Hickey et al.

Particle and Fibre Toxicology (2020) 17:28

https://doi.org/10.1186/s12989-020-00360-4

RESEARCH Open Access

Toxicity of particles emitted by fireworks

Christina Hickey1†, Christopher Gordon1†, Karen Galdanes1, Martin Blaustein1, Lori Horton1, Steven Chillrud2,
James Ross2, Lital Yinon1, Lung Chi Chen1 and Terry Gordon1*

Abstract
Background: Particle matter (PM) has been associated with increased morbidity and mortality rates across the
world. This study was designed to test the hypotheses that pyrotechnic firework displays introduce significant
amounts of toxic metals into the atmosphere and are hazardous to human health. Size-selective emissions from 10
different fireworks displays were collected during particle generation in a dynamic, stainless steel chamber and
tested for toxicity in cells. A subset of 2 particle types were tested in vivo in mice. At doses that did not produce
cytotoxicity in an LDH assay, in vitro reactive oxygen species (ROS) formation was measured in bronchial epithelial
airway (BEAS-2B) and human pulmonary microvascular endothelial (HPMEC-ST1.6R) cell lines treated with size-
fractionated particles from the emissions of fireworks.
Results: Significant increases in ROS, in both cell types, were dependent upon the type of firework but not particle
size. The in vitro ROS activity was correlated with lung inflammation produced in groups of mice treated by
oropharyngeal aspiration with 0, 50, or 100 μg fireworks PM10/mouse. Trace metal analyses of the PM10 samples
showed significant differences in metal content among fireworks type. Interestingly, the PM10 sample for the
fireworks type producing the greatest in vitro ROS response in BEAS-2B cells contained ~ 40,000 and ~ 12,000 ppm
of lead and copper, respectively. This sample also produced the greatest inflammatory response (i.e., increased
neutrophils in bronchoalveolar lavage fluid) in mice.
Conclusions: These findings demonstrate that pyrotechnic display particles can produce adverse effects in
mammalian cells and lungs, thus suggesting that further research is needed to expand our understanding of the
contribution of metal content to the adverse health effects of fireworks particles. This information will lead to the
manufacture of safer fireworks.
Keywords: Fireworks, Particles, Inhaled metals, In vitro, In vivo, Air pollution

Background the Olympics, amusement parks, and sports venues.

Throughout the year, numerous types of celebratory Amusement parks are the largest consumers of fireworks
fireworks or pyrotechnic displays are set off across the in the U.S., whereas the single largest fireworks show is
world. Often, the only limit on the size and number of the July 4th display sponsored by Macy’s [1]. In addition
the displays is cost. In the past, while U.S. pyrotechnics to these large public fireworks displays, small, and often
traditionally were reserved for special occasions such as illegal, fireworks are often ignited locally within residen-
July 4th and Chinese New Year, pyrotechnic displays are tial neighborhoods. According to the American Pyro-
now prevalent at rock concerts, opening ceremonies of technics Association, the amount of consumer fireworks
(258.4 million pounds) intended for use by the general
* Correspondence: [email protected] public (i.e., 1.4G explosives) and purchased in the U.S. is
†
Christina Hickey and Christopher Gordon contributed equally to this work. more than 10-fold greater than that used for large cele-
1
Department of Environmental Medicine, NYU School of Medicine, 341 East bratory fireworks (19.1 million pounds) displayed by
25th St, New York, NY 10010, USA
Full list of author information is available at the end of the article

© The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License,
which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if
changes were made. The images or other third party material in this article are included in the article's Creative Commons
licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons
licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain
permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the
data made available in this article, unless otherwise stated in a credit line to the data.
Hickey et al. Particle and Fibre Toxicology (2020) 17:28 Page 2 of 11

