- Email: [email protected]
Investigating airborne GSR particles by the application of impactor technology
Forensic Chemistry 8 (2018) 72–81
Contents lists available at ScienceDirect
Forensic Chemistry journal homepage: www.elsevier.com/locate/forc
Investigating airborne GSR particles by the application of impactor technology Rebecca Luten, Dieter Neimke, Martin Barth, Ludwig Niewoehner ⇑ Bundeskriminalamt (BKA), Forensic Science Institute KT23, Thaerstraße 11, D-65193 Wiesbaden, Germany
a r t i c l e
i n f o
Article history: Received 15 November 2017 Received in revised form 16 February 2018 Accepted 20 February 2018 Available online 24 February 2018 Keywords: Gunshot residue (GSR) Impactor Airborne particles Still room Contamination risk Particle counter GSR distribution Sedimentation SEM-EDS Propagation
a b s t r a c t Gunshot residues (GSR) are of interest when firearms are used in criminal cases. GSR analyses are usually based upon the elemental composition and morphological appearance of very minute particles by means of SEM-EDS. Based on these two parameters, GSR particles are divided into specified classes. The amount of detected GSR particles depends on the time since discharge of the weapon and the sampling position relative to the location of discharge. In this paper, the influence of time on the local concentration and the distribution of airborne GSR particles were investigated with impactor technology. The particle concentration is constant in the still room; changes in concentration are only related to the emission of GSR particles by the discharge of a firearm. Here we showed that large quantities (50% of max. concentration) of airborne GSR particles can be detected several hours after discharge and contamination can take place as much as three hours after discharge. This study is a first approach to describe the propagation and sedimentation of GSR particles. With respect to statistical confirmation further experiments are already projected in order to comprehend the well-known variability of GSR emission and behavior. Ó 2018 Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons. org/licenses/by-nc-nd/4.0/).
1. Introduction Gunshot residues (GSR) are often investigated when firearms are used in a criminal case. GSR consists of small particles in the submicron to 10 mm range, which can be inorganic or organic [1–3]. These particles originate from the primer and the propellant, as well as material that was re-condensed from formerly evaporated material from the bullet, the cartridge case, and previous residues in the barrel, as well as abrasion particles from the same sources. Lead, antimony, and barium mainly originate from the primer in contemporary ammunition where they are used in the form of lead styphnate, antimony trisulphide, and barium nitrate, e.g. in Sinoxid ammunition by Dynamit Nobel, and many other ammunition manufacturers [4]. GSR is classified according to the ENFSI (European Network of Forensic Science Institutes) and ASTM (American Society for Testing and Materials) guidelines: characteristic of GSR and consistent with GSR [5,6]. Characteristic particles are at least three-component particles (with characteristic morphology), which are normally only found when a firearm has been discharged. Consistent with particles are two-component particles ⇑ Corresponding author. E-mail address: [email protected] (L. Niewoehner).
that, although uncommon, can also have their origin in the environment. The main topic in this article is the variation in concentration of GSR particles versus time in a closed, still room after discharge of a firearm. The particle concentration is constant in such a still room prior to the discharge of a firearm. It can be reasonably assumed that concentration changes are only related to the emission of GSR particles by this discharge. A time model of the behavior of GSR in this situation would be desirable, but is difficult to create as there are many different effects that influence the concentration of airborne GSR. Fojtásek and Kmjec (2005) tried to develop a time model when looking at the sedimentation of GSR particles [7]. Their results show that the main sedimentation process comes to an end after 8 min for the investigated size classes. Our study works on the principle of collected airborne GSR instead of sedimented GSR. The most common method to sample and analyse GSR is tape-lifting and the use of a scanning electron microscope coupled with an energy dispersive X-ray spectrometer (SEM/EDS) [1,8]. Tape-lifting is done on clothing or the hands of the shooter and cannot be used to collect airborne particles. However, Andrasko and Pettersson (1991) reported the use of a double filtration system in combination with an ordinary vacuum cleaner for the collection of GSR [9]. Here, two filters of 20 and 0.8 mm pore size
https://doi.org/10.1016/j.forc.2018.02.005 2468-1709/Ó 2018 Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
R. Luten et al. / Forensic Chemistry 8 (2018) 72–81
73
environmental air, and the particle density can be calculated. This will aid significantly in understanding the distribution of particles. Additionally, semi-continuous measurements were performed with a particle counter because the only source of particle emission is the discharge of the firearm. The particle counter is not able to distinguish between GSR-related and not-related particle concentrations. However, if the formerly measured background is subtracted the particle concentration can be observed over time. 2. Experimental – materials and methods
Fig. 1. Overview of the shooting range.
were used for the separation of particles. The calculated flow rate in the vacuuming system was around 6 L/min. In the present study an impactor was used with a triple separation system with cut off sizes at 10, 2.5 and 0.4 mm [10–12]. This flow rate in this system was 33 L/min. The flow rate is six times greater than that of Andrasko and Pettersson [9]. The impactor was used in combination with a pump with a controlled volume passing through the system instead of an ordinary vacuum cleaner. In this way the particle concentration can be related to a volume of the
The experiments were conducted in a closed shooting range of 6.6 4.3 2.3 m (L W H, V 65 m3). This shooting range was especially designed for shooting distance tests (Fig. 1). It is not used for any ballistic investigations and firearms identification tests. The air conditioning was switched off at least 10 min before shooting, the lights were turned off after the shot had been fired, so the only turbulence inside the room came from the shooter walking out at a slow pace. A Glock 19 pistol (barrel length of 102 mm) with Geco 9 mm Luger ammunition (a full metal jacket lead projectile with an open base, 8.0 g) was used for all the experiments. In all experiments the shooter was the same person, firing the pistol with both hands in three different experimental setups. Setup 1, shooter stands in the middle of the room and the two suction tubes, one for impactor and one for the particle counter are behind the shooter in the corner of the room at 1 m height (Fig. 2). This corresponds to a distance of 3.25 m between the shooter and the suction tubes. Setup 2, two suction tubes, one for impactor and one for particle counter are 20 cm to the right and 80 cm to the front of the shooter (Fig. 2). Setup 3 uses the same position as in setup 1; however, here at 4 different heights (0.5 m, 1 m, 1.45 m and 1.9 m) samples were only taken with the particle counter. 2.1. Equipment 2.1.1. Particle counter An ACS-Plus 228 from KM OptoElectronic GmbH (Leonberg, Germany) with detection sizes from: 0.2 mm to 1.0 mm with steps of 0.1 mm; 1.0 to 4.0 mm with steps of 0.5 mm; and 5 mm was used during the experiments [13].
