of both dihydrobis(azolyl)borates were employed.

Pyrotechnic formulations The replacement of toxic and environmentally hazardous barium nitrate as green-burning colorant by boroncontaining colorants has been investigated. Several pyrotechnic formulations containing different nitrogenrich and metal-free dihydrobis(azolyl)borate salts as green colorants were investigated and compared to the US Army’s in-service green burning composition M125A1 (Ba-control) (Table 3). The pyrotechnic compositions were prepared by mixing all substances, except the binder, in a mortar. Then the binder, a solution of 25% vinyl alcohol acetate resin (VAAR), was added. After drying under high vacuum for 3 hours the mixtures were ground again. Pellets of 0.6 g each were pressed using a consolidation dead load of 2000 kg. The pellets were dried overnight at ambient temperature. The controlled burn down was filmed with a digital video camera recorder (SONY, DCR-HC37E). The performance of each composition has been evaluated with respect to color emission, smoke generation (perceived by the naked eye) and the amount of solid residues. Spectrometric measurements

The barium control M125A1 was mixed on a small scale (0.6 g) and the color properties as well as thermal and energetic properties were measured and are summarized in Tables 3–5. Table 3. US Army composition M125A1. Ba(NO3)2 [wt%] 46

Ba-control

Mg [wt%] 33

PVC [wt%] 16

VAAR [wt%] 5

Table 4. US Army composition M125A1. Burn time Dw [s] [nm] Ba-control

Sp [%]

558

LI [cd] LE [cd s g−1]

390

1950

Table 2. Overview of the physico-chemical properties of 4–18. Formula 4 5 6 7 8 9 10 14 15 16 17 18

MW [g mol–1]

Ωa [%]

Nb [%]

IS [J]

size FS [N] grain [µm]

ESD [J]

Tdehydr [°C] Tmelt [°C] Tdec [°C]

C3H12N13B C3H13N14B C3H14N15B C4H13N14OB C3H14N15O2B C2H14N11O2B C2H10N11OB C5H12N9B C5H15N10OB C5H20N11O3B

241.03 256.04 271.06 284.05 303.06 235.02 214.98 209.02 242.05 293.10

−89.61 −87.48 −85.59 −84.49 −65.99 −71.48 −70.70 −133.95 −118.98 −100.99

75.55 76.59 77.51 69.04 69.33 65.56 71.67 60.31 57.89 52.57

>40 >40 >40 >40 >40 >40 30 >40 >40 >40

>360 >360 324 >360 >360 >360 360 >360 >360 >360

100–500 100–500 100–500 <100 <100 <100 <100 100–500 100–500 100–500

1.5 1.0 0.3 1.5 1.0 0.5 1.0 1.0 1.0 1.0

252.05

−120.61

55.57

>40

>360

<100

1.5

—

139

>400

C4H10N7B

166.98

−138.93

58.72

288

100–500

0.7

—

154

285

C6H13N10OB

— — — — 82 86 — — 81 101

120 143 153 115 190 160 — 111 — 143

173 193 165 164 320 201 140 262 164 212

Oxygen balance assuming the formation of CO2. Nitrogen content.

Journal of Pyrotechnics, Issue 33, 2014

Page 29

Table 9. Color performance of formulations containing compound 4.

Table 5. US Army composition M125A1. Friction [N]

Impact [J] Ba-control

360

Grain size T [°C] dec [μm] <100

258

Table 6 shows the formulations containing guanidinium dihydrobis(5-aminotetrazolyl)borate (4) as colorant. In these two formulations the amounts of the colorant, ADN, boron, magnesium and binder were varied. Both compositions burned without smoke and no solid residues were observable. The spectral purity of GB5At_1 is 70% and lower than that of the barium control, whereas composition GB5At_2 shows with 80% a higher spectral purity than the Ba-control with 75% (Table 4). Two further formulations GB5At_3 and GB5At_4 containing compound 4 (Tables 7 and 8) have been investigated. Formulations GB5At_3 and GB5At_4 contain boron carbide instead of amorphous boron. In order to compare the influence of compound 4 on the combustion behavior, formulations A and B containing only B4C and ADN and NH4NO3, respectively (Tables 11 and 12) were examined. Formulation GB5At_3 burned with an intense green flame color and no smoke and no residues were observable. Composition GB5At_4 also burned smokeless, but the flame color was rather pale-green. This observation is in accordance with the spectral purities. While GB5At_3 exhibits a spectral purity of 85%, which also exceeds that of the barium control, GB5At_4 shows a spectral purity of only 45% (Table 9). Among these four formulations GB5At_3 shows the highest luminous intensity of 76 cd, which is however much lower than that of the barium control formulation (Table 4). Table 6. Formulations GB5At_1 and GB5At_2 containing compound 4.

GB5At_1 GB5At_2

4 [wt%]

ADN [wt%]

B [wt%]

Mg [wt%]

VAAR [wt%]

10 20

60 55

15 9

10 9

5 7

Table 7. Formulation GB5At_3 containing compound 4.

GB5At_3

4 [wt%]

ADN [wt%]

B4C [wt%]

Mg [wt%]

VAAR [wt%]

Table 8. Formulation GB5At_4 containing compound 4.

GB5At_4

Page 30

4 NH4NO3 [wt%] [wt%]

B4C [wt%]

Mg VAAR [wt%] [wt%]

Burn time Dw [s] [nm]

Sp [%] LI [cd]

LE [cd s g−1]

GB5At_1

565

340

GB5At_2

572

160

GB5At_3

562

380

GB5At_4

578

315

The energetic and thermal properties of these formulations were determined as well. Both formulations GB5At_1 and GB5At_2 using ADN, boron and magnesium show very high impact sensitivities of 1 J and 2 J, but are not (>360 N) or only slightly (324 N) friction sensitive (Table 10). Formulation GB5At_3 employing ADN, B4C and Mg shows a lower impact sensitivity of 10 J, but is at 192 N friction sensitive. The combination of NH4NO3, B4C and magnesium in formulation GB5At_4 reveals a composition which is friction and impact insensitive but however shows worse color properties. The decomposition temperatures are in the range between 156 °C and 179 °C in the case of the ADN-containing formulations. The NH4NO3-containing formulation shows a higher decomposition temperature of 201 °C (Table 10). In formulations A and B (Tables 11 and 12) no dihydrobis(azolyl)borate salt was employed. Composition A was compared with GB5At_3 and composition B with GB5At_4. While formulation B burned with no green flame color, formulation A was not ignitable, so that it can be concluded that compound 4 possesses a significant influence on the combustion behavior. Table 10. Energetic and thermal properties of formulations containing compound 4. size Impact [J] Friction [N] Grain [µm] GB5At_1 GB5At_2 GB5At_3 GB5At_4

1 2 10 >40

>360 324 192 >360

Tdec [°C]

<100 <100 <100 <100

156 164 179 201

Table 11. T Formulation A containing ADN and boron carbide. ADN [wt%] A

B4C [wt%] 10

Mg [wt%] 8

VAAR [wt%] 7

Table 12. Formulation B containing ammonium nitrate and boron carbide. NH4NO3 [wt%] B4C [wt%] B

Mg [wt%] 8

VAAR [wt%] 7

Journal of Pyrotechnics, Issue 33, 2014

In order to figure out to what extent compound 5, which compared to compound 4 differs only in the cation, behaves differently in a pyrotechnic formulation and therefore modifies the combustion behavior, three compositions containing 5 were investigated. Formulations AGB5At_1, AGB5At_2 and AGB5At_3 (Table 13) were prepared in the same ratios as the previously discussed compositions GB5At_1, GB5At_2 and GB5At_3. All three formulations revealed green flame colors and a smokeless and residue-free burn down. In Table 14 the performances are summarized. The spectral purities are 61% (AGB5At_1), 71% (AGB5At_2) and 77% (AGB5At_3). Only the spectral purity of AGB5At_3 exceeds the barium control. All three formulations show a burn time of 3 seconds, but different luminous intensities. The highest luminous intensity of 378 cd, which is in the range of the barium control formulation (Table 4), is shown by composition AGB5At_2. The lowest luminous intensity of 37 cd is revealed by formulation AGB5At_1 (Table 14). The comparison of formulations AGB5At_1, AGB5At_2 and AGB5At_3 with GB5At_1, GB5At_2 and GB5At_3 shows that although they only differ in the borate compound the performances are quite different. GB5At_1 and GB5At_2 show higher spectral purities than AGB5At_1 and AGB5At_2, but the latter ones show much higher luminous intensities (Tables 9 and 14). But all three AGB5At-formulations possess burn times of 3 seconds.

Only formulation AGB5At_2 shows a luminous intensity of 378 cd, which is comparable with the barium control (Table 4). In Table 15 the sensitivities towards impact and friction as well as the decomposition temperatures of AGB5Atcompositions are shown. All three formulations are impact but not friction sensitive and thus differ from the GB5At-formulations, which are friction sensitive (Table 10). The decomposition temperatures range from 145 °C to 180 °C (Table 15). Since there is a difference in combustion behavior, color, thermal and energetic properties when using different dihydrobis(5-aminotetrazolyl)borate salts, two further dihydrobis(5-aminotetrazolyl)borate salts 6 and 7 were tested in formulations employing the same components and ratios as the previously discussed formulations. Table 16 summarizes formulations using compound 6. All three formulations burned smokeless, residue-free and with a green flame color. The measured spectral purities (Table 17) and luminous intensities are lower compared with the spectral purity and luminous intensity of the barium control (Table 4). The comparison of the color performances and sensitivity data (Tables 17 and 18) with those of the GB5At-formulations (Tables 9 and 10) or AGB5Atformulations (Tables 14 and 15) reveals differences. Only the decomposition temperatures are comparable. The next compositions contain compound N-guanylurea

Table 13. Formulations AGB5At_1, AGB5At_2 and AGB5At_3 containing compound 5. 5 [wt%] AGB5At_1 AGB5At_2 AGB5At_3

10 10 20

ADN [wt%] 60 60 55

B [wt%] Mg [wt%] — 15 9

VAAR [wt%]

25 10 9

5 5 7

Table 14. Color performance of formulations containing compound 5. Burn time Dw [s] [nm] AGB5At_1 AGB5At_2 AGB5At_3

3 3 3

571 560 561

Sp [%] LI [cd] 61 71 77

37 378 135

LE [cdsg–1] 185 1890 675

Table 15. Energetic and thermal properties of formulations containing compound 5. Impact [J] AGB5At_1 AGB5At_2 AGB5At_3

3 2 4

6 [wt%] DAGB5At_1 DAGB5At_2 DAGB5At_3

10 10 20

ADN [wt%]

B [wt%]

60 60 55

— 15 9

Mg [wt%] 25 10 9

VAAR [wt%] 5 5 7

Table 17. Color performance of formulations containing compound 6. Burn time [s] Dw [nm] Sp [%] DAGB5At_1 DAGB5At_2 DAGB5At_3

3 3 3

580 571 572

42 74 60

LI [cd] LE cd s g−1] 55 139 63

275 695 315

Table 18. Energetic and thermal properties of formulations containing compound 6.

Friction [N]

Grain size Tdec [°C] [µm]

>360 >360 >360

<100 <100 <100

Journal of Pyrotechnics, Issue 33, 2014

Table 16. Formulations DAGB5At_1, DAGB5At_2 and DAGB5At_3 containing compound 6.

169 180 145

Impact [J] DAGB5At_1 DAGB5At_2 DAGB5At_3

2 1 3

Friction [N]

Grain size [μm]

120 >360 144

<100 <100 <100

Tdec [°C] 162 180 154

Page 31

dihydrobis(5-aminotetrazolyl)borate (7) as colorant. In Table 19 the content of formulations using ADN as the oxidizer is summarized. Table 20 shows the composition with ammonium nitrate used as the oxidizer. Formulations NGUB5At_1, NGUB5At_2 and NGUB5At_3 burned green, smokeless and no residues remained. Composition NGUB5At_4 also showed a green flame color and no smoke, but after the burn down some solid residue remained. Except for NGUB5At_3, revealing a spectral purity of 82% (Table 21), which exceeds that of the barium control formulation, all other formulations show a lower spectral purity than the barium control (Table 4). The luminous intensities of all four formulations are rather low. Formulations NGUB5At_1, NGUB5At_2 and NGUB5At_3 are friction and impact sensitive. Formulation NGUB5At_4 is friction insensitive but impact sensitive. As expected the highest decomposition temperature was achieved in formulation NGUB5At_4 using ammonium nitrate as the oxidizer. The other formulations show decomposition Table 19. Formulations NGUB5At_1, NGUB5At_2 and NGUB5At_3 containing ADN and compound 7. 7 ADN [wt%] [wt%] NGUB5At_1 NGUB5At_2 NGUB5At_3

10 10 20

B [wt%]

60 60 55

Mg [wt%]

VAAR [wt%]

25 10 9

5 5 7

— 15 9

Table 20. Formulation NGUB5At_4 containing NH4NO3 and compound 7. 7 [wt%] NGUB5At_4 10

NH4NO3 [wt%]

B [wt%] Mg [wt%]

VAAR [wt%]

Table 21. Color performance of formulations containing compound 7. burn time [s] NGUB5At_1 NGUB5At_2 NGUB5At_3 NGUB5At_4

3 4 5 3

Dw [nm]

Sp [%]

LI [cd]

578 571 573 566

55 54 82 70

35 54 24 30

LE [cd s g-1] 175 360 200 150

Table 22. Energetic and thermal properties of formulations containing compound 7.

NGUB5At _1 NGUB5At _2 NGUB5At _3 NGUB5At _4

Page 32

Impact [J]

Friction [N]

10 5 3 5

216 144 216 >360

Grain size [µm] <100 <100 <100 <100

temperatures between 165 °C and 170 °C (Table 22). In the following section formulations containing dihydrobis(1,2,4-triazolyl)borate compounds are investigated as colorants and compared among themselves and with the corresponding formulations containing dihydrobis(5-aminotetrazolyl)borates. In Table 23 formulations using compound 16 as colorant are summarized. Table 24 shows the composition employing ammonium nitrate as the oxidizer. All four formulations showed green flame colors and a smokeless burn down. While no residues could be observed for compositions DAGBTr_1 and DAGBTr_2, formulations DAGBTr_3 and DAGBTr_4 revealed small amounts of residues. However, the spectral purities are significantly lower than that of the barium control but in the range of the corresponding DAGB5At-formulations (Table 17). The highest spectral purity of 70% was obtained in formulation DAGBTr_4. In the case of formulation DAGBTr_2 a high luminous intensity of 334 cd was achieved, which is in the range of the barium control formulation (Table 25). The first three formulations are impact and friction sensitive. Only composition DAGBTr_4 containing NH4NO3 instead of ADN is friction insensitive. But in contrast to the impact-insensitive formulation GB5At_4 also containing NH4NO3, DAGBTr_4 is impact sensitive. As in the case of all ammonium nitrate containing formulations the highest decomposition temperature of 279 °C was observed for formulation DAGBTr_4 (Table 26). Table 23. Formulations DAGBTr_1, DAGBTr_2 and DAGBTr_3 containing ADN and compound 16.

DAGBTr_1 DAGBTr_2 DAGBTr_3

ADN [wt%]

10 10 20

60 60 55

B [wt%]

Mg [wt%]

VAAR [wt%]

25 10 9

5 5 7

— 15 9

Table 24. Formulation DAGBTr_4 containing NH4NO3 and compound 16. 16 [wt%] DAGBTr_4

NH4NO3 [wt%]

B Mg VAAR [wt%] [wt%] [wt%] 15

Table 25. Color performance of formulations containing compound 16. Burn time [s]

Tdec [°C] 165 169 170 210

16 [wt%]

DAGBTr_1 DAGBTr_2 DAGBTr_3 DAGBTr_4

3 3 4 5

Dw [nm] 577 560 572 564

Sp [%] 55 65 62 70

LI [cd] 118 334 46 75

LE [cd s g−1] 590 1670 307 625

Journal of Pyrotechnics, Issue 33, 2014

Table 26. Energetic and thermal properties of formulations containing compound 16. Impact [J] DAGBTr_1 DAGBTr_2 DAGBTr_3 DAGBTr_4

2 2 4 3

Friction [N]

Grain size [µm]

216 96 144 >360

<100 <100 <100 <100

160 165 173 279

While formulations NGUBTr_1 and NGUBTr_2 burned smokeless and no residues could be observed, composition NGUBTr_3 showed some solid residues and a smokeless burn down as well. Formulation NGUBTr_2 achieved a high spectral purity of 80% (Table 28), exceeding that of the barium control of 75% (Table 4). While in the case of formulations NGUBTr_1 and NGUBTr_3 spectral purities were measured which show values which are comparable to the DAGBTr-formulations (Table 25), NGUBTr_2 exhibits a deviant value (Table 28). All three formulations are friction and impact sensitive. The decomposition temperatures are in the range from 165 °C to 180 °C (Table 29). The comparison of either dihydrobis(5-aminotetrazolyl) borate-containing formulations or dihydrobis(1,2,4triazolyl)borate-containing formulations among themselves revealed that there are significant differences in color performance, impact and friction sensitivity as well as decomposition temperatures when varying the borate compound. All formulations have in common that they exhibit high impact sensitivity, when combining ADN, boron, magnesium and an azolylborate salt. When using NH4NO3 instead of ADN, only formulation GB5At_4 revealed friction and impact insensitivity. Table 27. Formulations NGUBTr_1, NGUBTr_2 and NGUBTr_3 containing compound 17.

NGUBTr_1 NGUBTr_2 NGUBTr_3

ADN [wt%]

B [wt%]

10 10 20

60 60 55

15 9

Mg [wt%]

VAAR [wt%]

25 10 9

5 5 7

—

Table 28. Color performance of formulations containing compound 17. burn time Dw [s] [nm] NGUBTr_1 NGUBTr_2 NGUBTr_3

4 4 3

573 573 574

Sp [%] 55 80 50

Journal of Pyrotechnics, Issue 33, 2014

LI [cd] 179 90 25

Impact [J]

Tdec [°C]

Among the dihydrobis(1,2,4-triazolyl)borates the N-guanylurea dihydrobis(1,2,4-triazolyl)borate salt (17) was tested as colorant as well. The content of the formulations is summarized in Table 27.

17 [wt%]

Table 29. Energetic and thermal properties of formulations containing compound 17.

