This guideline describes methods for determining the concentrations of major, minor, and trace elements in glass fragments. The methods described may be used to measure either absolute or relative element concentrations in glass fragments ranging in mass from gram(s) to less than one microgram.

The analytical considerations for the use of scanning electron microscopy-energy dispersive X-ray spectrometry (SEM-EDS), X-ray fluorescence spectrometry (XRF), inductively coupled plasma-optical emission spectrophotometry (ICP-OES), and inductively coupled plasma-mass spectrometry (ICP-MS) are described. Other analytical techniques, such as atomic absorption spectrophotometry may also be used but are not specifically included in this guideline because they are not as widely used as those listed.

Several of the analytical methods described are destructive. Therefore, all nondestructive examinations must be completed and legal considerations concerning the destruction of evidence must be satisfied prior to conducting these measurements.

This guideline does not purport to address all of the safety problems, if any, associated with its use. It is the responsibility of the user of this guideline to establish appropriate safety and health practices and to determine the applicability of regulatory limitations prior to use.

This guideline provides general considerations, rather than detailed instrumental operating instructions. Measurement of element concentrations in glass fragments can be reliably obtained using any of many makes and models of instruments, each with its own specific instructions. The examiner should use this guideline in conjunction with instructions provided by the manufacturer of the particular instrument being used and with validated internal laboratory procedures for the characterization of glass evidence.

Scientific Working Group for Materials Analysis Documents

Trace evidence recovery guidelines
Quality assurance guidelines

American Society for Testing and Materials Standards

D1193 Specifications for Deionized Water
E50 Reagent Purity
E135 Standard Terminology Relating to Emission Spectroscopy

Environmental Protection Agency Test Method

Method 200.7 Inductively Coupled Plasma-Atomic Emission Spectrometric Method for Trace Element Analysis of Water and Wastes

This guideline describes several techniques for determining selected major, minor, and trace elements in glass fragments. Each technique described may either be used to determine the concentrations of elements in a glass fragment or to determine the concentrations of several elements relative to each other. A brief introduction to the principles, analytical methodologies, uses, and advantages and limitations of scanning electron microscopy-energy dispersive X-ray spectrometry, energy dispersive X-ray fluorescence spectrometry, inductively coupled plasma-optical emission spectrophotometry, and inductively coupled plasma-mass spectrometry in forensic glass examination is presented. The reader is encouraged to see the cited references and bibliography for more specific information.

The concentrations of certain elements in glass serve to chemically characterize its source. The concentrations of several elements are intentionally controlled by the manufacturers to impart specific end-use properties to a particular glass product, and in some instances can be used to identify the product type of a recovered glass fragment. However, even individual glass objects that have major element concentrations within the manufacturer’s acceptable ranges display variations that can be measured and provide useful points for a forensic comparison. Glass manufacturers generally do not control the concentrations of trace elements, except as needed to impart color or to keep them below levels that would impart undesirable physical or optical properties to the glass. The differences in concentrations of manufacturer-controlled elements or uncontrolled trace elements may be used to differentiate sources when the variation among objects exceeds the variation within each object. Element concentrations may be used to differentiate among glasses made by different manufacturers, glasses from different production lines of a single manufacturer, specific production runs of glass from a single manufacturer, and in some instances individual glass objects produced at the same production facility.

The discrimination potential of element concentrations in glass was documented as early as 1973. Several instrumental methods have been used by forensic scientists including neutron activation analysis (Coleman and Goode 1973), flameless atomic absorption (Hughes et al. 1976), inductively coupled plasma-optical emission spectrophotometry techniques (Hickman 1987; Koons et al. 1988), energy dispersive X-ray fluorescence spectrometry (Andrasko and Machly 1978; Reeve et al. 1976), scanning electron microscopy-energy dispersive X-ray spectrometry (Ryland 1986), and inductively coupled plasma-mass spectrometry (Haney 1977; Parouchais et al. 1996).

Comprehensive reviews of the literature on the application and advantages and limitations of several of the analytical techniques used for the elemental analysis of forensic glass samples have been reported (Almirall 2001; Buscaglia 1994).

Several factors that can be considered in selecting the appropriate analytical method for the analysis of glass in the forensic laboratory are shown in the following table.

The selection of a sample preparation technique will depend on the particular method to be used for analysis, fragment size and shape, and the purpose of the examination.

