 |
 |
| |
FULLY
AUTOMATED GUNSHOT RESIDUE ANALYSIS
Introduction
When a gun is fired, hot gasses and small particles
are generated during the explosion of the primer and the propellant
in the cartridge casing. These particles, as well as those
formed by the condensation of some of the hot gases, are deposited
on parts of the hand and clothes of the person firing the
gun, as well as on other objects in the vicinity of the weapon.
|
|
|
| |
| These small particles are called
gunshot (GSR) or firearms residue. Because they have distinctive
characteristics, their presence forms evidence that examined
persons have been close to a gun being fired. This kind of
evidence is often used in criminal court cases involving the
use of firearms.The residues contain elements that result
from the propellant and primer decomposing, as well as from
the bullet and cartridge case. Samples of GSR are taken from
people suspected of firing a gun, by pressing adhesive material
to the areas of interest, for example, fingers, the top part
of the hand and clothes. In this way particles are transferred
from the area of interest onto a small sample for examination
in a scanning electron microscope (SEM) connected to an X-ray
micro-analysis system equipped with a GSR software package
developed by Eastern Analytical. This software package allows
automated, unattended classification of GSR particles.
Graphical user interface of the GSR
software, showing a typical spectrum GSR 3-component particle
in combination with the stage configuration set-up.
|
| |
Finding the GSR particles
The characteristic particles for GSR often
contain combinations of the elements Lead (Pb), Barium
(Ba) and Antimony (Sb). In general these particles are
very small - in the range of 0.5 -10 microns. In the
scanning electron microscope, these particles can be
found, magnified and analyzed chemically, and the presence
of GSR on a sample confirmed.
The elements constituting GSR can originate
from the primer, the bullet, a coating or jacket on
the bullet, cartridge components, and previous residues
in the barrel. |
|
Typical GSR three-component particle. |
| |
Residue particles are
arbitrarily divided into two types, depending on their
primary source: primer particles and bullet particles.
Particles containing Pb, Sb and Ba and those containing
Sb and Ba (also referred to as two-component particles)
are typical inorganic components from the primer, and
are classified as primer particles. Large numbers of
particles composed mainly of Pb originate from the bullet
and are termed bullet particles.
In normal GSR suspect-taping, the particles
on the adhesive tape are typically less than 10 microns
in size (usually 0.5 - 2 microns). There are relatively
few particles on the taping, divided over a relatively
large area. As the majority of the GSR particles have
a small diameter (less than 2 microns), a large surface
has to be analyzed in order to yield results of sufficient
statistical validity. Because the sample is often not
smooth and contains particles which are of no particular
interest, the task of locating and analyzing GSR particles
manually by operator control is not practical and is
a tedious and time-consuming task which runs a high
risk of introducing errors. For these reasons, an automatic
routine has been developed that allows unattended analysis
(overnight) followed by operator-assisted validation
and summarizing of results. |
|
| |
The automatic routine
The analysis of gunshot residue can be performed
using dedicated software such as GSR-XT, which ensures fast,
reliable and fully unattended analysis.
The GSR-XT software program also identifies other indicative
particles similar to the large amounts of Lead (Pb) particles
from the bullet, such as Titanium (Ti) and Zinc (Zn) which
are found in lead-free ammunition. Other environment-related
particles could also provide additional information about
the subject. |
|
| 
Low magnification image using the backscattered electron
detector thresholds to seperate and locate particles.
|
|
The steps to
perform an analysis are as follows:
Tape is mounted on the stubs
Stubs are placed on a multi-stub holder
Multi-stub holder is mounted on the SEM specimen stage
The stub is split into multiple fields which are scanned
one-by-one
Each area is scanned for particles using backscattered
imaging
The position and size of the candidate particles are
recorded
The particles are revisited by the electron beam to
perform X-ray analysis
The emitted X-ray spectrum from each individual particle
is compared to a classification scheme
All results are stored (on disc)
A report of all detected particles is produced after
each sample.
