Standard Guide
for Using Scanning Electron
Microscopy/X-ray Spectrometry in Forensic Paint Examinations
Scientific
Working Group on Materials Analysis (SWGMAT)
Scope.......Referenced.......Definitions.......Significance
and Use.......Sample Handling
Sample Preparation.......Analytical
Procedures.......Documentation.......References
1.
Scope
This document
is an outline of methods for scanning electron microscopy intended
for use by forensic paint examiners. The methods employed by each
examiner and/or laboratory depends upon sample size, sample suitability,
and laboratory equipment. The term scanning electron microscopy
occasionally refers to the entire analytical system, including energy
dispersive X-ray spectrometry and/or wavelength dispersive X-ray
spectrometry. This guide does not cover the theoretical aspects
of many of the topics presented.
This guide does
not purport to address all of the safety concerns, if any, associated
with this technology. It is the responsibility of the analyst to
establish appropriate safety and health practices and to determine
the applicability of regulatory limitations prior to its use.
2.
Referenced
2.1. ASTM. Guide
for Forensic Paint Analysis and Comparison, ASTM E1610-94, 1994.
2.2. ASTM. Practice
for Receiving, Documenting, Storing, and Retrieving Evidence in
a Forensic Laboratory, ASTM E1492-92, 1992.
2.3. Scientific
Working Group on Materials Analysis. Forensic paint analysis and
comparison guideline, Forensic Science Communications [Online].
(July 1999). Available: www.fbi.gov/hq/lab/fsc/backissu/july1999/painta.htm
2.4. Scientific
Working Group on Materials Analysis. Trace evidence quality assurance
guidelines, Forensic Science Communications [Online]. (January
2000). Available: www.fbi.gov/hq/lab/fsc/backissu/jan2000/swgmat.htm
2.5. Scientific
Working Group on Materials Analysis. Trace evidence recovery guidelines,
Forensic Science Communications [Online]. (October 1999).
Available: www.fbi.gov/hq/lab/fsc/backissu/oct1999/trace.htm
3.
Definitions
Background
X-rays (Bremsstrahlung, braking radiation, continuous spectrum):
Nonspecific X-ray radiation with a continuous energy range from
zero up to the beam voltage. Background radiation results from the
deceleration of beam electrons in the atomic Coulombic field. A
typical X-ray spectrum consists of both a continuous background
and peaks from characteristic X-rays.
Backscattered
electrons (BE): Primary beam electrons that are scattered from
the sample after undergoing few inelastic interactions. The probability
of backscattering is proportional to the atomic number.
Bulk analysis:
A type of scanning electron microscopy analysis that determines
the average elemental composition of a material. The area of analysis
is as large as possible and may be achieved by a single large area
raster or the summed results from multiple smaller area rasters.
Cathodoluminescence:
Emission of photons in the ultraviolet, visible, and infrared regions
of the electromagnetic spectrum as a result of electron beam interaction
with certain materials.
Characteristic
X-rays: X-ray emission resulting from de-excitation of an atom
following inner shell ionization. The energy of the X-rays is related
to the atomic number of the atom, providing the basis for energy
dispersive X-ray spectrometry. A typical X-ray spectrum consists
of both a continuous background and peaks from characteristic X-rays.
Charging:
Negative charge accumulation on either a nonconductive sample or
a sample that is not properly grounded. This effect may interfere
with image formation and X-ray analysis because of beam deflection.
It can usually be eliminated by the application of a conductive
coating.
Concentration:
For the purpose of this guide, the following ranges shall apply:
Major: greater than 10 percent; Minor: 1 to 10 percent; Trace: less
than 1 percent.
Detector
fluorescence peak (dead-layer peak, silicon internal fluorescence
peak): A peak resulting from the emission of characteristic X-rays
in a thin layer of inactive crystal area in the front of an energy
dispersive X-ray spectrometer detector. The peak is characteristic
of the type of detector, such as silicon for a lithium-drifted silicon
detector. In a silicon detector this peak may appear at 0.2 percent
apparent concentration.
Electron
probe microanalyzer (EPA, EPMA, EMMA): An electron beam instrument
designed for quantitative X-ray analysis (electron probe microanalysis).
It is related to scanning electron microscopy but with multiple
wavelength spectrometers and capable of operating at higher beam
currents. Electron probe microanalysis is the determination of elemental
concentration by X-ray emission from the microvolume of material
in which a static electron beam interacts.
Embedment:
The procedure of casting a sample in a block of material that polymerizes,
or otherwise hardens, to permit handling during further preparation.
Energy dispersive
X-ray spectroscopy (EDS, EDXA, EDX): X-ray spectroscopy based
on the measurement of the energy of X-rays. Energy dispersive X-ray
spectrometry is a complementary spectroscopy to wavelength dispersive
spectroscopy.
Escape peak:
A peak resulting from incomplete deposition of the energy of an
X-ray entering the energy dispersive X-ray spectrometer detector.
This peak is produced when an incoming X-ray excites a silicon atom
within the detector crystal, and the resulting Si K-alpha fluorescence
X-ray exits the detector crystal. It occurs at the principal peak
energy minus the energy of the Si K-alpha fluorescence X-ray (1.74KeV).
