“Matches” An Over inference of Data? A Giglio Obligation?: The case of Paint Examination
By: Frederic Whitehurst, J.D., Ph.D.  and Darlene R. Brezinski, Ph.D.
Forensic chemical analyses of complex matrices such as plastics, paints, or paper products, may lead to inconclusive results without analysts or reviewing counsel aware of the failing of the analyses. Complex matrices are materials which contain many components in varying mass ratios. These types of materials are regularly presented as evidence to the modern forensic chemical analysis laboratory. An example of such evidence is paint.
When vehicles collide there is often a transfer of paint from surface to surface. When homes are burglarized, paint pried from entrance points can adhere to tools used to make forced entry. Parts of home-made bombs at times are painted and, when those bombs explode, parts from the bombs still have paint adhering to them.
Paint is therefore evidence considered in modern forensic laboratories. Different chemical analysis protocols are followed in the analyses of this type of evidence. Many different opinions are rendered concerning the significance of the analytical chemical data from this type of evidence. In fact so many different opinions can be found in the legal literature concerning the significance of forensic paint evidence that one might begin to question the real value of such evidence. The questions arise not so much from a consideration of the protocols used but from the opinions rendered.
Certainly forensic paint evidence is not all the same. Some paint specimens are simply smears of paint on surfaces while other paint specimens are multi-layered specimens with many layers of paint of different color and type. One must keep this in mind when questioning the correctness of opinions rendered by forensic paint analysts.
This paper will deal with a subclass of forensic paint analysis and the opinions that may be rendered based upon the data gleaned from analysis. We will explore the possible over inferences of data that can occur as well as question the obligations of the forensic scientist to report alternative explanations for the data acquired, explanations which may prove to be exculpatory in nature. The subclass of paint evidence considered will be the single-layered paint system. These systems can be viewed as somewhere between the one extreme of paint smears which, due to their limited quantities, are very difficult to analyze and the other extreme of multi-layered paint chips, which can have a great deal of information suggesting uniqueness. Single-layered paint systems can appear in evidence which has been painted only once or where layers of paint have not adhered to each other during impact and only one layer is available for analysis.
The analytical scheme that will be referred to here as an example is that which has been utilized for a number of years in the forensic laboratory of the Federal Bureau of Investigation and followed somewhat in other forensic laboratories. The details of the protocol are not so important here as the implications of the data itself. The purpose of this paper is to give an understanding of those implications and possible limitations of the data.
Forensic scientists describe forensic evidence as class evidence and individual evidence. The FBI’s Handbook of Forensic Science describes individual evidence as “This evidence can be positively identified as having come from a specific source or person…” and class evidence as “This evidence, no matter how thoroughly examined, can only be placed into a class. A definite identification as to its source can never be made since there is the possibility of more than one source for the evidence found. (Examples are: soil, blood, hairs, fibers, single-layered paint…)” Many forensic investigations result in a tension between efforts to determine the degree of uniqueness of items of class evidence and the limitations of time, technology and knowledge about the particular type of material represented by that class. Forensic paint analysis is often in the center of this struggle.
Look around you. Paint and coating materials are virtually everywhere you look. To say that two items were coated with paint would not be very probative of their origins. To say that two items were coated with a particular color of paint, gray paint for instance, that one could not differentiate with the human eye under incandescent illumination, would be a bit more probative. But then many things (including virtually the whole U.S. Navy) are coated with gray paint. Therefore one would not significantly narrow the field of alternative explanations for sources of paints simply by describing paint as gray. Triers-of-fact recognize that. They deal with the practical side of recognizing paint every day and understand the limited probative value of two objects being painted the same color. For instance a finding that white paint was on a prying tool found in the possession of a defendant accused of breaking into a white-painted house would not be remarkable. There are many houses and objects painted white. And so the forensic scientist tries to narrow down the field.