Wavelength dispersive X-ray spectrometry sorts the generated characteristic and background X-ray emissions by their wavelengths using a crystal monochromator. Wavelength dispersive X-ray spectrometry offers the best spectral resolution of all of the X-ray emission methods. Due to the high resolution and low background, the lowest levels of detection and most reliable quantitation are attainable. Cost and complexity of instrumentation have limited its use with scanning electron microscopy in forensic science laboratories. For this reason, scanning electron microscopy-wavelength dispersive X-ray spectrometry is not discussed in further detail in this guideline.

Because a scanning electron microscope can scan a focused electron beam over a small area, scanning electron microscopy-energy dispersive X-ray spectrometry has the ability to detect the presence of elements in a small sample, such as a questioned glass fragment.

In most published studies, scanning electron microscopy-energy dispersive X-ray spectrometry measurements of element intensity ratios have been applied to classification of glass types (Ryland 1986; Terry et al. 1982). An analytical scheme that combines measurement of Ca/Mg intensity ratios obtained using scanning electron microscopy-energy dispersive X-ray spectrometry with Ca/Fe ratios obtained using X-ray fluorescence spectrometry has been used with good success by several forensic laboratories to classify glass fragments into sheet and container categories (Keeley and Christofides 1979; Ryland 1986).

For discrimination among glass sources, a scanning electron microscopy-energy dispersive X-ray spectrometry protocol was reported for determining the ratios of the intensities of Na/Mg, Na/Al, Mg/Al, Ca/Na, and Ca/K in glass fragments (Andrasko and Machly 1978). Measurement of these ratios by scanning electron microscopy-energy dispersive X-ray spectrometry was incorporated into a scheme with refractive index, density, and emission spectrography. Thirty-eight out of 40 window glasses analyzed by this scheme were found to be distinguishable. The variation in the measured element intensity ratios by scanning electron microscopy-energy dispersive X-ray spectrometry was found to be consistent across a new sheet, an old sheet, and within a single fragment of glass.

Samples should be cleaned and dried according to Section 6.2. prior to embedding. Small irregularly shaped samples may be analyzed, but flat sample surfaces are recommended whenever possible and are particularly important when accurate, precise quantitative results are desired. This may be achieved by embedding the sample in a resin that is subsequently cured and polished to provide a flat surface. The embedded sample may need to be coated with a thin conductive layer, such as carbon, to reduce charging from the electron beam. The glass sample is altered and partially consumed during the embedding and polishing process.

Scanning electron microscopy-energy dispersive X-ray spectrometry measurements are made on individual fragments of glass. The beam current and magnification used will depend on sample size and instrument capabilities. Magnifications on the order of 1,000X are adequate for most samples, but magnifications up to 10,000X can be used.

The operating current should be adjusted for optimal count rates so that analytical time is not excessive. It should be noted that Na migration could occur at high operating currents. Na migration is evidenced by a significant drop in
NaK_a intensity with time.

The detector-to-specimen distance and the takeoff angle should be optimized for each instrument.

Acquisition times are selected based on sample size and count rate. These can be either based on a fixed time during which the detector is acquiring data (e.g., a specified number of live seconds) or on achieving a specified intensity for a selected X-ray energy (e.g., acquiring spectral data until the intensity of the Ca K_a peak reaches 50,000 counts). The advantage of using the latter method is that it aids in normalizing the data and yields similar precision for samples of different sizes and shapes.

Spectra should be acquired from replicate samples in order to obtain a measure of the variability within the sample and specimen. The most frequently used methods of data interpretation are spectral comparison by simple overlaying of X-ray spectra of two fragments or by calculating the element intensity ratios.

Full quantitative analysis is possible but requires calibration with matrix-matched standards embedded in each stub along with the analytical samples under identical operating conditions.

The most significant advantages of scanning electron microscopy-energy dispersive X-ray spectrometry for determining element concentrations in glass fragments are that it is nondestructive, applicable to very small samples, and readily available to many forensic laboratories.

Scanning electron microscopy-energy dispersive X-ray spectrometry, when combined with energy dispersive X-ray fluorescence spectrometry, permits product-use classification of glass samples.

Scanning electron microscopy-energy dispersive X-ray spectrometry is useful in forensic laboratories as a rapid, screening method that can add some discrimination capability to optical and physical measurements.