The escape peak intensity is about 1-2 percent of the parent peak.
Extraneous
material (contaminant, foreign material): Material originating
from a source other than the specimen.
Final aperture:
The last beam-restricting orifice in an electron optical column.
The orifice diameter influences the beam current and depth of focus.
Interaction
volume: The sample volume in which the electron beam loses most
of its energy. It is generally thought of as the volume in which
detectable X-rays are produced. The actual volume varies depending
upon beam voltage, average atomic number, and density of the sample.
Live time:
The time over which the energy dispersive X-ray spectrometry electronics
are available to accept and process incoming X-rays. Live time is
often expressed as a percentage of real time.
Microtomy:
A sample preparation method that sequentially passes a blade at
a shallow depth through a sample, resulting in sections of selected
thickness as well as a flat block. Each may be used for the determination
of sample characteristics.
Particle
analysis: An analytical method intended to determine the elemental
composition of a single particle such as a pigment particle in a
paint layer. Usually performed with a static (non-scanning) electron
beam.
Pulse processor
time constant: Operator-selected value for pulse-processing
time. A higher value (longer time) results in a more accurate determination
of the detector amplifier pulse height (better spectral resolution).
A lower value results in a higher count rate but with reduced spectral
resolution.
Raster:
The rectangular pattern scanned by the electron beam on a sample.
The raster dimensions change inversely with magnification.
Sample
(representative sample): A representative portion of the specimen
selected and prepared for analysis that is believed to exhibit all
of the elemental characteristics of the parent specimen.
Sample polishing:
A sample preparation method using progressively finer abrasives
to achieve a flat, smooth sample surface. Generally this is required
for quantitative analysis.
Sample size:
For the purposes of this document, the following terms are used
to describe sample size. The actual size demarcation between each
is somewhat arbitrary.
- Fragment:
A sample or specimen smaller than approximately 0.2 millimeters.
If the material from which the fragment originated was layered,
then the fragment may also show a layered structure with light
microscopy inspection and scanning electron microscopy analysis.
A fragment is frequently not of sufficient size to permit multiple
tests.
- Particle:
A sample or specimen whose greatest dimension is less than approximately
50 micrometers. Material of this size generally has none of the
overall structural characteristics that can be associated with
the material from which the particle originated. A particle is
generally not of sufficient size to permit multiple tests.
- Piece:
A sample or specimen larger than approximately 0.2 millimeters.
If the material from which the piece originated was layered, then
the piece may show a layered structure. A sample of this size
is sufficient to perform all of the suggested cross-sectional
preparation and analytical methods.
Scanning
electron microscopy (SEM): A type of electron microscope in
which a focused electron beam is scanned in a raster on a solid
sample surface. The strength of resulting emissions of signals vary
according to sample characteristics such as composition or topography.
These signals directly modulate the intensity of the display cathode
ray tube. The electron beam of the scanning electron microscope
and the display cathode ray tube are scanned synchronously, resulting
in a two-dimensional image of the sample. By popular usage, the
term scanning electron microscopy may also include the analytical
techniques energy dispersive X-ray spectrometry and wavelength dispersive
X-ray spectrometry.
Secondary
electrons (SE): Low-energy electrons produced from the interaction
of beam electrons and conduction band electrons of atoms within
the interaction volume. They are produced throughout the interaction
volume, but only those near the surface have enough energy to escape.
The secondary electron signal is typically used to form topographic
images.
Smear:
A transfer of paint resulting from contact between two objects and
consisting of co-mingled particles, fragments, and possible pieces
of one or both surfaces.
Specimen:
Material submitted for examination. Samples are removed from a specimen
for analysis.
Spectral
artifacts: Spectral peaks other than characteristic peaks, produced
during the energy dispersive X-ray spectrometry detection process.
Examples are escape peaks and sum peaks.
Spectral
resolution: A measure of the ability to distinguish between
adjacent peaks in an X-ray spectrum. It is usually determined by
measuring peak width at half the maximum value of the peak height
or full-width-half-maximum.
Sum peak:
A peak occurring at the sum of the energy of two individual peaks.
System dead
time: The time during which the energy dispersive X-ray spectrometer
is not able to process X-rays. Dead time is typically expressed
as a percentage of real time.
System peaks
(stray radiation): Peaks that may occur in the X-ray spectrum resulting
from interaction of the electron beam or fluorescent radiation with
components of the scanning electron microscope itself.
Take-off
angle: Angle between the specimen surface and the detector axis.
Thick section:
For the purposes of this guide, a sample that is two micrometers
or thicker.
Thin section:
For the purposes of this guide, a sample with a thickness of less
than two micrometers.
Transmission
electron microscopy (TEM): A type of electron microscopy in
which an image of a sample prepared as a thin section is formed
by the interaction of the beam passing through the sample.
Variable
pressure scanning electron microscopy (LV, CP, VP, ESEM): A
type of scanning electron microscopy that is designed to operate
at higher chamber pressure than the conventional. The need for application
of a conductive coating is minimized when using a variable pressure
scanning electron microscope; however, energy dispersive X-ray spectrometry
may be complicated because of the electron beam spread experienced
at higher operating pressures.
