BY RONALD K. MARLEY
The evidentiary value of glass fragments has long been understood by the criminalist and forensic scientist but has often been underrated by the fire investigator. For the fire investigation process, glass can be broken down into two broad categories: glass that assists with the analysis of the fire`s origin and cause and glass that may link suspects to the scene or cause of the fire (this type of glass evidence is referred to as trace evidence). Glass is useful in origin and cause analysis and as evidence in general for a number of reasons, including the following:
Large quantities of glass are usually present at every structure fire. Wide variations in the compositions and properties of different types of glass items allow detailed comparisons to be made between samples.
The manner in which glass fractures, particularly when struck with an object, provides numerous fragments for forensic analysis.
The ease with which small glass particles transfer from the impact point to the clothing or skin of a nearby individual frequently results in glass fragments being transferred to the clothing of suspects who entered the structure by breaking a window.
Physically and chemically, glass is very stable and can be used for accurate comparisons, even years after the criminal event (without having to be placed in a complex or costly storage system).
Variations in the chemical and physical properties of glass, even glass that comes from the same batch, enhance its usefulness as comparative evidence.
The bottom line is that glass can be a very valuable item for fire investigators in trying to determine the origin and cause of a fire.
GLASS TYPES AND COMPOSITION
Glass is a solid material composed of mostly silicon dioxide (SiO2) and is blended with other materials (normally metal oxides such as calcium, magnesium, aluminum, or sodium), based on the desired characteristics or use of the final glass item. Glass is technically referred to as supercooled liquid, because the structure of the solidified material is noncrystalline (amorphous) in nature. Silicon dioxide in its pure state is found in quartz, the primary material from which glass is made. In quartz, silicon dioxide and oxygen atoms are arranged in a crystal lattice (uniform arrangement), which equates to a stable structure. These same atoms are randomly oriented, providing a nonuniform arrangement to the glass. This enables glass windows and other materials to “move” over time (a pane of glass that was relatively uniform in thickness when manufactured will end up thicker at the bottom of the pane years later because of the effects of gravity, a condition that does not occur in quartz crystals).
Note that under specific conditions, highly purified quartz can be melted and formed into a crystallized structure known as cristobalite, but these conditions do not occur during the normal glass-making processes. In manufacturing glass, quartz sand is mixed with calcium oxide (CaO = lime) to prevent it from dissolving in water, and sodium carbonate (Na2CO3 = soda) is added to lower its melting temperature and viscosity, thus improving its workability. This mixture is then melted at approximately 1,4007F), and molten glass is produced. This molten glass, termed soda-lime glass, represents one of the largest groups of glass types in use today.
Additional types of glass include lead crystal glass–made by adding lead oxide to the melt, which produces high refractive indexes (providing the sparkle in expensive crystal ware); optical glass; borosilicates; and tempered or safety glass. Borosilicates (the most common trade name being Pyrexw) are produced by adding boron oxide or sodium tetraborate (borax) to the mixture, which results in a glass that is very resistant to thermal shock.
Glasses can also be modified by changing the manner in which they are produced, the most notable being tempered glass. This type of glass is produced by rapidly heating and cooling the glass surface during the manufacturing process, resulting in a differential cooling rate between the surface and interior of the glass, which produces stresses and surface tension in the cooled glass. The final glass product is very strong and break-resistant but shatters into small, uniform particles when it does break. Tempered or safety glass is relatively common and can be found in vehicles, shower/bath doors, offices, and schools. Plate or window glass is frequently a factor in the origin and cause analysis process and as trace evidence with criminal investigations.
The most common process for producing plate glass is a relatively new technique called the tin float process. In this system, tin metal is heated to a liquid and the molten glass is poured onto the liquid tin surface. The liquid tin bath is isolated from all mechanical vibrations, thereby allowing both the glass-tin and glass-air surfaces to become exceptionally smooth as the molten glass is processed and cooled. This process produces glass that does not require additional polishing. Prior to the tin-float process, plate glass was formed by running the molten glass through a series of rollers until the glass plate was at the required thickness and width. This process required an additional (and expensive) mechanical polishing step before the glass could be used for window applications.
It is important for the investigator to keep in mind that each of these separate production processes produces glass that has a unique set of physical and chemical properties, which enables a forensic scientist to make detailed comparisons between glass found at the scene and glass fragments recovered from a suspect.
PHYSICAL PROPERTIES OF GLASS
Glass has a number of optical and nonoptical properties that make it very useful from the evidentiary standpoint. Some samples can be differentiated with simple visual comparison techniques. Properties such as surface finish (clear, frosted, textured), color, flat vs. curved (for large samples), thickness, and so on can be determined quickly in the laboratory. Examiners can use this information to weed out glass samples that are not related to each other. To make useful comparisons, other physical properties such as density, hardness, refractive index, fluorescence, and absorption must be extensively examined in the laboratory.