LE [cdsg−1] 1193 600 125

NGUBTr _1 NGUBTr _2 NGUBTr _3

3 2 2

Friction [N] 120 144 216

grain size [µm] <100 <100 <100

Tdec [°C] 180 165 171

Conclusions Metal-free and nitrogen-rich salts of dihydrobis(5aminotetrazolyl)borate and dihydrobis(1,2,4-triazolyl) borate have been synthesized and characterized by NMR and IR spectroscopy, elemental analysis and mass spectrometry. Crystal structures of compounds 5b, 6 and 13 were obtained. The impact and friction sensitivities as well as the sensitivity towards electrostatic discharge have been determined to be low or not sensitive for all compounds. Formulations using selected borate salts as colorants have been investigated with respect to their performance and sensitivity properties. Formulations employing the same content except for the borate salt have been compared among themselves and with the barium control formulation. It was observed that the use of different borate salts influences significantly the combustion behaviour and the performance and sensitivity properties. Additionally, formulations GB5At_2, GB5At_3, AGB5At_3, NGUB5At_3 and NGUBTr_2 using different borate salts as colorants reveal higher spectral purities than the barium control formulation.

Experimental part All reagents and solvents were used as received (SigmaAldrich, Fluka, Acros Organics) if not stated otherwise. Melting and decomposition points were measured with a Linseis PT10 DSC using heating rates of 5 °C min−1, which were checked with a Büchi Melting Point B-450 apparatus. 1 H, 13C and 14N NMR spectra were measured with a JEOL Eclipse 400, JEOL Eclipse 270 or JEOL EX400 instrument at an ambient temperature of 25 °C if not stated otherwise. All chemical shifts are quoted in ppm relative to TMS (1H, 13 C), nitromethane (14N) or boron trifluoroetherate (11B) as external standards. Infrared spectra were measured using a Perkin Elmer Spectrum BX-FTIR spectrometer with a Smiths DuraSampl IR-ATR unit. Elemental analyses were performed with an Elementar Vario EL or an Elementar Vario EL micro cube. Mass spectra were measured on a JEOL MS station JMS 700 instrument. The impact and friction sensitivity was determined using a BAM drophammer and a BAM friction tester.19 The sensitivities of the compounds are indicated according to the UN Recommendations on the Transport of Dangerous Goods (+):19 impact: insensitive >40 J, less sensitive >35 J, sensitive >4 J, very sensitive <4 J; friction: insensitive >360 N, less sensitive = 360 N, sensitive <360 N >80 N, very sensitive <80 N, extremely sensitive <10 N. The syntheses of potassium dihydrobis(azolyl)borate (1 and 11) salts were performed under an argon atmosphere, Page 33

using standard Schlenk techniques according to the literature procedure.5a,8 Guanidinium, aminoguanidinium, diaminoguanidinium and 5-aminotetrazolium sulfates were synthesized by combining the corresponding chlorides20,21 and silver nitrate in water and room temperature under exclusion of light.

found 75.1. MS (FAB−): m/z for C2H6N10B−: calculated 180.95; found 181.2. IS: >40 J (grain size 100–500 μm). FS: >360 N (grain size 100–500 μm). ESD: 1.0 J (grain size 100–500 μm). DSC: 193 °C (dec.).

CAUTION!

As product a colorless solid (78%) was gained. IR (ATR): ṽ [cm−1] = 3440 (m), 3357 (w), 3311 (m), 3214 (m), 2980 (w), 2480 (w), 2436 (w), 1667 (s), 1621 (s), 1546 (w), 1514 (w), 1500 (w), 1450 (w), 1353 (w), 1325 (w), 1298 (w), 1278 (w), 1199 (w), 1157 (w), 1133 (w), 1117 (w), 1060 (w), 993 (w), 950 (w), 862 (w), 757 (w), 683 (w). 1H NMR (DMSO-d6): δ: 3.54 (br s, 2H, BH2), 4.57 (s, 4H, NH2), 5.58 (s, 4H, C–NH2), 7.21 (s, 2H, NH2+), 7.68 (s, 2H, NH) ppm. 13 C NMR (DMSO-d6): δ: 159.2 (2C, C–NH2), 159.5 (1C, C– NH2) ppm. 11B NMR (DMSO-d6): δ: −14.3 (1B) ppm. EA: C3H14N15B (271.06): calculated N 77.51, C 13.29, H 5.21; found N 76.46, C 13.19, H 5.20%. MS (FAB+): m/z for CH8N5+: calculated 90.11; found 90.1. MS (FAB−): m/z for C2H6N10B−: calculated 180.95; found 181.3. IS: >40 J (grain size 100–500 μm). FS: 324 N (grain size 100–500 μm). ESD: 0.3 J (grain size 100–500 μm). DSC: 165 °C (dec.).

The compounds described in this work are potential explosives, which are sensitive to environmental stimuli such as impact, friction, heat or electrostatic discharge. While we encountered no issues in the handling of these materials, appropriate precautions and proper protective measures (safety glasses, face shields, leather coat, Kevlar gloves and ear protectors) should be taken when preparing and manipulating these materials. Salts of dihydrobis(5-aminotetrazolyl)borate All compounds 4–10 were obtained by the reaction of compound 3 with the corresponding nitrogen-rich sulfates. To a solution of 3 in water the sulfates were added in stoichiometric amounts. The reaction mixtures were stirred at room temperature for 15 minutes, whereupon barium sulfates precipitated. After filtration and evaporation of the solvent, colorless solids were yielded as products.

Diaminoguanidinium dihydrobis(5-aminotetrazolyl) borate (6)

N-Guanylurea dihydrobis(5-aminotetrazolyl)borate (7)

As product a colorless solid (94%) was gained. IR (ATR): ṽ [cm−1] = 3416 (m), 3338 (s), 3184 (s), 2775 (w), 2472 (m), 1778 (w), 1635 (s), 1563 (m), 1542 (m), 1514 (m), 1447 (m), 1296 (m), 1221 (w), 1155 (w), 1126 (w), 1060 (m), 1011 (w), 994 (w), 905 (w), 862 (w), 754 (w), 740 (w), 687 (w). 1 H NMR (DMSO-d6) δ: 3.37 (br s, 2H, BH2), 5.58 (s, 4H, C–NH2), 7.19 (s, 6H, 3∙NH2) ppm. 13C NMR (DMSO-d6): δ: 158.3 (1C, C-NH2), 159.3 (2C, C-NH2) ppm. 11B NMR (DMSO-d6): δ: −15.3 (1B) ppm. EA: C3H12N13B (241.03): calculated N 75.55, C 14.95, H 5.02; found N 73.00, C 14.84, H 5.17%. MS (FAB+): m/z for CH6N3+: calculated 60.08; found 60.1. MS (FAB−): m/z for C2H6N10B−: calculated 180.95; found 181.3. IS: >40 J (grain size 100–500 μm). FS: >360 N (grain size 100–500 μm). ESD: 1.5 J (grain size 100–500 μm). DSC: 173 °C (dec.).

As product a colorless crystals (83%) were obtained. IR (ATR): ṽ [cm−1] = 3458 (w), 3295 (s), 3129 (s), 3009 (m), 2477 (w), 2422 (w), 2356 (w), 2200 (w), 1718 (s), 1702 (s), 1601 (s), 1548 (s), 1462 (s), 1376 (m), 1353 (s), 1300 (m), 1293 (s), 1228 (m), 1191 (m), 1156 (m), 1124 (s), 1099 (s), 1079 (s), 1048 (s), 1025 (m), 976 (m), 933 (m), 868 (m), 781 (s), 749 (s), 730 (s), 702 (s). 1H NMR (DMSO-d6): δ: 3.30 (br s, 2H, BH2), 5.55 (s, 4H, C-NH2), 6.78 (s, 2H, NH2+), 8.15 (s, 4H, NH2), 10.07 (s, 1H, NH) ppm. 13C NMR (DMSO-d6): δ: 154.8 (1C), 155.6 (1C), 159.2 (2C, C-NH2) ppm. 11B NMR (DMSO-d6): δ: −14.6 (1B) ppm. EA: C4H13N14OB (284.05): calculated N 69.04, C 16.91, H 4.61; found N 66.51, C 17.18, H 4.46%. MS (FAB+): m/z for C2H7N4O+: calculated 103.10; found 103.1. MS (FAB−): m/z for C2H6N10B−: calculated 180.95; found 181.1. IS: >40 J (grain size ˂100 μm). FS: >360 N (grain size ˂100 μm). ESD: 1.5 J (grain size ˂100 μm). DSC: 164 °C (dec.).

Aminoguanidinium dihydrobis(5-aminotetrazolyl) borate (5)

5-Aminotetrazolium dihydrobis(5-aminotetrazolyl) borate dihydrate (8)

A colorless solid was obtained as product (60%). IR (ATR): ṽ [cm−1] = 3480 (w), 3420 (w), 3150 (m), 3063 (w), 2908 (w), 2817 (w), 2470 (w), 2348 (m), 1685 (m), 1661 (m), 1620 (s), 1561 (s), 1468 (m), 1455 (m), 1304 (w), 1295 (w), 1278 (w), 1236 (w), 1214 (w), 1189 (w), 1141 (m), 1117 (m), 1103 (m), 1071 (m), 1011 (w), 929 (w), 860 (w), 761 (s). 1H NMR (DMSO-d6): δ: 3.52 (br s, 2H, BH2), 4.63 (s, 2H, N– NH2), 5.59 (s, 4H, C–NH2), 6.74 (s, 2H, C–NH2) 7.22 (s, 2H, C–NH2), 8.55 (s, 1H, NH) ppm. 13C NMR (DMSO-d6): δ: 159.2 (2C, C-NH2), 159.3 (1C, C–NH2) ppm. 11B NMR (DMSO-d6): δ: −14.5 (1B) ppm. EA: C3H13N14B (256.04): calculated N 76.59, C 14.07, H 5.12; found N 74.03, C 14.49, H 4.98%. MS (FAB+): m/z for CH7N4+: calculated 75.09;

A colorless solid (95%) was gained. IR (ATR): ṽ [cm−1] = 3472 (m), 3374 (m), 3190 (m), 2360 (s), 1782 (w), 1638 (s), 1450 (m), 1297 (m), 1156 (m), 1057 (s), 994 (s), 907 (s), 755 (s), 740 (s), 668 (s). 1H NMR (DMSO-d6): δ: 3.17 (s, H2O), 3.27 (br s, 2H, BH2), 6.45 (br s, 4H, C-NH2), 8.10 (br s, 4H, C-NH2/NH) ppm. 13C NMR (DMSO-d6): δ: 156.7 (3C, C-NH2) ppm. 11B NMR (DMSO-d6): δ: −15.4 (1B) ppm. EA: C3H10N15B∙2H2O (303.06): calculated N 69.33, C 11.89, H 4.66; found N 64.71, C 11.68, H 4.56%. MS (FAB+): m/z for CH4N5+: calculated 86.08; found 86.10. MS (FAB−): m/z for C2H6N10B−: calculated 180.95; found 181.1. IS: >40 J (grain size ˂100 μm). FS: >360 N (grain size ˂100 μm). ESD: 1.0 J (grain size ˂100 μm). DSC: 320 °C (dec.).

Guanidinium dihydrobis(5-aminotetrazolyl)borate (4)

Page 34

Journal of Pyrotechnics, Issue 33, 2014

Ammonium dihydrobis(5-aminotetrazolyl)borate hydrate (9) A colorless solid (85%) was obtained. IR (ATR): ṽ [cm−1] = 3138 (w), 3122 (w), 3047 (w), 2439 (w), 2397 (w), 2348 (w), 2287 (w), 1783 (w), 1513 (m), 1428 (m), 1323 (m), 1280 (m), 1216 (m), 1183 (m), 1171 (s), 1157 (s), 1131 (s), 1113 (s), 1015 (m), 990 (s), 895 (m), 886 (m), 870 (s), 718 (m), 677 (s). 1H NMR (DMSO-d6): δ: 3.23 (s, H2O), 3.36 (br s, 2H, BH2), 5.57 (br s, 4H, C-NH2), 7.20 (4H, NH4+) ppm. 13C NMR (DMSO-d6): δ: 158.3 (3C, C-NH2) ppm. 11B NMR (DMSO-d6): δ: −15.2 (1B) ppm. EA: C2H10N11B∙H2O (217.00): calculated N 71.00, C 11.07, H 5.57; found N 70.94, C 11.13, H 5.47%. MS (FAB+): m/z for NH4+: calculated 18.04; found 18.10. MS (FAB−): m/z for C2H6N10B−: calculated 180.95; found 181.2. IS: >40 J (grain size ˂100 μm). FS: >360 N (grain size ˂100 μm). ESD: 0.5 J (grain size ˂100 μm). DSC: 201 °C (dec.). Hydroxylammonium dihydrobis(5-aminotetrazolyl) borate (10) As product a colorless solid (91%) was gained. IR (ATR): ṽ [cm−1] = 3436 (w), 3414 (w), 3323 (w), 3222 (w), 3166 (w), 3111 (w), 2936 (w), 2676 (w), 2489 (w), 2348 (m), 2129 (w), 1655 (m), 1625 (s), 1595 (m), 1563 (m), 1472 (m), 1455 (m), 1299 (m), 1282 (w), 1266 (w), 1218 (w), 1195 (w), 1149 (s), 1131 (s), 1113 (m), 1073 (s), 1024 (m), 992 (m), 877 (w), 867 (m), 762 (s), 750 (m). 1H NMR (DMSO-d6): δ: 3.41 (br s, 2H, BH2), 5.58 (br s, 4H, C-NH2), 10.04 (NH3OH+) ppm. 13C NMR (DMSO-d6): δ: 158.8 (3C, C-NH2) ppm. 11B NMR (DMSO-d6): δ: −15.2 (1B) ppm. EA: C2H10N11OB (214.99): calculated N 71.67, C 11.17, H 4.69; found N 68.81, C 11.13, H 4.78%. MS (FAB+): m/z for NH4O+: calculated 34.04; found 34.1. MS (FAB−): m/z for C2H6N10B−: calculated 180.95; found 181.1. IS: 30 J (grain size ˂100 μm). FS: 360 N (grain size ˂100 μm). ESD: 1.0 J (grain size ˂100 μm). DSC: 140 °C (dec.). Salts of dihydrobis(1,2,4-triazolyl)borate The compounds 14–18 were gained by the reaction of compound 13 with the corresponding nitrogen-rich sulfates. To a solution of 13 in water the sulfates were added in stoichiometric amounts. The reaction mixtures were stirred at room temperature for 15 minutes, whereupon the barium sulfates precipitated. After filtration and evaporation of the solvent, colorless solids were obtained as products. Guanidinium dihydrobis(1,2,4-triazolyl)borate (14) As product a colorless solid (90%) was obtained. IR (ATR): ṽ [cm−1] = 3585 (w), 3432 (m), 3326 (m), 3232 (m), 3151 (m), 3121 (m), 3103 (m), 2424 (w), 2402 (w), 2267 (w), 2183 (w), 2133 (w), 2040 (w), 2010 (w), 1997 (w), 1971 (w), 1957 (w), 1943 (w), 1774 (w), 1736 (w), 1646 (s), 1540 (w), 1503 (s), 1415 (m), 1317 (m), 1267 (m), 1216 (w), 1178 (m), 1160 (s), 1143 (s), 1128 (s), 1031 (w), 1020 (m), 970 (m), 902 (w), 88 (w), 872 (m), 863 (m), 735 (w), 723 (w), 705 (vw), 676 (m), 666 (s). 1H NMR (DMSO-d6) δ: 3.31 (br s, 2H, BH2), 7.60 (s, 2H, CH), 7.99 (s, 2H, CH), 7.43 (s, 6H, Journal of Pyrotechnics, Issue 33, 2014

3∙NH2) ppm. 13C NMR (DMSO-d6): δ: 147.2 (2C, CH), 151.5 (2C, CH), 158.4 (1C, C-NH2) ppm. 11B NMR (DMSO-d6): δ: −10.1 (1B) ppm. EA: C5H12N9B (209.02): calculated N 60.31, C 28.73, H 5.79; found N 58.14, C 27.86, H 5.61%. MS (FAB+): m/z for CH6N3+: calculated 60.08; found 60.1. MS (FAB−): m/z for C4H6N6B−: calculated 148.94; found 149.1. IS: >40 J (grain size 100–500 μm). FS: >360 N (grain size 100–500 μm). ESD: 1.0 J (grain size 100–500 μm). DSC: 262 °C (dec.). Aminoguanidinium dihydrobis(1,2,4-triazolyl)borate hydrate (15) A slightly yellow solid was obtained as product (70%). IR (ATR): ṽ [cm−1] = 3133 (w), 3121 (w), 3114 (w), 3045 (w), 2439 (w), 2398 (w), 2287 (w), 1783 (w), 1513 (m), 1428 (m), 1323 (m), 1280 (m), 1216 (w), 1183 (w), 1171 (m), 1157 (s), 1131 (s), 1113 (m), 1015 (m), 996 (s), 884 (m), 870 (s), 756 (m), 718 (m), 677 (m). 1H NMR (DMSO-d6): δ: 3.21 (s, H2O), 3.46 (br s, 2H, BH2), 4.68 (s, 2H, N–NH2), 6.93 (s, 2H, C–NH2) 7.28 (s, 2H, C–NH2), 7.66 (s, 2H, CH), 7.97 (s, 2H, CH), 8.28 (s, 1H, NH) ppm. 13C NMR (DMSO-d6): δ: 147.9 (2C, CH), 151.6 (2C, CH), 159.3 (1C, C–NH2) ppm. 11B NMR (DMSO-d6): δ: −9.9 (1B) ppm. EA: C5H13N10B (224.04): calculated N 62.52, C 26.81, H 5.85; found N 60.00, C 26.39, H 5.93%. MS (FAB+): m/z for CH7N4+: calculated 75.09; found 75.1. MS (FAB−): m/z for C4H6N6B−: calculated 148.94; found 149.1. IS: >40 J (grain size 100–500 μm). FS: >360 N (grain size 100–500 μm). ESD: 1.0 J (grain size 100–500 μm). DSC: 164 °C (dec.). Diaminoguanidinium dihydrobis(1,2,4-triazolyl)borate trihydrate (16) As product a colorless solid (99%) was gained. IR (ATR): ṽ [cm−1] = 3585 (w), 3432 (m), 3326 (m), 3232 (m), 3151 (m), 3121 (m), 3103 (m), 2424 (m), 2402 (m), 2267 (w), 2133 (w), 1774 (w) 1736 (w), 1646 (s),1540 (w), 1503 (s), 1415 (m), 1317 (m), 1267 (m), 1216 (w), 1178 (m), 1160 (s), 1143 (s), 1128 (s), 1031 (m), 1020 (m), 989 (w), 970 (m), 902 (w), 888 (w), 972 (m), 863 (m), 735 (w) 723 (w), 872 (m), 863 (m), 735 (w) 723 (m), 676 (m), 666 (s), 656 (m). 1H NMR (DMSO-d6): δ: 3.10 (s, H2O), 3.54 (br s, 2H, BH2), 4.53 (s, 4H, NH2), 7.29 (s, 2H, NH2+), 7.65 (s, 2H, CH), 7.98 (s, 2H, CH), 7.79 (s, 2H, NH) ppm. 13C NMR (DMSO-d6): δ: 147.9 (2C, CH), 151.6 (2C, CH), 159.2 (1C, C–NH2) ppm. 11B NMR (DMSO-d6): δ: −9.8 (1B) ppm. EA: C5H14N11B (239.05): calculated N 64.45, C 25.12, H 5.90; found N 58.67, C 25.19, H 5.98%. MS (FAB+): m/z for CH8N5+: calculated 90.11; found 90.1. MS (FAB−): m/z for C4H6N6B−: calculated 148.94; found 149.1. IS: >40 J (grain size 100–500 μm). FS: >360 N (grain size 100–500 μm). ESD: 1.0 J (grain size 100–500 μm). DSC: 212 °C (dec.). N-Guanylurea dihydrobis(1,2,4-triazolyl)borate (17) A colorless solid (82%) was obtained as product. IR (ATR): ṽ [cm−1] = 3447 (w), 3322 (w), 3130 (w), 3087 (w), 2920 (w), 2761 (w), 2356 (w), 2274 (w), 2042 (w), 1724 (m), 1681 (m), 1586 (m), 1541 (w), 1507 (s), 1453 (m), 1417 (w), 1395 (w), 1321 (m), 1270 (m), 1197 (w), 1185 (w), 1171 (m), 1161 (s), Page 35