The backscattered electron signal is used to detect
the particles. Typical GSR particles appear as bright
particles within the image. By applying a threshold
to the signal, the particles are distinguished and located
by the GSR software. |
|
| |
Description
The dedicated GSR analysis package is a fully
embedded software package for analytical scanning electron
microscopes of different manufacture.
An integral part of the software package is an advanced automatic
calibration procedure which allows complete set-up and dynamic
readjustment of the microscope column, motorized stage, backscattered
electron detector (BSD), and X-ray analysis system without
the need for operator intervention. The system is designed
to operate successfully using either the Robinson scintillator
or solid-state backscattered electron detectors.
If the intensity of the electron beam varies during a run,
GSR particles may no longer fall within the pre-selected threshold
level. To cope with this contingency, a calibration is performed
between samples. This is also helpful during long runs to
guarantee reliability of long-run analysis. Because of height
differences between samples, autofocus can be applied for
each stub or even for each field. Through calibration, the
SEM column as well as the BSD can be dynamically adjusted
to yield reproducible results. |
| |
| Set-up
and Calibration
X-ray excitation conditions (beam electron
intensity measured in terms of X-ray counts per second)
are calibrated first, using a standard to ensure reproducible
results. Before a run starts, the SEM column parameters
(kV, spot size, etc) are automatically adjusted. After
each sample stub is completed, the stage can relocate
to the standard and perform a new calibration. Following
X-ray count-rate adjustment, the background detector
is calibrated with relation to Z (atomic number), using
the calibration standard delivered with the system.
The purpose of this is to search for particles based
on the average Z, and not to rely on their video intensity
as in most other systems. This is a key feature of the
GSR software. Particles are detected based on their
average atomic number falling within a user-defined
Z-window. The SEM motorized stage is automatically driven
to each standard, and the video levels are measured
without changing any column conditions. The BSD response
curve is then used to set upper and lower threshold
criteria for particle detection by the BSD, based on
the average atomic number. This serves as input for
the preliminary search on particles in the automatic
routine. The calibration process may be performed unattended
during the course of a run in cases where the SEM gun
is unstable.
|
|
GSR Calibration Stub. |
|
| |
|
Substage layout page to store the 16 sample positions
(regular mode).
|
|
Analysis
The calibration procedure is one of the key
functions of the GSR-NT software. After calibration
of the SEM system, batches of GSR samples are configured
to begin a run.
Samples can be arranged on the stage in a regular array,
or in an irregular array of up to 16 samples. In both
cases, the operator can use the pre-defined substage
layout. This allows fast and accurate recording of the
sample positions without moving the stage.
|
|
| |
|
Sample configuration set-up page.
|
|
Information such as sample size, search area, sample
array, number of samples, minimum size of search particles,
and percentage of stubs to be analyzed are defined.
Search "stop" criteria and "run-time"
options such as maximum number of particles per sample
and "pause" are available. |
|
| |
| A typical GSR run requirement
is to search for particles 0.5 microns or greater in size
over the 100 mm² surface of a taping. The GSR software
optimizes the search by calculating the search magnification
based on the minimum particle size sought. For a search of
100 mm², an array of 130 fields is needed, using a typical
magnification of 120x, with a confidence level of 2x2 pixels
per particle.
"A confidence level of 2x2 pixels per particle"
means that in order for a group of bright neighboring particles
to be classified as a particle, there must be at least two
neighboring bright pixels in the X and two in the Y direction.
In order to search at as low a magnification as possible,
there is the option of defining a particle as one pixel (typical
magnification is 60x for 0.5 micron particles or 30x for 1.0
micron particles). The latter results in a search of only
9 fields for 100 mm² taping. Using the fast scan on the
SEM, the empty fields can be searched in as little as 60 minutes
for a 10 mm x 10 mm field, giving a fourfold improvement over
the generic scan rate. Operating at one pixel per particle
as a minimum particle-definition criterion will yield the
faster scan, but may not be as reliable if there are very
small GSR particles present. It can, however, be used as the
first fast indication of the suspect sample. |
| |
|
Classification scheme for fixed and user definable classes.
|
|
Classification
The position and size of the candidate particles
is recorded automatically. A revisit by the electron
beam is used to analyze the particle. The emitted X-ray
spectrum of each particle is classified using the choice
of either Regions of Interest (ROI) or Spectrum Stripping
spectrum analysis (STRIP) classification schemes.
