Fire Debris Analysis: GC-MS and ASTM E1618 Pattern Recognition

ASTM International has published separate standard practices for each of the three principal fire debris extraction methods. The choice between them depends on the volatility profile of the suspected accelerant, the matrix of the debris, and the quantity of sample available.

Passive headspace concentration (ASTM E1412) is the workhorse method for the majority of case samples and is the default across most accredited fire debris laboratories in the US, UK, and Australia. The sealed evidence container (the paint can collected at the scene) is placed in a laboratory oven at 60 to 80°C for a period typically between 2 and 16 hours. Volatile organic compounds from the debris vaporise into the headspace above the debris. A small activated charcoal strip (ACS), suspended on a metal hook inside the can without touching the debris, adsorbs these vapours onto its surface. After the heating period, the ACS is removed, the adsorbed compounds are eluted with a small volume of carbon disulfide (CS2) or diethyl ether, and the eluate is injected into the GC-MS. Because the charcoal concentrates the vapours, the method achieves excellent sensitivity: residues of gasoline at concentrations of a few micrograms per kilogram of debris are routinely detected. The primary limitation is that the adsorption characteristics of activated charcoal favour mid-range and moderately volatile compounds; very light components (C4 to C6, present in fresh gasoline) and very heavy components (above C20, present in heavy petroleum distillates) may be under-recovered.

Dynamic headspace extraction (ASTM E1413) uses an inert carrier gas (nitrogen or helium) pumped through the sealed container headspace and then through an adsorbent trap, typically Tenax TA or a multi-sorbent bed. The trapped vapours are thermally desorbed directly into the GC injection port. The method recovers lighter components more efficiently than passive headspace and is particularly useful for weathered or heavily suppressed samples where only the lighter fractions may remain. It is also the method of choice when the analyst wants a real-time vapour profile without removing the debris from the original evidence container. Equipment cost and the requirement for a thermal desorption unit mean that dynamic headspace is less universally available than passive charcoal strip methods.

Solvent extraction (ASTM E1386) involves direct extraction of the debris with a solvent, most commonly pentane, hexane, or diethyl ether, using agitation or ultrasonic bath to drive compound transfer into the liquid phase. The extract is then filtered, concentrated if necessary, and injected. Solvent extraction recovers heavier residues (diesel, fuel oil, heavy petroleum distillates above C15) more efficiently than headspace methods, because these high-boiling compounds do not partition significantly into the vapour phase at 60 to 80°C. The limitation is interference: the solvent itself, co-extracted polar compounds from char and ash, and very high-boiling substrate components all appear in the chromatogram and can complicate interpretation. Solvent extraction is typically reserved for debris suspected of containing heavier fuel products, or as a complementary method when headspace extraction has failed to produce a classifiable pattern.

The gas chromatograph separates compounds in the fire debris extract by their differential affinity for the stationary phase of the capillary column and the mobile phase (carrier gas). For fire debris analysis, the standard column is a 5 percent phenyl-95 percent dimethylpolysiloxane phase (DB-5 or equivalent), 30 metres long, 0.25 mm internal diameter, 0.25 micrometre film thickness. This non-polar to slightly polar stationary phase gives adequate resolution for the C8 to C25 carbon range that encompasses most ignitable liquid components, with elution in order of increasing boiling point within each compound class. The carrier gas is helium (preferred for sensitivity) or hydrogen (faster runs, higher linear velocity, increasingly common in laboratories managing helium supply costs under post-pandemic gas shortages).

The temperature programme is the primary instrument variable for fire debris work. A typical programme begins at 40 to 50°C (held for 2 to 4 minutes to allow the solvent front to elute before compounds of interest), ramps at 5 to 10°C per minute to a final temperature of 270 to 300°C (held for 5 to 10 minutes to elute heavy components). The ramp rate trades run time against resolution: a faster ramp (10°C/min) gives a 35-minute total run time adequate for pattern recognition but with less separation within congested regions; a slower ramp (5°C/min) gives better resolution of co-eluting components at the cost of 60 to 70 minutes per injection. Accredited fire debris laboratories typically use a validated method with a documented ramp rate and hold conditions; method changes require revalidation and proficiency testing.