Density. A nonoptical property of glass, density is the mass of an item per unit of volume and is the same regardless of the size of the object (it is an intensive property). Glass has a broad range of densities, but in every case it will have a density of two or
greater when compared with water. Densities can be measured in very minute pieces of glass taken from a suspect`s clothing, which can be compared with control samples taken from the scene. Densities can be determined by simply weighing the item and dividing the weight by the volume, but this method does not provide very precise results.
The techniques used in laboratories center around suspending the glass fragments in liquids of varying densities. Although this process is not complex, it is very time- consuming, since numerous tests have to be conducted and the temperatures of the various test liquids have to be very tightly controlled. The end result is an accurate measurement of the fragment`s density. Density is a broad class characteristic physical property and by itself has little value in linking a suspected glass fragment to the crime scene. When combined with other class and individual characteristics, however, it can be used to link one glass fragment recovered from a suspect to control samples recovered at the scene of a fire.
Hardness. Another nonoptical property that may be used in the identification of glass fragments, hardness can be measured by its ability to scratch other materials. Currently, there is no reliable system for ranking individual glass fragments with respect to hardness, but suspect samples can be tested against control samples taken from the scene. This process is probably more useful for eliminating a sample than for linking the sample to the fire scene.
Refractive index. Refraction is the bending of light caused by a change in velocity. The refraction index of a fragment is a constant physical property when the sample is measured at a fixed temperature and the wavelength of the light source is constant. As such, all fragment samples must be measured at the same temperature and with the same wavelength light source to have valid comparisons. Since it would be impractical or impossible to shape all glass fragments into a prism and measure the refractive index, an alternative system had to be established. Like density, refractive analysis centers around placing the fragments into liquids. These liquids have known refractive indexes and, by using the right light source, the glass fragment will “disappear” when it matches the refractive index of the liquid. The actual refractive index can be measured in this manner, or the samples can simply be compared with one another. Since the refractive index is a function of temperature and the wavelength of light, more precise comparisons can be made by comparing glass fragments while using a number of different light wavelengths.
GLASS FRACTURES
When subjected to an applied force, glass, which is very elastic, bends or deforms until it reaches its elastic limit, at which point it breaks or shatters in a very predictable manner. Criminalists or forensic scientists can study and interpret the fractured glass pieces and normally determine the direction from which the force was initially applied. For fire investigators, this may be useful during fires where the home- owner is fraudulently claiming that someone broke into the structure and set a fire. A
detailed laboratory examination of the glass fragments may indicate that the window was broken from the inside of the structure prior to the fire. Simply determining on which side of the frame most of the glass fell is not a reliable indicator of the direction from which the glass was broken. In the majority of situations, most of the glass will fall on the side opposing the direction of force, but this does not always occur. The type of frame, width, and slope of the windowsill; the type of window covering (either side); the size of the window; the position of the initial impact; the force of the initial impact; the weight of the impact object; and wind, fire behavior, and ventilation factors–all can affect the direction in which the glass fragments may fall.
The only reliable method for determining force direction is a physical examination of the fragments. If a broken window (and the direction of force) becomes an issue during the investigation, the investigator should carefully collect all of the glass that has fallen out of the window and all of the glass remaining in the frame (in some cases it will be advantageous to collect the whole window frame; keep as much of the glass in the frame as possible). Glass from fire scenes is typically more damaged than glass from other crime scenes due to thermal dynamics, impacts from falling debris, and the effects of suppression and overhaul crews. Once all of the large pieces have been collected and protected, the small pieces can be swept up and placed in a container. The glass still in the frame should be marked (showing which side faces out) and, if a large amount remains, it should be collected in mass. It is extremely important that as much of the glass in the frame be recovered as possible without causing additional damage. Each piece should be marked showing the direction that it faced in the frame; these will be the control pieces from which the examiner will rebuild the window and determine the original direction of force. As with any item of evidence, samples should be photographed prior to their collection. I will review the system used by examiners to determine impact direction.
GLASS AND FIRE INVESTIGATIONS
For the fire investigator, glass can provide a wealth of information even without laboratory assistance.