pyrotechnic professionals (i.e., 1.3G explosives) [2] and, Materials and methods
thus, are a significant concern for adverse health effects. Particle sampling
To date, the greatest health concern regarding fire- Ten types of pyrotechnic displays (Table 1) were ignited
works has been the potential for injury to life and limb in a 1 m3 stainless steel chamber which was modified so
due to the explosive force of fireworks. Each year, ap- that all incoming air was drawn in through a HEPA filter
proximately 10,000 to 25,000 people (predominantly attached to the top inlet of the chamber. Two collection
male teenagers) in the U.S. suffer physical and burn in- methods were used to sample airborne particles over a
juries, due to fireworks, which include the loss of fingers, 20 min period or until ‘smoke’ had cleared from the
limbs, eyesight, and sometimes, life [3–5]. The environ- chamber. A stainless steel cyclone was used to impact
mental effects are also a concern as evidenced by nu- and remove particles greater than 10 μm in aerodynamic
merous publications that describe the release and size, and PM10 was collected on an 8″ by 10″ polypro-
contamination of air and waterways with perchlorates pylene substrate (GM-3500, Manadnock Non-Wovens,
and other toxicants [6, 7]. Yet, even though there has Mount Pocono, PA, USA) using a high volume sampling
been a large increase in the amount and size of fireworks pump (500 L/min). The cyclone/filter system was cali-
events, little to no research has investigated the effect of brated with a Venturi flow meter. Two co-located Siou-
fireworks-generated particles, and their composition, on tas Cascade Impactors (SKC, Inc) were calibrated with a
human health [6, 8–12]. In fact, epidemiology studies bubble meter (Gilibrator-2, Sensidyne, Clearwater, FL,
examining the health effects of ambient particulate mat- USA) and operated at 9 L/min to collect size-separated
ter (PM) routinely remove the health and exposure as- coarse, fine, and pseudo-ultrafine (UF) particles from the
sessment data from celebratory periods, such as July 4th chamber (PM10–2.5, PM2.5–0.25, and PM0.25, respectively),
and New Year’s Eve, because of the vast change in for 3 fireworks types, onto 37 mm Teflon filters (Pall
source composition of PM at those times. In the present Corporation, Port Washington, NY, USA). Although this
study, the ROS responses of cells to a number of fire- paper refers to PM0.25 as UF, it must be noted that this
works PM was greater than the typical response to ambi- is due to the impactor specifications and does not match
ent PM ([13]. Interestingly, in our previous in vitro the widely accepted < 0.1 μm definition of UF PM. All
ambient PM study, the highest ROS response of 360 collection substrates were acclimated for at least 24 h be-
samples, collected in winter and summer at 5 sites fore being weighed pre- and post-sampling with a Met-
across the U.S., was produced by a PM2.5 sample col- tler 2500 scale (Toledo-Mettler) in a temperature- and
lected on July 4th, 2008 in Anaheim, California (courtesy humidity-controlled weighing room. After weighing, fil-
of Dr. Michael Kleinman, University of California, Irvine ters were stored under sterile conditions at − 80 °C. Par-
[13];). Of note, the collection site in Anaheim was lo- ticles were then sterilely scraped into pre-weighed sterile
cated near a major theme park which sponsors a large polypropylene tubes and diluted with sterile water to
firework holiday celebration. 250 μg/ml and 1 mg/ml for the in vitro and in vivo ex-
Emissions from pyrotechnic displays are composed periments, respectively.
of numerous organic compounds as well as metals.
The high temperature ignition of different metal com- Trace element analyses
pounds, which are purposefully added to fireworks, Particles were also scraped off the PM10 filters into poly-
produces different fireworks colors [14]. While re- propylene tubes and transferred to Lamont-Doherty
search has addressed the exposure assessment of or- Earth Observatory for elemental analysis. The powdered
ganic effluents [6, 15] and the contribution of samples were transferred into XRF cups normally used
fireworks celebrations to ambient PM composition for dried sediment samples. Samples were analyzed on
[16, 17], the toxicity of fireworks has not been ex- an energy dispersive, polarized excitation x-ray fluores-
plored. This paper describes the in vitro and in vivo cence spectrometer with a 50 W, 50 kV Pd tube as pri-
toxicity of PM10 produced by a number of differently mary source, the radiation of which was modified by
colored pyrotechnic firework displays. The generation four secondary targets to optimize the excitation across
of reactive oxygen species (ROS) by airway epithelial the entire elemental range (XEPOS by Spectro Analyt-
and vascular endothelial cell lines was examined ical, Kieve, Germany). Helium gas was used to flush the
in vitro and a subset of fireworks emissions were spectrometer allowing light elements to be included with
tested for their ability to produce pulmonary inflam- total elemental range from Na to U. Resulting spectra
mation and injury in an in vivo mouse model. Be- were processed by the manufacturer’s Turboquant™ cali-
cause the physical-chemical properties of inhaled PM bration software which allows for semi-quantitative ana-
can modify pulmonary toxicity, both trace element lysis of a wide range of sample matrices. Without a
composition and particle size were also investigated matrix specific calibration, resulting concentrations are
in this study. expected to be within 30% of true concentrations
Hickey et al. Particle and Fibre Toxicology (2020) 17:28 Page 3 of 11