Limitations of scanning electron microscopy-energy dispersive X-ray spectrometry for glass examinations

Several characteristics of scanning electron microscopy-energy dispersive X-ray spectrometry have limited its widespread application in forensic laboratories for the comparison of glass fragments.

The irregular shape of some small glass fragments makes precise and accurate quantitative determination of element concentrations difficult. Variations in fragment surface orientation contribute to relatively poor precision and accuracy of results obtained by scanning electron microscopy-energy dispersive X-ray spectrometry compared with other methods of analysis.

The compositional differences distinguishable by scanning electron microscopy-energy dispersive X-ray spectrometry will, in most instances, manifest themselves in readily distinguishable refractive index and/or density values.

The detection limits on a concentration basis of scanning electron microscopy-energy dispersive X-ray spectrometry are poor, typically in the 0.1 percent range; and therefore, the number of elements that can be determined is limited. Many elements, such as Ba and Sr, which are useful for both classification and source discrimination, are present in most glasses at levels that are not detectable using scanning electron microscopy-energy dispersive X-ray spectrometry.

The detection limits for scanning electron microscopy-energy dispersive X-ray spectrometry are approximately an order of magnitude poorer than for scanning electron microscopy-wavelength dispersive X-ray spectrometry.

X-ray fluorescence spectrometry is an elemental analysis technique based upon the measurement of characteristic X-rays emitted from a sample following excitation by an X-ray source. The energies or wavelengths of the detected X-rays are used to identify the elements, and the intensities of the X-ray peaks in the measured spectrum correlate with the quantities of each element present in the sample area exposed to the X-ray beam.

Wavelength dispersive X-ray spectrometry (see Section 7.1.1.2.1. for a description) offers the best spectral resolution of all of the X-ray fluorescence methods. The glass manufacturing industry for quality control purposes uses wavelength dispersive X-ray spectrometry extensively. However, in addition to the high cost of instrumentation, wavelength dispersive X-ray spectrometry requires relatively large, flat samples, both of which severely limit its forensic use. Wavelength dispersive X-ray fluorescence spectrometry is not discussed in further detail in this guideline.

Energy dispersive X-ray fluorescence spectrometry (see Section 7.1.1.2.2. for a description) methods are considered nondestructive, surface or near surface techniques. Fluorescent X-rays that reach the detector typically originate from a glass sample region no deeper than about 100µm. To analyze small, irregularly shaped fragments common to forensic casework, micro- and capillary X-ray fluorescence spectrometry techniques are appropriate. In these techniques, the beam is collimated down to micrometer size beam diameters, typically 100 to 300µm for forensic glass analyses.

Energy dispersive X-ray fluorescence spectrometry has been used to classify unknown source glass samples as to their product-use types (Dudley et al. 1980; Howden et al. 1977; Ryland 1986).

Energy dispersive X-ray fluorescence spectrometry is capable of distinguishing glass samples of different product-use type whose refractive indices are indistinguishable (Dudley et al. 1980).

Using element intensity ratios determined by energy dispersive X-ray fluorescence spectrometry, sheet and container sources were correctly classified in 93 percent of the cases tested (Ryland 1986). A combined energy dispersive X-ray fluorescence spectrometry and scanning electron microscopy-energy dispersive X-ray spectrometry procedure was reported to be useful for classification of modern sheet and container glasses. These samples could not be separated by scanning electron microscopy-energy dispersive X-ray spectrometry determination of Ca/Mg alone. Using a 35 kilovolts accelerating voltage in energy dispersive X-ray fluorescence spectrometry, a low Ca/Fe ratio is indicative of sheet glass and a high Ca/Fe ratio indicates a container or tableware source. Some samples had intermediate values and were unclassifiable with this technique.

Discrimination among sources of glass may be accomplished by energy dispersive X-ray fluorescence spectrometry using both qualitative and semiquantitative methods.

Energy dispersive X-ray fluorescence spectrometry measurements are generally made on individual fragments of glass using a beam collimator of 3mm or less, depending on sample size and instrument capabilities.

An accelerating voltage of 35 kilovolts can be used for end-use classification purposes, and a fixed setting in the range of 35 to 50 kilovolts is suitable for discrimination purposes.