Narrowing Down the Field
One method of narrowing down the field of sources of paint is to look at surface texture of the paint particles. Scratches or surface blemishes which appear to be the same on two paint specimens are consistent with though not proof of the fact that paints could have originated from the same source. Of course if a paint specimen which has broken from a painted substrate can be refitted exactly into the pattern that particle left when it was detached then a positive identification of two paints as originating from the same source can be made.[7,8]
Another method of narrowing the field of paints down is to subject paint specimens to chemical spot tests. Paint particles react with various solvents and chemicals. Visible manifestations of those reactions are observable under a microscope. Paint particles that react differently to chemicals can be reasonably determined to have not originated from the same source. However two paint particles that react in the same manner to chemicals can only be said to have originated from paint types that react in the same manner to those chemicals. Without an understanding of the significance of reactions of paints to particular chemicals, one can not be sure even if paints that react the same are of the same type. Stoecklein suggests that a thorough understanding of paint chemistry is necessary before one renders opinions in this very complex field.
The microscopic and chemical spot test work do not narrow down the field of possible sources of paint very far and can leave the analyst with less than probative results at times when single layered paint chips are analyzed. Therefore an attempt to further characterize the uniqueness of paint specimens then enters into chemical instrumental analysis.
Modern forensic chemical analysis of paint evidence depends very heavily on a number of types of technology to assist in identifications of materials. X-ray powder diffractometry, x-ray fluorescence spectroscopy, energy dispersive x-ray analysis coupled with scanning electron microscopy, Fourier transform infrared spectrometry and pyrolysis/gas chromatography/mass spectrometry are among the more popular technologies in use today. Stocklein also recommends a number of other instrumental techniques that are not in wide use today in forensic paint analysis laboratories. The x-ray powder diffractometer confirmed with the energy dispersive x-ray analysis or x-ray fluorescence can be used to identify major crystalline components of paint matrices. The Fourier transform infrared spectrometer can be used to characterize major organic components of the paint matrix. And pyrolysis/gas chromatography/mass spectrometry can also be used to characterize intact paint components as well as decomposition products of paint matrices subjected to extremely high heats as part of the analysis scheme. The decomposition products hypothetically can be related back to the original components but such hypotheses must be tested and validated.
There are known limitations to each of these instrumental techniques. For instance, all these analytical techniques have minimum detectable limits. If components placed in paints are of such a small quantity that one can not detect them then one can not reasonably say whether they are present, whether they differ between samples, or even whether they effect the analytical data. This situation might at first be addressed with an argument that, if components are in such small mass quantities that they can not be detected, they are not important in differentiating paint samples. However that argument flies in the face of the reality of the economics of manufacturing commercial products. No manufacturer will put components into a product that cost money without reason. Even minor components in terms of mass cost money to place in products. Therefore such components are important whether we can detect them or not. Again, if we can not detect them, we do not know if they are different in different specimens and therefore we can not narrow our paint sources beyond that level.
Another complication is in the use of pyrolysis/gc/ms. This analysis technique analyzes pyrolyzates, not unpyrolyzed materials. Paint samples are literally burned (pyrolyzed) in a tiny vile and the gases from the burned material, pyrolyzates, are analyzed. Though one can hypothesize from the materials which are left after the pyrolysis what the original material was, validation of such hypotheses are necessary. Because even as far back as 1982 we find Thornton noting that “the paint industry is one of the most complex segments of the total chemical industry, utilizing over 600 different kinds of raw materials and intermediates, and because paint formulas are being changed so quickly to keep up with environmental regulations validations of such hypotheses can be very time consuming and expensive.
A third complication arises in the use of the Fourier transform infrared spectrometer. This technology actually characterizes groups of chemical structures which are found in these complex matrices. From the data derived from a pure material one may hypothesize and be somewhat sure of validating a hypothesis that a particular material is present. However unless one is very skilled in the use of this technique, putting all the chemical groups or parts of molecules together in the correct manner can be difficult, especially when there are many materials combined in one complex matrix. One can easily misinterpret the data from this technique unless one has a solid foundation in molecular spectroscopy.