Table 1 X-ray fluorescence analysis of fireworks PM10. Up to the top 5 trace elements (ppm) are presented for each display type
Product Name Major Trace Elementsa Concentration (ppm) in Collected PM10
Purple Colorful Storm Ti 22,000
Cu 44,000
Al 8600
Yellow Colorful Storm Cu 14,000
Sr 1200
Zn 730
Blue Colorful Storm Cu 53,000
Ti 11,000
Al 5200
Sr 4800
Red Colorful Storm Al 40,000
Ba 10,000
Sr 6200
Color Changing Wheel Ba 11,000
Sr 6300
Ti 3800
Tiger Roaring 1 Ba 430
Tiger Roaring 2 Fe 1200
Ba 460
Black Cuckoo Pb 40,000
Cu 12,000
Ba 5300
Al 3100
Sr 3000
Saturn Missiles 1 Fe 3100
Pb 1600
Br 850
Co 150
Saturn Missiles 2 Fe 4200
Br 850
Co 170
Bottle Rocket Not determined
Firecrackers Al 95,000
Fe 6300
Zn 2000
a
Excluding sulfur, potassium, and chloride which were present in all samples at high concentrations (typically greater than 20,000 ppm). Duplicate experiments
with 2 fireworks (Tiger Roaring and Saturn Missiles) were conducted to examine variability between commercial samples

although larger differences have been observed between 10’s of ppm for light elements such as Na and Mg to
Turboquant concentrations and matrix specific calibra- around 1–2 ppm or less for a wide range of metals.
tions [18]. However, relative differences of samples with
similar matrix are well constrained by the Turboquant Cell culture
software (e.g., can determine that sample A is 2.1 times To assess the toxicity on likely lung cell targets of in-
the concentration of sample B even though the accuracy haled fireworks particles, 2 cell lines were tested with
of the concentration of both samples may both be too the PM10 emitted by the fireworks. A human pulmonary
low). Detection limits for powdered samples range from microvascular endothelial cell line (HPMEC-ST1.6R)
Hickey et al. Particle and Fibre Toxicology (2020) 17:28 Page 4 of 11

was provided by Drs. James Kirkpatrick and Vera calculated as the net increase in fluorescence intensity,
Krump-Konvalinkova (Johannes Gutenberg University, over time, by subtracting the mean media vehicle blank
Mainz, Germany). As previously described [19], HPMEC from the change in fluorescence intensity for each treat-
cells were maintained at 37 °C in a humidified atmos- ment and as the fold increase in ROS activity over the
phere of 5% carbon dioxide and grown in Endothelial response to control media alone.
Growth Medium (EGM-2) containing 1% penicillin/
streptomycin (Gibco, Grand Island, NY, USA) and sup-
plemented with an EGM-2 BulletKit and 5% fetal bovine Animal study
serum (Lonza, Switzerland). A bronchial epithelial cell Eight to 10 week old male and female FVB/N mice
line (BEAS-2B) was obtained from the American Type were purchased from The Jackson Laboratory (Bar
Culture Collection (ATCC, Rockville, MD, USA) and Harbor, ME, USA) and bred at NYU. Mice were
maintained in DMEM medium (Dulbecco’s Modified housed in polycarbonate cages with corn-cob bedding
Eagle Medium; Gibco) with 10% fetal bovine serum in temperature and humidity controlled rooms with a
(FBS) (Gemini Bio Products, Calasas, CA, USA) and 1% 12 h light/dark cycle. Animals were provided standard
penicillin/streptomycin (Gibco). chow and water ad libitum. All animal procedures
and handling were performed under the National In-
Lactate dehydrogenase (LDH) and reactive oxygen stitutes of Health and Animal Welfare Act guidelines
species (ROS) assays for the ethical treatment of animals using an ap-
Cells were seeded onto a COSTAR 96 well plate (Fisher proved NYU School of Medicine Institutional Animal
Scientific, Pittsburgh, PA, USA) and the LDH (Takara Care and Use Committee protocol.
Bio Inc., Madison, WI) and ROS assays were performed A subset of 2 of the 10 extracted PM10 samples was
as previously described [19]. After thawing, particles selected for an in vivo bioassay of lung inflammation
were diluted in cell culture media and initially dispersed and injury in mice. These samples represented fireworks
in an ultrasonic water bath at a maintained temperature PM which produced the highest and lowest in vitro ROS
of < 28 °C, for at least 20 min. Immediately before apply- response: Black Cuckoo and Roaring Tiger, respectively.
ing to wells, PM samples were vortexed for 10 s to sus- An oropharyngeal aspiration technique [21] was
pend the particles. Sufficient material was available from used to disperse PM into the lungs of mice. Briefly,
the impactor sampling of only 3 types of fireworks to mice were anesthetized with isofluorane (Abbott La-
examine size-dependent differences in toxicity. For the boratories, King of Prussia, PA, USA), placed on a 45°
ROS and LDH assays, cells were washed with PBS (phos- board, and aspirated with 50 μL of sterile water (nega-
phate buffered saline), and the media was changed to tive control) or 50 μL of PM suspended in sterile
DMEM/F12 (Dulbecco’s Modified Eagles Medium/Nu- water (vortexed immediately before delivery) for a
trient Mixture F-12, Gibco) containing no phenol red, total dose of 0, 50, or 100 mg/mouse (n = 5 per
2% FBS, and 1% penicillin/streptomycin prior to treat- group). Twenty-four hours after aspiration, animals
ment with PM. At the tested final PM concentrations of were euthanized with 0.26 mg/g sodium pentobarbital
50 and 100 μg/ml, no cytotoxicity was observed in the (ip). The lungs were lavaged twice using PBS (Invitro-
LDH assay, so treatment concentrations of 10, 50, and gen) and cell counts, cell differentials, and total pro-
100 μg/ml were used in the ROS assay. Briefly, for the tein (BCA Protein Assay Kit, Thermo Fisher
measurement of ROS, cells were loaded with dichloro- Scientific, Pittsburgh, PA, USA) were measured in lav-
fluorescein prior to treatment with particle extracts. age fluid.
After dye removal and washing, the pyrotechnic particles
were sonicated for 30 min and added to the cell culture
media in each well for a final concentration of 10, 50, Statistics
and 100 μg/ml, for each particle type (in triplicate). Van- Prism 5.0 for Windows (GraphPad Software, San Diego,
adium and carbon particles with 5% iron (generated by a CA, USA) was used for the analysis of data, which are
high voltage spark, a gift of Drs. Oberdorster and Elder, reported as means + standard error (SE). Where appro-
U of Rochester [20];) were used as positive control parti- priate, data were analyzed using unpaired t-tests. For
cles, whereas carbon black (Printex, Orion, Houston, in vitro experiments using more than two groups for
TX) served as a negative control particle. The plates analysis, a one-way ANOVA, followed by Tukey’s Mul-
were maintained at 37 °C and 5% CO2 and monitored in tiple Comparison Test, was used. For in vivo experi-
a fluorescence microplate reader (HTS 7000, Perkin ments using more than two groups for analysis, one-way
Elmer, Waltham, MA, USA with HTSoft software) every ANOVA, followed by Dunnett’s Multiple Comparison
30 min for 3 h using excitation and emission filters of Test, was used. The statistical significance was set for
480 nm and 535 nm, respectively. ROS production was p < 0.05.
Hickey et al. Particle and Fibre Toxicology (2020) 17:28 Page 5 of 11