Glass Fragments
Glass fragments scattered any significant distance (more than 10 to 15 feet) from the building will probably indicate that an explosion has occurred. If the glass fragments are all relatively clean (devoid of combustion products), the explosion was likely prefire, such as a natural gas/propane explosion. Glass fragments heavily stained with combustion products will most likely be from postfire development explosions. A postfire combustion explosion in one part of the building can blow out the windows in the entire structure, even if other sections of the building were free from fire or smoke at the time of the explosion. If only the glass from the nonfire sections of the structure is examined, it would initially appear to the investigator that the explosion could have been caused by a gas leak, when in fact it was a postfire development combustion explosion. All of the glass surrounding the structure has to be examined before a conclusion can be reached. The distance glass fragments travel outward from the structure may be helpful in determining the area of origin of the explosion/fire. But, without other indicators, origin location determinations based on glass fragment locations should not be made without consulting an engineer with experience in this type of evaluation. I am repeatedly amazed by the actual blast dynamics observed during test explosions in structures in relation to the final resting place of glass fragments. Areas of the building I expected to be significantly damaged at times had been largely untouched, whereas remote sections of the building had sustained extreme damage. The ability of “blast waves” to travel through the structure and be concentrated, reflected, or dissipated by the design of the building means that detailed engineering analysis might have to be performed before origin locations can be inferred from glass fragments.
The original height of the glass is very important in these evaluations. Glass falling from two or more stories and striking a solid surface may spread considerable distances outward from the structure, independently of an explosive-related event. Additionally, glass fragments may be projected considerable distances during interior suppression, ventilation, and overhaul operations. In these cases, the majority of the glass will normally be at the base of the building, and it should be relatively easy to differentiate between the events by conducting detailed interviews with the suppression crews. If the glass fragments are going to be an issue relating to the fire or explosion`s origin and cause, then the majority of the locations of the glass fragments will have to be charted and photographed. In some cases, all of the pieces will have to be collected.
Temperature Indicator
Window glass starts to soften or melt at around 1,4007F. Melted glass can be used as a broad temperature indicator, but keep in mind that melting temperature variances will occur due to the composition, arrangement, and thickness differences in the glass. The glass surface facing the fire will be heated first, and, in the case of window glass, this surface will expand slightly more than the “cool” side, causing the window to bend inward toward the fire. As the window pane continues to be heated, it eventually will melt and flow on the side of the frame facing the fire (ventilation dynamics may override this phenomenon and push the melted window away from the fire).
Crazing
Window glass that is crazed, or covered with small tightly spaced cracks, is often reported to be the result of a very rapid heat buildup and is frequently attributed to an accelerated fire. Although there is no statistical evidence to support this position, a significant amount of evidence is available to disprove this theory. After the 1991 Oakland Hills fire in California, a group of investigators examined 50 residential structures that had burned to completion.1 They concluded that glass in fires at which the structure had burned to total completion (black holes) provided “little value in determining the behavior” of the fire and that “crazing is more likely to be found when glass is rapidly cooled than when it is rapidly heated.” In a controlled behavior study, glass was rapidly heated and cooled to various temperatures to determine if crazing was in fact caused by the heating of glass surfaces.2 In all of the tests, crazing was observed only when the glass was rapidly cooled with a water spray. No crazing was ever observed through heating or cooling alone (this observation occurred in all of the tests). Additionally, I have often observed crazing in test burns once the suppression crew applied water to the room or window, even though no crazing occurred during the fire`s development (even when the fire was started with flammable accelerants).
Soot Deposits
Heavy deposits of oily soot on the inside surfaces of window glass were once thought to have been the result of a hydrocarbon-accelerated fire. Like crazing, there is limited statistical evidence to support this position, even though it may have had some validity at one time (older all-wood homes with natural furnishings). Currently, the extensive use of petroleum-based synthetics in home furnishings, carpeting, wall coverings, and building materials may result in the presence of oily soot deposits within the structure. It might be possible to recover petroleum-based combustion by-products from the soot deposits on window glass. But, the time and money involved in correctly identifying and linking these deposits to a flammable accelerant and eliminating the possibility that the deposits could not have come from another source make the process questionable and costly.
Lightbulbs
Lightbulbs can sometimes be used to determine the initial direction of the fire`s spread. Most lightbulbs are filled with an inert gas that expands as it is heated. The expanding gases will “push out” the softened glass on the side of the lightbulb facing the oncoming fire. As the fire develops, the entire lightbulb will eventually melt and lose its original shape and ultimately fall completely out of the fixture, at which point no inference should be made related to the fire`s direction of travel. In some cases, the lightbulb will be intact except for the “pushed-out” section, which will clearly show the direction of the fire`s approach. Even if the bulb is completely melted and resolidified in the sagging and elongated shape, the direction of fire spread may still be determinable. On close examination, the initial “pushed-out” area of the bulb can sometimes be seen within the melted mass. In these situations, the investigator will have to visualize the bulb at the time the side was “pushed out,” since the final position of the totally melted lightbulb may not reflect the bulb`s position at the time the side failed.