The fact that one can not narrow down the size of the population of alternative sources beyond a particular level can be viewed to be exculpatory to defendants in courts of law. Reasonable alternative explanations for data which do not point toward guilt could be used to prove defendants innocent. What obligation then do forensic scientists have to explain fully the inability of their analysis protocols to narrow down to unique sources their paint specimens? We can be sure that this obligation is addressed by Giglio and progeny. One could reasonably expect that alternative explanations for data from paint analysis would be an expected part of any expert opinion because it is virtually impossible to source-specify paint beyond a certain point. The breadth of the spread of alternative sources of paint could be viewed in some matters as exculpatory information.
With such considerations as described above in mind, the authors attempted a blind analysis of a known paint standard and an interpretation of the data from those analyses.
The paint was purchased in October 1996 from Davis Frost, Inc. for actual use in painting a part of a building. The paint was labelled “I-47 Battleship Gray.” The authors did not know the chemical composition of the paint when the can was purchased. A few weeks after the paint was consumed, dry paint film on the inner walls of the paint can was removed for analysis.
The paint was initially microscopically observed under an incandescent light source. Then chemical spot tests were conducted. Neither of these analyses, of course, determined what the chemical composition of the paint was. These analyses determined that the paint was gray, that it was not a nitrocellulose-based paint and that it was composed in part of a cross-linked binder (a binder that would not dissolve in the solvents applied to it.) The subsequent instrumental analysis included energy dispersive x-ray analysis coupled with scanning electron microscopy, x-ray powder diffractometry, Fourier transform infrared spectrometry and pyrolysis\gas chromatography\mass spectrometry. The data was interpreted by the authors and analysts who provided the data. The interpretation of the data indicated the following:
The elemental analysis conducted with energy dispersive x-ray analysis in conjunction with scanning electron microscopy indicated the presence of aluminum, silicon, zirconium, molybdenum or sulfur, titanium, calcium and iron. The x-ray powder diffraction data was consistent with the presence of titanium dioxide, and possibly iron oxide. The Fourier transform infrared (FTIR) analysis data was consistent with the presence of an alkyd binder (possibly a medium oil alkyd), titanium dioxide mainly in the form of rutile, and ferric oxide. The FTIR could not determine driers and minor components if any existed. The pyrolysis/gas chromatography/mass spectrometry data (Py/GC/MS) data was consistent with the presence of an alkyd binder.
Essentially then what we know from these analyses is that the paint was a gray alkyd-based paint which contained titanium dioxide and iron oxide pigments. We do not understand the presence of zirconium, aluminum, silicon, or calcium which was detected by the energy dispersive x-ray analysis but was not identified by the x-ray powder diffractometer. The data then narrows this gray paint down to an alkyd-type paint with titanium dioxide pigment (both types of titanium dioxide, rutile and anatase, are present) and iron oxide pigment. We must ask how many gallons of such paint are manufactured which fit this description in order to understand the probative value of this evidence. Following the analysis and interpretation of the data the manufacturer of the paint  provided a list of the components that she combined to form the product. That list contained the following components as well as rough percentages of the components present:
The reader can already see that the forensic analysis did not detect all of the components present in the paint sample but only major components.
The manufacturer also provided the Material Safety Data Sheets (MSDS) for the components and samples of the materials that her company combined to make the paint. The MSDS data lists those components of materials that might create a safety hazard for personnel handling the materials. Therefore all components of such materials are not necessarily listed. However it is instructive to list here the materials which were indicated on the MSDS documentation:
As one can see, the MSDS sheets indicate the presence of at least 38 chemicals in the original paint formula of which forensic analysis “identified” only about four or five. The reader should understand that many of the components listed in the MSDS sheets are volatile and evaporate once the paint is applied. Therefore these materials would not be detected during analysis of a dried paint film. However this complicates the process even further. If one does not know what volatile materials were used in the original paint formula and can not determine this due to their absence, then one can not further narrow down the list of possible sources of wet paint simply from an analysis of dried paint film.
One can quickly see that the forensic analysis protocol that was followed did not determine the chemical composition of the paint specimen.
It is also very important to understand that even if one could determine the qualitative chemical composition of these complex chemical matrices, without a determination of the quantitative chemical composition, one could still not say that paints were chemically identical. These issues are very important when one hears a forensic paint examiner opine that paints analyzed had identical compositions or were “alike in chemical composition.”