The mass spectrometer is operated in full-scan mode (typically m/z 35 to 350 or 40 to 550) to capture the entire mass spectral fingerprint of each eluting compound. Full-scan acquisition is essential for pattern recognition: the analyst needs to reconstruct extracted ion profiles (EIPs) at specific m/z values characteristic of compound classes. Selected ion monitoring (SIM) mode, which detects only pre-specified ions and achieves lower detection limits, is not appropriate for pattern recognition because it sacrifices the compound-class information needed for E1618 classification. Some laboratories use full-scan acquisition with simultaneous library searching against the NIST/EPA/NIH Mass Spectral Library for individual compound identification, which supports but does not replace the pattern-based classification.

The injector is typically a split/splitless inlet, operated in splitless mode for maximum sensitivity. Injection volume is 1 to 2 microlitres of the concentrated CS2 or solvent eluate. The CS2 solvent is preferred for charcoal strip eluates because its mass spectrum is relatively simple and its main fragment ions (m/z 76 and 44) do not overlap with the characteristic ions of most ignitable liquid compound classes. The solvent delay (ion source filament off during solvent elution) is typically 2 to 4 minutes.

ASTM E1618 (current edition 2022) classifies ignitable liquids into eight categories based on the chromatographic and mass spectral pattern of their components. The classification rests on extracted ion profiles (EIPs) at specific m/z values that are characteristic of particular compound classes present in petroleum products.

The eight E1618 categories are:

1. Gasoline. The most common accelerant in arson cases worldwide. Identified by its characteristic combination of the aromatic compounds (benzene, toluene, ethylbenzene, xylenes, C3-benzenes, C4-benzenes and naphthalenes) in the EIP at m/z 91, 105, 119, 128, 142; together with C6 to C12 aliphatic and cycloaliphatic compounds in the total ion chromatogram. Fresh gasoline produces a distinctive "humped" TIC envelope in the C6 to C12 region; weathered gasoline loses the lighter fractions and may resemble a medium petroleum distillate if substantially evaporated. The forensic chemistry treatment of petroleum products and lubricants provides a complementary perspective on fuel composition that aids inter-laboratory comparisons.
2. Petroleum distillates (light). Carbon range approximately C6 to C10. Examples: mineral spirits (white spirit in UK/EU nomenclature), naphtha, lighter fluid. EIP at m/z 57 (branched alkanes) and m/z 71 (normal and branched alkanes) shows a smooth distribution without the aromatic spike pattern of gasoline.
3. Petroleum distillates (medium). Carbon range approximately C8 to C13. Examples: kerosene (paraffin in UK), jet fuel (Jet A, JP-8), charcoal lighter fluid. The C8 to C12 normal alkane series shows as a characteristic picket fence of evenly spaced peaks on the TIC; EIP at m/z 57 and 71 confirms the aliphatic character. Kerosene and jet fuel are distinguished from gasoline by the dominance of the normal and branched alkane series over the aromatic fraction.
4. Petroleum distillates (heavy). Carbon range approximately C9 to C24 and above. Examples: diesel, fuel oil, lubricating oil. The TIC shows a broad envelope of overlapping n-alkane peaks in the C10 to C24 range; below-baseline unresolved complex mixture (UCM) hump is characteristic. The EIP at m/z 57 shows the n-alkane series extending to C24 or beyond.
5. Isoparaffinic products. Dominated by branched alkanes with little or no aromatic or n-alkane content. Examples: Isopar solvents (ExxonMobil), paint thinners formulated from isoparaffinic hydrocrackate. EIP at m/z 57 shows a smooth envelope; m/z 91 (aromatic) is essentially absent; the normal alkane picket-fence pattern at m/z 57 is absent or very weak.
6. Aromatic products. Dominated by C6 to C9 aromatic compounds. Examples: xylene-based solvents, aromatic naphtha, some paint thinners. EIP at m/z 91, 105, 106 shows a pattern of alkylbenzenes; the aliphatic fractions at m/z 57 and 71 are weak.
7. Naphthenic-paraffinic products. A blend of cycloalkanes (naphthenes) and alkanes. Examples: de-aromatised white spirit, some dry-cleaning solvents. EIP at m/z 55 and 69 (cycloalkane fragments) combined with m/z 57 and 71 (alkane fragments) with minimal aromatic content.
8. Oxygenated solvents. Alcohols, ketones, esters, and glycol ethers that do not fit the petroleum distillate categories. Examples: acetone, methanol, ethanol, methyl ethyl ketone, ethyl acetate. These are identified primarily by the TIC pattern, the presence of characteristic oxygen-containing fragment ions (m/z 31 for methanol, m/z 43 for acetone and many ketones), and the absence of the alkylbenzene and n-alkane patterns that define the petroleum categories.