1132 (m), 1074 (m), 1019 (m), 997 (w), 972 (s), 899 (m), 886 (m), 801 (w), 784 (w), 719 (m). 1H NMR (DMSO-d6): δ: 3.46 (br s, 2H, BH2), 7.67 (s, 2H, CH), 6.81 (s, 2H, NH2+), 7.87 (s, 2H, CH), 8.34 (s, 4H, NH2), 10.03 (s, 1H, NH) ppm. 13 C NMR (DMSO-d6): δ: 147.3 (2C, CH), 152.1 (2C, CH), 154.4 (1C), 155.1 (1C) ppm. 11B NMR (DMSO-d6): δ: −9.5 (1B) ppm. EA: C6H13N10OB (252.05): calculated N 55.57, C 28.59, H 5.20; found N 51.61, C 28.39, H 5.15%. MS (FAB+): m/z for C2H7N4O+: calculated 103.10; found 103.1. MS (FAB−): m/z for C4H6N6B−: calculated 148.94; found 149.1. IS: >40 J (grain size ˂100 μm). FS: >360 N (grain size ˂100 μm). ESD: 1.5 J (grain size ˂100 μm). DSC: >400 °C (dec.).

Chen, P. Chu, R. Damavarapu and T. M. Klapötke, Chemistry – A European Journal, Vol. 18, 2012, p. 628; (d) J. J. Sabatini, J. M. Raab, R. K. Hann Jr. and C. T. Freeman, Zeitschrift für Anorganische und Allgemeine Chemie, Vol. 639, 2013, p. 25. 2

(a) J. A. Conkling, Chemistry of Pyrotechnics: Basic Principles and Theory, Taylor & Francis Group, New York, 1985, 156; (b) J. J. Sabatini, J. C. Poret and R. N. Broad, Angewandte Chemie International Edition, Vol. 50, 2011, p. 4624.

B. E. Douda, DTIC Report, Accession Number 447410, 1969.

(a) D. E. Chavez, M. A. Hiskey and D. L. Naud, Journal of Pyrotechnics, Issue 7, 1998, p. 11; (b) D. E. Chavez and M. A. Hiskey, Journal of Pyrotechnics, Issue 10, 1999, p. 17.

(a) T. J. Groshens, Journal of Coordination Chemistry, Vol. 63, 2010, p. 1882; (b) D. Lu and C. H. Winter, Inorganic Chemistry, Vol. 49, 2010, p. 5795; (c) C. Janiak, Journal of the Chemical Society, Chemical Communications, 1994, p. 545; (d) C. Janiak and T. G. Scharmann, Polyhedron, Vol. 22, 2003, p. 1123.

D. Lu and C. H. Winter, Inorganic Chemistry, Vol. 49, 2010, p. 5795.

C. J. Snyder, P. D. Martin, M. J. Heeg and C. H. Winter, Chemistry – A European Journal, Vol. 19, 2013, p. 3306.

Z. Zeng, B. Twamley and J. M. Shreeve, Organometallics, Vol. 26, 2007, p. 1782.

G. G. Lobbia, M. Pellei, C. Pettinari, C. Santini, B. W. Skelton and A. H. White, Inorganica Chimica Acta, Vol. 358, 2005, p. 1162.

Ammonium dihydrobis(1,2,4-triazolyl)borate (18) As product a colorless solid (91%) was obtained. IR (ATR): ṽ [cm−1] = 3138 (w), 3122 (w), 3047 (w), 2439 (w), 2397 (w), 2287 (w), 1783 (w), 1513 (m), 1428 (m), 1323 (m), 1280 (m), 1216 (w), 1183 (w), 1171 (m), 1157 (s), 1131 (s), 1113 (m), 1015 (m), 990 (s), 886 (m), 870 (s), 718 (m), 677 (m). 1H NMR (DMSO-d6): δ: 3.51 (br s, 2H, BH2), 7.25 (4H, NH4+), 7.51 (s, 2H, CH), 7.61 (s, 2H, CH) ppm. 13C NMR (DMSO-d6): δ: 147.9 (2C, CH), 152.6 (2C, CH) ppm. 11B NMR (DMSO-d6): δ: −9.9 (1B) ppm. EA: C4H10N7B (116.98): calculated N 58.72, C 28.77, H 6.04; found N 53.81, C 28.59, H 5.97%. MS (FAB+): m/z for NH4+: calculated 18.04; found 18.10. MS (FAB−): m/z for C4H6N6B−: calculated 148.94; found 149.1. IS: 40 J (grain size 100–500 μm). FS: 288 N (grain size 100–500 μm). ESD: 0.7 J (grain size 100–500 μm). DSC: 285 °C (dec.).

Acknowledgements Financial support of this work by the Ludwig-Maximilian University of Munich (LMU), the U.S. Army Research Laboratory (ARL) under grant no. W911NF-09-2-0018, the Armament Research, Development and Engineering Center (ARDEC) under grant no. W911NF-12-1-0467, and the Office of Naval Research (ONR) under grant nos. ONR.N00014-10-1-0535 and ONR.N00014-12-1-0538 is gratefully acknowledged. The authors acknowledge collaborations with Dr. Mila Krupka (OZM Research, Czech Republic) in the development of new testing and evaluation methods for energetic materials and with Dr. Muhamed Suceska (Brodarski Institute, Croatia) in the development of new computational codes to predict the detonation and propulsion parameters of novel explosives. We are indebted to and thank Drs. Betsy M. Rice and Brad Forch (ARL, Aberdeen, Proving Ground, MD) for many inspired discussions. The Cusanuswerk is gratefully acknowledged for the award of a PhD scholarship (M. Rusan).

References 1

(a) G. Steinhauser and T. M. Klapötke, Angewandte Chemie International Edition, Vol. 47, 2008, p. 3330; (b) J. J. Sabatini, A. V. Nagori, E. A. Latalladi, J. C. Poret, G. Chen, R. Damavarapu and T. M. Klapötke, Propellants, Explosives, Pyrotechnics, Vol. 36, 2011, p. 373; (c) J. J. Sabatini, A. V. Nagori, G.

Page 36

10 G. Socrates, Infrared and Raman Characteristic Group Frequencies – Tables and Charts, 3rd edn, John Wiley & Sons, Chichester, 2004. 11 SIR-92, 1993, A program for crystal structure solution, A. Altomare, G. Cascarano, C. Giacovazzo and A. Guagliardi, Journal of Applied Crystallography, Vol. 26, 1993, p. 343. 12 G. M. Sheldrick SHELXS-97, Program for Crystal Structure Solution, Universität Göttingen, 1997. 13 PLATON, A Multipurpose Crystallographic Tool, Utrecht University, Utrecht, The Netherlands, A. L. Spek, 1998. 14 L. J. Farrugia, Journal of Applied Crystallography, Vol. 32, 1999, p. 837. 15 Empirical absorption correction using spherical harmonics, implemented in SCALE3 ABSPACK scaling algorithm (CrysAlisPro Oxford Diffraction Ltd., Version 171.33.41, 2009). 16 Crystallographic data for the structure(s) have been deposited with the Cambridge Crystallographic Data Centre. Copies of the data can be obtained free of charge on application to The Director, CCDC, 12 Union Road, Cambridge CB2 1EZ, UK (Fax: int.code Journal of Pyrotechnics, Issue 33, 2014

(1223)336-033; e-mail for inquiry: fileserv@ccdc. cam.ac.uk; e-mail for deposition: deposit-@ccdc.cam. ac.uk). 17 V. Ernst, T. M. Klapötke and J. Stierstorfer, Zeitschrift für Anorganische und Allgemeine Chemie, Vol. 633, 2007, p. 879. 18 R. Damavarapu, T. M. Klapötke, J. Stierstorfer and K. R. Tarantik, Propellants, Explosives, Pyrotechnics, Vol. 35, 2010, p. 395. 19 UN Recommendations on the Transport of Dangerous Goods. Model Regulations 15th edn, United Nations, New York a. Geneva, 2007. 20 T. M. Klapötke, P. Mayer and J. Stierstorfer, Phosphorus, Sulfur, and Silicon and the Related Elements, Vol. 184, Issue 9, 2009, p. 2393. 21 T. M. Klapötke and C. Miró Sabaté, Heteroatom Chemistry, Vol. 63, 2008, p. 301.

Journal of Pyrotechnics, Issue 33, 2014

Page 37

Editorial Policy

Information for Readers and Authors

Articles accepted for publication in the Journal of Pyrotechnics can be on any technical subject in pyrotechnics. However, a strong preference will be given to articles reporting on research (conducted by professional or serious individual experimenters) and to review articles (either at an advanced or tutorial level). Both long and short articles will be gladly accepted. Also, responsible letters commenting on past Journal articles will be published along with responses by the authors as well as “Opinion” pieces. Manner of Submission Submissions should be made directly to the Publisher at the address at the bottom of this page preferably by email. Upon receipt of an article, we will send an acknowledgment and the draft paper will be sent to referees. For specific requests regarding referees etc. please include a note with this information. Editing The Journal of Pyrotechnics is refereed. However, the editing style is friendly, and the author makes the final decision regarding what editing suggestions are accepted. Back Issues Back issues of the Journal will be kept in print permanently as reference material. In addition all past articles and issues are available at the Journal of Pyrotechnics - http://archive.jpyro.com. Publication Routes From the next issue (2015 articles) we will be changing the format and production of the Journal. Following the current issue (2014) we do not intend to routinely produce hard copies of the Journal but will print on demand. We will of course print enough copies to service existing hard copy orders. Times to publication – these have slipped although publishing them online as soon as they are complete means articles are available quicker than when the Journal was published exclusively in hard copy. In order to facilitate swift publishing we are instigating three routes to acceptance and processing of articles. We have prepared a comprehensive template in Microsoft Word which, for the time being, is available by emailing the publisher at jpyro@carndu.com. The template will also be available online at the JPyro archive pages once it is complete and tested. •

Articles submitted using the template will have a projected publication date within 21 days after receipt of the final version from the authors.

•

We will also offer an “expedited” publication – 10 days from receipt of the final version – for payment of a fee of £100. We do not expect this to be used frequently, but recognise that there are occassions when expedited publication is important to authors.

•

Submissions not using the template will be done in approx. 8–12 weeks from receipt of the final version.

Publication Frequency Articles in the Journal of Pyrotechnics will appear first electronically at http://archive.jpyro.com and will subsequently be produced in a complete form annually in the early spring of the year following publication. Subscriptions The following electronic only subscription types are available •

Single article access – indefinite access to the specified article.

•

Issue access – indefinite acccess to a single or multiple issues.

•

Access to the entire archive – annual access to the full archive.

Page 38

Journal of Pyrotechnics, Issue 33, 2014

Display Fireworks And Stage Pyrotechnics In Use – Which Distances Are ‘Safe’ In Germany And Other Parts Of the EU? Christian Lohrer

Federal Institute for Materials Research and Testing (BAM) - Unter den Eichen 87, 12205 Berlin, Germany Email: christian.lohrer@bam.de

Abstract: Display fireworks and theatrical pyrotechnic articles are widely used in the EU by persons with specialist knowledge for local festivities, events, concerts and various music shows. According to the European Directive 2013/29/EU relating to the making available on the Union market of pyrotechnic articles, these articles are categorized as fireworks of category F4 and theatrical pyrotechnic articles of category T2, respectively. Before these pyrotechnic articles may be made available on the market, manufacturers must ensure that they satisfy the essential safety requirements (ESR) of this Directive. By application of the standard series EN 16261 for F4 articles and EN 16256 for T2 articles an assumption of conformity to the ESR is triggered. Both standards do not specify minimum safety distances to the spectators or to the audience, but give guidance to the Member States for setting up their own regulations for defining the safety distances by means of measured article-dependent performance parameters or construction properties. These safety distances differ between the Member States due to the cultural differences and various methods of calculation. This paper explains the procedures for defining safety distances for F4 and T2 in Germany. Respective advantages and disadvantages are pointed out, and results for identical items, categorized as F4 and T2, illustrate the current measures. In addition, a brief overview of the corresponding regulations regarding display fireworks in some other European Member states is presented. The different approaches are compared with each other by calculating the respective safety distances for identical articles.

Introduction With the coming into force of the new European Directive 2013/29/EU1 of the European Parliament and of the Council of 12 June 2013 on the harmonisation of the laws of the Member States relating to the making available on the market of pyrotechnic articles (recast), fireworks and theatrical pyrotechnic articles are subject to certain conformity assessment procedures to demonstrate conformity to the essential safety requirements (ESR) of this Directive. An overview on these general procedures, though under the former Directive 2007/23/EC2, was already given by Kurth3, Ramón4 and Lohrer et al.5,6 Fireworks and theatrical pyrotechnic articles are categorized depending on the potential hazard and intended use into the categories F1–F4 and T1–T2. Article 10 of the Directive 2013/29/EU1 specifies the minimum label requirements for fireworks articles. For category F4 fireworks, which present a high hazard, which are intended for use only by persons with specialist knowledge (commonly known as fireworks for professional

use) and whose noise level is not harmful to human health, the information regarding the minimum safety distance(s) is mandatory on the label. The same requirements exist for theatrical pyrotechnic articles of the category T2 (for stage use which are intended for use only by persons with specialist knowledge). In the recent years the standard series EN 162617 for display fireworks (F4) and EN 162568 for theatrical pyrotechnic articles (T1/T2) were developed by the corresponding working groups (WGs) 2 and 3 of the Technical Committee CEN/TC 212 of the European Committee for Standardization (CEN). By application of these standard series a formal assumption of conformity to the ESR is triggered. However, the corresponding standard series don’t give guidance on how to set or calculate the required minimum safety distances for F4 and T2 articles. These decisions taken by the WGs 2 and 3 at that time were mainly due to differing safety philosophies in the EU Member States: It was challenging at that time to harmonize the understandings of a ‘safe’ distance (if existing at all) throughout all

Article Details

Article No: - 0107

Manuscript Received:-17/07/2014

Final Revisions:-12/11/2014

Publication Date:-12/11/2014

Archive Reference:-1724

Journal of Pyrotechnics, Issue 33, 2014

Page 39

participating Member States. One simple example illustrating this situation is the assessment of flash bangers. Whereas some countries allow the use of such articles only for persons with a special permit, others countries accept their use even in close proximity to persons. Furthermore, in some countries it is common and part of the festivity to run through a curtain of sparks emitted by articles, whereas in other countries people would try to sue the manufacturers or display operators if sparks should hit the audience. The general use of fireworks and the corresponding measures for safe use depend strongly on cultural aspects specific to all different regions in Europe. After long discussions a consensus regarding a specific approach or calculation procedure applicable to all EU Member States was not reached. It was decided by the WG experts to measure and display the relevant performance parameters on the label of the pyrotechnic articles concerned and leave it up to the Member States to define their own regulations for minimum safety distances by means of measured type- and article-dependent performance parameter or construction aspects. It is therefore expected that these calculation procedures will differ between the Member States due to the cultural differences and various methods of calculation. According to the standard series EN 162617 and EN 16256,8 the following performance parameters were generally identified to be relevant for the determination of the minimum safety distances within the Member States in Europe for articles of the categories F4 and T2: •

effect distance (or burst height),

•

sound pressure level (SPL) including the measuring distance,

•

the hazardous debris distance (if existing), and

•

information on incandescent particles returning to the level from which the device was fired (if existing).

In addition to these four general performance parameters, the following category- and typedependent parameters are given: •

radial effect distance (only for T2, and where applicable per type),

•

effect range (only for aquatic fireworks of the category F4),

•

overall duration (only for aerial wheels of the category F4), and

Page 40

•

maximum firing angle (only for articles/tubes placed at an angle other than vertical of the category F4).

The focus of this work is to present the respective requirements for setting safety distances for display fireworks and theatrical pyrotechnic articles according to the German regulations and guidelines. In addition, a brief overview on the safety distance procedures within some other European Member States regarding display fireworks and a comparison between these procedures is illustrated, as well.