One could only know that to be true if one of three conditions applied:
1.) The quantitative and qualitative chemical analyses were possible and were conducted and data was derived and presented which indicated the paints were identical. Lower limits of detection for analytical equipment might very well stand in the way of establishing this data.
2.) Paints of chemical composition known to be the same as the unknown specimens were analyzed and the data from the analysis of the known was consistent with the data from the unknowns. But then the unknowns are not known so how could one test paints of chemical composition known to be the same as the unknown specimens?
3.) The paints had a unique chemical marker in them which was only placed in that particular batch of paint and would definitely indicate that the paints were the same and therefore had the same chemical composition.
Forensic paint analysts who opine in reports and courts of law that chemical compositions of paint specimens or any other complex chemical mixtures are identical should be closely questioned about their interpretation of their data. There are limitations to technology, as well as time and expense limitations which dictate that forensic analysts might not be able to establish these identities as easily as some may think. There is a huge difference between saying that no differences were found between paint samples as a result of forensic chemical tests and saying that there were no differences, that paints samples were identical. In the first statement, that no differences were found, one is left with the very real fact that the analyst may not know if there were chemical differences or not if the technology used did not detect all the components present. By correctly saying that no differences were found as opposed to saying that there were no differences, the expert could very well open the door to valuable examination. That examination could allow counsel to elicit the fact the paint analyst does not know if two paint samples are different or not. That could be very important exculpatory information. The second opinion closes that door, very possibly, inappropriately.
 Executive Director, Forensic Justice Project, Washington, D.C., B.S. Chemistry, 1974, East Carolina University, Ph.D. in Chemistry, 1980, Duke University, J.D., 1996, Georgetown University School of Law. (202)342-6980.
 President, Consolidated Research, Inc. Kingsford, MI, B.S. Chemistry & Biology, 1964, Mundelein College, MS, Chemistry, Iowa State Univ., 1967, Ph.D. Chemistry, Iowa State Univ., 1969. (906) 779-9498.
 A WestLaw computer data base analysis of 1966 to 1995 federal and state cases which contained the words “forensic” and “paint” within the same paragraph noted a spectrum of reported opinions concerning the significance of forensic paint analysis data. The opinions described the level of certainty of paint sample matches as varying from a “correspondence” between paint samples to opinions that samples originated from the same manufacturer and batch of paint, up to paint samples being chemically “identical” to each other.
 Dr. Whitehurst was taught this analytical scheme as a forensic paint analysis trainee in the FBI Laboratory in 1994 through 1996.
 S.G. Ryland & R.J. Kopec, The Evidential Value of Automobile Paint Chips, 24 J. Forens. Sci. 140 (1979).
 Federal Bureau of Investigation, U.S. Department of Justice, Handbook of Forensic Science, 1984.
 Federal Bureau of Investigation, U.S. Department of Justice, Paint Examination Techniques Utilized in the FBI Laboratory, 2,1980.
 John I. Thornton, D. Crim., Forensic Paint Examination, FORENSIC SCIENCE HANDBOOK 529 547 (1982).
 Id. at 550.
 W. Stoecklein, Paints, varnishes and lacquers, THE ENCYCLOPEDIA OF ANALYTICAL SCIENCE (Academic Press Ltd. 1995).
 American Society of Testing and Materials, Standard Guide for Forensic Paint Analysis and Comparison, 1993.
 Supra note 10. 7
 Roy-Keith Smith, Handbook of Environmental Analysis, 89-91(Genium Publishing Corporation, 1994).
 Supra note 8, at 1.
 Marc. S. Reisch, Paints and Coatings, CHEMICAL & ENGINEERING NEWS, September 25, 1995, at 30.
 Giglio v. U.S., 405 U.S. 150 (1972) spells out the obligation of the prosecutor to know of evidence in the possession of the prosecution or members of the prosecution team. Exculpatory information withheld by forensic scientists would be in violation of Giglio.
 Davis Frost, Inc., 3416 Candler’s Mountain Rd. Lynchburg, Virginia, 24502. (804) 846-5277.
 Denise Henning, Vice President, Research and Development, Davis Frost Inc..