Fig. 2. Setups of the experiments in the shooting range, X is the position of the suction tubes. Experimental setup 3 uses the position of setup 2.
74
R. Luten et al. / Forensic Chemistry 8 (2018) 72–81
Fig. 3. Particle distribution versus time; the area between the dotted lines indicates the major changes in particle concentration (experiment 1, setup 1, measured with particle counter).
Fig. 4. Particle distribution versus time; the dotted lines denote the maximum and the half maximum values of the particle concentration (experiment 1, setup 1, measured with particle counter).
The ACS-Plus from KM OptoElectronic is a desktop device. It counts particles in air with respect to their sizes. The particles have to cross a laser beam, where they scatter the light related to their sizes. A light sensitive diode placed orthogonal to the laser beam detects the intensity of the scattered light. By a comparator the signal height is related to a particle size range and counted as one event in that size channel. The system displays the number of detected particles per size class.
2.1.2. Impactor A Dekati PM10 Impactor in combination with a Satorius MD8 airscan pump was used during this project. The flowrate of the pump was fixed to 2.0 m3 per hour, which was six times higher than the calculated value mentioned in the paper by Andrasko and Pettersson [9]. The impactor was equipped with an antistatic tube. The impactor is a particle separation system with three collection stages widely used in aerosol investigations [10–12]. It sizes the
R. Luten et al. / Forensic Chemistry 8 (2018) 72–81
75
P Fig. 5. Size class distribution ( = 100%) versus time after discharge for PbBaSb particles (experiment 1, setup 1, collected with the impactor, measured with SEM/EDS); the diagram in the lower right corner shows the percentage ratio of three selected size classes.
particles in the air depending on their inertia. Particles with the highest inertia are deposited on the first impaction stage, the others follow the air stream to the next impaction stage. It has its own cutoff size, so it is known which particles will be detected, e.g. a cut-off of 2.5 lm means that everything larger will be expected to impact on it [10–12]. This cut-off value is based on completely spherical particles of low specific weight. Each stage is covered with a 1 inch double-sided adhesive tape to collect the impacting particles. 2.1.3. SEM/EDS Due to the large number of samples, two scanning electron microscopes were employed during this project. A FEI Quanta 400 FEG and a Zeiss Gemini 500, both equipped with an Oxford AZtecÒ EDS system. Both SEMs used an automatic particle search to analyse the impactor samples. All impactor-SEM samples were carbon-coated with a layer of 3–8 nm before being analysed with the SEM/EDS systems to avoid charging effects. The automated GSR analysis system performs the investigation in several subsequent steps: the sample area is divided into a grid of mechanical fields and the stage moves to the first field. The software acquires a backscattered electron (BSE) image of the field. In the BSE image, particles appear with different brightness depend-
ing on their material contrast (z-contrast). An image analysis is conducted by binarising the grey scale image using a predefined threshold. On all regions exceeding the threshold the electron beam will be positioned, inducing the emission of elemental specific X-rays. The X-rays are analysed using an EDS detector. The auto-ID function assigns the spectral peaks to the corresponding element and a quantification routine calculates the chemical composition of the particle. Subsequently, the particle is classified according to its composition to a predefined class. Finally the matched particles are summarised in a list. For the evaluation of this project only characteristic of (PbBaSb) and consistent with particles (PbBa, PbSb, and BaSb) and single lead particles were considered. The samples were also measured with a lower threshold than in normal case scenarios. This was an effort to be more sensitive to small particles and not to exclude low contrast GSR particles.
2.1.4. Diluter A VD100 diluter from KM OptoElektronik (Leonberg, Germany) was used to dilute the particle concentration of the cloud by 1:100 for the impactor for experiments with setup 2 [14].
76
R. Luten et al. / Forensic Chemistry 8 (2018) 72–81
Fig. 6. Particle distribution versus time; the area between the dotted lines indicates the major changes in particle concentration (experiment 2, setup 1, measured with particle counter).
Fig. 7. Particle distribution versus time; the dotted lines denote the maximum and the half maximum values of the particle concentration (experiment 2, setup 1, measured with particle counter).
2.2. Behavior of GSR particles in a closed room after discharge For the examination of the GSR distribution in the air experimental setup 1 was used. The air temperature and humidity in the room were monitored during the experiments. No relevant changes were observed. A series of experiments was carried out with the particle counter to determine the time for complete propagation of the GSR cloud. Because of the absence of a peak shortly after the discharge (Figs. 3 and 6) when the shock wave has reached the wall/ceiling, it can be concluded that no surface contaminations of previous discharge events were released.