Legal requirements in Germany General aspects The approaches for setting safety distances in Germany are purely deterministic depending on major expected hazards, without consideration of any probabilistic behavior. In other words: The safety distances are based on the idea that the article will malfunction (with a probability of 100%), and in case of a shell will fall to the ground without bursting in the sky (‘blind shell’). Comparable deterministic concepts can be found in several other EU Member States, such as France, Italy, The Netherlands, Portugal and Spain. However, these methods are in contrast for example to the assessment philosophy in the United Kingdom, where risk based systems (i.e. hazards in combination with their expected frequencies) are taken into account. The following approaches are published by BAM as application guidelines.9,10 They originate from long-term experience, R&D activities and comprehensive accident investigations, carried out with all involved parties such as manufacturers, pyrotechnicians, enforcement bodies and the notified body BAM. Even though not mandatory, they are currently applied by the vast majority of persons with specialist knowledge and the enforcement bodies in the field, sometimes with slight technical changes. Nevertheless, it is expected that future legislation in Germany regarding the safety distances for display fireworks and theatrical pyrotechnics will adopt these approaches in the same way without major technical changes. The resulting values of the calculations of the minimum safety distances at ‘standardized’ conditions of use and testing (see below) have to be displayed on the label of the F4 and T2 articles by the manufacturers (refer to article 10 of 2013/29/ EU1 and article 12 of 2007/23/EC2, respectively). Journal of Pyrotechnics, Issue 33, 2014

However, the applicable distances determined by persons with specialist knowledge in Germany may deviate from those depending on the specific conditions of use (wind, inclination angle etc.).

Influence of wind Fireworks articles or respective units which are ejected into the air are prone to drift due to wind force. In case of wind speeds >9 m s−1 but ≤13 m s−1 the safety distances in the direction of the wind (including those after assessment of inclination angle) must be increased by 100%. If the wind speed exceeds 13 m s−1 the safety distances in the direction of the wind (including the ones after assessment of inclination angle) must be increased by 200%.

Display fireworks of the category F4 Standardized conditions of use and testing The distances displayed in Table 1 refer only to a ‘standard’ use in which they were tested, i.e. vertical orientation of the mortars, firing level equal to level of audience and wind less than 9 m s−1 at a representative location at a height of 2 m.

These increased values however don’t apply to ground fireworks (refer to entry no. 1 of Table 1).

As already pointed out earlier, conditions of use might lead to different values in application, e.g. depending on wind, angle of orientation, elevated shooting locations etc. These influences on the safety distances are illustrated in the following.

In the direction opposite to the wind direction the safety distance may be reduced by a maximum of 40%. Influence of the height of buildings If display fireworks are fired from the roofs of buildings, then the respective height is added to the burst and effect heights of the articles falling under no. 2–4 and 8 of Table 1. The corresponding adjustments due to inclination angle and wind have to be made in addition afterwards.

Influence of inclination angle If the fireworks articles are used at an angle, the safety distances according to Table 2 apply: In the direction opposite to the firing direction the safety distance may be reduced by a maximum of 40%.

Table 1. Safety distances of display fireworks in Germany under standardized conditions No.

Fireworks articles

Safety distances to the audience

Ground fireworks

20 m; in case of lance work the maximum distance of the single articles applies

Shells and bombettes with caliber ≥50 mm (also out of mines, Roman candles, shot tubes) Shells and bombettes with report as main effect (also out of mines, Roman candles, shot tubes)

80% of the burst height, at least 800 × (caliber in mm)

3 4 5 6 7 8

Rockets and aerial wheels

100% of the burst height, at least 1000 × (caliber in mm) 200 m in shooting direction, 125 m to the audience and all other directions

Miscellaneous fireworks with effect-/burst heights <30 m which do NOT belong to numbers 30 m 2–4 Miscellaneous fireworks with effect-/burst heights >30 m which do NOT belong to numbers 50 m 2–4 1.5 × maximal effect distance plus 2 × radial effect Aquatic shells width Daylight shells without burning matter/effects 80% of the burst height

Table 2. Safety distances for display fireworks in dependence on the inclination angle for fireworks falling under the numbers 2–4 and 8 of Table 1. Inclination angle from the vertical

Increase of safety distance

5–10° 11–15° 16–20% >20%

40% 60% 80% Single case evaluation necessary

Journal of Pyrotechnics, Issue 33, 2014

Page 41

Influence of terrain elevations

theatrical venue, e.g. indoors/outdoors).

If the display fireworks are fired from elevated terrains with slopes ≥20%, the safety distance of the articles falling under no. 2–4 and 8 of Table 1 has to be increased by 20%. The corresponding adjustments due to inclination angle and wind have to be made in addition afterwards. In case of almost vertical slopes (such as buildings), the previous clause applies.

The minimum safety distances based on effect dimensions (MSDeffect; contains the distances in effect and radial effect directions) for articles of category T2 shall be calculated in effect and radial effect directions based on formula (1):

Theatrical pyrotechnics of the category T2 The basis for the calculation of the safety distances for theatrical pyrotechnic articles of the category T2 is the assessment of the effect dimensions (e.g. how far do the sparks go) and the sound pressure level. The effect dimensions comprise the effect distance and the radial effect distance. According to EN 162568, the following definitions apply: Effect distance: maximum distance of the effect in the direction of firing measured from the base of the article. Radial effect distance: maximum distance of the effect in any direction except in the direction of firing. Minimum safety distance (MSDED): minimum distance of persons or flammable materials or obstructions from the article in the direction of the effect to reduce the risk to as low as reasonably practicable. Minimum radial safety distance (MSDRED): minimum distance of persons or flammable materials or obstructions to the article, in any direction except the direction of firing, to reduce the risk to as low as reasonably practicable. NOTE: There might be different safety distances for persons who are protected from the effects by protective equipment and those without protective equipment. This is often the case on stages where the artists are closer to the effect than the audience and therefore need special protection. The T2 procedure is generally identical with the setting of safety distances for theatrical pyrotechnic articles of the category T1 within conformity assessment procedures applying EN 162568. Standardized conditions of use and testing The minimum safety distances given in the following refer again only to a ‘standard’ use in which they were tested, i.e. vertical orientation of the articles, firing level equal to level of audience and wind less than 9 m s−1 at a representative location at a height of 2 m (if at all applicable at a Page 42

MSDeffect = 1.3 × L (1) where L is the maximum length measured during testing based on effect dimensions, debris, and burning or incandescent matter. In addition, the minimum safety distance based on sound pressure level (MSDSPL; contains the distances in effect and radial effect directions) can be estimated by using the formula (2), based on the requirement that 120 dB (AImax) shall not be exceeded outside the safety distance:

MSDSPL  10

(log( rmeasurement ) 

(SPL threshold SPL measurement ) ) 20

(2)

where rmeasurement = distance of sound pressure meter to the pyrotechnic article [m] SPLthreshold = sound pressure level threshold of = 120 dB(AImax) SPLmeasurement = sound pressure level measured at rmeasurement in dB(AImax) The resulting minimum safety distance (in effect and radial effect direction) equals the maximum of both values MSDeffect and MSDSPL. Influence of inclination angle The placement of the articles under an angle towards the audience is permitted under the following conditions as displayed in Figure 1 in combination with formula (3): MSDresulting = MSDED × cos(α)

+ MSDRED × cos(90° − α) (3)

MSDresulting = resulting minimum safety distance to the audience and α = inclination angle [°]. Influence of wind The effects and ejected units from theatrical pyrotechnic articles are prone to drift due to wind force. In case of wind speeds >9 m s−1 but ≤13 m s−1 the safety distances in the direction of the wind (including those after assessment of inclination angle) must be increased by 100% for articles having an effect or burst height of >30 m under standardized conditions of use. If the wind speed exceeds 13 m s−1 the safety distances in Journal of Pyrotechnics, Issue 33, 2014

Figure 1. Resulting minimum safety distances to the audience for theatrical pyrotechnic articles T2 placed at an angle towards the audience. the direction of the wind (including those after assessment of inclination angle) must be increased by 200% for articles having an effect or burst height of >30 m under standardized conditions of use. Specific bouquet effects For articles that produce wide bouquet effects at large altitudes (e.g. Crossette) it is permitted to reduce the resulting minimum safety distance to the audience near the ground down to 2 m. This, however, is only after thorough case-by-

case assessments by the persons with specialist knowledge of the possible resulting hazards, such as falling debris or burning/incandescent matter, blind stars, low burst heights, radial effect distances at that height and sound pressure level (the requirement of 120 dB(AImax) still applies, as well).

Figure 2. Exemplary safety distances of a mine – based on effect dimensions for display fireworks F4 and theatrical pyrotechnic articles T2 in Germany. Journal of Pyrotechnics, Issue 33, 2014

Page 43

Conceptual differences for F4 and T2 in Germany

distances for theatrical pyrotechnics are always defined in each single direction (i.e. firing direction and radial). For example: A F4 mine (caliber >50 mm) consisting of bombettes would have a safety distance to the audience equal to 80% of the burst heights of the bombettes (though, at least 800 × caliber in mm). However, another mine categorized as a T2 theatrical pyrotechnic article would have a radial safety distance to the audience depending on the maximum radial effect

As displayed above, the concepts of safety distances for theatrical pyrotechnic articles and display fireworks differ significantly. Where in the latter case the safety distance to the audience is predominantly based on effect/burst height (direction shifted by 90° from the horizontal in case of standardized use) and caliber, the safety

Table 3. Safety distances to the audience for identical articles categorized asT2 and F4 in direct comparison Type of article

Effect/burst height [m]

Comet (star)

100

Radial effect distance [m]

MSDRED for T2 [m]

MSD for F4 [m]

f(SPL) ~10 m*

Comet (bombette)

100

f(SPL) ~20 m

Mine (stars only)

Fountain/jet

Airburst

Rotating effect (saxon)

*Estimated distance at which 120 dB(AImax) are reached.

Table 4. Specific testing requirements for theatrical pyrotechnic articles and display fireworks in the EU Specific standard testing requirements Minimum number of items to be type tested on function (destructive) Minimum number of items to be dismantled during type tests (destructive) Measurement of effect width/radial effect distance during type and batch tests

Minimum number of items to be batch tested on function (sample size depending on lot size)

T2: refer to EN 16256 (T2)8

F4: refer to EN16261 (F4)7

30 pcs.

9 pcs.

Yes

Lot size 2–15 16–25 26–90 91–150 151–500 501–1200

Sample size 1 2 3 5 8 13

Lot size

Sample size

2–25

26–150

151–500 501–1200

3 8

Relevant performance parameter values of the single results within a permitted tolerance (in %) of the measured average during batch tests

±20%

±30%*

Functioning of articles

Complete and 100%

‘As intended’ by manufacturer† Not all variants of a product family need to be tested (e.g. min and max)

family variants are tested Criteria for testing ‘similar’ items (variants) within All for function unless they differ ‡ a product family only in color effects

* Except otherwise justified by manufacturer; therefore > ±30 % possible. † Therefore <100% possible. ‡ Guidelines published by the European Commission for grouping articles that are similar in design, function or behavior into product families.11

Page 44

Journal of Pyrotechnics, Issue 33, 2014

distance, which is usually much smaller than the burst height (considering the sound pressure level is not dominating, see Figure 2). Table 3 illustrates this discrepancy between safety distances to the audiences for further articles that could likely be categorized either as F4 or T2. In addition, the permitted inclination of articles and their effects towards the audience in case of T2 is different to the conditions of use for display fireworks. These differences in the assessment of articles falling under F4 and T2 can be justified with the tightened testing criteria for T2 in comparison with F4. This is displayed in Table 4. As displayed above, the criteria for passing the type and batch tests are significantly stricter for T2 articles in comparison with F4 articles. Therefore, and as a rule of thumb, theatrical pyrotechnics might be generally considered to be of a better quality than regular display fireworks due to the intended close proximity use. In consequence, it appears to be justifiable to apply smaller ‘safe’ distances to the audience in comparison with display fireworks.

The situation for display fireworks in other EU Member States General remarks The procedures that are briefly described in the following for the countries France (FR), Spain (ES), The Netherlands (NL), Italy (IT) and Portugal (PT) are neither complete nor all-embracing. The focus of this work was on some specific

article types used and tested under ‘standardized conditions’. The safety distances displayed here are defined from the article/effect to the audience. Safety distances to buildings or other objects, which might deviate, are not taken into account. The assessment of adjustments which might be necessary for the respective ‘non-standard use’, such as tilted mortars, wind, elevated shooting locations etc. was also not part of this study. These values and procedures presented below were to the best knowledge of the author up-todate when writing this contribution. With the exception of the UK, all other countries which are part of this study apply deterministic approaches for setting safety distances. In these cases, the potential hazards (the article will malfunction in a certain way) are taken as the only basis for setting the respective safety distances. Any assessment regarding a certain probability of malfunctioning is not carried out. France The French system12 for determining safety distances is quite comparable to the described German procedures. It is also based on performance data (effect/burst height), construction aspects (caliber), effects (stars/visual and report/aural) and the actual generic type of the fireworks. A compendium of some relevant requirements is given in Table 5. Portugal In contrast to the systems in Germany and France, the procedures in Portugal13 to set the safety distances for display fireworks comprise only

Table 5. Safety distances of display fireworks in France under standardized conditions No.

Fireworks articles

Safety distances to the audience

Shells (except report shells)

80% of the burst height, at least 800 × (caliber in mm)

Shells with report as main effect

100% of the burst height, at least 800 × (caliber in mm)

Rockets

150 % of burst height. If these results are in between the following ranges, then the upper range limit applies as the MSD: 0–25 m: 25 m 25–50 m: 50 m 50–100 m: 100 m >100 m: values rounded to the next 10 m.

Mines (stars only)

100% of the effect height, at least 800 × (caliber in mm)

Roman candles (stars only)

50% of effect height

Fountain (small effect width)

effect distance, at least 15 m

Shot tube (comet)

50% of effect height

Shot tube (bombette)

80% of the burst height, at least 800 × (caliber in mm)

Ground fireworks (without movement)

effect distance, at least 15 m

Journal of Pyrotechnics, Issue 33, 2014

Page 45

the information regarding the firework types and respective calibers. For rockets, the calibers apply to the rocket motors. Performance parameters such as effect/burst heights are not taken into account. The relevant type-dependient safety distances are given in Table 6 and Table 7. Italy The Italian system14 distinguishes between ground and aerial effects depending on the type of fireworks and the respective caliber. It is therefore comparable with the Portuguese system. Some relevant details for this study are presented in Table 8. The Netherlands The Dutch regulations15 also depend on the firework type/effect and the relevant construction property caliber. Table 9 displays an extract of these approaches. Spain The concept idea in Spain16 is more or less in line with the regulations in Italy, Portugal and The Netherlands, as it considers the types/effects and the caliber of the display fireworks. Performance parameters like effect/burst height are not taken into account. A summary is given in Table 10.

declined to define fixed distances or specific procedures on how to calculate these values for display fireworks. Whereas in the other countries deterministic approaches are the preferred choice (purely focusing on some expected hazards), risk based assessments have to be carried out by the persons with specialist knowledge on a case-bycases basis in the UK. Within a risk assessment, not only the expected hazards are taken into account, but also their expected frequencies (probability of occurrence). In other words: ‘what could happen and how often’? The Health and Safety Executive (HSE) requires each person with specialist knowledge to make an appropriate risk assessment based on things like: properties of the fireworks (e.g. type, effect, caliber, debris, long-burning stars,), environmental conditions (e.g. wind, rain/humidity) proximity to flammable materials, audience or performers, Table 7. Safety distances of display rockets in Portugal under standardized conditions No.

Caliber of the rocket motors [mm]

Safety distances to the audience [m]

<15

United Kingdom (UK)

As already pointed out, the situation in the UK differs significantly from the other countries presented in work. The legislator in the UK

100

110

125

130

140

Table 6. Safety distances of display fireworks (except rockets) in Portugal under standardized conditions No.

Caliber of fireworks Safety distances to (except rockets) [mm] the audience [m]

100

125

100

150

120

200

160

250

200

Page 46

Table 8. Safety distances of display fireworks in Italy under standardized conditions No.

Type of fireworks

Safety distances to the audience [m]

Waterfalls, fountains, wheels

Shot tubes: caliber <25 mm caliber 25–50 mm caliber 50–110 mm

40 50 100

Cylindrical shells and rockets: caliber <110 mm caliber 110–130 mm caliber 130–210 mm

100 150 200

Spherical shells: caliber <130 mm caliber 130–220 mm caliber 220–400 mm

100 150 200

Journal of Pyrotechnics, Issue 33, 2014

Table 9. Safety distances of display fireworks in the Netherlands under standardized conditions No.

Type of fireworks

Safety distances to the audience [m]

1 2 3 4

Rockets (away from audience) Lance work Ground fireworks (e.g. fountains) Roman candles (≤2 inch ≈ 50.8 mm) Mines: caliber ≤4 inch (101.6 mm) caliber 4–6 inch (101.6–152.4 mm) Shells: caliber <3 inch (<76.2 mm) caliber >3 inch (>76.2 mm) caliber >4 inch (>101.6 mm) caliber >5 inch (>127 mm) caliber >6 inch (>152.4 mm) caliber >8 inch (>203.2 mm) caliber >10 inch (>254 mm) caliber >12 inch (>304.8 mm) caliber >18 inch (>457.2 mm) caliber >24 inch (>609.6 mm)

125 15 30 75

consideration of buildings and amenities close by etc. Based on that the appropriate safety distances are identified and applied by the persons with specialist knowledge. This decision can be based on the experience of that person, or other measures might be taken into account (e.g. using

60 100 120 165 200 230 265 325 390 455 645 845

specialized software like ShellCalc©17,18 or noise level calculations), as well.

Table 10. Safety distances of display fireworks in Spain under standardized conditions No.

