by Wal Stern
Accurate detection of an accelerant can help police in arson investigation. Prof. Wal Stern explains how, by using modern methods of chemical analysis, even a small quantity from a dismal fire scene can be detected, recovered and identified.
The detection of an accelerant at a fire scene, where one is not normally present, is a very strong indication of arson (a deliberate fire). Cooke and Ide's text "Principles of Fire Investigation" states that their presence "is almost conclusive evidence of intent and preparation provided they are not normal to the premises".
An accelerant in this context is any substance, nearly always a liquid, which has been placed at the fire scene to facilitate the spread of a fierce and fast blaze. The most common accelerants found are petrol, kerosene, mineral turps and diesel - all mixtures of hydrocarbons derived from all mixtures of hydrocarbons derived from petroleum. Other accelerants found include ethanol or methylated spirits, and acetone.
It might be assumed that most, if not all, accelerants would be consumed in a fire or evaporate. However, modern methods of chemical analysis can detect very small quantities of residual material, even in the dismal scene of a fire. Detection at the parts per million level can fairly readily be achieved.
The steps which need to he taken to recover and identify an accelerant are:
- To collect samples;
- To extract the fire debris and thereby obtain a sample fit for instrumental analysis;
- To carry out the instrumental analysis; and
- To interpret the results obtained.
The locations from which samples are taken should he based upon the physical evidence such as burn patterns, V-patterns, hum-throughs, trails and so on. Some investigators use electronic sniffers to assist and in recent years trained dogs have also been used. They may be useful in determining which sites to sample but they can never be a replacement for laboratory analysis.
Samples should be of materials which are absorbent or adsorbent. Timber, cloth, carpet, paper and soil are good in this respect. Charcoal is a very good adsorbent, as anybody who has taken charcoal pills for a stomach disorder knows. On the other hand, materials such as glass, plastic or concrete are not good adsorbents and are less likely to give positive results.
The samples should be stored in containers where they will not be contaminated. The best containers are unlined, clean, metal paint cans. Nylon bags can also be used and glass jars can be used if nothing else is readily available.
The aim of extracting the fire debris in the chemical laboratory is to separate and concentrate the accelerant from other debris such as burnt timber, paper, plastic, carpet and so on. Many extraction methods have been used over the years including distillation and solvent extraction but they are not commonly used today since they lack sensitivity.
The methods now commonly used are:
- to place a charcoal or Tenax-coated wire or strip into a can of fire debris at normal or elevated temperature (passive absorption), or
- to draw vapour from the sample container through charcoal or Tenax (headspace adsorption) at normal or elevated temperature, or
- to heat the can containing debris and sweep the vapours with an inert gas through a charcoal or Tenax plug (dynamic headspace adsorption).
Different laboratories favour different methods. In each case, the volatile components would have been highly concentrated and would be ready for instrumental analysis. They can be washed off with a solvent such as carbon disulphide or heated off the absorbent and injected into a chromatogram for analysis.
Gas chromatography is ubiquitously used as the preferred technique in this type of analysis. It is an analytical method which separates mixtures and indicates the relative quantity of each component on the basis of the component's volatility, solubility and absorption. In simple terms, they separate liquids on the basis of their boiling points.
The results are displayed as a graph showing a number of peaks. One scale depicts the amount of each constituent, the other the time taken for it to emerge from the instrument. The chromatogram of a single component should yield a single peak. Under the same conditions, the time taken for the component to emerge will always be the same and as such it may be identified.
In the case of a mixture of two components, one of which is present three times as much as the other, one should then obtain two peaks - one three times as large as the other. On this basis, identification of accelerants is based upon pattern recognition of the number of peaks, the position of peaks and their relative sizes. To achieve the best results, the use of the right columns, care of columns, pure gases, rigid temperature control and other parameters are essential. Improvements are being made to give unambiguous results. For example, long capillary columns have replaced in most cases the wide bore columns used years ago.
If the conditions are carefully controlled, it becomes relatively easy to identify pure petroleum fractions since they are normally composed of a large number of identifiable components. In the Geronimo Laboratory at the University of Technology, Sydney, a dual plot of the sample against a similar reference standard is always run.
Interpretation of results
Analysis of a standard kerosene sample shows eight evenly-spaced major peaks, with a number of minor intermediate peaks. From other analyses, the identity of each of the eight components, their formula and chemical structure are known.
These eight peaks are also present in a chromatogram of diesel but in different proportions. Diesel also contains an additional eight evenly-spaced peaks which emerge later from the instrument (higher boiling components).
The spectrum of pure petrol shows a multitude of peaks of different proportions from which at least seven known components must be present, in correct relative proportions, to be identified as petrol.
In a similar way, other accelerants can also be identified by recognising a number of components and the relative proportions of these components. The American Society for Testing and Materials(ASTM) has a standard test method for dealing with fire debris, separating accelerants into five classes and defining the chromatographic characteristic of each class. Analysts following this method should always agree on a finding. The ASTM method is not mandatory in Australia, the United Kingdom or Singapore. It can be argued that a Standard should be introduced.
Chromatographic analysis of fire debris would not normally be of pure accelerant. The material would probably have evaporated, partially distilled off and been contaminated with other substances.
Many experiments have been conducted to evaporate and burn accelerants, to recover known accelerants from test fires, and to contaminate them with likely impurities. The chromatograms vary as samples evaporate or burn, in a predictable manner (the more volatile components decrease in proportion, the less volatile components increase).
In the case of petrol, it is still often possible to identify its components. Evaporated kerosene can look like diesel (they may contain the same components and they may be present in similar proportions, the lighter components having been evaporated).
The presence of other substances, however, can make the task difficult. Over the years, one learns to recognise the presence of burnt carpet, burnt plastic or wood oils, but they can confuse and distract. The ASTM standard quite clearly defines when one can confidently define an accelerant (for example, no intermediate component can be missing). Notwithstanding this, doubts remain, particularly in samples which are contaminated or highly evaporated, or present in only trace amounts. Contamination occurs from materials present before the fire such as wood oils and from decomposition products formed in the fire.
In instances where contamination peaks are in the same region as suspected accelerant peaks, it is imperative to identify and quantify each significant component by mass spectrometry. Chromatography only characterises components on the basis of the time they take to emerge from the instrument. It does not positively identify their formula and chemical structure. Computer techniques are also used, in conjunction with mass spectrometry, to visualise certain types of hydrocarbons.
The use of mass spectrometry doubles the cost (from around $100 per sample to around $200) but it does provide the formula and structure of each component which may not be necessary for all samples but it is certainly worthwhile for some. It can be carried out months after the initial testing if the sample extracts are properly stored.
On the other hand, if one extracts by passive adsorption and automatically carries out gas chromatography and mass spectrometry for all samples, the cost could probably be reduced to around $150. Some American laboratories have moved along these lines but many still rely heavily on just gas chromatography. ¤