The ROI classification scheme uses ratios of ROI counts
data to determine the elements present in the spectrum.
The ROI scheme classifies each spectrum into one of
35 categories in accordance with the Metropolitan Police
GSR classification scheme. It is possible to select
different classes or configure the selected classes.
Up to 35 user-definable classes can be defined for each
run. Saving the settings as different configuration
files can increase the number of definable classes.
|
| |
The STRIP classification scheme performs
full background subtraction and iterative peak stripping
to derive intensity data for the elements present in the
spectrum. This classification scheme allows classes to
be constructed, by the user, according to tight or wide
chemical criteria, so that both fixed stoichiometric compounds
can be identified, as well as continuous-composition phases
typical of GSR particles.
After a run, particles can be revisited and reanalyzed
using a new, perhaps longer, acquisition-time spectrum,
and then reclassified. The reclassified data are merged
with the original data but preserving the original classification.
In addition, whole runs can be reclassified automatically
should the user decide to change or add a class definition
to the classification scheme. |
|
| |
Reporting
Once the run has started, critical data on each detected
particle are stored in spreadsheet format to yield a complete
results file. This file can be reviewed, particle by particle,
directly after the run has been completed or at a later date.
A single-page, easy-to-read report is automatically generated
by the GSR software after each sample has been analyzed. |
| |
|
Stub particle map to show all detected particles per
sample.
|
|
A complete GSR Parameter
List or Stub Particle Map can also be printed.
The Stub Particle Map shows an overview of all detected
particles, classified by type. Using the special GSR
Spectra Utility software, all detected and classified
particles can be investigated and printed in a handy
format, together with particle data. This fast
spectra-viewing utility helps to save time reviewing
the many individual spectra generated during the unattended
run. Fast single page report generation is possible,
including an image of the particle, its spectrum and
text. |
|
| |
| The ROI
classification scheme uses ratios of ROI counts data
to determine the elements present in the spectrum. The
ROI scheme classifies each spectrum into one of 35 categories
in accordance with the Metropolitan Police GSR classification
scheme. It is possible to select different classes or
configure the selected classes. Up to 35 user-definable
classes can be defined for each run. Saving the settings
as different configuration files can increase the number
of definable classes. |
|
GSR Spectra Utility for fast investigation of the
stored spectra.
|
|
| |
Relocation
Relocation and reconfirmation of class is often
required. Particles can be revisited easily for further
imaging and verification by a simple double click on the
particle number. The high-precision motorized stage of
a motorized SEM will center and zoom the particle automatically.
In this way each particle will be easily relocated. The
GSR analysis can be started within the GSR application
to acquire a spectrum over longer time, which can be saved
on disc and used in the final report. The accuracy of
the relocation is very important because of the small
size of many GSR particles (approximately 0.5 - 2 microns).
After relocation, the analysis for confirmation is made
at a magnification of about 2000 - 5000x. The motorized
stage of the SEM in a busy crime lab will make a few million
movements per year. A high-precision motorized stage is,
therefore, a must. The ability to search automatically
at a low magnification is a standard procedure included
in the GSR software program. This will help to reduce
the number of stage movements and reduce SEM maintenance
costs. |
| |
| 
Easy relocation of a particle is included in the GRS-NT
software.
|
| |
| Please note this GSR software
may not achieve full specifications on all models of SEM.
For further information on specific SEMs, or details on
the GSR software, please contact Eastern Analytical,
EDAX, FEI or your SEM supplier. |
|
|
|
|
|