Wavelength
dispersive X-ray spectroscopy (WDS, WDXA): X-ray spectroscopy
that separates and identifies X-rays based on their differences
in wavelength. Wavelength dispersive X-ray spectrometry is a complementary
spectroscopy to energy dispersive X-ray spectrometry.
4.
Significance and Use
4.1. The scanning
electron microscope can be used to define and compare the layer
structure of multilayered samples, the structure of individual layers,
the bulk elemental composition of individual layers, and the elemental
composition of individual particulate components within paints and
coatings.
4.2. The methods
described in this guide may have some limitations. They include
the inability to detect elements in trace concentrations, the need
for a conductive coating of the sample, the inability to remove
a sample from most embedding materials after analysis, and the discoloration
of materials by irradiation.
4.3. Although
quantitative and semiquantitative methods are available for energy
dispersive X-ray spectrometry, they are not appropriate for most
paint analyses because of the typical heterogeneity of paint. Application
of quantitative methods is further complicated by an inability to
predict what compounds may be present (see Section 7.12).
4.4. The information
available from a specimen may diminish as its size is reduced, and
its condition degrades. The smaller a specimen is, the less valuable
it becomes for association with a known because it may contain fewer
characteristics of the original material. As specimen size is reduced,
it may no longer be representative of the original material. This
may also be true of a degraded sample.
4.5. This guide
is intended to advise and assist laboratory analysts in the effective
application of scanning electron microscopy to the analysis of paint
evidence. It is intended to be applicable to most modern scanning
electron microscopes typically used in the forensic laboratory.
4.6. It is not
the intention of this guide to present comprehensive methods of
scanning electron microscopy. It is necessary that the analyst have
an understanding of scanning electron microscopy operation and general
concepts of specimen preparation prior to using this guide. This
information is available from manufacturers' reference materials,
training courses, and references such as Scanning Electron Microscopy
and X-ray Microanalysis: A Text for Biologists, Materials Scientists,
and Geologists (Goldstein et al. 1992).
5.
Sample Handling
5.1. The general
collection, handling, and tracking of samples shall meet or exceed
the requirements of ASTM 1492-92 as well as the relevant portions
of the SWGMAT's Trace Evidence Quality Assurance Guidelines (www.fbi.gov/hq/lab/fsc/backissu/jan2000/swgmat.htm)
and Trace Evidence Recovery Guidelines (www.fbi.gov/hq/lab/fsc/backissu/oct1999/trace.htm).
5.2. The work
area and tools used for the preparation of samples must be free
of all materials that could transfer to the sample. Samples prepared
for scanning electron microscopy analysis should be maintained in
a protective container, such as a petri dish or box.
5.3. When samples
are prepared for scanning electron microscopy, a map identifying
sample location should be constructed. This may be in the form of
a sketch, a photomicrograph, or a captured video image and should
include an index mark on the mount.
6.
Sample Preparation
6.1. Samples
should first be examined with a stereomicroscope, noting size, structure,
overall homogeneity, and any material adhering to the sample.
6.2. The choice
of a specific method for sample preparation depends on the size,
nature, and condition of the specimen, as well as the analytical
request. It may be necessary to use multiple preparation methods
in order to analyze all sample characteristics.
6.3. In developing
a strategy for analysis, the following should be considered:
- determination
of the presence of extraneous materials and a strategy for removal
- method of
attachment to a scanning electron microscopy mount
- method(s)
for exposing internal structure, if the specimen is inhomogeneous
- method(s)
for producing a uniform geometry
- necessity
of applying a conductive coating to the prepared samples
- determination
of the presence of surface features of analytical interest
6.3.1. If an
analytical goal is to determine elemental composition, then any
possible contribution from extraneous materials must be eliminated.
6.3.2. If an
analytical goal is to determine structure, then the internal structure
must be exposed using an appropriate method.
6.3.3. For the
accurate comparison of elemental composition and structure, samples
must be prepared in the same manner.
6.3.4. Although
embedment with subsequent polishing or microtomy may be considered
labor-intensive, these methods permit precise, reproducible sample
preparation.
6.3.5. If sufficient
sample size permits, mounting flat, intact specimens may allow visualization
and analysis of surface features.
6.4. Recognition
and removal of extraneous materials
It is not unusual
for extraneous materials to be present on the surface of a specimen
submitted for analysis. Because the scanning electron microscopy
method is a surface analysis, the presence of even a small amount
of this material can prevent an accurate determination and comparison
of composition. Therefore, a strategy for the recognition and removal
or visualization and abatement of this material must be employed.
6.4.1. Depending
on sample size and type, extraneous material may be physically removed
with a brush, probe, or fine blade. Debris can also be lifted off
the sample with tape. Samples that are too small to be effectively
taped can be rolled on a thin adhesive layer. Care should be taken
that the adhesive does not adhere to the sample surface, which might
interfere with any subsequent organic or inorganic analysis. If
necessary, a fresh surface may be exposed by scraping or cutting
with a fine scalpel blade.
6.4.2. To immobilize
extraneous materials, the technique of embedment described in Section
6.6.1.4 is effective. Subsequent processing of the sample may then
proceed without direct concern for the extraneous materials.