Results 3 fireworks types, ROS activity in BEAS-2B cells was

PM characterization assessed after treatment with Coarse, Fine, or UF PM
The total amount of PM10 collected varied among fire- (Fig. 2). Particle size did affect ROS activity, but this
works display type, likely due to the differences in dis- size-dependence response was influenced by fireworks
play size and purpose (e.g., small missiles flares vs. type. For example, the UF particle size induced less ROS
firecrackers vs. colorful display units) of the purchased activity than the Coarse and Fine particle sizes in 1 fire-
products (30–60 g were generated from each fireworks works type (Purple Colorful Storm), whereas the UF par-
type and PM10 comprised approximately 1/3 of this ticles produced the most ROS activity for the other 2
weight; data not shown). Interestingly, the relative mass fireworks types. Regardless of particle size, the Black
of coarse, fine, and UF particles collected varied among Cuckoo and Saturn Missiles fireworks produced greater
fireworks types (Fig. 1). For example, the Back Cuckoo ROS activity than did the Purple Colorful Storm
and Saturn Missiles samples produced predominantly fireworks.
UF particles (~ 90%), whereas the Purple Colorful Storm The type of fireworks display significantly influenced the
display had a more even distribution of particles with ROS activity in BEAS-2B and HPMEC cells challenged
coarse PM accounting for ~ 50% of the collected mass. with PM10 (only BEAS-2B results are shown although the
The trace element composition varied significantly ROS responses in HPMEC cells [Supplemental Fig. 1]
among the PM10 collected from the fireworks display were similar). The oxidative stress responsiveness of the
types. Whereas sulfur, potassium, and chlorine were cells was confirmed by the significant increases in ROS
common major components of the particle emissions of observed after treatment with the positive control particles
each firework type (data not shown), there were signifi- comprised of vanadium or carbon with 5% iron. No sig-
cant differences in the levels of toxic metals (Table 1). In nificant effect on ROS activity was observed for the nega-
general, high amounts of Fe, Al, Cu, Ba, Ti, and Sr were tive control carbon particles. A subset of the types of
measured in the PM10 of multiple display types. Surpris- fireworks caused a significant increase in ROS response,
ingly, Pb was present in the PM10 of 2 display types with while other particle types caused only a low or no increase
> 40,000 ppm Pb in the Black Cuckoo sample. in ROS (Fig. 3). This was best illustrated by the group of 4
Colorful Storm pyrotechnic displays which were identical
In vitro study in construction (i.e., the size and shape of the firework dis-
Cytotoxicity, as measured by the LDH assay, did not play), but differed in the color and the composition of par-
occur in either cell line at 100 μg/ml for each display ticles (Table 1) produced by ignition. 100 μg/ml of PM10
type (data not shown), so this non-cytotoxic concentra- produced a significant increase in ROS activity in BEAS-
tion was chosen as the high dose in the ROS assay. The 2B cells for 3 Colorful Storm samples, whereas the
ROS results demonstrated that the fireworks type and response to the Red Colorful Storm sample was not sig-
presumably its composition, played a more important nificantly different from cell culture media alone. It should
role in ROS activity than did particle size. In a subset of be noted that this Red Colorful Storm PM10 sample had a