Trace Evidence
For fire investigators, glass is most valuable as trace evidence when persons suspected of setting a fire have shattered a pane of glass to enter the structure. Research has demonstrated that a considerable amount of small glass particles is ejected backward toward an individual breaking a window.3 The location of the individual breaking the window has a bearing on the number of particles likely to be found on his clothing. But, in almost every case involving the striking of a window with an object in the individual`s hand, glass particles will be found on that person`s clothing. The transfer of glass particles can occur even when the person throws an object through the window`s surface while positioned a number of feet from the window. The type of clothing material and the tightness of the weave have significance for the number of particles retained within the clothing. If investigators believe that an individual broke a glass window prior to the fire, all of the person`s clothing (hats to shoes) should be collected. If the clothes cannot be gotten voluntarily, a warrant should be obtained. Each item of clothing should be stored in a regular brown shopping bag unless there are other evidentiary considerations (such as biological or chemical analysis). No attempt should be made to remove glass fragments from clothing in the field. Glass fragments may stay on the surface of the clothing for weeks. Time delays should not prevent investigators from collecting these clothing items if they believe glass breaking/forced entry was an issue related to the fire`s origin and cause. In cases where vandalism occurred prior to setting the fire, glass fragments may be recovered from the soles of the shoes of all the individuals involved (whether or not they initially broke the glass item).
At this point it should be clear that glass has a number of possible uses in the origin and cause process. Individuals with unique experiences related to origin and cause investigations and glass are encouraged to send information to me in care of Fire Engineering. If possible, please include event dates; copies of reports, if available; photographs; drawings; and any other materials pertinent to the particular investigation (please do not send original materials). Forensic glass analyses are helpful to investigators and are used as a resource to further enhance and support physical evidence. Obviously, each case is unique, and these techniques may not be applicable in every case. n
References
1. Lentini, John J.; David M. Smith; and Richard W. Henderson, “Baseline Characteristics of Residential Structures Which Have Burned to Completion: The Oakland Experience,” Fire Technology, the National Fire Protection Association (NFPA), Quincy, Mass., Aug. 1992, 195-214.
2. Lentini, John J., “Behavior of Glass at Elevated Temperatures,” Journal of Forensic Science, Sept. 1992, 1358-1362.
3. Luce, R.J.W.; J.L. Buckle; and I. McInnis, “A Study on the Backward Fragmentation of Window Glass and the Transfer of Glass Fragments to Individual`s Clothing,” Canadian Journal of Forensic Science. Paper presented at the 33rd meeting of the Canadian Society of Forensic Science, 1986.
Additional References
Andrasko, J. and A.C. Maehly, “The Discrimination Between Samples of Window Glass by Combining Physical and Chemical Techniques,” Journal of Forensic Sciences, Apr. 1978, 250-262.
Brewster, Fay, “The Retention of Glass Particles on Woven Fabrics,” Journal of Forensic Sciences, July 1985, 30:3,789-805.
Brown, Guy A., “Factors Affecting the Refraction Index Distribution of Window Glass,” Journal of Forensic Sciences, July 1985, 806-813.
Calloway, A.R. and P. F. Jones, “Enhanced Discrimination of Glass Samples by Phosphorescence Analysis,” Journal of Forensic Sciences, Jan. 1978, 263-273.
Carrol, John R. Physical and Technical Aspects of Fire and Arson Investigation. 1979. Thomas Books, Springfield, Ill.
Cole, Lee S. Investigation of Motor Vehicle Fires, Third Edition. 1992. Lee Books, Novato, Calif.
Cote, Arthur E. Fire Protection Handbook, 17th Edition. 1991. NFPA, Quincy, Mass.
DeHaan, John D. Kirk`s Fire Investigation.1991. Brady, Englewood Cliffs, N.J.
NFPA 921, Standard on Fire and Explosion Invest-igations–1992 and 1955, NFPA.
Mcgill, Lamont “Monty,” unpublished teaching document, 1994.
Rodes, E.F. and J. I. Thornton, “The Interpretation of Impact Fractures in Glassy Polymers,” Journal of Forensic Sciences, Sept. 1974.
Ryland, Scott G., “Sheet or Container?–Forensic Glass Comparisons with an Emphasis on Source Classification,” Journal of Forensic Sciences, Oct. 1986, 1314-1329.
Thornton, John I. and Paul J. Cashman, “Glass Fracture Mechanism–A Rethinking,” Journal of Forensic Sciences, July 1986, 818-824.
RONALD K. MARLEY is fire chief of the Shasta College Fire Department in Redding, California, and an instructor in the college`s fire technology program. He has 13 years of explosive ordnance disposal experience within the military and local law enforcement community. Involved with fire investigation since 1984, he previously served as the chief fire investigator for the City of Dixon (CA) Fire Department. Marley is a certified fire officer, fire instructor, and fire investigator and an accredited master instructor in the California State Fire Marshal`s Office. He has various degrees including a bachelor`s degree in fire and explosion origin and cause investigation and is currently conducting a variety of long-term research projects related to fire and explosion analysis.