6.4.3. When
extraneous materials cannot be removed and the sample is not embedded,
their location should be noted during light microscopy and/or backscatter
electron scanning electron microscopy observations. During analysis,
areas with extraneous material should be avoided. Note that some
surface extraneous materials may not be visible by light microscopy
alone.
6.5. Methods
of attaching a sample to a scanning electron microscope mount
6.5.1. All samples
to be analyzed in the scanning electron microscope must be attached
to some form of a scanning electron microscope mount. These mounts
are usually made of aluminum, carbon, beryllium, or brass. Because
the presence of a carbon peak in the spectrum does not usually interfere
with elemental comparisons, mounts constructed of carbon are preferred.
Carbon mounts are available either as spectroscopically pure or
pyrolytic. Pyrolytic carbon offers the advantages of a hard, flat,
glasslike surface that results in a featureless background when
imaged. Samples may be attached directly to a scanning electron
microscope mount, with the prior application of an adhesive layer.
Ideally, the adhesive should be organic with minimal inorganic content
and soluble in a solvent that evaporates rapidly. The adhesive may
be applied to the mount dropwise by a micropipette or spread into
a thin film by drawing out the drop with a coverslip. The thickness
of adhesive may be adjusted by regulating the size of the drop (Ward
1999).
6.5.2. Electrically
conductive carbon paints are commercially available and may be used
for directly attaching samples onto the surface of a scanning electron
microscope mount. The paints typically consist of micronized carbon
suspended in an organic solvent. A small streak of carbon paint
can be placed on the mount using a fine tipped brush while viewing
under a stereomicroscope at low magnification. The sample may then
be touched to the surface of the paint just before it goes to dryness
causing it to adhere to the surface of the mount with an electrically
conductive attachment.
6.5.3. Various
carbon conductive adhesives and double-sided tapes are commercially
available and may be used. Their elemental compositional purity
should be characterized prior to use (Wrobel et al. 1998).
6.6. Demonstration
of internal structure
6.6.1. For characterization,
the sample must be prepared so that the internal structure is exposed.
A variety of methods are presented in Sections 6.6.1.1 through 6.6.1.4.
If the specimen is too small to carve manually, pieces and fragments
may be prepared in cross section by freehand cutting, polishing,
or microtomy of the sample following embedment in a supporting material.
6.6.1.1. The
sample may be cut and attached on edge to a scanning electron microscope
mount or shaved after attaching to a scanning electron microscope
mount. This method is suitable only for large samples. This method
can be performed rapidly; however, layers can separate, extraneous
materials can be dragged onto the surface to be analyzed, and the
geometry between samples may not be consistent.
6.6.1.2. Some
samples may be slowly carved, exposing each individual layer. This
may be done by holding the sample in place, either with forceps
or in some other manner, then peeling the layers away with a clean,
sharp scalpel blade or diamond knife. The cutting tool must be held
at a very low angle to produce thin peels and to avoid excessive
pressure on the sample. Thin peels of the individual layers may
then be harvested and mounted by one of the techniques described
in Section 6.5. This method does require substantial sample manipulation
but provides the advantages of reproducible flat sample geometry,
no potential for the interaction volume to extend into neighboring
layers, and the availability of large analytical surface areas.
Sample size must be relatively large and preparation by this method
does not provide an opportunity to image a cross-section of the
specimen. Furthermore, detection of minor elemental constituents
requires longer analytical acquisition times owing to the reduced
analytical volume afforded by the thin peel.
6.6.1.3. Some
samples may be stair-stepped by cutting a layered structure on intralayer
planes and peeling to expose underlying layers for analysis. Although
this method can expose a large area of each layer for X-ray analysis
and potentially avoid spectral variations due to inhomogeneity,
the interaction volume may extend into an underlying layer. Sample
size must be relatively large, and preparation by this method does
not provide an opportunity to image a cross section of the specimen.
6.6.1.4.
Embedment
Prior to microtomy
or polishing, a sample is embedded in order to provide support.
The sample is placed in a mold with an identifying label, and the
mold is filled with embedding material that is allowed to polymerize
or harden. Several mold types are available, such as a silicone
flat holder, Beem® (Better Equipment for Electron Microscopy,
Bronx, New York) capsules, slotted stub, and ring mounts. Embedment
and subsequent exposure of the specimen's cross section offers the
advantages of abating extraneous materials, providing precise control
and manipulation of samples smaller than 0.2 millimeters, and processing
of several samples simultaneously. Disadvantages are the possibility
of selective removal of soft or soluble layers, trapping of polishing
materials, and extension of the beam interaction volume into the
adjacent layers when thin layers are encountered.
6.6.1.4.1.