Fig. 1 The contribution of each particle size to the total mass, collected by a 3 stage Sioutas cascade impactor, for 3 fireworks types
Hickey et al. Particle and Fibre Toxicology (2020) 17:28 Page 6 of 11

Fig. 2 The effect of particle size on ROS activity (measured in triplicate) in BEAS2-B cells (fold increase in fluorescence intensity over media
control) treated with 100 μg/ml of 3 selected fireworks particle types for 3.5 h

much lower copper concentration than the 3 other Color- correlated with an increase in ROS response in BEAS-2B
ful Storm samples. Of the 10 product types (2 types were (r2 = 0.43; p = 0.03) and HPMEC (r2 = 0.65; p = 0.03) cell
tested in duplicate), the Black Cuckoo PM10 produced the lines.
largest ROS response in BEAS-2B cells and an extended
dose-response curve demonstrated that increased ROS ac- In vivo study
tivity occurred at 5 μg/ml (Fig. 4). A regression analysis To test whether an in vivo response to fireworks parti-
demonstrated that Cu was the only trace metal statistically cles was similar to that seen in vitro, using a subset of

Fig. 3 The effect of 12 fireworks types on the fold increase (over media control) in ROS activity in BEAS2-B cells treated with 100 μg/ml (PM10).
The columns and error bars represent the mean and SEM, respectively. * p < 0.05 compared to media control
Hickey et al. Particle and Fibre Toxicology (2020) 17:28 Page 7 of 11

Fig. 4 Dose-response change in ROS activity generated in BEAS-2B cells (fold increase in fluorescence intensity over media control) treated with
the Black Cuckoo PM10 sample. Squares and error bars represent means ± SE. * p < 0.05

the fireworks particles tested in vitro, mice were treated in vivo. No significant increases in protein or other cell
with 50 or 100 μg of Tiger Roaring (produced a low types were observed (data not shown).
in vitro ROS response) or Black Cuckoo (high ROS re-
sponse) PM10 via oropharyngeal aspiration. Twenty-four Discussion
hours later, lung inflammation (denoted as an increase The excitement and pleasure associated with the colors
in neutrophils in lavage fluid) was significantly increased and sounds generated by fireworks are important cele-
in mice treated with 50 and 100 μg of the Black Cuckoo, bratory components of cultures throughout the world.
but not the Tiger Roaring, particles compared to the re- Yet, the adverse health effects of fireworks have been
sponse in vehicle-treated (i.e., sterile water) control mice evaluated almost exclusively for: 1) the injuries and
(Fig. 5). These results suggest that the in vitro ROS re- burns produced by the rapid changes in the physical en-
sponse has potential as a predictor of what to expect vironment attributed to the explosive forces of fireworks;

Fig. 5 The effect of 2 fireworks types on the influx of inflammatory neutrophil (PMNs) in the lavage fluid of mice (n = 5/group) treated with sterile
saline or 50 or 100 μg PM10 generated from the Black Cuckoo PM10 fireworks. The columns and error bars represent the mean and SEM,
respectively. * p < 0.05 compared to saline control group
Hickey et al. Particle and Fibre Toxicology (2020) 17:28 Page 8 of 11