The impactor was additionally used for sampling at discrete times. Figs. 3–8 show two examples. Measurements were started before the shot was fired. The shooter waited in the room until given the signal for shooting and afterwards walked out of the room at a slow pace. Impactor samples were taken at 5, 10, 20, and 60 min and 21 h after the shot in experiment 1 (Fig. 5) and at 20, 60, 100, 170, 250, and 340 min after discharge (Fig. 8) in experiment 2. As a part of experiment 2 (setup 1), a further test was carried out to check if a person gets contaminated with GSR after walking into a room where a firearm was discharged 200 min prior to entry. A piece of cotton (24 20 cm), which was checked for contamination before the test, was placed on the shoulder of the person, who then walked
R. Luten et al. / Forensic Chemistry 8 (2018) 72–81
77
P Fig. 8. Size class distribution ( = 100%) versus time after discharge for PbBaSb particles (experiment 2, setup 1, collected with the impactor, measured with SEM/EDS).
into the shooting range for two minutes at a slow pace. This piece of clothing was investigated using the tape-lift technique and measured with the automatic SEM/EDS analysis (Fig. 9). 2.3. Density of a GSR cloud A series of experiments was conducted to look at the density, composition, and distribution of a GSR cloud. Experimental setup 2 was used to sample directly after discharge. Samples were taken both with the impactor and the particle counter (Figs. 10–13). 2.4. Influence of height on the GSR propagation In order to evaluate the influence of height on the particle concentration and distribution another experiment was conducted.
Experimental setup 3 was used and particle counter measurements were performed (Fig. 14).
3. Results and discussion 3.1. Behavior of GSR particles in a closed room after discharge The distance between the muzzle of the Glock pistol and the location of the suction tubes of the particle counter and the impactor was 3.25 m, which was the maximum distance to all sidewalls and the ceiling (exp. setup 1). To clarify the physical processes and time constants of the particle transport mechanism by diffusion, a series of experiments using the particle counter in short term intervals was conducted (Figs. 3, 4 and 6, 7). The results indicate that the frontline of the expanding GSR cloud driven by diffusion
78
R. Luten et al. / Forensic Chemistry 8 (2018) 72–81
Fig. 9. Size distribution plot of PbBaSb particles detected on fabric, worn for 2 min in the shooting range, 3 h after a shot, collected by tape-lift method, measured with SEM/ EDS.
Fig. 10. Image of a typical GSR cloud after app. 0.5 s (end of phase 1).
processes appears approximately 3 min earlier for the small particles (<1 mm) than for the larger ones (1 mm). The size of the particles was found to affect the time constants of distribution and sedimentation after a discharge. Smaller particles (e.g. 0.3 mm) take a longer time to fully sediment than the larger (e.g. 1.0 mm) particles (Figs. 4 and 7). However, the position
where the measurement was taken influenced the results achieved: these measurements were taken at a height of 1 m above the floor. Generally, the propagation process of the GSR cloud can be divided into three phases. In phase 1 the emitted GSR cloud expands as long after discharge as the kinetic energy associated with the shot had dissipated. After 0.5 s the elongation of the cloud ends at approximately 2 m in direction of the shot with a diameter of circa 1.2 m (a nearly cylindrical shape, Fig. 10). In phase 2 the GSR particles in this cylinder are distributed into the entire room volume by diffusion. Depending on the particle size and density, or physically the inertia, the diffusion rate is different. The dilution of particles in the whole room ends when all regional concentration differences are equalised. The phase 3 is the reduction of airborne particles by sedimentation. The three phases take place not successively but simultaneously, making the physical description complicated. The results do not follow an easy mathematic model like an exponential function. That is the reason why we introduced only a ‘‘1st half-life” instead of a constant half-life. The maximum of the smaller particles shows a delay to all others because the sedimentation takes longer and the aerosol concentration increases in the lower third
Fig. 11. Particle size distribution of GSR cloud measured with particle counter directly at place of discharge (experimental setup 2).
R. Luten et al. / Forensic Chemistry 8 (2018) 72–81
79
Fig. 12. Particle size distribution histogram of PbBaSb particles collected at place of discharge (0 min, experimental setup 2) and 3.25 m away (20 min after discharge, experimental setup 1), both collected by impactor and measured with SEM/EDS (left ordinate: place of discharge, right ordinate: 3.25 m away).
of the room where the nozzles of the suction tubes were located. The particle concentration is constant in the still room without a discharge. It can be assumed that concentration changes are only related to the emission of GSR particles by the discharge of a firearm. The behavior of the GSR particle size versus time has been investigated with the impactor in two different experiments and is shown in Figs. 5 and 8. It can be clearly seen that the width of the particle size distribution decreases with increasing time. This confirms that smaller particles (<1.0 mm) take longer time to fully sediment than larger ones. The results from the contamination experiment (see 2.2) show more than 300 characteristic GSR particles on the fabric placed on the shoulder, most of them smaller than 1 mm (Fig. 9). This confirms that contamination can occur even after three hours which is longer than previously assumed [7]. 3.2. GSR cloud The composition of the cloud was measured at two different positions in the room (experimental setup 1 and 2, Fig. 2). It can be observed that more characteristic particles were found at the place of discharge in comparison to a position of 3.25 m away (Fig. 13). However, at the place of discharge only the middle part of the cloud was sampled (see Fig. 10 for the different cloud parts). At a distance of 3.25 m from the discharge the particles had time enough to distribute throughout the room and obtain an equal concentration, resulting in a more accurate representation of the GSR population at this sampling location. However, these results come from two different shots which might also affect the different GSR composition. Particle sizes were also compared at both positions (experimental setups 1 and 2). The results for PbBaSb particles are given in Fig. 12. More particles are found at the place of discharge (experimental setup 2) than at a distance of 3.25 m (experimental setup 1). The particle counter shows a similar size distribution for experimental setup 2 compared to the impactor at setup 2 (Fig. 11). At the place of discharge the measurement is not influenced by separation processes including sedimentation and diffusion. By con-
trast the second impactor sample taken at 3.25 m from the shooter’s position is affected by all processes that separate different size classes. Different diffusion velocities and sedimentation tendencies are responsible for a different size distribution histogram. It can be presumed that large particles (1 mm) are not completely diluted in the whole room volume 20 min after the shot. Due to the layering effect, the equalisation process of the GSR distribution is incomplete. In Fig. 14 the effect of different propagation velocities is obvious. Near the floor the fastest propagation is observed. For an aerosol including large particles, the tendency of sinking down is responsible for a faster propagation near the ground. All particles near the ceiling propagate quite fast because of convection e.g. due to room lighting. For particle sizes <1 mm an equilibrium of the concentration is obtained after appr. 20 min (Fig. 14). For particles >1 mm a different behavior of the particle concentration versus time regarding the four sampling heights is found. At heights of 1 m and 1.45 m, resp., it can be assumed that the propagation of the particles is mainly driven by diffusion processes. A rough estimation of particle numbers emitted by a single shot was performed based on the measurement at experimental setup 1 (3.25 m distance and 20 min after discharge). Exemplarily for sizes between 0.5 and 3.0 mm of all types of particles, a total number of particles for a shot was calculated. The values of the particle counter are related to 1 m3 of ambient air. When calculated for the entire room (65 m3) the values of the particle counter reveal 4.5 * 109 particles. For the impactor measurement – same position and same time after shot – 30 500 PbBaSb particles were found in the same size range. The sample volume of 33L related to the room volume of 65 m3 results in a factor of 2000. Considering this factor, a total number of 6.6 * 107 characteristic particles can be derived. This corresponds to 1.5% of PbBaSb particles as a portion of total particle population emitted by a single shot. 4. Conclusion This project was carried out to investigate the behavior of GSR particles in a closed room after discharge of a firearm. All results are related to the experimental setup: a Glock 19 pistol and Geco
80
R. Luten et al. / Forensic Chemistry 8 (2018) 72–81
Fig. 13. Comparison of the GSR composition at different positions in the shooting range; A): measured directly at the place of discharge (0 min, experimental setup 2), B) measured at 3.25 m away from and 20 min after the discharge (experimental setup 1), both collected by impactor and measured with SEM/EDS. The color code for the particle classes in the pie charts are the same as in the bar charts.