An ignitable liquid undergoes continuous chemical change from the moment it is poured until the laboratory analyst injects the extract. Three degradation processes are relevant to pattern recognition: evaporative weathering, combustion, and water leaching.

Evaporative weathering removes the low-boiling fraction preferentially. Gasoline, which has components from approximately C4 to C12, loses its C4 to C7 fraction rapidly, especially in the high-temperature environment of a fire. A gasoline that has lost 60 to 70 percent of its mass by evaporation will show a TIC pattern shifted to heavier components, the C9 to C12 aromatic fraction dominating, with the light C6 to C8 aliphatics reduced to trace levels. This pattern can visually resemble a medium petroleum distillate. The E1618 guidance notes that significantly weathered gasoline is identifiable by the presence of the characteristic high-carbon aromatic compounds (C3-benzenes, C4-benzenes, naphthalene, methylnaphthalenes) that persist even after heavy weathering; a medium petroleum distillate contains far less aromatic compound in that range.

Combustion directly destroys hydrocarbon molecules. Ignitable liquid that is within the flame zone during combustion is largely consumed; residue survives at the margins of pour areas, beneath objects that shielded the floor from the flame, and absorbed into porous substrate below the combustion zone. The residue that survives is a depleted mixture that may be dominated by the least volatile compounds and may show the overall carbon number distribution shifted upward by one to three carbons compared with the original product. Pattern recognition under these conditions relies on the relative ratios of the surviving compounds and their characteristic fragment ions, not on the absolute pattern of the original product.

Water from fire suppression can dissolve or displace water-soluble components (primarily the lower-molecular-weight aromatics including benzene and toluene) and physically mobilise the remaining residue across the debris. In concrete floors with cracks or joints, water-driven migration can move accelerant residue metres from the pour location. This means that a positive GC-MS result in one sample from a given location does not establish that accelerant was poured at that specific location; it establishes that ignitable liquid residue was present in that sample, which may or may not correspond to the original pour area.

Fire debris analysts account for weathering by applying the full E1618 criteria systematically rather than pattern-matching to ideal reference spectra. The SWGFEX best practices document (Section 6, Interpretation) recommends that analysts document which E1618 criteria are met and which are not, and that the classification opinion specify whether the pattern is consistent with a fresh or a weathered/degraded form of the classified liquid. The UK Forensic Science Regulator's guidance on fire debris similarly endorses this practice as producing a more informative and defensible opinion than a bare classification without qualification.

The Organisation of Scientific Area Committees (OSAC) for Forensic Science, convened by NIST, has developed specific criteria for the language that fire debris analysts use in reports and testimony. OSAC's Fire and Explosion Investigation Subcommittee published a needs statement on opinion language in 2021 that is incorporated by reference in several US state accreditation programmes and referenced in the UK Forensic Science Regulator's 2023 guidance update.

The OSAC-endorsed opinion for a positive identification has three required components. First, the identification of the class of ignitable liquid detected, using E1618 category nomenclature (for example, "medium petroleum distillate" rather than "kerosene" unless a specific product match has been established by quantitative comparison with a reference sample of the actual product). Second, a statement of the analytical basis: the extracted ion profiles and pattern recognition criteria that support the classification. Third, a statement addressing the comparison sample, either confirming that the debris pattern was not explainable by substrate background or identifying the features that remained after background subtraction.

A SWGFEX-compliant identification opinion reads approximately: "The analytical results are consistent with the presence of a medium petroleum distillate in the fire debris sample. The pattern of straight-chain alkanes in the C8 to C13 carbon range, confirmed by extracted ion profiles at m/z 57 and 71, is consistent with a kerosene-range petroleum product. The comparison sample from unburned substrate in the same room did not produce a comparable pattern, and the chromatographic features observed in the debris sample are not attributable to substrate background alone."