and 2) the release of chemicals that pollute waterways. chemical properties to the toxicity of particles emitted
Injuries to limbs and fingers occur predominantly in by pyrotechnic displays. The study was focused on
males during the teen years and such accidents are re- the pulmonary toxicity of particles generated by fire-
corded and reported at many levels of the health care works and therefore, thoracic and respirable particle
system in the U.S. [3–5, 22–24]. These acute injuries size were collected for the in vitro and in vivo stud-
and burns are significant adverse health effects and ies. The stability of fireworks particles, by size or
largely preventable by the use of proper safety proce- otherwise, in the generation chamber were not
dures. Little to no research, however, has addressed the assessed and nor was the effect of climatic conditions
potential for the particles and gases released during fire- on the particle stability.
works celebrations to cause adverse cardiopulmonary ef- The results of this study suggest that in addition to
fects via the inhalation exposure route [25]. This organic pollutants [6], significant amounts of toxic
research study addressed this knowledge gap by examin- metals are released into the ambient environment
ing the in vitro and in vivo toxicity of PM generated by from both home pyrotechnic displays and larger com-
a selection of fireworks displays that are commonly used mercial fireworks. These releases, as demonstrated by
by individuals at home. exposure assessment air pollution studies [16, 25], are
The in vitro experiments clearly demonstrated that episodic in nature but can potentially be significant
particle size had an effect on the ROS activity response emission sources of metals. We have confirmed this
in both airway epithelial and vascular endothelial cell time-dependent increase in airborne metal concentra-
lines. Presumably, this was due to differences in particle tions throughout the U.S. by using a 10 year period of
composition among the ultrafine, fine, and coarse par- data from EPA’s speciation network [27]. During that
ticle sizes. This particle size-dependent effect on cell- period, for example, 19 of the 22 highest peaks (i.e.,
based ROS production was accompanied however, by an greater than 0.15 μg/m3) for strontium (Sr) in air-
even greater effect of the type of fireworks being tested. borne particles in the U.S. occurred on the days sur-
Five of the 10 fireworks products (2 products were rounding July 4th and New Year’s Day (red circles in
assessed in duplicate) had significant increases in ROS Fig. 6). Similar celebration-associated episodic results
activity (Fig. 3) and as shown in Table 1, these (Fig. 6) occurred for some (i.e., Ti), but not all (Fe
fireworks-dependent differences in response were ac- and Pb), of the major trace elements associated with
companied by significant differences in trace element the in vitro and in vivo toxicity observed in this
composition of the PM10 collected after ignition of each study.
pyrotechnic display in a stainless steel chamber. On a Given the increases in ROS production observed
mass concentration (μg/ml) basis, the ROS effect of in vitro and the correlated in vivo changes, as well
these fireworks particles was greater than that of ambi- as the episodic elevations in fireworks-associated
ent PM collected in the NYC metropolitan area (data metals associated with July 4th and New Year’s Eve
not shown). Because of the well established role of celebrations (Fig. 6), targeted interventions in fire-
metals in oxidative stress [13, 26], the observed differ- works manufacturing can reduce the potential for
ences in metal content among the fireworks types were adverse cardiopulmonary effects in the U.S. and glo-
likely responsible for the fireworks type-dependent dif- bally. The Walt Disney Company, for example, has
ferences in ROS activity. addressed contamination concerns in Florida’s Ever-
Importantly, fireworks manufacturers adjust the metal glades National Park by substituting pneumatic
content of the each fireworks to produce the desired launching systems for explosives charges in their
color based upon the high temperature oxidation of daily fireworks displays. Also, chlorine-free fireworks
metals. A particular concern in our findings is the dis- have recently been developed to replace the blue,
turbing amount of Pb in 2 of the tested fireworks. At 40, green, and red colors produced by chlorine-based
000 ppm, it is likely that Pb was not an inadvertent con- stable metal compounds [28]. Similarly, Gluck [28]
taminant in the Black Cuckoo PM10 sample, but pur- has suggested that lithium colorant displays can be
posely added to the product for achieving the desired utilized to replace the red color produced by Sr in
effect in the fireworks. Because the 10 fireworks chosen fireworks, whereas Han [29] has demonstrated that a
for this study were all pyrotechnic displays that can be new unpacking powder containing ‘micronano’ sili-
purchased by individuals and used at home, the potential con can reduce the amount of PM released into the
for exposure of children to significant amounts of Pb atmosphere. In addition, although primarily for rea-
and other toxic metals from such products is unwar- sons of cost, some fireworks exhibitions have been
ranted yet preventable. replaced by light shows that encompass laser and
There were limitations to the study which impair a LED displays accompanied by music and explosive
full understanding of the contribution of physical and sounds.
Hickey et al. Particle and Fibre Toxicology (2020) 17:28 Page 9 of 11

Fig. 6 The levels of strontium (SR), barium (Ba), lead (Pb), and copper (Cu) as measured at EPA speciation sites across the U.S. from 1999 to 2014
[27]. The Y-axis is μg/m3 as measured in the ambient air. Red circles indicate samples taken on or around holidays (i.e., July 4 or 5; December 31
or January 1)

Conclusions species present in the air across the U.S. demonstrated

Our research has demonstrated that fireworks particulate that the metals associated with increased toxicity in
emissions are more toxic in vitro than typical urban par- our ground-based pyrotechnic displays are elevated in
ticulate matter. This in vitro toxicity was dependent on samples taken around the holiday celebrations of July
the composition of the particulate emissions as shown by 4 and New Years. These findings bring up the obvi-
the large range in toxicity among the fireworks types ex- ous question of whether adverse cardiopulmonary ef-
amined in this study. The in vivo studies in mice validated fects are associated with exposure to fireworks-linked
the in vitro findings for a subset of particles and suggest metals during these holiday periods. Responsible
that the in vitro results are translatable to the mammalian manufacturing can have a major impact on reducing
lung. Surprisingly, highly toxic metals, such as Pb, were toxic metals in both commercial and residential pyro-
present at exceedingly high levels in the emissions of some technics displays and their potential for producing ad-
of the tested fireworks. Our temporal survey of the metal verse health effects.
Hickey et al. Particle and Fibre Toxicology (2020) 17:28 Page 10 of 11