R. Luten et al. / Forensic Chemistry 8 (2018) 72–81
81
shooting range, showed more than 300 (mostly small) threecomponent GSR particles. Presently, small particles (0.6 mm e.g.) are not reliably detected in SEM-EDS measurements due to the compromise between suppressing environmental particles and being sensitive for gunshot residue. With respect to statistical confirmation, further experiments are already projected in order to comprehend the well-known variability of GSR emission and behavior. Acknowledgements This work was partly carried out as part of a project placement by Rebecca Luten in partial fulfilment of her MRes Forensic Science, Department of Forensic & Analytical Science, King’s College London. We are very grateful to Mr. Luke Haag for thorough proofreading this article. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.forc.2018.02.005. References
Fig. 14. Comparison between 0.3 mm, 1.0 mm, and 4.0 mm particles at different heights versus time after discharge (experimental setup 3), measured with particle counter.
9 mm Luger FMJ ammunition, discharged in a still room of 65 m3 volume. It is expected that other weapons and/or ammunitions will give similar results regarding the propagation parameters of the GSR cloud. The main conclusion is that airborne GSR particles can be found much longer in the air than previously assumed [7]. Our findings show that several hours after discharge GSR particles are still airborne. The sedimentation rate, given as the first half-life time, differs for small particles and larger ones. In contrast to Ref. [7], this study shows that there is a risk of contamination three hours after discharge. A piece of clean fabric, worn on the shoulder of a person walking for two minutes in the
[1] O. Dalby, D. Butler, J. Birkett, Analysis of gunshot residue and associated materials-a review, J. Forensic Sci. 55 (4) (2010) 924–943. [2] J. Yang, A. Gunn, T. Palmbach, D. Wei, S. Sinha, Nano forensics: nanoparticles in gun-shot-residue, in: IEEE Conference on Emerging Technologies – Nanoelectronics, 2006, pp. 269–272. [3] W.B. Hill, Transmission electron microscopy study of gunshot-residue nanoparticles collected in air samples, Microscope 60 (3) (2012) 133–137. [4] U. Brede, R. Hagel, K.H. Redecker, W. Weuter, Primer compositions in the course of time: from black powder and SINOXID to SINTOX compositions and SINCO booster, Propellants Explos. Pyrotech. 21 (1996) 113–117. [5] ASTM Standard E1588-17, Standard guide for gunshot residue analysis by scanning electron microscopy/energy dispersive X-ray spectrometry https:// www.astm.org/Standards/E1588.htm (Jan. 30, 2018) [6] ENFSI Working Group Firearms: ENFSI- Guide for Gunshot Residue Analysis by Scanning Electron Microscopy/ Energy-Dispersive X-ray Spectrometry, 1-11. https://epe.europol.europa.eu/documents/2632715/3098991/ENFSI-Guide_ V2.pdf/5784f76f-8458-4515-96f7-38566daa6e01?version=1.0 (Jan. 30, 2018) [7] L. Fojtásek, T. Kmjec, Time periods of GSR particles deposition after dischargefinal results, Forensic Sci. Int. 153 (2005) 132–135. [8] F.S. Romolo, P. Margot, Identification of gunshot residue: a critical review, Forensic Sci. Int. 119 (2001) 195–211. [9] J. Andrasko, S. Pettersson, A simple method for collection of gunshot residues from clothing, J. Forensic Sci. Soc. 31 (1991) 321–330. [10] V. Marple, B. Olson, Aerosol Measurement: Principles, Techniques, and Applications, in: P. Kulkarni, P. Baron, K. Willeke (Eds.), Sampling and measurement using inertial, gravitational, centrifugal, and thermal techniques, John Wiley & Sons, Inc, 2011 (Ch. 8). [11] G.G. Newton, O.G. Raabe, B.B. Mokler, Cascade impactor design and performance, J. Aerosol Sci. 8 (1977) 339–347. [12] O.G. Raabe, D.A. Braaten, R.L. Axelbaum, S.V. Teague, T.A. Cahill, Calibration studies of the drum impactor, J. Aerosol Sci. 19 (2) (1988) 183–195. [13] http://www.km-optoelektronik.de/downloads/Acsplus_English.pdf (Jan. 30, 2018) [14] http://www.km-optoelektronik.de/downloads/Vd100_English.pdf (Jan. 30, 2018)
Contents lists available at ScienceDirect
Forensic Chemistry journal homepage: www.elsevier.com/locate/forc
Investigating airborne GSR particles by the application of impactor technology Rebecca Luten, Dieter Neimke, Martin Barth, Ludwig Niewoehner ⇑ Bundeskriminalamt (BKA), Forensic Science Institute KT23, Thaerstraße 11, D-65193 Wiesbaden, Germany
a r t i c l e
i n f o
Article history: Received 15 November 2017 Received in revised form 16 February 2018 Accepted 20 February 2018 Available online 24 February 2018 Keywords: Gunshot residue (GSR) Impactor Airborne particles Still room Contamination risk Particle counter GSR distribution Sedimentation SEM-EDS Propagation
a b s t r a c t Gunshot residues (GSR) are of interest when firearms are used in criminal cases. GSR analyses are usually based upon the elemental composition and morphological appearance of very minute particles by means of SEM-EDS. Based on these two parameters, GSR particles are divided into specified classes. The amount of detected GSR particles depends on the time since discharge of the weapon and the sampling position relative to the location of discharge. In this paper, the influence of time on the local concentration and the distribution of airborne GSR particles were investigated with impactor technology. The particle concentration is constant in the still room; changes in concentration are only related to the emission of GSR particles by the discharge of a firearm. Here we showed that large quantities (50% of max. concentration) of airborne GSR particles can be detected several hours after discharge and contamination can take place as much as three hours after discharge. This study is a first approach to describe the propagation and sedimentation of GSR particles. With respect to statistical confirmation further experiments are already projected in order to comprehend the well-known variability of GSR emission and behavior. Ó 2018 Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons. org/licenses/by-nc-nd/4.0/).