Microtomy of embedded samples
Microtomes are
generally of two types: histomicrotome and ultramicrotome, either
of which may be used for the preparation of paints. A glass knife
is usually used in an ultramicrotome although diamond or tungsten
carbide knives may be used for hard materials. A steel or tungsten
carbide knife is used in a histomicrotome. In addition to producing
a flat sample block for subsequent scanning electron microscopy
analysis, sections may be cut for light microscopy, ultraviolet-visible
microspectrophotometry, and infrared microspectroscopy. Multiple
samples may be embedded in the same mold for microtomy. Their relative
positions must be indexed such that their cross sections may be
identified in the sample block. This may be accomplished by mounting
a taggant fragment in the mold and noting its position relative
to the questioned and known samples prior to microtomy. Microtomy
produces a sample block that is flat across the entire face. Slight
variations in take-off angle may exist between samples if embedded
in separate molds. Mounting samples for comparison in the same mold
minimizes these variations. In doing so, however, care must be taken
to assure that the paint fragments lie parallel to one another in
order that the beam/sample geometry between samples does not vary.
6.6.1.4.2.
Polishing of embedded samples
Polishing is
a process by which the embedded sample is exposed to a series of
successively finer abrasives. Individual paint samples may be embedded
in a single block or embedded individually and mounted in a holder
containing multiple sample slots. Individual embedding permits individual
sample height adjustment, whereas if several samples are mounted
simultaneously, only one final polishing plane is possible. Various
types and combinations of polishing materials are available and
suitable. Diamond abrasives, however, are recommended for the final
polish step because they do not leave particle residues that may
be mistaken for paint components. When paints are simultaneously
polished, the analyst is assured that each has been prepared in
the same manner. Each is equally flat, scratch-free, and in the
same plane. However, edge rounding may occur between areas of differing
hardness.
6.7. Uniform
geometry
6.7.1. If samples
are to be compared, the take-off angle of each specimen must be
similar. Only then are spectral differences indicative of differences
in the chemistry of the samples.
6.7.2. Similar
geometry can be achieved if the samples are microtomed or polished
simultaneously.
6.7.3. If microtomy
is selected as a preparation method and multiple blocks are used,
each block should be microtomed at a similar angle.
6.8. Generally it is necessary to apply a conductive layer to the
sample surface to eliminate charging. Carbon is preferred, because
the presence of a carbon peak in the spectrum usually does not interfere
with elemental comparisons.
7.
Analytical Procedures
7.1. Instrument
calibration
7.1.1. Prior
to beginning an analysis, verification of the operational condition
of the scanning electron microscope must be established. This includes
presence of system peaks, accuracy of magnification, and determination
of spectral energy calibration and resolution.
7.1.2. The presence
of system peaks is generally determined upon installation of the
scanning electron microscope or following a modification or addition
of accessories. Goldstein (Goldstein et al. 1992) describes a procedure.
7.1.3. For a
determination of accuracy of magnification, a percentage of error
of magnification must be calculated. A scanning electron microscope's
indicated value of magnification (such as a measurement marker)
is compared to a measurement of a certified standard (such as NIST
SRM 484D). A calibration check of the primary image output device
to the certified standard must be performed periodically and a record
kept in a permanent log. Relationships of measurements on display
monitors, as well as any other image capture applications to the
primary image output device, should also be recorded. Magnification
standards for scanning electron microscopes are commercially available,
with errors of less than five percent generally achievable.
7.1.4. Energy
calibration must be established frequently for the energy dispersive
X-ray spectrometer, including zero offset and gain, and a record
kept in a permanent log. Energy calibration may be determined directly
by measuring the centroid energy of a low- and high-energy peak
or determined automatically using software provided by the instrument
manufacturer. If automated methods are used, measured spectral energies
typically do not exceed 10eV from that of actual energies. Automatic
methods for calibration are described in documentation from the
manufacturer.
7.1.5. Spectral
resolution for the energy dispersive X-ray spectrometer must be
determined regularly and a record kept in a permanent log. This
may be determined automatically or can be determined manually by
measuring the width of the Mn K-alpha peak at half the maximum peak
height. Automatic methods for calibration as well as recommended
performance limits are often available from the manufacturer.
7.2. Structural
imaging
7.2.1. Light
microscopy is useful for defining layers and structures based on
color characteristics. The end-on view can be an edge-mounted, microtomed,
or embedded sample observed with reflected light. When the samples
have been prepared by a method producing a flat surface, such as
polishing or microtomy, the entire field of view is in focus at
high magnification. Mounted on a glass slide, the thick section
is observed with transmitted light. Light microscopy demonstrates
layer structure as well as some structural detail within each layer.
7.2.2. A backscatter
electron image is useful for defining layers and structures based
on the average atomic number of the matrix. For comparison purposes,
a magnification similar to that of the light microscopy image is
suggested. Higher magnification images may be useful for demonstrating
structural details.
7.2.3. A cathodoluminescence
image can also be used to provide structural information and discrimination
and is complementary to the light microscope and backscatter electron
images (Stoecklein and Goebel 1992).
7.2.4. Scanning
electron microscopy micrographs should include a measuring scale
or magnification scale or both. The micrograph should also display
which detector was used to produce the image (BE detector or SE
detector).
7.3. Selection
of scanning electron microscopy/energy dispersive X-ray spectrometry
operating conditions
7.3.1. The following
suggested operating conditions are meant as general guides for starting
conditions. As the analyst determines specific analytical needs,
actual requirements may vary.