Supplementary information 2014;114:158–64. https://doi.org/10.1016/j.chemosphere.2014.03.088 PMID:

Supplementary information accompanies this paper at https://doi.org/10. 25113197.
1186/s12989-020-00360-4. 7. Sijimol MR, Mohan M. Environmental impacts of perchlorate with
special reference to fireworks-a review. Environ Monit Assess. 2014;
186(11):7203–10. https://doi.org/10.1007/s10661-014-3921-4 Review.
Additional file 1: Supplemental Figure 1. The effect of 12 fireworks
PMID:25004859.
types on the fold increase (over media control) in ROS activity in HPMEC
cells treated with 100 μg/ml (PM10). The columns and error bars 8. Burki TK. Are fireworks a hazard for respiratory health? Lancet Respir Med.
2017;5(2):103–4. https://doi.org/10.1016/S2213-2600(17)30016-4.
represent the mean and SEM, respectively. * p < 0.05 compared to media
control. 9. Garaga R, Kota SH. Characterization of PM10 and Impact on Human Health
During the Annual Festival of Lights (Diwali). J Health Pollut. 2018;8(20):
181206 doi: 105696/2156-9614-820181206 eCollection 2018 Dec.
Acknowledgements 10. Greven FE, Vonk JM, Fischer P, Duijm F, Vink NM, Brunekreef B. Air pollution
Mr. Decker and Mrs. Flannagan of the Science Research Program of Monroe- during new Year's fireworks and daily mortality in the Netherlands. Sci Rep.
Woodbury (NY) Central School District for their guidance and mentoring. The 2019;9(1):5735. https://doi.org/10.1038/s41598-019-42080-6.
authors thank Dr. Audrey Smargiassi, University of Montreal, for advisement 11. Gouder C, Montefort S. Potential impact of fireworks on respiratory health.
on the project. Lung India. 2014;31(4):375–9. https://doi.org/10.4103/0970-2113.142124
Review. PMID:25378846.
Authors’ contributions 12. Licudine JA, Yee H, Chang WL, Whelen AC. Hazardous metals in ambient air
TG and CG initiated the study idea. TG, LCC, and CG developed the study due to new year fireworks during 2004–2011 celebrations in Pearl City,
design. CH participated only in the in vitro experiments and statistical Hawaii. Public Health Rep. 2012;127(4):440–50 No abstract available. PMID:
analyses. SC and JR conducted the chemical analysis of the test samples. LH, 22753988.
MB, and KG contributed to the sample preparation and the in vivo 13. Lippmann M, Chen LC, Gordon T, Ito K, Thurston GD. National Particle
experiments. LY acquired the trace element data from EPA and analyzed the Component Toxicity (NPACT) initiative: integrated epidemiologic and
metal speciation data and prepared the resulting figure. The author(s) read toxicologic studies of the health effects of particulate matter components.
and approved the final manuscript. Res Rep Health Eff Inst. 2013;177:5–13 PMID:24377209.
14. Conkling JA. Chemistry of pyrotechnics. New York: Marcel Dekker, Inc.; 1985.
Authors’ information 15. Webster HA III. Propellants Explos. Pyrotech. 1985;10:1–4.
Not Applicable. 16. Wang Y, Hopke PK, Rattigan OV. A new indicator of fireworks emissions in
Rochester. New York Environ Monit Assess. 2012;184(12):7293–7. https://doi.
Funding org/10.1007/s10661-011-2497-5 PMID:22222633.
This project was supported by NIEHS’ Core Center grant ES000260, T32 17. Yao L, Wang D, Fu Q, Qiao L, Wang H, Li L, Sun W, Li Q, Wang L, Yang X,
Training grant ES07324 and Columbia University's Center grant ES009089. Zhao Z, Kan H, Xian A, Wang G, Xiao H, Chen J. The effects of firework
regulation on air quality and public health during the Chinese spring
Availability of data and materials festival from 2013 to 2017 in a Chinese megacity. Environ Int. 2019;126:96–
The datasets used and/or analysed during the current study are available 106. https://doi.org/10.1016/j.envint.2019.01.037.
from the corresponding author on reasonable request. 18. Wovkulich K, Stute M, Mailloux BJ, Keimowitz AR, Ross J, Bostick B, Su JN,
Chillrud SN. In Situ oxalic acid injection to accelerate arsenic remediation at
Ethics approval and consent to participate a SuperfundsSite in NJ. Environmental Chemistry. 2014;11:525–37 PMCID:
Not applicable. PMC4294815.
19. Mirowsky J, Hickey C, Horton L, Blaustein M, Galdanes K, Peltier RE, Chillrud
Consent for publication S, Chen LC, Ross J, Nadas A, Lippmann M, Gordon T. The effect of particle
Not applicable. size, location and season on the toxicity of urban and rural particulate
matter. Inhal Toxicol. 2013;25:747–57.
Competing interests 20. Oberdörster G, Finkelstein JN, Johnston C, Gelein R, Cox C, Baggs R, Elder
The authors declare that they have no competing interests. AC. Acute pulmonary effects of ultrafine particles in rats and mice. Res Rep
Health Eff Inst. 2000;(96):5–74; disc. 75–86. PMID:11205815.
Author details 21. Rao GV, Tinkle S, Weissman DN, Antonini JM, Kashon ML, Salmen R, Battelli
1 LA, Willard PA, Hoover MD, Hubbs AF. Efficacy of a technique for exposing
Department of Environmental Medicine, NYU School of Medicine, 341 East
25th St, New York, NY 10010, USA. 2Lamont-Doherty Earth Observatory of the mouse lung to particles aspirated from the pharynx [published
Columbia University, Palisades, NY 10983, USA. correction appears in J Toxicol Environ Health A. 2003;66(24):2365]. J Toxicol
Environ Health A. 2003;66(15):1441–52. https://doi.org/10.1080/
Received: 16 January 2020 Accepted: 17 June 2020 15287390306417.
22. Billock RM, Chounthirath T, Smith GA. Pediatric Firework-Related Injuries
Presenting to United States Emergency Departments, 1990–2014. Clin
References Pediatr (Phila). 2016. [Epub ahead of print]. PMID:27550873.
1. US Today. 2016. https://www.usatoday.com/story/news/nation-now/2016/ 23. Dowd MD. Fireworks-related injury in children. Pediatr Ann. 2018;
07/03/how-america-celebrates-4th-july-numbers/86602246/. 47(6):e227–9. https://doi.org/10.3928/19382359-20180522-02 PMID:
2. APA (American Pyrotechnics Association). 2019. https://www.americanpyro. 29898232.
com/industry-facts-figures. 24. Sandvall BK, Keys KA, Friedrich JB. Severe Hand Injuries From Fireworks:
3. Canner JK, Haider AH, Selvarajah S, Hui X, Wang H, Efron DT, Haut ER, Injury Patterns, Outcomes, and Fireworks Types. J Hand Surg Am. 2017;42(5):
Velopulos CG, Schwartz DA, Chi A, Schneider EB. US emergency department 385.e1–385.e8. doi: https://doi.org/10.1016/j.jhsa.2017.01.028. Epub 2017 Mar
visits for fireworks injuries, 2006-2010. J Surg Res. 2014;190(1):305–11. 22. PMID:28341070.
https://doi.org/10.1016/j.jss.2014.03.066 PMID:24766725. 25. Delira CA, Andrade MS, Campos MH, Vancini RL. Fireworks in soccer
4. Moore JX, McGwin G Jr, Griffin RL. The epidemiology of firework-related stadiums: an amazing spectacle for fans but terrible for athletes' health and
injuries in the United States: 2000-2010. Injury. 2014;45(11):1704–9. https:// sport performance. J Sports Med Phys Fitness. 2016;56(10):1260.
doi.org/10.1016/j.injury.2014.06.024 PMID:25047335. 26. Ghio AJ. Biological effects of Utah Valley ambient air particles in humans: a
5. Patel R, Mukherjee B. Crash and burn: ocular injuries due to fireworks. Semin review. J Aerosol Med. 2004;17(2):157–64.
Ophthalmol. 2016;31(3):243–8. https://doi.org/10.3109/08820538.2014. 27. Environmental Protection Agency, 2015. Particulate Matter. [Cited 2015 Mar
962157 PMID:25310140. 31]. Available from. https://www3.epa.gov/pm/.
6. Schmid P, Bogdal C, Wang Z, Azara V, Haag R, von Arx U. Releases of 28. Glück J, Klapötke TM, Rusan M, Sabatini JJ, Stierstorfer J. A strontium- and
chlorobenzenes, chlorophenols and dioxins during fireworks. Chemosphere. chlorine-free pyrotechnic illuminant of high color purity. Angew Chem Int
Hickey et al. Particle and Fibre Toxicology (2020) 17:28 Page 11 of 11

Ed Engl. 2017;56(52):16507–9. https://doi.org/10.1002/anie.201710746 Epub

2017 Dec 5. PMID:29144054.
29. Han Z, Jiang Q, Du Z, Huey Hoon H, Yu Y, Zhang Y, Li G, Sun Y. A novel
environmental-friendly and safe unpacking powder without magnesium,
aluminum and Sulphur for fireworks. J Hazard Mater. 2019;373:835–43.
https://doi.org/10.1016/j.jhazmat.2019.04.001 Epub 2019 Apr 2.

Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.