1. Introduction Gunshot residues (GSR) are often investigated when firearms are used in a criminal case. GSR consists of small particles in the submicron to 10 mm range, which can be inorganic or organic [1–3]. These particles originate from the primer and the propellant, as well as material that was re-condensed from formerly evaporated material from the bullet, the cartridge case, and previous residues in the barrel, as well as abrasion particles from the same sources. Lead, antimony, and barium mainly originate from the primer in contemporary ammunition where they are used in the form of lead styphnate, antimony trisulphide, and barium nitrate, e.g. in Sinoxid ammunition by Dynamit Nobel, and many other ammunition manufacturers [4]. GSR is classified according to the ENFSI (European Network of Forensic Science Institutes) and ASTM (American Society for Testing and Materials) guidelines: characteristic of GSR and consistent with GSR [5,6]. Characteristic particles are at least three-component particles (with characteristic morphology), which are normally only found when a firearm has been discharged. Consistent with particles are two-component particles ⇑ Corresponding author. E-mail address: [email protected] (L. Niewoehner).
that, although uncommon, can also have their origin in the environment. The main topic in this article is the variation in concentration of GSR particles versus time in a closed, still room after discharge of a firearm. The particle concentration is constant in such a still room prior to the discharge of a firearm. It can be reasonably assumed that concentration changes are only related to the emission of GSR particles by this discharge. A time model of the behavior of GSR in this situation would be desirable, but is difficult to create as there are many different effects that influence the concentration of airborne GSR. Fojtásek and Kmjec (2005) tried to develop a time model when looking at the sedimentation of GSR particles [7]. Their results show that the main sedimentation process comes to an end after 8 min for the investigated size classes. Our study works on the principle of collected airborne GSR instead of sedimented GSR. The most common method to sample and analyse GSR is tape-lifting and the use of a scanning electron microscope coupled with an energy dispersive X-ray spectrometer (SEM/EDS) [1,8]. Tape-lifting is done on clothing or the hands of the shooter and cannot be used to collect airborne particles. However, Andrasko and Pettersson (1991) reported the use of a double filtration system in combination with an ordinary vacuum cleaner for the collection of GSR [9]. Here, two filters of 20 and 0.8 mm pore size
https://doi.org/10.1016/j.forc.2018.02.005 2468-1709/Ó 2018 Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
R. Luten et al. / Forensic Chemistry 8 (2018) 72–81
73
environmental air, and the particle density can be calculated. This will aid significantly in understanding the distribution of particles. Additionally, semi-continuous measurements were performed with a particle counter because the only source of particle emission is the discharge of the firearm. The particle counter is not able to distinguish between GSR-related and not-related particle concentrations. However, if the formerly measured background is subtracted the particle concentration can be observed over time. 2. Experimental – materials and methods
Fig. 1. Overview of the shooting range.
were used for the separation of particles. The calculated flow rate in the vacuuming system was around 6 L/min. In the present study an impactor was used with a triple separation system with cut off sizes at 10, 2.5 and 0.4 mm [10–12]. This flow rate in this system was 33 L/min. The flow rate is six times greater than that of Andrasko and Pettersson [9]. The impactor was used in combination with a pump with a controlled volume passing through the system instead of an ordinary vacuum cleaner. In this way the particle concentration can be related to a volume of the
The experiments were conducted in a closed shooting range of 6.6 4.3 2.3 m (L W H, V 65 m3). This shooting range was especially designed for shooting distance tests (Fig. 1). It is not used for any ballistic investigations and firearms identification tests. The air conditioning was switched off at least 10 min before shooting, the lights were turned off after the shot had been fired, so the only turbulence inside the room came from the shooter walking out at a slow pace. A Glock 19 pistol (barrel length of 102 mm) with Geco 9 mm Luger ammunition (a full metal jacket lead projectile with an open base, 8.0 g) was used for all the experiments. In all experiments the shooter was the same person, firing the pistol with both hands in three different experimental setups. Setup 1, shooter stands in the middle of the room and the two suction tubes, one for impactor and one for the particle counter are behind the shooter in the corner of the room at 1 m height (Fig. 2). This corresponds to a distance of 3.25 m between the shooter and the suction tubes. Setup 2, two suction tubes, one for impactor and one for particle counter are 20 cm to the right and 80 cm to the front of the shooter (Fig. 2). Setup 3 uses the same position as in setup 1; however, here at 4 different heights (0.5 m, 1 m, 1.45 m and 1.9 m) samples were only taken with the particle counter. 2.1. Equipment 2.1.1. Particle counter An ACS-Plus 228 from KM OptoElectronic GmbH (Leonberg, Germany) with detection sizes from: 0.2 mm to 1.0 mm with steps of 0.1 mm; 1.0 to 4.0 mm with steps of 0.5 mm; and 5 mm was used during the experiments [13].