7.3.1.1. A beam
voltage of 20-25KeV is an adequate compromise between the need for
sufficient over-voltage necessary for efficient X-ray excitation
and X-ray spatial resolution. Most of the X-ray lines produced may
be displayed with an energy range of 0 to 20KeV. The pulse processor
time constant should be set at a midrange value, which is a compromise
between maximum count rate and maximum spectral resolution. The
beam current should be adjusted to yield an X-ray detector dead
time of approximately 30 percent. A live time of 100-200 seconds
is usually sufficient to provide reasonable counting statistics
for minor peaks.
7.3.1.2. The
beam/sample/X-ray detector geometry should be optimized for X-ray
collection efficiency, particularly when attempting to analyze nonflat
samples.
7.3.1.3. Generally,
changes in the suggested initial conditions are required under the
following circumstances:
- Beam voltage
is increased when higher energy line excitation is required.
- Beam voltage
is decreased when greater spatial resolution is required.
- Pulse processor
time constant is lengthened when greater spectral resolution is
required.
- Pulse processor
time constant is shortened when a greater count rate is required
(e.g., for trace element analysis or construction of elemental
distribution maps).
- Detector
to sample distance can be reduced to increase X-ray collection
efficiency.
- Spectral
energy display scale is expanded when sufficient detail is not
evident.
- Beam current
is increased when the X-ray count rate is too low. Decreasing
the condenser lens current and/or increasing the final aperture
size may increase beam current.
- Beam current
is decreased when the X-ray count rate is too high. Increasing
the condenser lens current and/or decreasing the final aperture
size may decrease beam current.
7.4. Bulk
spectra collection
7.4.1. Once
the structure of the material is defined, the average or bulk elemental
composition of each layer is determined. The raster should include
as much area of a layer as possible. This may be achieved by analyzing
a single large area or summing the spectra from several smaller
areas.
7.4.2. Because
the X-ray analytical volume may be larger than the raster and because
X-ray fluorescence may be significant, the analyst must consider
the possibility of the contribution of X-rays produced in adjacent
layers.
7.5. Qualitative
analysis
7.5.1. Once
an X-ray spectrum is collected, a qualitative analysis is performed
in order to determine the elements present. The process is straightforward
for the peaks of elements present in major amounts and those not
overlapping. Misidentifications or omissions of minor components
are possible, however, unless a systematic approach to elemental
identification is used which includes consideration of X-ray line
families, spectral artifacts, escape peaks, sum peaks, and overlaps.
7.5.2. Reference
lines, or energies, may be obtained from several sources, including
energy slide rules, published tables, and computer-generated KLM
markers that may be superimposed on the spectrum. Additionally,
manufacturers often provide an automatic element identification
application. These aids often are used in complementary fashion.
7.5.3. Identification
begins with high-energy peaks and major peaks. High-energy peaks
are generally less likely to overlap than lower energy peaks. If
a major peak is present, generally a complete family of peaks can
also be identified. Each line within the family is labeled with
elemental symbols. Spectral artifacts, including sum peaks and escape
peaks associated with major peaks, should be evaluated and labeled.
7.5.4. As spectral
interpretation alternates between the identification of major and
minor peaks, the vertical (counts) scale should be adjusted to reveal
required detail. In addition to the higher energy peaks, the presence
of any lower energy families and their expected relative intensities
should be noted. Individual asymmetric peaks and inconsistent peak
ratios within a family may indicate a peak overlap. Elemental identification
is aided by superimposing and scaling KLM reference lines on the
spectrum or referencing the actual spectrum of an elemental standard.
The analyst should become familiar with the characteristic pattern
and relative intensities of peaks of various atomic numbers. The
identification of major components is usually straightforward.
7.5.5. Following
the identification of major components, lower intensity peaks and
overlapped peaks are identified. The number of characteristic peaks
present in a spectrum often limits minor element identification.
7.5.6. The presence
of an element can be considered unequivocal only when a distinctive,
unique set of lines is produced or when a single peak occurs at
an energy where it cannot be mistaken for another element or spectral
artifact. Unequivocal identification may not be possible if an element
is present in low concentration or if lines required for confirmation
are overlapped with the lines of other elements.
7.5.7. If an
identification is unequivocal, each individual peak is labeled with
the corresponding elemental symbols (and X-ray line if the software
permits). If the identification is probable but not unequivocal,
the peak label should indicate so (such as by parenthesizing the
elemental symbols).
7.5.8. Spectra
must be displayed on a scale that clearly demonstrates the peaks
identified. In order to display peaks from elements with significant
differences in concentration, the (small) peaks from the elements
in low concentration may be viewed by displaying the spectra separately
on different display scales.
7.5.9. If an
automatic identification application is used, the analyst must confirm
peak identification.
7.5.10. Although
the natural X-ray line width is approximately 2eV, energy dispersive
X-ray spectrometry resolution is generally no better than approximately
140eV. As a result, there may be an overlap of peaks in the energy
dispersive X-ray spectrometry spectrum of materials containing several
elements. Some commonly occurring overlaps encountered in energy
dispersive X-ray spectrometry are as follows: Ti K beta/V K-alpha,
V K-beta/Cr K-alpha, Cr K-beta/Mn K-alpha, Mn K-beta/Fe K-alpha,
Fe K-beta/Co K-alpha, Pb M-alpha/S K- alpha/Mo L-alpha, Ba L-alpha/Ti
K-alpha, K K-beta/Ca K-alpha, Zn L-alpha/Na K-alpha, P K-alpha/Zr
L-alpha, and Al K-alpha/Br L-alpha.