Type of fireworks

Ground fireworks (no projected items): caliber 20 mm caliber 30 mm caliber 40 mm caliber 50 mm caliber 60 mm caliber 70 mm Shells, mines (with report effects): caliber 50 mm caliber 60 mm caliber 75 mm caliber 100 mm caliber 120 mm caliber 125 mm caliber 150 mm caliber 175 mm caliber 180 mm caliber 200 mm caliber 250 mm caliber 300 mm caliber 350 mm Mines (stars only, no report effect): aliber < 50 mm caliber < 75 mm caliber < 100 mm caliber < 120 mm caliber < 150 mm Roman candles, shot tubes: caliber < 50 mm caliber < 60 mm caliber < 70 mm Rockets

Journal of Pyrotechnics, Issue 33, 2014

Safety distances to the audience [m] 10 12 14 20 30 40 25 36 45 60 72 75 120 140 145 200 250 300 350 25 35 50 60 75 25 48 56 50

Page 47

Comparison between the presented deterministic approaches for display fireworks – discussion and interpretation Based on the information provided in the previous clauses, a comparison between the deterministic procedures of the countries Germany, France, Spain, The Netherlands, Italy and Portugal is displayed for some relevant display fireworks articles (category F4). The detailed information is given in Table 11. The resulting safety distances to the audience are based on ‘standardized’ conditions of testing and use as described earlier in the text. Relevant and frequently used fireworks types were chosen combined with exemplary calibers and estimated/calculated effect and burst

heights. For each fireworks type and caliber or effect/burst height, the respective extreme values are highlighted (maximum in bold face; minimum in italic) in Table 11. For the case of spherical shells of Peony type, the single values are in addition displayed in Figure 3. With regards to Figure 3 one has to keep in mind that in some systems (France and Germany) the influence of the respective caliber might be dominated by the burst height, which is not displayed on the abscissa. However, this information can be taken from Table 11. The differences between the single approaches are remarkable. This, however, was expected due to the different cultural events and histories regarding festivities and fireworks displays within the countries concerned.

Table 11. Safety distances of display fireworks in some European countries under standardized conditions

Type of article

Resulting safety distances to the audience in ‘standard’ conditions DE FR ES NL IT PT

Nominal caliber

Burst/effect height*

Spherical Shell-Peony

[mm] 75

[m] 136

[m] 109

[m] 45

[m] 120

[m] 100

(stars as main effect)

100

184

147

165

100

150

240

192

120

230

150

200

285

228

200

265

150

Spherical Shell- Salute

136

120

100

(report as main effect)

100

184

165

100

150

240

120

230

150

200

285

200

265

150

Mine

(stars as effect)

100

Roman candle

(stars as effect)

120

100

Fountain

(small effect width)

Shot tube

(Comet/star effect)

Shot tube

(Bombette/report effect)

Rocket

125

110

125

100

(caliber for rocket motor)

150

125

230

125

100

Ground fireworks

[m] 60 80 120 160 60 80 120 160 25 60 9 48 9 20 9 25 9 25 75 100 6

(no movement) * Estimated; for shells using ShellCalc© v5.1.8 (0 m s−1 wind, 0 m elevation, terrain 2, other entries default or zero)

Page 48

Journal of Pyrotechnics, Issue 33, 2014

Figure 3. Resulting safety distances to the audience for spherical shells (type Peony) of different calibers in comparison between the individual deterministic approaches. By taking a closer look at TableÂ 11 it is noticeable that the approaches of Spain and Portugal in no cases revealed a specific maximum of the single safety distances for the chosen fireworks articles. However, the approach from The Netherlands showed in 14 out of these 21 exemplary cases maximal values for the safety distances in comparison with the other systems (in some cases together with other countries). In contrast, both systems of The Netherlands and Germany revealed in no cases respective minimal values for the safety distances in comparison with the other systems. This, however, was reached by Spain in 13 cases and by Portugal in 10 cases (in some cases parallel). These findings might lead to the conclusion that the system from The Netherlands could be evaluated as the most conservative approach for defining safety distances for display fireworks between these investigated countries under standardized conditions. In consequence, the reverse finding might also be justifiable, in a way the safety distances in Spain and Portugal in many cases are not considered as conservative. This is most likely due to the very specific fireworks culture in these countries, combined with the significantly different social acceptance of fireworks and their Journal of Pyrotechnics, Issue 33, 2014

effects in these countries. The presented approaches for the determination of safety distances under standardized conditions of use and testing (e.g. vertical orientation, no wind) might be put into three different general groups, depending on the type and number of parameters influencing the respective safety distances (TableÂ 12). The necessary information needed to apply the respective systems generally increases with these group numbers. All of the three systems from The Netherlands, Spain and Portugal that were identified in this work that lead to extreme values (min and max) for safety distances as presented above would belong to the first group, which generally requires the least amount of information necessary for the assessment and calculation. The main advantage of the approaches of this group, and partly of the second group as well, is the fact that these approaches are simple and handy to apply by the persons with specialist knowledge and in the same way by the enforcement bodies responsible for the granting of permission of the fireworks displays. Another important advantage of this philosophy of group 1 is the fact that these approaches are independent from any performance parameters that Page 49

Table 12. Overview of the relevant influence parameters for setting safety distances in the selected European countries Group

Relevant parameter

Applied by

Type of article/effect (e.g. shell) and constructional properties (e.g. caliber) Type of article/effect, constructional properties and performance parameter (e.g. burst/effect height) Type of article/effect, constructional properties, several performance parameters (e.g. burst/effect heights, debris, long burning stars) and risk assessment (hazard x frequency)

Italy, Spain, Portugal and The Netherlands

2 3

need to be measured during type and batch testing. Recent investigations regarding the measurement of burst/effect heights for display fireworks revealed significant discrepancies between several test methods, procedures and equipment that are commonly used in the field.19 In these cases one has to expect comparatively high measurement uncertainties for the values of burst/effect heights in general. However, on the negative side, one might identify the possibility of over- and underestimating certain hazards in comparison with risk based approaches that require more information on the articles and surrounding conditions. This might be due to neglecting certain hazards (such as explosion of the shell on the ground after falling down or explosion directly inside the mortar, burning matter on ground) or considering unrealistic functioning or behavior. In conclusion, all approaches for setting safety distances for display fireworks have advantages and weak points in their application in daily life. These respective systems must, however, be seen in the context of the situation in each single country due to different cultural aspects and social acceptance of fireworks and their corresponding effects. The term ‘safe’ is and will always be assessed differently by the respective countries. Simple examples highlighting these differences are the question of what is ‘loud’ (being exposed to a certain sound pressure level) and the presence of sparks (being exposed to effect sparks or some incandescent matter/fallout). Each system and approach presented in this work has its own eligibility and justification for being applied in the respective countries.

Mr. L. Aufauvre, Mr. N. Beato, Ms. B. Collado Page 50

United Kingdom

Aguilar, Mr. E. de Jong, Ms. M. Douet, Mr. S. Myatt, Ms. S. Parente, Mr. Á. Santamaría Martín, Ms. F. Signorini, Dr. T. Smith, and Mr. P. Thebault. Their kind, competent and prompt support at short notice was very much appreciated. Any mistake or incorrect information that might be displayed in this work regarding the calculation procedures of safety distances is the result of a false interpretation or implementation into this paper by the author alone.

References 1

Directive 2013/29/EU of the European Parliament and of the Council of 12 June 2013 on the harmonisation of the laws of the Member States relating to the making available on the market of pyrotechnic articles (recast), Official Journal of the European Union, L 178/27, 28.6.2013.

Directive 2007/23/EC of the European Parliament and of the Council of 23 May 2007 on the placing on the market of pyrotechnic articles, Official Journal of the European Union, 14.6.2007.

L. Kurth, “Display fireworks under the European pyrotechnics directive – how BAM gets prepared”, Proc. 10th International Symposium of Fireworks in Montréal, Canada, pp. 122–130, 2007.

C.F. Ramón, “The new “European manufacturer” of pyrotechnic articles”, Proc., 12th International Symposium of Fireworks in Porto/Gaia, Portugal, pp. 250–257, 2010.

C. Lohrer, H. Fink and L. Kurth, “Quality systems for fireworks – auditing experiences by the notified body BAM”, Proc. 13th International Symposium of Fireworks in Malta, pp. 160–172, 2012.

C. Lohrer, H. Fink and L. Kurth, “Failure rates of consumer fireworks –testing experiences by the notified body BAM”, Proc. 14th International Symposium of

Acknowledgements The authr would like to thank the following persons for supplying the respective national regulations on how to determine the safety distances for display fireworks (in alphabetical order):

France and Germany

Journal of Pyrotechnics, Issue 33, 2014

Fireworks in Changsha/PR China, pp. 166– 180, 2013. 7

EN 16261:2013 Standard series for Pyrotechnic articles – Fireworks, Category 4, consisting of four parts, CEN/TC 212 WG2, 2013. EN 16256:2012 Standard series for Pyrotechnic articles – Theatrical pyrotechnic articles, consisting of five parts, CEN/TC 212 WG3, 2012. Guidelines for safety distances for display fireworks F4 in Germany: http://www.bam.de/ de/service/amtl_mitteilungen/sprengstoffrecht/ sprengstoffrecht_medien/leitf_abbrennen_ von_feuerwerk_f4.pdf; Bundesanstalt für

17 J. Harradine and T. Smith, “A Prediction of Aerial Shell and Comet Trajectories Using ShellCalc©”, Journal of Pyrotechnics, Issue 22, pp. 9–15, 2005. 18 T. Smith, “ShellCalc© – 10 years on. A review of progress and application”, Journal of Pyrotechnics, Issue 32, pp. 3–22, 2013. 19 M. Douet, C. Lohrer and M. Lefebvre, “Round robin test for height measurements of various pyrotechnics within the forum of notified bodies in Europe”, Proc. 14th International Symposium of Fireworks in Changsha/PR China, pp. 216–228, 2013.

Materialforschung und -prüfung (BAM), accessed June 16th, 2014.

10 Guidelines for safety distances for theatrical pyrotechnics T2 in Germany: http:// www.bam.de/de/service/amtl_mitteilungen/ sprengstoffrecht/sprengstoffrecht_medien/ leitf_verwendung_theaterpyrotechnik_t2.pdf;

Bundesanstalt für Materialforschung und -prüfung (BAM), accessed June 16th, 2014. 11 Guidelines for grouping of articles that are similar in design, function or behavior into product families for F4 and T1/T2: http://

ec.europa.eu/enterprise/sectors/chemicals/ documents/specific-chemicals/pyrotechnicarticles/index_en.htm; European Commission,

accessed June 25th, 2014.

12 Safety distance regulations – France: http://

www.bulletin-officiel.developpement-durable. gouv.fr/fiches/BO201017/DEVP1020325S%20 recueil%20juillet%202010_vdef.pdf; accessed

July 1st, 2014.

13 Safety distance regulations – Portugal: http:// www.apipe.org/docs/INSTRUCAO%20PSP%20 FOGOS%20ARTIFICO%20aprovada.pdf;

accessed July 1st, 2014.

14 Safety distance regulations – Italy:

http://www.gazzettaufficiale.it/eli/ id/2014/06/09/14A04345/sg in connection with http://www.indicenormativa.it/sites/default/files/ Interno_20110111_Circolare_FuochiArtificio. pdf; accessed July 1st, 2014.

15 Safety distance regulations – The Netherlands: http://wetten.overheid.nl/

BWBR0031686/geldigheidsdatum_02-07-2014;

IENM/BSK-1012/109051; accessed July 2nd, 2014.

16 Safety distance regulations – Spain: http://

www.elcampello.es/upload/contratos_ficheros/ microsoft_word__ppt_pirotecnia_121.pdf;

accessed July 1st, 2014.

Journal of Pyrotechnics, Issue 33, 2014

Page 51

The Journal of Pyrotechnics Archive on the web The primary means of publication of the Journal is now electronic. Not only does this allow for faster publication times, but it presents many opportunities for online notification of new articles and for discussion. The archive webpage http://archive.jpyro.com contains all the articles published in the Journal, together with other publications from the Journal of Pyrotechnics Inc. and Pyrolabs Inc. It also contains some supplementary information to existing papers. Articles are generally â&#x20AC;&#x153;pay to downloadâ&#x20AC;? - registration (free) and completing a purchase allows indefinite access to the articles. (Please follow the instructions carefully). New articles will usually be notified to all those who have registered on the site.

The JPyro Archive Facebook page http://www.facebook.com/ jpyroarchive allows users (be they subscribers or not) to obtain information on recently published articles.

Page 52

Journal of Pyrotechnics, Issue 33, 2014

Comparison of national “safety distances” at professionally fired firework displays and distances derived from ShellCalc© Dr Tom Smitha and Dr Christian Lohrerb

CarnDu Ltd, 8 Aragon Place, Kimbolton, Huntingdon, Cambs UK PE28 0JD. Email: tom@carndu.com Federal Institute for Materials Research and Testing (BAM), Unter den Eichen 87, 12205 Berlin, Germany Email: christian.lohrer@bam.de a

Abstract: There is a wide variety of approaches across the world in determining appropriate “safety distances” for firework displays. Comparison of the different national approaches and distances for shells derived from ShellCalc© highlights the variety and derivation of the “safety distances” adopted and these distances are related to the various failure modes of shells that may affect the audience. It is not intended that this paper should encourage the maximum distances derived always to be adopted, but that an appreciation of the probabilities and hazards of various accident scenarios and therefore the risks involved be part of the decision making process for designers and commissioners of displays as well as enforcing authorities.

Introduction The adoption of diverse “Safety Distances” for professionally fired firework displays around the world illustrates the different appreciation of what a “safety” distance really means. To some it is an inviolable distance beyond which people are 100% “safe” – that is, not subject to any risk, whereas to others it is no more than a recommended distance based on history, custom and practice which bears little relationship to the risks posed. Far too often we hear after a display “we got away with it”. This generally means that there was no major incident or misfiring and that the debris from the show fell into an area which had been loosely described as the “fallout” area. How this actual area was determined in practice in relation to the site and conditions prevailing at the specific display is often little more than arbitrary. This paper attempts to provide some quantification of the distances that may be rationally applied, and to examine the risks posed and to compare these mathematically derived risk minimisation approaches based on shell trajectory modelling distances – using the ShellCalc© program1 developed by John Harradine and extended by Tom Smith – to the “safety distance” regimes from a number of countries. A recent paper by Lohrer2 has already made comparisons of the “safety” distances applied to a wide variety of fireworks across a number of European member states. This paper concentrates on what we consider to be the highest hazard items (shells) given that shells are usually the determinant of the overall site suitability and layout. However for smaller shows or shows on restricted sites other firework types may become the determinants of appropriate “safety distances” and the same general principles outlined here may also be applied to such fireworks. ShellCalc© now models a variety

of firework types including •

shells

•

Roman candles

•

mines

•

fountains (gerbs)

and it is apparent that on some sites, especially with Roman candles or mines fired at low trajectories, these may pose greater risks than smaller calibre shells fired vertically at the same display. This paper highlights the effect of firing angles and wind strength/direction on shell trajectories and derived “safety distances”. As Lohrer points out, there is a great diversity of approaches and therefore derived distances throughout Europe, with the UK adopting an unusual approach in that “safety distances” are calculated on the basis of site and product specific risk assessment rather than using “fixed” distances related to, for example, calibre. It is important to appreciate that risk is, in simple terms, the product of the frequency (likelihood) of an event and the hazard of that event – i.e. the consequences. Events can be high risk because they have a high frequency or high hazard, or both. We have sought advice from practitioners in the various countries commented on and acknowledge their help and expertise. However any misinterpretations are entirely our own.

What is a “safety” distance anyway? The main issue we encounter when discussing “safety” distances is what people actually mean by the term. Is a safety distance a distance at which people are “safe” – that

Article Details

Article No:- 108

Manuscript Received:- 18/09/2014

Final Revisions:- 10/11/2014

Publication Date:-

Archive Reference:- 1728

Journal of Pyrotechnics, Issue 33, 2014

Page 53

is, at no risk, or is it a distance at which people are at an acceptably low risk? The difference is important and has implications not only for event design, but for litigation following any incident. We do not live in a risk free world, and although obviously anyone attending a firework display should expect not to be injured by the fireworks from the display, accidents will continue to occur because of the nature of the products that are used. In essence fireworks are relatively cheap items with a very low (but quantifiable) failure rate but are items that cannot be subject to individual testing. By necessity fireworks may (and should) be subject to both a “Type” approval and “Batch” approval regime – it simply is impossible to test every item, as testing it necessarily causes it to function irreversibly! The Type/Batch approach is that adopted in essence by the European Standards3 for professional display fireworks, although its history is in similar regimes developed in the UK for the British Standard 71144 for consumer fireworks. The major difference between Category 1/2/3 Standards (for consumer fireworks) and Category 4 (for professional fireworks) is that professionals may manipulate and fire fireworks in creative ways using their professional skills and training to determine both the suitability of a particular firework for a display (at a particular site under a particular set of conditions) and the manner in which it is fired. The latter is the biggest contrast between consumer fireworks (fired by non-professionals using equipment provided and according to fixed instructions) and professional ones and may involve consideration of: •

Fusing methods

•

Firing positions

•

Firing angles

•

Mortar construction (e.g. for shells)

•

Mortar rack construction (e.g. for shells)

•

Supporting methods

•

Risk reduction methods

However, it is also clear that despite their greater knowledge professional display operators sometimes find themselves in the unenviable and unjustified position of having to apply greater “safety” distances than non-professionals for what are, in performance terms at least, identical items. Lohrer states that in Germany, for example, the distances are derived on the basis of hazard – i.e. that an assumption is made that the item will fail. However this paper demonstrates that these distances do not, in fact, represent such an approach – for example ignoring any ground burst of a “blind” shell. Other European countries with significantly lower prescribed “safety” distances must therefore be applying a risk based approach (albeit potentially unwittingly) rating high frequency events (such as “normal” fallout) greater than low frequency events (such as blind shells). Page 54

For simplicity this paper will concentrate on aspects of the firing of a small variety of shells, but the principles can be extended across a much larger range of fireworks. For instance mines, Roman candles or single shot devices fired from elevated structures at angles far from vertically upwards may actually be the biggest determinant of appropriate “safety” distances in shows with limited use of shells. In this paper we will also concentrate on the effect on members of the audience and the “safety distance” that is appropriate to them. However in most cases there are other potential areas affected by the display: •

The firers – these may be at a considerably higher level of risk but this may be an acceptable position because ◦◦ They are at work and cognisant of the risks involved ◦◦ There are only a limited number of firers at risk, as opposed to large numbers of audience

•

Structures – particularly where the display is fired from a structure itself, or is in close proximity to other structures (e.g. rooftop firing)

•

Other hazards – e.g. car parks, power lines etc

Mortar angles and the effects of wind It is commonplace to find shells fired from angled mortars. Modern display design often uses angled firing to maximise the spread of shell bursts in the sky and to create patterns from “tailed” shells and to attempt to fire shells “away” from the audience. Firing angles for aesthetic reasons typically range up to 30° from the vertical and are generally angled pto create teh greatest spread for the majority of the audeince. Firing angles for safety reasons are usually away from the audience for obvious reasons. The problems come when the audience is not simply on one side of the firing site, or when the conditions prevailing at the time of the display take debris from the display towards the audience. A previous paper attempted to quantify the risks from firing shells and noted that if an audience subtends 360° around the firing area then the overall risk of a “blind” shell or debris falling on the audience is necessarily increased from the situation where the audience only subtends, say, 36° – i.e. by a factor of 10! A maximum wind speed of 20 km h−1 was chosen for this study because in our experience this wind speed can be considered as that under which most shows in Europe will be able to be fired (or should be able to be fired) without significant modification. It simply is not sensible to plan a display assuming the wind speed will be significantly lower than this (except where local conditions indicate that such conditions are the norm). On the other hand, when the wind speed is above 20 km h−1 then we would expect that significant portions of the display would have to be removed if the wind is in a direction towards the audience or Journal of Pyrotechnics, Issue 33, 2014

other hazards, simply because the various fallout distances increase significantly. ShellCalc© is a useful tool therefore, for companies to produce curtailment or cancellation criteria based on their own range of fireworks and their own display designs. Some “fixed rule” approaches extend the wind strength considerably above 20 km h−1 but it is our belief that at such speeds and in directions which are specific to each site it is likely that removal of certain fireworks or types may be the most appropriate response.