Fig. 2. Setups of the experiments in the shooting range, X is the position of the suction tubes. Experimental setup 3 uses the position of setup 2.
74
R. Luten et al. / Forensic Chemistry 8 (2018) 72–81
Fig. 3. Particle distribution versus time; the area between the dotted lines indicates the major changes in particle concentration (experiment 1, setup 1, measured with particle counter).
Fig. 4. Particle distribution versus time; the dotted lines denote the maximum and the half maximum values of the particle concentration (experiment 1, setup 1, measured with particle counter).
The ACS-Plus from KM OptoElectronic is a desktop device. It counts particles in air with respect to their sizes. The particles have to cross a laser beam, where they scatter the light related to their sizes. A light sensitive diode placed orthogonal to the laser beam detects the intensity of the scattered light. By a comparator the signal height is related to a particle size range and counted as one event in that size channel. The system displays the number of detected particles per size class.
2.1.2. Impactor A Dekati PM10 Impactor in combination with a Satorius MD8 airscan pump was used during this project. The flowrate of the pump was fixed to 2.0 m3 per hour, which was six times higher than the calculated value mentioned in the paper by Andrasko and Pettersson [9]. The impactor was equipped with an antistatic tube. The impactor is a particle separation system with three collection stages widely used in aerosol investigations [10–12]. It sizes the
R. Luten et al. / Forensic Chemistry 8 (2018) 72–81
75
P Fig. 5. Size class distribution ( = 100%) versus time after discharge for PbBaSb particles (experiment 1, setup 1, collected with the impactor, measured with SEM/EDS); the diagram in the lower right corner shows the percentage ratio of three selected size classes.
particles in the air depending on their inertia. Particles with the highest inertia are deposited on the first impaction stage, the others follow the air stream to the next impaction stage. It has its own cutoff size, so it is known which particles will be detected, e.g. a cut-off of 2.5 lm means that everything larger will be expected to impact on it [10–12]. This cut-off value is based on completely spherical particles of low specific weight. Each stage is covered with a 1 inch double-sided adhesive tape to collect the impacting particles. 2.1.3. SEM/EDS Due to the large number of samples, two scanning electron microscopes were employed during this project. A FEI Quanta 400 FEG and a Zeiss Gemini 500, both equipped with an Oxford AZtecÒ EDS system. Both SEMs used an automatic particle search to analyse the impactor samples. All impactor-SEM samples were carbon-coated with a layer of 3–8 nm before being analysed with the SEM/EDS systems to avoid charging effects. The automated GSR analysis system performs the investigation in several subsequent steps: the sample area is divided into a grid of mechanical fields and the stage moves to the first field. The software acquires a backscattered electron (BSE) image of the field. In the BSE image, particles appear with different brightness depend-
ing on their material contrast (z-contrast). An image analysis is conducted by binarising the grey scale image using a predefined threshold. On all regions exceeding the threshold the electron beam will be positioned, inducing the emission of elemental specific X-rays. The X-rays are analysed using an EDS detector. The auto-ID function assigns the spectral peaks to the corresponding element and a quantification routine calculates the chemical composition of the particle. Subsequently, the particle is classified according to its composition to a predefined class. Finally the matched particles are summarised in a list. For the evaluation of this project only characteristic of (PbBaSb) and consistent with particles (PbBa, PbSb, and BaSb) and single lead particles were considered. The samples were also measured with a lower threshold than in normal case scenarios. This was an effort to be more sensitive to small particles and not to exclude low contrast GSR particles.
2.1.4. Diluter A VD100 diluter from KM OptoElektronik (Leonberg, Germany) was used to dilute the particle concentration of the cloud by 1:100 for the impactor for experiments with setup 2 [14].
76
R. Luten et al. / Forensic Chemistry 8 (2018) 72–81
Fig. 6. Particle distribution versus time; the area between the dotted lines indicates the major changes in particle concentration (experiment 2, setup 1, measured with particle counter).
Fig. 7. Particle distribution versus time; the dotted lines denote the maximum and the half maximum values of the particle concentration (experiment 2, setup 1, measured with particle counter).
2.2. Behavior of GSR particles in a closed room after discharge For the examination of the GSR distribution in the air experimental setup 1 was used. The air temperature and humidity in the room were monitored during the experiments. No relevant changes were observed. A series of experiments was carried out with the particle counter to determine the time for complete propagation of the GSR cloud. Because of the absence of a peak shortly after the discharge (Figs. 3 and 6) when the shock wave has reached the wall/ceiling, it can be concluded that no surface contaminations of previous discharge events were released.
The impactor was additionally used for sampling at discrete times. Figs. 3–8 show two examples. Measurements were started before the shot was fired. The shooter waited in the room until given the signal for shooting and afterwards walked out of the room at a slow pace. Impactor samples were taken at 5, 10, 20, and 60 min and 21 h after the shot in experiment 1 (Fig. 5) and at 20, 60, 100, 170, 250, and 340 min after discharge (Fig. 8) in experiment 2. As a part of experiment 2 (setup 1), a further test was carried out to check if a person gets contaminated with GSR after walking into a room where a firearm was discharged 200 min prior to entry. A piece of cotton (24 20 cm), which was checked for contamination before the test, was placed on the shoulder of the person, who then walked
R. Luten et al. / Forensic Chemistry 8 (2018) 72–81
77
P Fig. 8. Size class distribution ( = 100%) versus time after discharge for PbBaSb particles (experiment 2, setup 1, collected with the impactor, measured with SEM/EDS).
into the shooting range for two minutes at a slow pace. This piece of clothing was investigated using the tape-lift technique and measured with the automatic SEM/EDS analysis (Fig. 9). 2.3. Density of a GSR cloud A series of experiments was conducted to look at the density, composition, and distribution of a GSR cloud. Experimental setup 2 was used to sample directly after discharge. Samples were taken both with the impactor and the particle counter (Figs. 10–13). 2.4. Influence of height on the GSR propagation In order to evaluate the influence of height on the particle concentration and distribution another experiment was conducted.