In order to
resolve these overlaps, several methods may be employed.
- The processing
time of the pulse processor may be increased to improve spectral
resolution.
- Mathematical
spectral subtraction (deconvolution) methods supplied by the energy
dispersive X-ray spectrometer manufacturer can be employed.
- A wavelength
dispersive X-ray spectrometer scan can be performed over the questioned
energy range.
- An alternative
method of elemental analysis or X-ray diffraction may be used.
7.5.11. Should
peaks remain that are not identified by the above scheme, a comprehensive
tabulation of X-ray line energies may be consulted.
7.6. Individual
component analysis
7.6.1. Additional
evaluation of composition may be achieved by the spot (nonrastered)
analysis of specific particles within layers. Generally, these particles
appear bright in the backscattered electrons image. Such an analysis
may improve the detection limit beyond that achievable by a bulk
analysis, as well as serve to associate elements detected by a bulk
analysis. For example, the bulk analysis of a paint layer may reveal
the presence of Al, Si, Mg, Ba, S, and O. Specific particle analysis
may associate the elements Si, Mg, and O as being present in one
type of particle, Ba, S, and O present in a second type, and Al,
Si, and O in a third type. These associated elemental compositions
would then indicate these particles to be consistent with talc,
barium sulfate, and clay, respectively. Polarized light microscopy,
infrared spectroscopy, or X-ray diffractometry can be used to confirm
the presence of some of the compounds.
7.6.2. Because
the beam interaction volume may be considerably larger than an individual
particle, inclusion of other matrix components may be expected in
the spectrum from an individual particle. Lower beam voltages may
be used to confine more of the interaction volume to the particle.
7.7. Wavelength
dispersive X-ray spectrometry analysis
7.7.1. Wavelength
dispersive X-ray spectrometry may be used in conjunction with energy
dispersive X-ray spectrometry to provide greater sensitivity (x10)
and resolution (5eV vs.140eV) of characteristic peaks (Bearden 1967).
Stand-alone spectrometers with multiple diffracting crystals can
be interfaced with to many scanning electron microscopes. Some manufacturers
provide software to integrate the wavelength dispersive X-ray spectrometer
functions with energy dispersive X-ray spectrometry in such a way
that qualitative scans may be performed from the energy dispersive
X-ray spectrometer screen as a graphic-user interface.
7.7.2. Wavelength
dispersive X-ray spectrometry generally requires flat samples for
analysis. Higher beam current and higher magnification than are
frequently used for energy dispersive X-ray spectrometry are required
to provide a sufficient count rate. However, the thermal energy
produced from this higher beam current has the potential for damaging
the sample.
7.8. Comparison of a small particle to a bulk sample
7.8.1. Any individual
particle or fragment from an inhomogeneous paint may not be compositionally
representative of the bulk and therefore would not be expected to
produce spectra similar to the bulk material. In order to compare
the compositional characteristics of a particle with a bulk sample,
the spectrum of the particle should be compared to the non-representative
spectra obtained from multiple, randomly selected areas having a
similar analytical volume to that of the particle.
7.9. Analysis
of a smear
7.9.1. A smear
is composed of co-mingled particles, fragments, and possible pieces
of one or both original contact surfaces. Initially a light microscopical
search is made for the largest fragments present. Particles that
are approximately 50 micrometers can often be selected individually
and attached to a mount for analysis. It is also possible to lift
a collection of deposited particles with a sticky lift, attach the
lift to a scanning electron microscope mount, and analyze the particles
directly.
7.10. Analysis
of a primarily organic layer
7.10.1. Analysis
of a layer that is primarily organic, such as a clear coat or varnish,
may be useful for comparison. Within such a layer, the interaction
volume is significantly larger than that of a layer containing pigments.
This is a result of a lower average atomic number of the matrix.
In order to reduce the interaction volume, the beam voltage may
be reduced; however, the voltage must be sufficient to produce X-rays
from all lines of analytical utility.
7.10.2. Because
an organic layer may contain small amounts of some elements, the
counting time should be extended.
7.11. Inhomogeneity
versus analytical area
7.11.1. In order
to compare the average composition of structures, the spectrum used
for comparison must come from an area of the structure sufficient
to produce an average composition.
7.11.2. The
representative nature of a spectrum can be determined by the critical
comparison of spectra from adjacent areas. If no differences are
evident, the structure is homogeneous at that magnification. A representative
bulk analysis can be achieved by rastering the beam across as large
an area as the sample permits.
7.11.3. A representative
bulk analysis is most difficult to achieve when analyzing a thin,
inhomogeneous layer in a cross-sectioned specimen. In these types
of samples, a representative spectrum may not be achieved until
the individual spectra from 10 or even 20 areas have been summed.