•

The operators

•

Local hazards (e.g. car parks)

•

The audience

•

Other people who may not be the “intended” audience

It is extremely difficult to compare the “fixed rules” approaches under a variety of conditions of firing angle and wind conditions as each system applies its own criteria. We have attempted to make comparisons between the systems below.

•

Low hazard/high frequency events – such as lightweight debris or sparks carried by the wind and which could cause minor injuries and which will be generated at almost every show (i.e. the frequency is very high)

•

High hazard/low frequency events – such as a shell falling directly into the crowd and potentially functioning on impact

As noted above this paper concentrates on the audience, but the principles apply to all. The two main sources of risk from shells at a display may be considered as

Vertical mortars falling over One failing of many systems is the assumption that mortars will fire in their design orientation – and determination of the relevant “safety” distances as a result. Even the most pessimistic regime does not appear to consider the low frequency/high consequence failure in which one shell displaces an adjacent mortar from which a second shell is then fired. It is simply not realistic to apply a reduced set of distances for mortars which “cannot fall over” unless extreme measures have been put in place to completely remove such a risk, or to safeguard the audience completely (by use of, for instance, “catchers”) in the event that they do. For a complete assessment of risks this situation must be considered. It may be rated low risk (because of the very low frequency) but it is still a risk. Figure 1 shows the ShellCalc© plot for a 100 mm shell fired in 21 km h−1 wind. This paper therefore also considers the risk from such events.

Failure modes and types of incident In any process of risk assessment the various failure modes should be assessed and rated for their effects on all potential hazard areas. These may include, for example: •

The structure from which the fireworks are fired or which may be affected by impact (e.g. a building)

These risks could be regarded as equally important, and equally in need of risk reduction measures. Obviously a high hazard/high frequency event poses an unacceptable risk and should not be continued or contemplated until sufficiently robust risk reduction measures have been put into place and the risks reassessed. Hence there are a variety of incident types which should be considered as part of an overall assessment of risk. In approximate order of increasing hazard (and decreasing frequency) to the audience these are shown in Table 1. A previous paper by Smith5 has attempted to quantify these risks and concluded that the overall risk to the audience or operators at any display is very low. However it is still a quantifiable risk and it is obvious that incidents and accidents do still occur. The job of the display company is to minimise these risks by minimising the frequency or the potential hazard (or both) and it is the job of event organisers to guide the display company in determining what risks are acceptable.

Types of approach In the first instance therefore we must decide which are the relevant hazards to the audience, and which will determine the appropriate distance to them or limit the variety of fireworks used. This generally should be a decision made

Table 1. Hazards from shells Hazard

Comments

Long duration sparks from normally fired shell (e.g. Kamuro)

2 3 4 5

Long duration sparks from unintentionally angled mortar “Normal” debris from normally fired shell “Normal” debris from unintentionally angled mortar Shell “blind” from normally fired shell

6 7 8

Shell “blind” from unintentionally angled mortar Shell “blind” and ground burst from normally fired shell Shell “blind” and ground burst from unintentionally angled mortar

Lightweight sparks that will not cause major injury but may be a source of ignition E.g. if mortar is disrupted by previous shell misfire Shell fragments E.g. if mortar is disrupted by previous shell misfire Except in high wind conditions or from angled mortar will most likely not reach audience E.g. if mortar is disrupted by previous shell misfire

Journal of Pyrotechnics, Issue 33, 2014

E.g. if mortar is disrupted by previous shell misfire

Page 55

Figure 1. Shellcalc© plot showing Normal, Long duration, Blind and Blind+burst for vertical and 45° launching of a 100 mm shell in 21 km h−1 wind. in partnership between the event organiser and the display company •

•

Is the display to pose a low (but generally acceptable) risk either from low frequency/high hazard events (such as shell “blinds”) or high frequency/low hazard events (such as long duration stars)? Is the display to pose NO or at least very near-zero risk? This may be required for a very large televised event such as an Olympic opening ceremony, where the consequences of even a very low risk incident could be significant, not least in terms of adverse publicity from the world’s watching media.

In addition there are two basic approaches to determining “safety distances”. The first is what we will term a “fixed rules” approach – it is based on some rules which may include multipliers of shell diameter (for convenience) or shell apogee (which is a better measure) and consists of a set of values or tables which can be applied by the display designer and the enforcing authorities to provide “standard” distances which, although they may not truly be “safe”, provide an acceptable level of risk. The second is based on specific modelling of shell trajectories and fallout patterns on a shell-by-shell basis taking into account firing angles and wind strengths and directions.

How do you determine a safety distance? The determination of an appropriate “safety distance” Page 56

depends on many factors. The problem with most “fixed rules” systems is that they do not address all of these factors adequately. Features which may affect what is an appropriate “safety distance” include: •

Shell calibres

•

Shell types

•

Firing angles

•

Position of the audience in relation to the mortars (i.e. are the audience areas restricted)

In addition a number or features related for the rigging and firing may need to be considered: •

Mortar construction and possible failures

•

Effect on other fireworks (e.g. rack construction)

•

Mortar support methods (and “fail-safe” issues)

Finally the environment and meteorology at the time of firing (which may be quite different from that during site surveys or even during rigging): •

Wind speed

•

Wind direction (especially in relation to firing angles)

•

Firing elevation

•

Topography

In addition it is apparent that the basis of many “fixed rule” systems differ – presumably because of the history, custom and practice prevailing within each country and probably Journal of Pyrotechnics, Issue 33, 2014

Table 2. Comparison of systems US Canada Australia

Name

Basis

NFPA 11236 Display Fireworks Manual7 Safe use of outdoor fireworks in Western Australia8

Fixed distance per shell diameter Fixed distances dependent on shell diameter, site layout and firing angles Fixed distances up to 300 mm shells Vulnerable sites require 2× based on dud shells landing on the ground distance

France

Comments

Fixed distance per shell diameter or apogee and effect

Differentiation between normal shell and “report shell” (aural main effect) Fixed distance per shell diameter or Differentiation between normal apogee and effect. Marking of shell burst shell and “report shell” (aural height on the shell label is mandatory in main effect) Germany and usually is the predominant determinant of distance There are no fixed distances for professional users – distances are determined by companies on the basis of their own site and product risk assessments

Germany

as a result of recommendations made after investigations of accidents. In general the approaches seem to be based on either the expected “normal” debris arising from the “normal” firing of shells, or from a calculation of the maximum range a shell could achieve if it failed to burst at the design height (i.e. a “blind”). The reasons for this diversity are not important in the discussions that follow, but suffice to say we believe neither represents all the risks that arise from both the “normal” firing of shells at a variety of angles and in varying meteorological conditions and in particular the risks arising from the abnormal functioning of shells (for instance where a shell is fired at an unplanned trajectory because of disruption of the mortar from a previous shell failure).

Table 2 outlines the basic methodology of the various systems in use throughout the world. Table 3 outlines the basic methodology for dealing with firing angles and wind speed/direction of these systems.

Distances Table 4 highlights the derived distances for a variety of scenarios across the various countries examined for a variety of similar situations. In some cases various assumptions have been made which are expanded below. The ShellCalc© distances are shown as those calculated for the intended firing angle in the relevant wind, and for a displaced mortar angled at 45° with a “tailwind” – i.e. the greatest possible combination of effects.

Table 3. Wind and angled firing Angled shells US

Canada

Australia

France Germany

Offset changes position of mortars within secured diameter (which does not change) Does not appear to have fixed rules

Table considers various launch angles up to 45°and increases distances based on dud shells Does not appear to have fixed rules Distances increased depending on firing angles (distances may be decreased in opposite direction)

Journal of Pyrotechnics, Issue 33, 2014

Wind

Comments

Fixed increase dependent on wind strength or reduction in maximum shell calibre or angling mortars into wind

Canada has two types of site defined “Oblong” and “Circular” but these extremes are not really representative of modern display scenarios We presume the minimum distances apply unless the “shell drift” distance is greater

Operators should consider the effect of wind to increase flight distances – table illustrates values for vertically fired shells Does not appear to have fixed rules Distances increased depending on wind strength in direction of wind(distances may be decreased in opposite direction)

Page 57

Table 4. Derived distances “Safety” distances (m) Shell calibre

Firing angle

Wind speed

(mm)

(°from vertical)

(km h−1)

75 100 125 150 200 75 100 125 150 200 75 100 125 150 200 75 100 125 150 200 75 100 125 150 200 75 100 125 150 200

0 0 0 0 0 10 10 10 10 10 20 20 20 20 20 0 0 0 0 0 0 0 0 0 0 20 20 20 20 20

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 10 10 10 10 10 20 20 20 20 20 20 20 20 20 20

USA

AUS

Blind

Blind + Burst

Fallout

LD fallout

Blind

Blind + Burst

Fallout

LD Fallout

64 85 107 128 171 64 85 107 128 171 64 85 107 128 171 64 85 107 128 171 64 85 107 128 171 64 85 107 128 171

95 115 145 175 230 95 115 145 175 230 95 115 145 175 230 95 115 145 175 230 125 145 175 205 260 125 145 175 205 260

45 75 100 150 200 64 79 100 150 200 119 145 169 191 230 45 75 100 150 200 59 75 100 150 200 179 210 236 261 302

60 80 100 120 160 60 80 100 120 160 60 80 100 120 160 60 80 100 120 160 60 80 100 120 160 60 80 100 120 160

60 80 100 120 160 84 112 140 168 224 108 144 180 216 288 60 80 100 120 160 60 80 100 120 160 108 144 180 216 288

31 40 46 49 59 96 126 146 161 194 149 196 227 252 304 52 66 74 81 93 76 95 105 113 126 197 254 290 321 377

69 90 109 124 159 134 176 209 236 294 187 246 290 327 404 90 116 137 156 193 114 145 168 188 226 235 304 353 396 477

16 21 24 27 33 62 80 92 94 113 99 128 147 157 187 69 87 98 107 124 121 152 171 187 215 192 244 277 305 357

54 71 87 102 133 103 134 160 169 213 140 182 215 232 287 141 181 210 237 288 228 291 334 373 444 293 374 430 480 572

214 283 327 370 442 214 283 327 370 442 214 283 327 370 442 234 309 356 400 473 256 334 384 427 505 256 334 384 427 505

252 333 390 445 542 252 333 390 445 542 252 333 390 445 542 272 359 419 475 573 294 384 447 502 605 294 384 447 502 605

167 216 250 280 335 167 216 250 280 335 167 216 250 280 335 205 263 301 335 396 244 310 353 391 458 244 310 353 391 458

204 266 313 355 435 204 266 313 355 435 204 266 313 355 435 264 339 394 442 533 324 414 476 531 633 324 414 476 531 633

Blind = shell fails to burst and lands on ground. Blind + Burst = shell fails to burst in air and ignites on impact with ground, diameter of shell burst as designed. Fallout = “normal” fallout – shell fragments etc. LD Fallout – long duration stars – e.g. Kamuro. CA = Canada, AUS = Australia, FR = France, DE = Germany. Shellcalc© calculations done with “typical” mortar barrelling/tumbling enabled. Complex (i.e. wind and angle) distances are taken as cumulative – i.e. angled mortars in wind need both terms applied.

Given the various systems and the relationship between metric and imperial measurements we have attempted to make “like for like” comparisons but in some cases the derived distances may vary slightly. The following graphs illustrate the differences between the different approaches. Figure 2 shows plots for vertically fired mortars with Page 58

zero wind. In this situation the distances calculated using ShellCalc© are generally less than the “fixed rule” approaches. This is to be expected, but zero wind situations are not realistic. Obviously the 45° ShellCalc© values are greater than both the undisturbed mortar values from ShellCalc© and the “fixed rule” distances – but they represent, for example, “worst case” failures of mortars in racks. Journal of Pyrotechnics, Issue 33, 2014

Figure 2. Mortars vertical, wind 0 km h−1.

Figure 3. Mortars vertical, wind 20 km h−1.

Journal of Pyrotechnics, Issue 33, 2014

Page 59

Figure 4. Mortars 20 degrees, wind 20 km h−1. Key to Figures 2, 3 and 4

SC Blind = “blind shell” SC Fallout = “normal” debris SC Blind + Burst = “blind” shell bursting on impact SC LD = Lightweight (or long duration) debris and sparks SC 45 = figures if shell discharged at 45°.

The low ShellCalc© values for vertically fired mortars are not surprising – in a truly zero wind situation a shell fired vertically which failed to burst would, theoretically, land back in the mortar! ShellCalc© factors in shell tumbling and mortar barrelling to more accurately reflect this situation. Figure 3 shows plots for vertically fired mortars with 20 km h−1 wind. In this situation the ShellCalc© distances approximate to the Canadian distances. This is to be expected, since the Canadian “fixed rule” approach pays more attention to wind speed (at 20 km h−1) than other systems. Again the 45° ShellCalc© values are greater than both the undisturbed mortar values from ShellCalc© and the “fixed rule” distances Figure 4 shows plots for mortars angled at 20° to the vertical Page 60

with a 20 km h−1 “tailwind” – that is, in the same direction. Here both the “normal” (i.e. 20° angled) and 45° angled ShellCalc© values exceed any of the “fixed rule” approaches and although the German system, which takes into account both factors, now exceeds the other “fixed rule” approaches it still does not approach the ShellCalc© values.

Assumptions made and observations of the various systems We have had to make some assumptions in our analysis of systems in other countries and we are grateful to colleagues in those countries for correcting us where necessary. US

Journal of Pyrotechnics, Issue 33, 2014

•

The distances are not increased for varying wind strengths

•

1/3 offset distances for angled shells are not applied to angled firing for aesthetic effect

•

It is acknowledged that the figures are a compromise between enforcers and industry and not based on tests or modelling

•

Does not deal with fireworks angled in fan shapes. It seems confusing what distances should be applied using the 1/3 offset principles – are these “correspondingly increased” in the direction of the angle for all angles?

•

No quantification of “angle”

•

No apparent increased distances for wind

Australia •

The maximum of the “minimum separation distance” and the shell “dud” distances are used where mortars are angled in varying wind strengths

Canada •

Does not address angled fireworks adequately

•

Proposes that on circular sites shells are fired into the wind (even if this is towards the audience) – in the case of a dud shell this could be a high hazard event

•

Proposes reducing shell size in windy conditions – but the proposals are arbitrary (alternatively increases distance by fixed amount)

France •

Considers malfunction of the shell burst only (i.e. a “blind” shell) but ignores crucial aspects of shell/mortar failure (see below)

•

Does not consider mortar angles or wind

Germany •

Considers malfunction of the shell burst only (i.e. a “blind” shell) but ignores crucial aspects of shell/mortar failure (see below)

General points In general “fixed rule” systems do not attempt to reflect different performance parameters of shells, nor their construction. The German and French systems allow for calculations to be made on the basis of shell apogee as well as shell diameter and in Germany it is mandatory to mark this value (the “shell burst height), on the shell itself. In practice therefore this is a better determinant of “safety distance” than suggested by the table above. We believe that the approach of the European Standard for shells (part of the Category 4 standard3) which will label shells with their performance characteristics including •

Burst height (mandatory)

•

Burst diameter (optional)

Journal of Pyrotechnics, Issue 33, 2014

under “standard” conditions will go a long way to providing relevant information to users to enable them to calculate appropriate distances at a display. However it is recognised that such “Standard” conditions may not reflect the practices used by individual display companies – for instance mortar lengths and inside diameters may vary – in realistic conditions on display sites. However the “Standard” information provided will allow display designers a reference point in their calculations using a variety of methods. Future versions of ShellCalc© will enable such reference points to be added to the input criteria for trajectory modelling, together with parameters to reflect each company’s individual experience.

Comparison of the German system and distances derived from ShellCalc© Following Lohrer’s paper on comparison of European Safety distances and the conclusion that the Dutch and German systems generally set the greatest “safety distances”, the remainder of this paper compares the German distances with those derived from ShellCalc© for a variety of scenarios. The comparison of distances with the German system, based on “100%” failure rate, are interesting. We do not believe that the German system actually represents the worst case scenarios as outlined above, for two reasons: 1. For “blind” shells it fails to allow for a shell bursting upon impact with the ground 2. It fails to allow for accidental disruption of a mortar by an adjacent shell failure It is not alone in these two shortcomings, but we do not believe that the system can truly be described as representing the worst case scenarios or to assume 100% failure. The two points above are both extremely rare – but they still are realistic possible failure modes which have caused accidents in the past.9,10 The German system also draws a distinction between colour shells (e.g. peony) and report shells and in general increases the “safety distance” for the latter. However we do not believe that this accurately reflects the similarities in hazards from the two types of shell in a “blind” situation. ShellCalc© has been used extensively at some of the largest displays in the UK and the rest of the world to model fallout and shell failures under a wide variety of wind conditions (ShellCalc© allows wind strength and relative direction to be set when, for instance, the mortar orientation is fixed due to design or site constraints) and to develop objective cancellation or curtailment criteria at these shows. Some examples are given in Tom Smith’s book Firework Displays: Explosive Entertainment.11

Responsibilities at a display The discussion necessarily highlights the roles and responsibilities of the various parties involved. We believe that the increasing threat of litigation and the public perception of risk have altered the relationship between Page 61

Table 5. Illustrative roles Organisation The display company (contractor)

Role and responsibilities To provide a spectacular, appropriate and low risk display.