Experimental setup 3 was used and particle counter measurements were performed (Fig. 14).
3. Results and discussion 3.1. Behavior of GSR particles in a closed room after discharge The distance between the muzzle of the Glock pistol and the location of the suction tubes of the particle counter and the impactor was 3.25 m, which was the maximum distance to all sidewalls and the ceiling (exp. setup 1). To clarify the physical processes and time constants of the particle transport mechanism by diffusion, a series of experiments using the particle counter in short term intervals was conducted (Figs. 3, 4 and 6, 7). The results indicate that the frontline of the expanding GSR cloud driven by diffusion
78
R. Luten et al. / Forensic Chemistry 8 (2018) 72–81
Fig. 9. Size distribution plot of PbBaSb particles detected on fabric, worn for 2 min in the shooting range, 3 h after a shot, collected by tape-lift method, measured with SEM/ EDS.
Fig. 10. Image of a typical GSR cloud after app. 0.5 s (end of phase 1).
processes appears approximately 3 min earlier for the small particles (<1 mm) than for the larger ones (1 mm). The size of the particles was found to affect the time constants of distribution and sedimentation after a discharge. Smaller particles (e.g. 0.3 mm) take a longer time to fully sediment than the larger (e.g. 1.0 mm) particles (Figs. 4 and 7). However, the position
where the measurement was taken influenced the results achieved: these measurements were taken at a height of 1 m above the floor. Generally, the propagation process of the GSR cloud can be divided into three phases. In phase 1 the emitted GSR cloud expands as long after discharge as the kinetic energy associated with the shot had dissipated. After 0.5 s the elongation of the cloud ends at approximately 2 m in direction of the shot with a diameter of circa 1.2 m (a nearly cylindrical shape, Fig. 10). In phase 2 the GSR particles in this cylinder are distributed into the entire room volume by diffusion. Depending on the particle size and density, or physically the inertia, the diffusion rate is different. The dilution of particles in the whole room ends when all regional concentration differences are equalised. The phase 3 is the reduction of airborne particles by sedimentation. The three phases take place not successively but simultaneously, making the physical description complicated. The results do not follow an easy mathematic model like an exponential function. That is the reason why we introduced only a ‘‘1st half-life” instead of a constant half-life. The maximum of the smaller particles shows a delay to all others because the sedimentation takes longer and the aerosol concentration increases in the lower third
Fig. 11. Particle size distribution of GSR cloud measured with particle counter directly at place of discharge (experimental setup 2).
R. Luten et al. / Forensic Chemistry 8 (2018) 72–81
79
Fig. 12. Particle size distribution histogram of PbBaSb particles collected at place of discharge (0 min, experimental setup 2) and 3.25 m away (20 min after discharge, experimental setup 1), both collected by impactor and measured with SEM/EDS (left ordinate: place of discharge, right ordinate: 3.25 m away).
of the room where the nozzles of the suction tubes were located. The particle concentration is constant in the still room without a discharge. It can be assumed that concentration changes are only related to the emission of GSR particles by the discharge of a firearm. The behavior of the GSR particle size versus time has been investigated with the impactor in two different experiments and is shown in Figs. 5 and 8. It can be clearly seen that the width of the particle size distribution decreases with increasing time. This confirms that smaller particles (<1.0 mm) take longer time to fully sediment than larger ones. The results from the contamination experiment (see 2.2) show more than 300 characteristic GSR particles on the fabric placed on the shoulder, most of them smaller than 1 mm (Fig. 9). This confirms that contamination can occur even after three hours which is longer than previously assumed [7]. 3.2. GSR cloud The composition of the cloud was measured at two different positions in the room (experimental setup 1 and 2, Fig. 2). It can be observed that more characteristic particles were found at the place of discharge in comparison to a position of 3.25 m away (Fig. 13). However, at the place of discharge only the middle part of the cloud was sampled (see Fig. 10 for the different cloud parts). At a distance of 3.25 m from the discharge the particles had time enough to distribute throughout the room and obtain an equal concentration, resulting in a more accurate representation of the GSR population at this sampling location. However, these results come from two different shots which might also affect the different GSR composition. Particle sizes were also compared at both positions (experimental setups 1 and 2). The results for PbBaSb particles are given in Fig. 12. More particles are found at the place of discharge (experimental setup 2) than at a distance of 3.25 m (experimental setup 1). The particle counter shows a similar size distribution for experimental setup 2 compared to the impactor at setup 2 (Fig. 11). At the place of discharge the measurement is not influenced by separation processes including sedimentation and diffusion. By con-
trast the second impactor sample taken at 3.25 m from the shooter’s position is affected by all processes that separate different size classes. Different diffusion velocities and sedimentation tendencies are responsible for a different size distribution histogram. It can be presumed that large particles (1 mm) are not completely diluted in the whole room volume 20 min after the shot. Due to the layering effect, the equalisation process of the GSR distribution is incomplete. In Fig. 14 the effect of different propagation velocities is obvious. Near the floor the fastest propagation is observed. For an aerosol including large particles, the tendency of sinking down is responsible for a faster propagation near the ground. All particles near the ceiling propagate quite fast because of convection e.g. due to room lighting. For particle sizes <1 mm an equilibrium of the concentration is obtained after appr. 20 min (Fig. 14). For particles >1 mm a different behavior of the particle concentration versus time regarding the four sampling heights is found. At heights of 1 m and 1.45 m, resp., it can be assumed that the propagation of the particles is mainly driven by diffusion processes. A rough estimation of particle numbers emitted by a single shot was performed based on the measurement at experimental setup 1 (3.25 m distance and 20 min after discharge). Exemplarily for sizes between 0.5 and 3.0 mm of all types of particles, a total number of particles for a shot was calculated. The values of the particle counter are related to 1 m3 of ambient air. When calculated for the entire room (65 m3) the values of the particle counter reveal 4.5 * 109 particles. For the impactor measurement – same position and same time after shot – 30 500 PbBaSb particles were found in the same size range. The sample volume of 33L related to the room volume of 65 m3 results in a factor of 2000. Considering this factor, a total number of 6.6 * 107 characteristic particles can be derived. This corresponds to 1.5% of PbBaSb particles as a portion of total particle population emitted by a single shot. 4. Conclusion This project was carried out to investigate the behavior of GSR particles in a closed room after discharge of a firearm. All results are related to the experimental setup: a Glock 19 pistol and Geco
80
R. Luten et al. / Forensic Chemistry 8 (2018) 72–81
Fig. 13. Comparison of the GSR composition at different positions in the shooting range; A): measured directly at the place of discharge (0 min, experimental setup 2), B) measured at 3.25 m away from and 20 min after the discharge (experimental setup 1), both collected by impactor and measured with SEM/EDS. The color code for the particle classes in the pie charts are the same as in the bar charts.