A SWGFEX-compliant non-identification opinion reads: "No ignitable liquid residues were identified in the fire debris sample. Chromatographic features present in the sample are consistent with pyrolysis products of the substrate as demonstrated by comparison with the control sample from unburned substrate."

In India, the CFSL Standard Operating Procedure for fire debris analysis (updated 2019) references ASTM E1618 as the interpretive framework but notes that the specific product categories in E1618 may require supplementation with reference chromatograms of locally available petroleum products (Indian-grade kerosene, mineral turpentine oil, and white petroleum naphtha) because formulations differ from the US and European reference products used to develop the E1618 category descriptions. The Central Forensic Science Laboratories at Hyderabad, Mumbai, and Chandigarh maintain product reference libraries for Indian commercial fuel formulations.

Fire debris laboratories in all major jurisdictions operate under ISO 17025 accreditation, which requires documented quality assurance procedures including method validation, proficiency testing, instrument calibration, and control chart maintenance.

Reagent blanks are run with every batch of passive headspace extractions to confirm that the charcoal strips, the CS2 eluent, and the glass vials used for elution do not contribute detectable hydrocarbon contamination to the extract. A reagent blank that shows peaks at known ignitable liquid retention times invalidates the entire batch. The source of the contamination must be traced before any case samples from that batch can be reported.

Positive controls are extracts of reference ignitable liquid in a known substrate matrix (typically laboratory sand or cellulose), extracted by the same method as case samples and carried through the same analytical sequence. Positive controls verify that the extraction procedure recovered the target compound class at the expected pattern and concentration. Falling positive control response may indicate degraded charcoal strip performance, oven temperature drift, or MS sensitivity decline.

Proficiency testing for fire debris analysts is provided in the US by the Collaborative Testing Services (CTS) Fire Debris Proficiency Programme, which distributes known-content samples to participating laboratories twice annually. The programme has been running since 1990 and the results are publicly available; it is one of the longest-running proficiency datasets in forensic chemistry. The UK Forensic Science Regulator requires participation in UKAS-accredited external quality assurance schemes as a condition of accreditation under the Codes of Practice and Conduct. In the European Union, the ENFSI (European Network of Forensic Science Institutes) Fire Investigation Working Group coordinates proficiency testing across member laboratories.

1. Prepare the passive headspace batch
   Label new activated charcoal strips with case and sample numbers. Prepare a reagent blank strip in a clean vial. Prepare a positive control extract from reference ignitable liquid in sand matrix.
2. Load the evidence container and heat
   Suspend the labelled ACS on the hook inside the evidence can. Do not allow the strip to contact debris. Seal and place in the laboratory oven at 65 to 70°C for the validated extraction period (typically 16 hours overnight).
3. Elute and prepare injection vials
   Remove the ACS using clean forceps. Place in a clean glass vial. Add 0.25 to 1.0 mL CS2. Cap, vortex for 30 seconds, centrifuge briefly. Transfer eluate to a GC autosampler vial. Include reagent blank and positive control eluates in the same sequence.
4. GC-MS acquisition
   Inject 1 microlitre splitless onto DB-5 column. Run validated temperature programme (40°C hold 2 min, ramp at 8°C/min to 280°C, hold 10 min). Acquire full-scan mass spectra m/z 40 to 350 throughout.
5. Extract ion profiles and compare comparison sample
   Generate EIPs at m/z 57, 71, 91, 105, 119, 128, 142, 55, 69 as applicable. Overlay the debris sample EIPs with the comparison sample EIPs. Identify features present in the debris sample that are absent from the comparison sample.
6. Classify per E1618 and draft the opinion
   Apply E1618 pattern recognition criteria to the background-subtracted pattern. Document which criteria are met. Draft the opinion in OSAC/SWGFEX-compliant language, specifying the category identified and whether the pattern is consistent with a fresh or weathered form.

Documentation of instrument performance is a routine laboratory obligation that can become critical in litigation. Defence experts in major arson trials routinely subpoena instrument maintenance logs, column replacement records, detector calibration histories, and reagent lot records. A laboratory that cannot produce these records for the period when a case sample was analysed faces a significant evidentiary challenge independent of the quality of the actual analysis. The FBI, the Metropolitan Police Forensic Services, and the major accredited private laboratories in India (such as those under NABL accreditation from the Bureau of Indian Standards) all maintain searchable electronic audit trails for their forensic chemistry instruments.