7.12. The
assessment of results during analysis
7.12.1. Generally,
comparisons are facilitated by direct spectral comparison.
7.12.2. If spectral
differences are not detected, it is likely that the materials are
similar in composition.
7.12.3. If spectral
differences are detected, it is likely that the materials are not
similar in composition; however, several alternative explanations
are possible and must be evaluated. Only after the considerations
in Sections 7.12.3.1 through 7.12.3.5 have been addressed can the
analyst confirm that spectral differences are indicative of compositional
differences.
7.12.3.1. Differences
in background shape may result from dissimilar sample geometry.
To correct, see Section 6.7.
7.12.3.2. Differences
in the composition of major peaks may indicate that the spectra
are not representative of the bulk composition of an inhomogeneous
specimen. This could occur as a result of the analysis of a sample
too small to be representative or the analysis of a raster area
too small to be representative. To correct, see Section 7.8.
7.12.3.3. Compositional
differences may result from unequal contributions from elements
in adjacent layers. To correct, see Section 7.4.
7.12.3.4. If
there are no differences in major peak ratios, differences in minor/trace
components may result from the presence of extraneous materials.
If the sample was a fragment or a smear and unable to be cleaned,
a small amount of foreign material may have been present during
the analysis. Consequently, some of the minor elemental peaks in
the spectrum may have been produced from elements in the extraneous
material. To correct, see Section 6.4.
7.12.3.5. Differences
in carbon intensity may result from a contribution of carbon from
the mount.
7.13. Interpretation
of scanning electron microscopy/energy dispersive X-ray spectrometry
data
7.13.1. A conclusion
regarding similarity results from the comparison of images and elemental
composition of individual layers. Collect all data and micrographs
in the same manner over as short a time as reasonable. Spectra may
be critically compared by overlaying them.
7.13.2. Compositional
comparisons may be made by methods such as calculating the ratios
of the integrated peak intensities of a spectrum to that of a reference
peak and comparing the values to those of the comparison sample.
If a peak ratio method of comparison is used, the ratios should
be displayed in tabular form.
7.13.3. If a
comparative analysis did not demonstrate significant differences,
then no differences were indicated in structure and composition
within the limits of the analytical capability of scanning electron
microscopy/energy dispersive X-ray spectrometry.
7.13.4. If a
comparative analysis demonstrates significant differences between
samples regarding structure and composition, then it can be concluded
that the samples are different.
7.13.5. If ratio
differences between peaks exist, it can be concluded that these
differences may result from either actual differences in the bulk
composition of the materials or from the analysis of a small sample
(or area) whose chemistry is not representative of the bulk composition
of an inhomogeneous specimen. The latter should only be concluded
following an extensive investigation of the heterogeneity of the
known specimens and demonstration that the range of variation present
in the known sample encompasses that observed in the questioned
sample.
7.13.6. If there
are no differences in major peak ratios but there are differences
in minor/trace peaks, it can be concluded that no differences in
major elemental constituents are indicated, although some differences
in the bulk composition are evident. For example, if the sample
was a fragment or smear and unable to be adequately cleaned, a small
amount of foreign material may have been present during the analysis.
Consequently, some of the minor elemental peaks present in the spectrum
may have been produced from elements in the foreign material and
not from elements in the questioned material. Equally so, the observed
differences may be due to actual differences in the composition
of the samples. Therefore, with respect to the elemental composition
of these samples, an inconclusive result for this technique is indicated.
8.
Documentation
8.1. Case notes
must include a copy of all of the instrumental data that was used
to reach a conclusion. All hard copies must include a unique sample
designation, the operator's name/initials, and the date of analysis.
8.2. Case notes
must also include a description of the evidence analyzed by scanning
electron microscopy, the method of sample preparation, the analytical
instrumentation used, and its operating parameters. The case notes
should also include a statement or data confirming system calibration,
as detailed in Section 7.1.
8.3. See SWGMAT's
Trace Evidence Quality Assurance Guidelines for further requirements
(www.fbi.gov/hq/lab/fsc/backissu/jan2000/swgmat.htm).
References
Bearden, J.
A. X-ray wavelengths, Reviews of Modern Physics (1967) 39(1):78.
Goldstein, J.
I., Newbury, D. E., Echlin, P., Joy, D. C., Romig, A. D., Lyman,
C. E., Fiori, C., and Lifshin, E. Scanning Electron Microscopy
and X-ray Microanalysis: A Text for Biologists, Materials
Scientists, and Geologists. 2nd ed., Plenum Press, New York,
1992.
Stoecklein,
W. and Goebel, R. Application of cathodoluminescence in paint analyses,
Scanning Electron Microscopy (1992) 6(3):669-678.
Ward, D. C.
A small sample mounting technique for scanning electron microscopy
and X-ray analysis, Forensic Science Communications [Online].
(July 1999). Available: www.fbi.gov/hq/lab/fsc/backissu/july1999/ward.htm
Wrobel, H. A.,
Millar, J. J., and Kijek, M. Comparison of properties of adhesive
tapes, tabs, and liquids used for the collection of gunshot residue
and other trace materials for SEM analysis, Journal of Forensic
Sciences (1998) 43(1):178-181.
Top
of the page
|