The operating current should be adjusted as needed to obtain good count rates, optimally less than 50 percent detector dead time.

Acquisition times can be selected based on sample size and count rate, as with scanning electron microscopy-energy dispersive X-ray spectrometry. See Section 7.1.3.7. for more details.

Quantitative concentrations are routinely determined with energy dispersive X-ray fluorescence spectrometry on comparatively large, flat samples. For the most accurate quantitative determinations, calibration with matrix-matched standards is required. Standardless quantitative mathematical models for calibration do exist; however, they may result in poorer accuracy. Because primary and fluorescent X-rays do not reproducibly penetrate the small size and the irregular shape of forensic glass fragments, quantitative elemental concentrations generally cannot be determined with adequate precision and accuracy without sample preparation to provide a flat surface. However, ratios of the intensities of two X-ray lines of similar energies are reasonably constant. Therefore, comparison of the elemental compositions of glass fragments of varying masses and shapes can be performed with good precision using intensity ratios. Intensity data for selected elements should be ratioed after baseline subtraction and escape peak corrections have been performed. Like scanning electron microscopy-energy dispersive X-ray spectrometry spectra, energy dispersive X-ray fluorescence spectra are best compared either by overlaying spectral images or by comparing calculated peak intensity ratios.

In most inductively coupled plasmas, an electrical discharge is initiated in a flowing stream of inert gas, usually argon, and then sustained by a surrounding radio frequency field. The resulting stable discharge, or plasma, has the appearance of a small continuously glowing flame, with temperatures in the range of 7,000-10,000K. When a sample is introduced into the plasma, extensive atomization, ionization, and excitation of the sample atoms occur.

As the ions and atoms present in the sample enter cooler portions of the plasma and drop to lower excited states, they emit light at characteristic wavelengths. In an inductively coupled plasma-optical emission spectrophotometer, this emission is dispersed with a spectrophotometer and its intensity is measured. Comparison of the emission intensities of a sample with those of standard solutions is used to determine the concentration of the elements in the sample.

Glass samples can be introduced into the plasma either by dissolving them and nebulizing the resulting solution or by direct solid sampling. Only solution methods are discussed in the inductively coupled plasma-optical emission spectrophotometry portion of this guideline. Direct solid sampling is outlined in the inductively coupled plasma-mass spectrometry portion of the guideline because solid sampling is more commonly used with mass spectrometry than with optical emission spectophotometry.

The initial inductively coupled plasma-optical emission spectrophotometry methods developed for glass analysis were primarily designed for purposes of classification. An inductively coupled plasma-optical emission spectrophotometry analytical method was developed to determine the concentrations of Mn, Fe, Mg, Al, and Ba in glass fragments (Catterick and Hickman 1981). Over the next several years, the concentrations of additional elements in glass by inductively coupled plasma-optical emission spectrophotometry were determined, and 6 to 10 element classification schemes based on comparison with a glass database divided into nine product categories were developed (Hickman 1981; Hickman et al. 1983). Currently, the protocol most widely used for casework was developed for determining the concentrations of 10 elements (Al, Ba, Ca, Fe, Mg, Mn, Na, Ti, Sr, and Zr) with excellent analytical precision in milligram-sized glass fragments (Koons et al. 1988). A combination of five of these elements was shown to provide good classification into the two categories of sheet and container glass. Inductively coupled plasma-optical emission spectrophotometry has also been used to associate food containers to the manufacturing plants in which they were made and to identify sources of contaminant glass in cases involving product tampering (Wolnik et al. 1989).

In further studies, the distributions of up to 22 elements, most measured by inductively coupled plasma-optical emission spectrophotometry, in various glasses were shown to provide excellent discrimination capability among sources within a product class (Hickman 1983; Hickman et al. 1983). In a study measuring the concentrations of 10 elements in automobile side-window glasses, the probability that two glasses from different vehicles would be indistinguishable was reported to be one in 1,080, compared with one in five for refractive index alone and one in ten for energy dispersive X-ray fluorescence spectrometry analysis alone (Koons et al. 1991). Studies have shown that using inductively coupled plasma-optical emission spectrophotometry sheets of glass produced within minutes of each other in a single float-glass production line can be differentiated. In a recent study using statistical analysis of samples collected in casework, it was reported that inductively coupled plasma-optical emission spectrophotometry measurements provide very high discrimination capability. The probability that two glass fragments from different sources will have indistinguishable concentrations of ten elements is extremely small (Koons and Buscaglia 1999).