The event organiser (creative design or production)

To carry out site and product specific risk assessment to determine which fireworks are appropriate for the site and expected conditions and to provide objective cancellation or curtailment criteria (this may be done in conjunction with other bodies or consultants) To provide clear instructions as to level of acceptable risk (or no risk) depending on venue and event

The enforcing or licensing authority

To facilitate the safe rigging up, firing and de-rigging of the event by, for example, providing sterile areas and maintaining the fallout area clear from people To understand the requirements of the event organiser and the limitations and opportunities the site may provide

Consultants

To enforce consistently To act, perhaps, as “broker” between event organisers and contractors to ensure realist expectations are met and low risk as appropriate To prepare an objective set of cancellation/curtailment criteria

these organisations in a positive way. However we believe it is important to emphasise the role each must take and illustrations of such roles are given in Table 5.

Risk assessment aspects The above discussions highlight, we hope, the merits of the UK approach to “safety distances” at displays. We believe a site and product specific risk assessment, leading to objective criteria for cancellation or curtailment of a display, is the best way to ensure both a low (or near-zero) risk display and realistic expectations for the event organisers and, perhaps, the media without the pressure on the display company to carry on firing in a “the show must go on” approach. “Fixed rule” regimes undoubtedly have the benefit of simplicity, but we do not believe that they always highlight all of the risks involved, be they high hazard/low frequency or low hazard/high frequency. Of course, the obvious question that arises from this is “What is an acceptable level of risk?”. This depends on several factors and it would be presumptuous to dictate what levels of risk are acceptable in different countries and at different events. However the following should be considered by event organisers and enforcing authorities, and hence by display designers in determining appropriate fireworks for the specific event and the curtailment/cancellation criteria that are appropriate: •

The risk of fatalities (e.g. from shell failures)

•

The risk of major injuries (e.g. from normal fallout)

•

The risk of minor injuries (e.g. from long burning debris)

•

The type of event, the location of the audience and, perhaps, the media coverage

•

The scale of the event (e.g. how many shells of various calibres, fired at various angles may contribute to the risks)

•

The flexibility (or not) of the site to maximise the “safety” distances

Page 62

•

The types and sizes of fireworks from an artistic and practical point of view.

What we are not advocating It is important to appreciate that a risk based determination of “safety distances” does not mean that the greatest possible calculated distance from trajectory and fallout modelling must be used in deciding where an audience should be positioned (or more likely what maximum shell calibre and firing angles are appropriate). A risk based approach seeks to quantify the risks and to accept that the risks are reduced to acceptable levels – but recognises that they are not eliminated. For some shows it may be necessary to reduce the risks to as near zero as possible, but this is a fundamental decision which must be taken early by the event organisers and the contractors. It is not acceptable to make late changes under pressure to perform. Sensible and systematic assessment of the risks from a display must be done at the planning stages, and ideally contingency planning built into the display so that clear objective criteria may be developed and “signed off” to curtail or even cancel a display. For small events simple distance tables may be appropriate (but necessarily pessimistic) but for the largest displays, where the planning timescales and the nature of the event allow proper risk assessments to determine the distances to address both high frequency/low hazard events and low frequency/high hazard events, a more rigorous approach is both necessary and justified. It is neither practical nor sensible to require display operators to be running ShellCalc© or similar programs at the display site just as the show is about to be fired. Planning is necessarily before the display even starts and ideally all parties are “signed up” to what curtailment or cancellation criteria are appropriate for the site and the display in question. However we recognise that for smaller Journal of Pyrotechnics, Issue 33, 2014

Figure 4. Simple spreadsheet to highlight “safety distances” on site. displays, where there is not the requirement for such indepth planning, the display operators could use a simple tool on site given the ubiquity of smartphones and tablets. Figure 4 shows a simple Excel© spreadsheet developed by the authors for such a purpose. It takes values from trials and ShellCalc© and allows user-input of a single parameter – the “Safety Distance”. It then colours (by conditional formatting) those items whose distances exceed the input value.

Christian Lohrer, JPyro, Issue 32, 2014, pp. 38–51.

EN 16261:2013 Standard series for Pyrotechnic articles – Fireworks, Category 4, consisting of four parts, CEN/TC 212 WG2, 2013.

BS7714 – available from BSI Publications.

Tom Smith, JPyro, Issue 29, 2010, pp. 12–31 (http:// www.jpyro.com/wp/?p=1164)

NFPA 1123 is available from http://www.nfpa. org/codes-and-standards/document-informationpages?mode=code&code=1123

Display Fireworks Manual 2010 available from http:// www.nrcan.gc.ca/explosives/fireworks/9903

Safe use of outdoor fireworks in Western Australia is available from http://www.dmp.wa.gov.au/documents/ Code_of_Practice/DGS_COP_SafeOutdoorFireworks. pdf

See for example http://www.huffingtonpost. com/2013/07/08/simi-valley-fireworksaccident_n_3561811.html

Conclusions The variety of systems in use throughout the world inevitably is led by varying historic custom and practice as well as revision post-incident investigation. In general we have identified that there are low frequency/ high hazard events which exceed the distances in most countries, sometimes significantly. However we believe that a proper appreciation of the potential risks, and a sound relationship between event organisers, enforcing authorities and the display companies, do not mean that the maximum distances have always to be applied.

References 1

John Harradine and Tom Smith, JPyro, Issue 22, 2005, pp. 9–15 (http://www.jpyro.com/wp/?p=23) and Tom Smith, JPyro, Issue 32, 2013, pp. 3–21 (http://www.jpyro.com/wp/?p=1545). See also http:// www.shellcalc.co.uk and http://www.facebook.com/ shellcalc

Journal of Pyrotechnics, Issue 33, 2014

10 See for example http://www.eig.org.uk/eig2007/wpcontent/uploads/tom_CBI%20presentation.pdf 11 T. Smith, Firework Displays: Explosive Entertainment, ISBN: 0820600903; see http://www. fd-ee.com

Page 63

Events and Sponsors Now that the Journal of Pyrotecnics is published, primarily, electronically the forthcoming events are available on the archive website http://archives.jpyro.com. If you have any events you would like included in that list please contact the publisher.

Please note An up-to-date list of current sponsors on the JPyro archive website http://archive.jpyro.com. Sponsors of the Journal get unlimted electronic access to the issues which they have sponsored, details online and in the hard copy and a discounted rate for hard copy editions. Please contact the publisher for more details. Information The Journal of Pyrotechnics Facebook page (http://www.facebook.com/jpyroarchive) will announce new articles and also provide space for discussions.

Caution The experimentation with, and the use of, pyrotechnic materials can be dangerous and may require licences or permits in certain countries; it is felt to be important for the reader to be duly cautioned. Without the proper training and experience no one should ever experiment with or use pyrotechnic materials. Also, the amount of information presented in this Journal is not a substitute for necessary training and experience, nor does it remove the relevant application of national or local laws and regulations. A major effort has been undertaken to review all articles for correctness. However it is possible that errors remain. It is the responsibility of the reader to verify any information herein before applying that information in situations where death, injury or property damage could result.

Page 64

Journal of Pyrotechnics, Issue 33, 2014

Quantification of Visible Aerosols from Pyrotechnics: The Effect of Relative Humidity Rene Yo Abe,a Yoshiaki Akutsu,a Akihiro Shimadab and Takehiro Matsunaga b

Abstract: Light transmission and particle size distribution measurements of aerosol (smoke) emissions from pyrotechnic compositions have been performed in a combustion chamber at atmospheric pressure over a wide range of relative humidity. Maximum light extinction over time is proportional to weight of incinerated sample and shows a characteristic curve over relative humidity for each pyrolant. When critical relative humidity is exceeded, a distinct increase in light extinction is observed, which can mostly be attributed to enhancement of the light-scattering efficiencies of submicrometer particles by hygroscopic growth thereof. The proposed measurement method allows humidity characteristics of pyrotechnic smoke emissions to be compared.

1. Introduction Through interaction with humidity in ambient air, combustion products of pyrotechnics can produce dense smoke-plumes, i.e. aerosols. For fireworks, reduction of adverse effect on visibility is of particular interest, because ongoing displays can be obscured under high humidity and low wind conditions. Even supposedly smoke-free compositions like ammonium perchlorate (AP)/ hydroxyl-terminated polybutadiene (HTPB), which is used as propellant in some rocket motors, develop visually obscuring aerosols from the gaseous combustion products under high humidity conditions.1 Though the production of visible aerosol is intended in some applications (e.g. smoke screens or signaling),2,3 it is undesirable in firework scenes especially if increased by humidity. Especially in the Asian region, very high humidity conditions tend to occur often during the warm season which is when firework festivals are mostly held. Solid hygroscopic particles in aerosols can grow by condensation of water to form aqueous droplets when critical relative humidity is exceeded. The effect of relative humidity on particle growth is well known for a number of substances.4,5 HCl emissions from chlorine based oxidants like perchlorates, which are commonly used in pyrotechnics, also cause mists to be produced in high humidity or low temperature conditions.1,6 In the presence of HCl vapor, condensation on hygroscopic particles is further promoted to produce even denser smoke clouds. Light scattering by aerosol particles reduces contrast and brilliance, i.e. visibility of pyrotechnics. Attenuation of straight light transmittance through aerosol clouds and addition of scattered light to an otherwise generally dark background both degrade the visual impression of

pyrotechnics. Most conceivable combustion products except soot can be described as transparent particles with refractive indices of about 1.33 to 1.5. For such particles, scattering efficiency increases greatly for visible light (λ ≈ 0.5 µm) when diameters grow to around 1 µm.7,8 Particle size and growth characteristics of therefore make up important factors in describing the optical depth of aerosol emissions. Although efforts are made to develop compositions and substances that produce reduced visible aerosol (smokeless) emissions,1,2,9,10 the effect of relative humidity has not yet adequately been characterized. Studies on aerosol emissions from pyrotechnics based on field measurements11–13 at outdoor pyrotechnic displays, where relative humidity cannot adequately be adjusted, or on confined volume experiments,14,15 mostly do not include experiments at high humidity conditions. Particle sizes reported in the range of a few hundred nm can however increase significantly by uptake of water vapor from humid ambient air. Visible aerosol development after combustion of pyrotechnics in ambient air of arbitrary relative humidity can be simulated in chamber experiments. Gas and particulate emissions are quickly diluted and cooled in air and undergo nucleation, coagulation and condensation processes. In this work, a measurement method using a small-scale combustion-chamber is proposed for quantification of visual obstruction by aerosols. This allows comparison of visible aerosol development after combustion of different types of pyrotechnic compositions over a range of controlled humidity conditions. Humidity characteristics of the following three types of pyrotechnic compositions were measured to illustrate typical differences in aerosol visibility: ammonium and potassium perchlorate (AP, KP) based compositions which are widely used as a basis for

Article Details

Article No:- 109

Manuscript Received:- 07/07/2014

Final Revisions:- 07/12/2014

Publication Date:- 07/12/2014

Archive Reference:- 1735

Journal of Pyrotechnics, Issue 33, 2014

Page 65

pyrotechnic compositions;16 black powder (BP) which is still the predominant composition used as lifting charge, although it is well known for producing relatively dense aerosols.

2. Experimental 2.1 Pyrotechnic compositions As a representation of pyrotechnic compositions used in fireworks, the perchlorate based composite formulations presented in Table 1 were used. Application of ammonium perchlorate (AP) and potassium perchlorate (KP) mixed with hydroxyl-terminated polybutadiene (HTPB) binder enabled easy laboratory scale production. Further, grain black powder (75% potassium nitrate, 15% charcoal, 10% Sulfur; Kayaku Japan Co., Ltd) was used as a representative chlorine free pyrotechnic composition. Combustion times varied from instantaneous combustion for black powder up to ten seconds for KP composites. The temperature inside the combustion chamber increased less than 1.0 K per Table 1. Formulations of composites based on perchlorates Symbol A K

Oxidant AP 82%

KP 84%

Binder (fuel) HTPB + Curing agent 18% 16%

gram combusted sample after combustion, but returned to within 0.2 K above initial conditions in a few minutes due to radiative and convective heat transfer to the surrounding room. 2.2 Combustion chamber A chamber consisting of a cubic steel frame covered with acrylic sheets at the top and three sides, a stainless steel floor and a detachable soft PVC sheet at the front side was constructed as shown in Figure 1. Room temperature was controlled by air conditioning. After initial temperature and humidity conditions were applied with a humidity controllable type air-conditioning unit (Apiste PAU300SHC), the chamber was sealed. Combustion experiments were mainly performed at 20 °C, but measurements were also made at 30 °C and for BP also at 10 °C. The air in the chamber was mixed using an electric fan so that a sufficiently homogeneous aerosol was yielded within about ten seconds. Air temperature and relative humidity inside the chamber were measured near the circulation fan using a psychrometer (measurement of dry-bulb and wet-bulb temperature). This measurement method was easy to maintain and avoided any drifting effects which corrosive gases or aerosol particles emitted by most composites may cause in other humidity sensors. A continuous drip of distilled water kept the wet bulb free of contamination and the whole assembly could be cleaned quickly. Calculation formulae published by the World Meteorological Organization17 were used to obtain relative humidity values. 2.3 Aerosol opacity and particle size distribution To evaluate the opacity of the aerosol, light transmittance was monitored during combustion experiments with two laser sensors (Keyence LX-100, λ = 670 nm) installed in the middle of the chamber at distances of 77.5 and 10 cm between emitter and sensor, respectively. The sensor and emitter windows were protected from any aerosol depositions by a flow of clean air which had been passed through a HEPA filter by a membrane pump (RH <60% from outside of the chamber). Two grams of sample per cubic meter were found to be optimal for producing an adequate response in the scattering coefficient over the whole humidity range while not inflating the front chamber cover too much.

The visibility of an aerosol is caused by light scattering or absorption and is dependent on its particle size and number concentration. Combustion products found in aerosol particles can be considered mainly to consist of solid salts or oxides and aqueous solutions of salts and acids which have refractive indices in the range of nr = 1.33 to 1.5 (real part) for visible light (λ ≈ 0.5 µm). Such particles only show minimal light absorption (imaginary part of refractive index), unless soot particles or special dye substances (colored smoke compositions) are present, or they grow very large as in precipitating clouds. Single particle scattering efficiencies can be described by Mie scattering theory for Journal of Pyrotechnics, Issue 33, 2014

small spherical particles (0.03λ < d < 32λ) and scatter visible light (λ ≈ 0.5 µm) most efficiently with particle diameters between 0.3 and 2 µm.7,8 Particles which are much smaller than the wavelength of light (d < 0.03λ) are in the Rayleigh regime and absorb or scatter light only weakly. For larger particles, scattering efficiencies also decrease with number concentrations because mass concentrations are limited by condensable vapors. Particle size distributions (PSD) of aerosols generated by pyrotechnics were reported in previous research to be found in the range of several hundreds of nm.11,13,14 The visible increase of smoke development in high humidity conditions, however, suggests further growth of hygroscopic particles. Because aerosol concentrations quickly exceeded the measurable range of 106 particles per liter, PSD measurements using an optical particle counter (RION KC-11) were only possible with extremely small amounts of sample in the order of tens of mg. This made complete combustion difficult and changes in light-transmittance hardly measurable under conditions which would allow the simultaneous application of particle counting techniques as was previously also observed in large scale chamber experiments.15 To measure the PSD of highly concentrated aerosols generated in the chamber, a HORIBA LA-920 particle size distribution analyzer was modified to introduce the aerosol in a sheath flow into a flow cell for measurement of its Mie light scattering pattern. This allowed direct measurement of the aerosol’s PSD without prior dilution. Calculation parameters for Mie scattering calculations in particle size measurements with the particle size distribution analyzer were performed using the refractive indices of water and air at 20 °C (nr = 1.33 ni = 1.0 × 10−9). Simple analysis of scattering patterns with Mie scattering codes18,19 shows that scattering patterns change greatly with particle size (d = 0.1–10 µm), but not noticeably with refractive indices in the expected range (nr = 1.33–1.50). Therefore, deviation of actual refractive indices from those used for Mie-scattering pattern analysis does not largely affect calculated PSDs. Also, overall scattering efficiencies can be considered practically constant in the range of refractive indices expected.20

Table 2. Variables used in the scattering relation Symbol

Description

Light attenuation

Initial light beam intensity

Light intensity at the detector

Sample mass

Length of light path Light scattering coefficient of the aerosol (smoke) Maximum light scattering coefficient

b bmax b*max

coefficient values (b*; b*max) which are independent of the two parameters. Symbols used in this relation are summarized in Table 2. Note that mass here does not refer to aerosol mass (mass concentration of particles in the chamber) but to mass of incinerated sample. Because some products remain gaseous or quickly precipitate as solid ashes, while water vapor can be absorbed by aerosol particles, mass of aerosol generated by pyrotechnics is not easily predictable, whereas sample mass is a convenient reference. Only for scattering coefficients greater than 2 m−1, this relation shows a non-linear behavior (Figure 2) and massspecific scattering coefficients decrease. Processes like coagulation and deposition, which are dependent on aerosol concentration and particle size, reduce particle numbers and thereby overall light scattering. The linear region towards lower sample mass best represents real firework displays, where aerosols are quickly dispersed over a large volume. A series of similar measurements on smoke screen compositions has been performed by Harkoma.21 In his measurements, which were performed in a smaller chamber (0.15 m3), scattering (extinction) coefficients much more quickly decreased after combustion than observed in this study. This may be an indication of increased coagulation 4.5

3. Results and discussion

4.0 3.5

⎛I ⎞ A = log10 ⎜ 0 ⎟ = b ⋅ l = b * ⋅ m ⋅ l ⎝ I ⎠ (1) Based on equation (1) analogous to the Lambert–Beer law9 they can be converted to sample-mass specific scattering Journal of Pyrotechnics, Issue 33, 2014

bmax (m-1)

3.1 Light-transmission measurements Measurements of light transmission showed that light attenuation (A) and maximum scattering (attenuation) coefficient (b) occurring in an experiment (bmax) show linear dependences on transmission length (l) and sample mass (m), respectively, even at high humidity settings (90% RH).

