Ed. Note: Chris Halkides has been kind enough to try to make us lawyers smarter by dumbing down science enough that we have a small chance of understanding how it’s being used to wrongfully convict and, in some cases, execute defendants. Chris graduated from the University of Wisconsin-Madison with a Ph.D. in biochemistry, and teaches biochemistry, organic chemistry, and forensic chemistry at the University of North Carolina, Wilmington.
In 1989, Han Tak Lee and his daughter were staying in a 1000 square foot cabin in Pennsylvania when a fire destroyed it, claiming his daughter’s life. In 1990, Mr. Lee was convicted of first degree murder and arson, and sentenced to life without parole. The evidence against him included identifying one ignition point of the fire by the presence of crazed glass, thin, irregular fractures. The collapsed furniture and bed springs, spill patterns (pour patterns?), and deep charring and alligatoring of wood were also taken as evidence of a very hot fire; therefore a deliberately set fire. Crazed glass is created by cold water hitting hot windowpanes. Nor are collapsed furniture springs, pour patterns, alligator patterns or depth of charring of wood accepted any longer as evidence of arson.
In order to understand the laboratory evidence against Mr. Lee, one must have an understanding of what gasoline is and how it is detected. Hydrocarbons are a class of chemical compounds that contain hydrogen, carbon, and no other elements. Gasoline is a mixture of hydrocarbons having 7-12 carbon atoms. About 40% of gasoline consists of isoalkanes, compounds of the formula CNH2N+2, in which the carbon chain is branched, not linear. 20-50% of gasoline consists of alkylbenzenes, many of which have the formula C8H10. Many common household items (turpentine, shoe soles, etc.) have hydrocarbons in them that are not gasoline.
Gas chromatography (GC) uses a column with a waxy liquid to separate molecules on the basis of the fraction of time they spend stationary (dissolved in the waxy liquid) versus the fraction of time they spend moving alongside a carrier gas. A gas chromatogram is a trace in which time is then displayed along x-axis and detector output is displayed along the y-axis. The time at which the signal reaches its maximum detected intensity is the retention time of that compound. Compounds with a greater number of carbon atoms generally have higher boiling points and longer retention times than compounds with fewer carbons. The relative signal areas within a mixture are related to the relative amounts of the compounds.
Some detectors respond to many classes of compounds without giving further indication of their identities. Therefore, the equality of the retention times between a standard versus a questioned compound may occur because they are identical compounds or different compounds that emerged from the column at the same time coincidentally. When GC is immediately followed by mass spectrometry (MS), the combined discriminatory power is greatly increased. Mass spectrometry gives the mass of each compound and its fragments, which often leads to its unambiguous identification. The Patricia Stallings case would probably not have come to trial if GC-MS had been used instead of GC alone.
Mr. Lee’s case returned to the courtroom in 2014, and some of the focus then turned to the other testimony that was used to convict him. One of the investigators performed calculations indicating that 62 gallons of home heating fuel and 12.2 pounds of gasoline had been used to set the fire. Yet the report stated that only a furnace fuel filter and a container next to the fuel tank by the cabin produced evidence of an ignitable substance. The prosecution used results from gas chromatography to allege that Mr. Lee’s clothing contained “a volatile substance with a hydrocarbon range of C7-C22,” meaning that the substance was a mixture of molecules with 7 to 22 carbon atoms. By the time that the case was appealed, the gas chromatograms were no longer available.
The laboratory chemist who testified suggested the presence of a substantial quantity gasoline or Coleman fuel, despite no mention of this in the report. Paradoxically, he also testified that there wasn’t a sufficient quantity of gasoline to confirm by his instrument. He claimed that the range of hydrocarbons (meaning the number of carbon atoms) found on Mr. Lee’s clothing was the same as found on plastic items (probably a glove and a plastic jug) in the bathroom. Mark Hansen wrote that the chemist testified that, “his analysis of the evidence revealed chemical hydrocarbon profiles that were consistent with the same mixture of home heating fuel and gasoline or Coleman fuel” that had been claimed to be the accelerants by a different witness. The putative presence of hydrocarbons with up to 22 carbon atoms led the chemist to suggest that two accelerants (gasoline and fuel oil) had been used. Yet he denied under oath that mass spectrometry would have helped identify the other liquid.
The contention that 62 gallons of fuel oil could be used yet none detected in the fire debris beggars belief. The calculations suggesting this quantity of accelerant were fallacious. In the absence of data from mass spectrometry, the claim that hydrocarbons were present on Mr. Lee’s clothing is highly dubious. Aldehydes, for example, are sometimes found on clothing, and aldehydes have at least one oxygen atom, putting them into a different class from hydrocarbons. Nor would finding hydrocarbons on his clothing be the same thing as finding gasoline. A claim of identity of two mixtures based upon the similarity in the number of carbon atoms is nonsense. For one thing, the heights of the peaks in the chromatogram would have to be similar; the components of turpentine have a similar range of retention times compared to the components of gasoline, yet they are very different hydrocarbons. The claim that MS would not have been helpful is almost impossible to ascribe to incompetence; this and other problems in the chemist’s testimony probably originate from his consciously attempting to bolster the prosecution’s case. This problem is distinct from unconscious bias that plagued the Brandon Mayfield fingerprint case or the fire aboard the USS Bonhomme Richard.
The contention that 62 gallons of fuel oil could be used, yet none detected in the fire debris, beggars belief. The calculations suggesting this quantity of accelerant were almost certainly fallacious. In the absence of data from mass spectrometry, the claim that hydrocarbons were present on Mr. Lee’s clothing is dubious. Aldehydes, for example, are sometimes found on clothing, and aldehydes have at least one oxygen atom, putting them into a different class from hydrocarbons. Nor would finding hydrocarbons on his clothing be the same thing as finding gasoline. A claim of identity of two mixtures based upon the similarity in the number of carbon atoms is nonsense. For one thing, the heights of the peaks in the chromatogram would have to be similar; the components of turpentine have a similar range of retention times compared to the components of gasoline, yet they are very different hydrocarbons. The claim that MS would not have been helpful is remarkable.
Mr. Lee was released circa 2014 and remained free despite an appeal by the prosecution. Given the era, the use of crazed glass as a putative indicator of arson is not surprising. The degree to which laboratory results were misinterpreted to fit the prosecution’s narrative, discrepancies between the report and the testimony, and the lack of availability of the GC chromatograms for defense analysis, are equally disconcerting.
For further reading
Valena E. Beety & Jennifer D. Oliva, Evidence on Fire, 97 N.C. L. Rev. 483 (2019).
Mark Hansen “Long-held beliefs about arson science have been debunked after decades of misuse” ABA Journal 1 December 2015.
John Lentini “A Calculated Arson,” Fire & Arson Investigator April 1999 20-25.