Prior to analysis, each fragment must be washed to remove surface contamination and residual fluids from refractive index and density measurements. Typically, this is done by soaking in HNO₃, rinsing repeatedly with deionized water, then ethanol, followed by drying.

Fragments, typically weighing 0.2 to 8mg, are obtained by crushing the specimen between sheets of clean plastic. Prior to dissolution, analytical samples must be weighed to the nearest 0.01mg, or better, on a microbalance, whose calibration is checked immediately preceding each use.

Dissolution procedures involve the use of high purity HF, HNO₃, and, in some cases, HCl (Catterick and Hickman 1981; Koons et al. 1988). A typical procedure consists of the digestion of glass fragments in polypropylene tubes by the addition of concentrated HF and other mineral acids, taking the solution to dryness (to remove all residual HF) by heating the tubes, and complete redissolution of the residue in a strong acid, such as HCl. This dissolution procedure typically requires several days. The complete removal of HF is required to prevent deterioration of glass or quartz spray chambers, nebulizers, and torches present on most inductively coupled plasma instruments.

An internal standard, usually scandium, must be added to each sample and standard solution in order to correct for minor differences in analytical sensitivity among samples and among samples and standards.

The analytical curve for each element is prepared using a minimum of four calibration standards, including a calibration blank. Multielement calibration standards containing the internal standard and the diluting acids in the same concentration as the sample solutions are used. The concentration ranges for each analyte element must span the concentrations in the sample solution.

Analytical characteristics of inductively coupled plasma-optical emission spectrophotometry instruments include the capability to determine a wide range of elements, long linear response ranges, a limited number of spectral and matrix interferences, low detection limits for many elements of interest, and ease of automation of data handling.

The detection limits of this method vary slightly from day to day. Typical values are 0.01 to 0.1µg/g in the glass. For samples smaller than 5mg, these detection limits are raised somewhat because of increased dilution factors. Precision measurements for all elements when present in the middle of their respective concentration ranges are generally better than two percent relative standard deviation.

The useful features of inductively coupled plasma-mass spectrometry compared with inductively coupled plasma-optical emission spectrophotometry are a smaller and better-defined set of spectral interferences, detection limits roughly 1,000-fold lower, longer linear working ranges, and more rapid scanning capability. The better sensitivity of mass spectrometry compared to optical emission spectrophotometry permits use of a smaller sample size and the quantitative determination of some trace elements not detectable by inductively coupled plasma-optical emission spectrophotometry. Using solution inductively coupled plasma-mass spectrometry, approximately 40 elements can be determined in a 1mg glass fragment.

Another characteristic of inductively coupled plasma-mass spectrometry is that samples need not be introduced as solutions. The rapid scanning speed of the mass spectrometer allows measurement of transient signals, such as those produced with laser ablation of solid samples used to sweep a vapor into the plasma for analysis (Montaser et al. 1998).

The first reported studies of the use of inductively coupled plasma-mass spectrometry for the forensic comparison of glass fragments indicated an ability to quantitatively determine the concentrations of 40 to 62 elements in glass fragments as small as 500µg (Parouchais et al. 1996; Zurhaar and Mullings 1990). For any given glass, approximately 40 of these elements are likely to be present at detectable concentrations.

The dissolution procedures for inductively coupled plasma-mass spectrometry are similar to the procedures used for inductively coupled plasma-optical emission spectrophotometry, except that HNO₃ is preferred over HCl for the final analytical solution to minimize or eliminate interferences from polyatomic chlorides. Because of the high sensitivity and low detection limits of inductively coupled plasma-mass spectrometry, it is essential that contamination from such sources as digestion reagents and the laboratory environment be controlled. The range of major, minor, and trace element concentrations in glass digest solutions is so great that most inductively coupled plasma-mass spectrometry protocols require analysis of multiple sample dilutions to measure the concentrations of all analyte elements.

Interferences when using a quadrupole instrument, though generally well known, in some instances are not easily corrected and prevent accurate quantitation of a few elements. Removal of some interference effects, particularly isobaric mass overlaps, can be accomplished using a high-resolution spectrometer or by collision cell technology, both at added cost and complexity of instrument operation.