R. Luten et al. / Forensic Chemistry 8 (2018) 72–81
81
shooting range, showed more than 300 (mostly small) threecomponent GSR particles. Presently, small particles (0.6 mm e.g.) are not reliably detected in SEM-EDS measurements due to the compromise between suppressing environmental particles and being sensitive for gunshot residue. With respect to statistical confirmation, further experiments are already projected in order to comprehend the well-known variability of GSR emission and behavior. Acknowledgements This work was partly carried out as part of a project placement by Rebecca Luten in partial fulfilment of her MRes Forensic Science, Department of Forensic & Analytical Science, King’s College London. We are very grateful to Mr. Luke Haag for thorough proofreading this article. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.forc.2018.02.005. References
Fig. 14. Comparison between 0.3 mm, 1.0 mm, and 4.0 mm particles at different heights versus time after discharge (experimental setup 3), measured with particle counter.
9 mm Luger FMJ ammunition, discharged in a still room of 65 m3 volume. It is expected that other weapons and/or ammunitions will give similar results regarding the propagation parameters of the GSR cloud. The main conclusion is that airborne GSR particles can be found much longer in the air than previously assumed [7]. Our findings show that several hours after discharge GSR particles are still airborne. The sedimentation rate, given as the first half-life time, differs for small particles and larger ones. In contrast to Ref. [7], this study shows that there is a risk of contamination three hours after discharge. A piece of clean fabric, worn on the shoulder of a person walking for two minutes in the
[1] O. Dalby, D. Butler, J. Birkett, Analysis of gunshot residue and associated materials-a review, J. Forensic Sci. 55 (4) (2010) 924–943. [2] J. Yang, A. Gunn, T. Palmbach, D. Wei, S. Sinha, Nano forensics: nanoparticles in gun-shot-residue, in: IEEE Conference on Emerging Technologies – Nanoelectronics, 2006, pp. 269–272. [3] W.B. Hill, Transmission electron microscopy study of gunshot-residue nanoparticles collected in air samples, Microscope 60 (3) (2012) 133–137. [4] U. Brede, R. Hagel, K.H. Redecker, W. Weuter, Primer compositions in the course of time: from black powder and SINOXID to SINTOX compositions and SINCO booster, Propellants Explos. Pyrotech. 21 (1996) 113–117. [5] ASTM Standard E1588-17, Standard guide for gunshot residue analysis by scanning electron microscopy/energy dispersive X-ray spectrometry https:// www.astm.org/Standards/E1588.htm (Jan. 30, 2018) [6] ENFSI Working Group Firearms: ENFSI- Guide for Gunshot Residue Analysis by Scanning Electron Microscopy/ Energy-Dispersive X-ray Spectrometry, 1-11. https://epe.europol.europa.eu/documents/2632715/3098991/ENFSI-Guide_ V2.pdf/5784f76f-8458-4515-96f7-38566daa6e01?version=1.0 (Jan. 30, 2018) [7] L. Fojtásek, T. Kmjec, Time periods of GSR particles deposition after dischargefinal results, Forensic Sci. Int. 153 (2005) 132–135. [8] F.S. Romolo, P. Margot, Identification of gunshot residue: a critical review, Forensic Sci. Int. 119 (2001) 195–211. [9] J. Andrasko, S. Pettersson, A simple method for collection of gunshot residues from clothing, J. Forensic Sci. Soc. 31 (1991) 321–330. [10] V. Marple, B. Olson, Aerosol Measurement: Principles, Techniques, and Applications, in: P. Kulkarni, P. Baron, K. Willeke (Eds.), Sampling and measurement using inertial, gravitational, centrifugal, and thermal techniques, John Wiley & Sons, Inc, 2011 (Ch. 8). [11] G.G. Newton, O.G. Raabe, B.B. Mokler, Cascade impactor design and performance, J. Aerosol Sci. 8 (1977) 339–347. [12] O.G. Raabe, D.A. Braaten, R.L. Axelbaum, S.V. Teague, T.A. Cahill, Calibration studies of the drum impactor, J. Aerosol Sci. 19 (2) (1988) 183–195. [13] http://www.km-optoelektronik.de/downloads/Acsplus_English.pdf (Jan. 30, 2018) [14] http://www.km-optoelektronik.de/downloads/Vd100_English.pdf (Jan. 30, 2018)