Sample-mass specific scattering coefficient Maximum sample-mass specific scattering coefficient

3.0 2.5 2.0 1.5 1.0 0.5 0.0 0.0

1.0

2.0

3.0

4.0

msample (g)

Figure 2. Maximum scattering coefficient after combustion of composite-K at 90% RH plotted against sample mass. Page 67

Phase II coagula�on & deposi�on

1.2

1.0

0.8

0.6

0.4

0.2

0.0

Mode Diameter (μm)

b* (m2∙g-1)

1.2

Phase I par�cle growth

b (87%RH) *

b (23%RH) mode diameter (87%RH) mode diameter (23%RH)

0.0 0

Time (min)

Figure 3. b* and mode diameter of aerosol generated by combustion of 2 g composite-K at 87% RH (○) and 23% RH (□). and wall deposition effects. The prior is favored by higher aerosol concentrations or temperatures22 resulting from aerosols emitted by the same amount of sample (2 g) only being diluted to a smaller chamber volume (0.15 m3), and the latter by the chamber’s higher surface-to-volume ratio.23 Further, differences in type and resulting size distributions of the aerosols produced can also play a role. Despite its similarity to the Beer–Lambert law, b* gained by this relation displayed a time dependence caused by particle growth and deposition in the generated aerosol, especially at high relative humidity conditions. Composite-K shows typical characteristics as seen in time curves of b* in Figure 3 for 87% and 23% RH. After combustion under high humidity conditions, particle growth (mode diameter: ○) was observed for 4 to 6 minutes in phase I and reached a maximum after which b* declined slowly. Continuing growth in mode diameter of the aerosol observed in phase II suggested a coagulation processes, but increased losses due to particle deposition on wall and fan surfaces were also possible with larger particles. Under dry conditions up to 50% RH (23% in Figure 3), b* reached a nearly constant value immediately after initial fluctuations due to fan circulation. After slight growth (mode diameter: □) in phase I the measured PSDs also remain constant, indicating negligible coagulation and deposition for particle modes up to a diameter of 0.6 µm during the experiment. Increase of aerosol visibility of Composite-K at high RH can be attributed to condensation of water vapor and HCl onto KCl particles which leads to high number concentrations of particles with diameters around 1 µm, which have high scattering coefficients. Somewhat different results were obtained for the remaining two compositions: in experiments with composite-A, b* either reached its maximum even more quickly within 1 minute at humid conditions (above 80% RH) and declined within 5 to 10 minutes almost to complete transparency, or, Page 68

when dry conditions (below 80% RH) were applied, did not show any measurable light attenuation at all. This is because combustion products of composite A only consist of gases like CO2, H2O and HCl, of which the latter induces aerosol generation at high relative humidity. Although particle size could not be measured with the particle size distribution analyzer, because scattered light intensity was too low and PSD changed too quickly, visible particle depositions on chamber walls and floor were observed and hydrochloric acid could be captured in open containers containing a layer of water placed on the chamber floor. This indicates particle loss due to precipitation of large particles which cannot be sustained in a small agitated chamber. Black powder showed time curves of b* similar to those of composite-K, only reaching its maximum more quickly 2 to 3 minutes after combustion. 3.2 Particle-size measurements In the current series of combustion experiments, measured PSDs were mostly found to consist of a single mode. PSDs of aerosols generated by composite-K and black powder under humid conditions displayed a mode diameter shifting to larger sizes over time, indicating particle growth which may occur by coagulation of particles or absorption of water and other vapors. The absence of particle growth under dry conditions below 40% RH suggests that coagulation over a timescale of minutes can be neglected for small particles, but may play a role after particle growth to diameters in the near-µm range. In experiments with composite-A, particle size grew so quickly that the aerosol disappeared by deposition on chamber surfaces as indicated in the previous section and measurements with the Mie scattering PSD analyzer did not succeed due to low time resolution (≈30 s) and insufficient scattering intensity.

Journal of Pyrotechnics, Issue 33, 2014

Volume concentration (mm 3·m-3)

0.80 0.70

0.60

size class

0.50

total

0.40

<5.0µm

0.30

<2.0µm

0.20

<1.0µm <0.5µm

0.10 0.00

Figure 4. Particle volume distribution of aerosol 5 minutes after combustion of 2 g composite-K at 87% RH. In Figure 3, mode diameter is plotted along with b*max for composite-K at 87%. Mode diameter grew from 0.55 to around 1.05 µm within a quarter of an hour, passing 0.8 µm where b* reached its maximum. Figure 4 shows the PSD measured at that point, displaying its single mode diameter (maximum abundance). Only in a few experiments a second coarse mode of a larger particle size ranging from 4 to 9 µm could be observed because response with respect to aerosol mass was relatively low. Even larger particles are further quickly deposited before they can accumulate to sufficiently high concentrations necessary for measurement with this method. Particles larger than 4 µm proved difficult to measure using scattered light analysis, because aerosol-mass specific light scattering intensities rapidly decrease above that size (and are lost in the light scattering background of smaller particles). In total, these particles should therefore not play a major role in aerosol visibility. For detailed analysis of large particles in such aerosols, more advanced measurement methods are required which, without further countermeasures, however, may not withstand the corrosive combustion products over prolonged times. Supplementary measurements with an optical particle counter after combustion of very small amounts of sample as shown in Figure 5 confirmed the presence of particles exceeding diameters of 2 µm. Because the particle size resolution of the instrument used was low, the presence of multiple modes, particle size changes or precise volume concentrations, mass concentrations or diameters could not be determined with the instrument used. Volume concentrations in Figure 5 were estimated assuming spherical particles and constant number distributions in each size class. For Composite K, particles with diameters in the range of 2–5 µm were detected, but make up only a small fraction as number concentrations. Thus they only minimally contribute to light scattering and were not detected in the measurements using the PSD analyzer. Quick disappearance of large particles formed after combustion of composite-A Journal of Pyrotechnics, Issue 33, 2014

Volume concentration (mm 3·m-3)

0.80 0.70

0.60

size class

0.50

total

0.40

<5.0µm

0.30

<2.0µm

0.20

<1.0µm <0.5µm

0.10 0.00

Time (min)

Figure 5. Particle counter results converted to volume concentrations after combustion of 23 mg composite-K at 80% RH and 41 mg composite-A at 88% RH. confirmed loss of aerosol particles by deposition. Other particle measurement methods (e.g. mobility particle sizer) may provide better size resolution and enable measurement of the highly concentrated aerosols generated in chamber experiments, but require measures against particle size drifts caused by changes in relative humidity (e.g. by dilution) and corrosion of instrument parts by the combustion products. 3.3 Humidity dependence 3.3.1 Maximum mass-specific scattering coefficient To characterize the aerosol’s opacity at given conditions in simple charts, the maximum reading of b* was extracted for each experiment. Values of b*max plotted against humidity show characteristic curves, depicting aerosol properties for each composite (Figure 6). These curves are independent of temperature and coincide for measurements performed in this study at 20 °C and 30 °C (for BP also at 10 °C). Combustion products forming solids by a phase transition from the gas-phase typically produce fine aerosol particles in the submicrometer size-range and show a base-line at dry conditions. These particles can further grow when relative humidity exceeds a characteristic threshold to form larger particles displaying higher scattering efficiencies, if they contain a hygroscopic combustion product. Critical relative humidity, at which it starts absorbing water vapor to form a liquid phase, is easily recognizable by an abrupt rise in the curve’s slope and can be found in the literature for many compounds.

Page 69

1.4

Composite-A

b*max (m -1 g -1 )

1.2

20°C

1.0

30°C

0.8 0.6 0.4 0.2 0.0 1.4

Composite-K

b*max (m -1 g -1 )

1.2

20°C

1.0

30°C

0.8

Figure 7. Aerosol mode diameter at the time of b*max plotted against initial relative humidity.

0.6 0.4 0.2

Combustion products of black powder consist of a complex mixture of solids and gases. Main components found in the solid fraction include K2SO4 and K2CO3, among others24 of which only K2CO3 is hygroscopic with a critical relative humidity of 43%. Its high solubility and low abundance result in a moderate rise of aerosol opacity in respect to relative humidity. The final steeper rise at near saturation humidity can be attributed to K2SO4 (critical relative humidty: 97%).

0.0 1.4

Black Powder

b*max (m -1 g -1 )

1.2

10°C

1.0

20°C

30°C

0.8 0.6 0.4 0.2

3.3.2 Mode diameter at maximum attenuation

0.0 0%

20%

40%

60%

80%

100%

Relative Humidity (%)

Figure 6. b*max of composite-K, composite-A measured at 20 °C and 30 °C, of BP also at 10 °C plotted against initial relative humidity. (For composite-K values go up to 2.48 at 96% RH, but are not plotted.). Composite-A developed no visible aerosol at all for humidity settings below 80% RH, because its primary combustion products only consist of gases. Hydrochloric acid, one of its combustion products, produced particles which rapidly grow by absorbing water vapor at highly humid conditions causing a rise in b*max. Under experimental conditions, however, the particles grew too large to stay suspended in the confined volume of the chamber and were quickly deposited on chamber surfaces. Therefore, b* only briefly rose and the air in the chamber cleared almost completely within a few minutes. This kind of deposition may not occur quickly in the open air. Therefore aerosols may be persistent and reach higher opacities than these measurements suggest. Aerosols emitted by combustion of composite-K consist of KCl. Therefore, the plot of b*max is constant until critical relative humidity of KCl is reached at 85% where the plot shows a sharp rise. Unlike in composite-A, particle size does not grow so large that particles are deposited quickly.

Page 70

PSDs measured with the Mie scattering analyzer showed only a single mode for the composites measured in this work. Other modes do obviously exist, as indicated by particle counter measurements, but light scattering properties can approximately be represented by the single mode PSD measured. The size effect can be visualized more easily by plotting mode diameters of the PSDs at maximal attenuation. These plots (Figure 7) look remarkably similar to b*max (Figure 6), but rise less intensely at high humidity conditions. The significant increase in particle size from a few hundred nm to the µm-range causes the higher b*max in composite K at high RH conditions. KCl particles undergo hygroscopic growth by reduced vapor-pressure of water. For black powder, however, no rise in particle diameter is observed at all with increasing RH, although b*max clearly shows a rising tendency. Here, the main aerosol constituent K2SO4 does not display a simple hygroscopic growth mechanism as with composite K. The increase of b*max in this case may be indicating growth of the hygroscopic subcomponent K2CO3 existing as separate particles of different sizes, which could not be resolved in the presence of large amounts of K2SO4 particles. Also the PSD analyzer is not able to measure volume or number concentration values, which would provide more insight. To properly explain complex growth phenomena like those occurring in the example of black powder, more sophisticated measurement methods with a wider dynamic range are required.

Journal of Pyrotechnics, Issue 33, 2014

4. Conclusions Measurement of maximum light attenuation after combustion of pyrolant composites at controlled relative humidity conditions in an unpressurized combustion chamber were found a reproducible and handy method to characterize dependence of aerosol visibility on relative humidity. By conversion to composite-mass specific scattering coefficients, a plot independent of temperature against relative humidity can be obtained. The effect of humidity on total visible aerosol development is visualized in a simple plot and is quantitatively intercomparable between different composite types. Relative humidity limits for low smoke application can be investigated (critical relative humidity of hygroscopic combustion products) and visible aerosol development per pyrotechnic unit shot (smokeintensiveness) can be predicted for arbitrary humidity conditions. Aerosols which were previously reported in the sub µm-range were observed to grow to diameters of around 1 µm and above by absorption of water vapor from ambient air under high relative humidity, causing aerosol opacity to increase drastically. The proposed measurement method allows effectiveness of smoke-reducing techniques in pyrotechnics to be tested in small scale experiments and can elucidate chemical components which increase visibility of aerosol emissions.

References 1

N. Kubota, “Propellant chemistry”, in Pyrotechnic Chemistry, Journal of Pyrotechnics Inc., Whitewater, 2004, Chapter 11, pp. 7–11.

J. A. Conkling, “Smoke and sound”, in Chemistry of Pyrotechnics, Marcel Dekker Inc., New York, 1985, pp. 167–179.

A. P. Hardt, “Generation of smoke”, in Pyrotechnics, Pyrotechnica Publications, Post Falls, 2001, pp. 318– 332.

D. J. Cziczo and J. P. D. Abbatt, “Infrared observations of the response of NaCl, MgCl2, NH4HSO4, and NH4NO3 aerosols to changes in relative humidity from 298 to 238 K”, Journal of Physical Chemistry, Vol. 104, 2000, pp. 2038–2047; doi: 10.1021/jp9931408.

scattering of light by small particles. WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, 2004, ISBN 9780471293408. 9

P. F. Pagoria, G. S. Lee, A. R. Mitchell and R. D. Schmidt, “A review of energetic materials synthesis”, Thermochimica Acta, Vol. 384, Number 1–2, 2002, pp. 187–204; doi: 10.1016/S00406031(01)00805-X.

10 D. Badgujar, M. Talawar, S. Asthana and P. Mahulikar, “Advances in science and technology of modern energetic materials: An overview”, Journal of Hazardous Materials, Vol. 151, Number 2-3, 2008, pp. 289–305. 11 B. Wehner, A. Wiedensohler and J. Heintzenberg, “Submicrometer aerosol size distributions and mass concentration of the millennium fireworks 2000 in Leipzig, Germany”, Journal of Aerosol Science, Vol. 31, Number 12, 2000, pp. 1489–1493. 12 Y. G. Wang, Zhuang, C. Xu and Z. An, “The air pollution caused by the burning of fireworks during the lantern festival in Beijing”, Atmospheric Environment, Vol 41, Number 2, 2007, pp. 417–431. 13 A. Dutschke, C. Lohrer, L. Kurth, S. Seeger, M. Barthel and U. Panne, “Aerosol Emissions from Outdoor Firework Displays”, Chemical Engineering & Technology, Vol. 34, Number 12, 2011, pp. 2044– 2050; doi: 10.1002/ceat.201100080. 14 A. Dutschke, C. Lohrer, S. Seeger and L. Kurth, “Gasförmige und feste Reaktionsprodukte beim Abbrand von Indoor-Feuerwerk“, Chemie Ingenieur Technik, Vol 81, Number 1–2, 2009, pp. 167–176; doi: 10.1002/cite.200800122 15 J. T. Hanley and E. J. Mack, “A laboratory investigation of aerosol and extinction characteristics for Salty Dog, NWC 29 and NWC 78 pyrotechnics”, Calspan Report No. 6665-M1, Department of the Navy Naval Air Systems Command, Washington DC, 1980. http://oai.dtic. mil/oai/ oai?verb=getRecord&metadataPrefix=html &identifier=ADA093098 16 K. L. Kosanke and B. J. Kosanke, “The chemistry of colored flame”, In Pyrotechnic Chemistry, Journal of Pyrotechnics Inc., Whitewater, 2004, Chapter 9, pp. 25–49.

A. Schenkel and K. Schaber, “Growth of salt and acid aerosol particles in humid air”, Journal of Aerosol Science, Vol. 26, No. 7, 1995, pp. 1029–1039; doi: 10.1016/0021-8502(95)00038-E.

A. T. Cocks and R. P. Fernando, “The rates of formation of hydrogen chloride / water aerosols by homogeneous nucleation”, Atmospheric Environment, Vol. 15, Number 7, 1981, pp. 1293–1299; doi: 10.1016/0004-6981(81)90323-1.

18 S. Prahl, “Mie Scattering Calculator”, Oregon Medical Laser Center, 2012. http://omlc.ogi.edu/calc/ mie_calc.html

M. Z. Jacobson, “Particle absorption and scattering”, In Fundamentals of atmospheric modelling, second edition, Cambridge University Press, New York, 2005, pp. 301–313.

C. F. Bohren and D. R. Huffman, Absorption and

20 F. Peng and S. W. Effler, “Mass-specific scattering coefficient for natural minerogenic particle populations: particle size distribution effect and closure analyses”, Applied Optics, Vol. 51, Number 13, 2012, pp. 2236–2249.

Journal of Pyrotechnics, Issue 33, 2014

17 Guide to Meteorological Instruments and Methods of Observation, Seventh edition, World Meteorological Organization, Geneva, 2008.

19 P. J. Flatau, “scatterlib”, Google Project Hosting, 2011. https://code.google.com/p/scatterlib/

Page 71

21 M. Harkoma, “Visual and Near Infrared Mass Extinction Coefficient of Five Pyrotechnic Screening Smokes”, Journal of Pyrotechnics, Vol. 32, 2013, pp. 67–77. 22 K. Okuyama, Y. Kousaka, Y. Kida and T. Yoshida, “Turbulent coagulation of aerosols in a stirred tank”, Journal of Chemical Engineering of Japan, Vol. 10, Number 2, 1977, pp. 142–147. 23 J. G. Crump, R. C. Flagan, J. H. Seinfeld, “Particle wall loss in vessels”, Aerosol Science and Technology, Vol. 2, Number 3, 1983, pp. 303–309. 24 I. von Maltitz, “Our present knowledge of the chemistry of black powder”, in Pyrotechnic Chemistry, Journal of Pyrotechnics Inc, Whitewater, 2004, Chapter 6, pp. 1–12.

Page 72

Journal of Pyrotechnics, Issue 33, 2014

Other titles available from the Journal of Pyrotechnics Pyrotechnic Reference Series No 1

The Illustrated Dictionary of Pyrotechnics, 1995 [ISBN 1-889526-01-0]

No 2

Lecture Notes for Pyrotechnic Chemistry, 2004 [ISBN 1-889526-09-6]

No 3

Lecture Notes for Fireworks Display Practices, 2006 [ISBN 978-1-88952617-1]

No 4

Pyrotechnic Chemistry, 2013 [ISBN 1-889526-31-7 – Version 1.1].

No 5

Pyrotechnic Literature Series (Please note new versions which replace previous individual publications) No 13 No 14

Selected Pyrotechnic Publications of K.L. and B.J. Kosanke, (1981 through 2009), 2012 [ISBN 1-889526-30-0]. Selected Pyrotechnic Publications of Dr. Takeo Shimizu, 2013 [ISBN 978-1-889526-29-4].

Journal of

Issue 33, 2014 $100.00

Pyrotechnics

Fireworks - Pyrotechnic Special Effects - Propellants & Rocketry - Civilian Pyrotechnics

Full papers:

Information for Readers and Authors Journal of Pyrotechnics Archive on the web Events, Sponsors and Caution

ks or

Ro ck

ew Fir

Journal of Pyrotechnics

d an lE ia

la n

3 9 16

38 52 64

s ct

ffe

op el

iv il ia C

Dedicated to the Advancement of Pyrotechnics Through the Sharing of Information

Dedicated To The Advancement Of Pyrotechnics Through The Sharing Of Information