Introduction
Arguably the most critical point in a fire investigation is the interface between the fire scene investigator and the forensic science laboratory. This is because, in most instances, the scene investigator does not conduct the laboratory testing or present the laboratory results. This is particularly so in cases involving criminalproceedings.
As a consequence, in these cases the presiding magistrate or members of a jury must determine the significance of the test results in the context of evidence given by the scene examiner. Predictably, other evidence, such as motive, will also become available during the hearing but the scientific evidence concerning the actualstarting of the fire should not require the support of other types of evidence. Obviously, other matters, such as who was responsible for the starting of the fire, will require other evidence.
The Purposes of Collecting Samples
The collection of samples and items at fire scenes can serve several purposes. Undoubtedly, most items are taken for subsequent examination by the fire scene investigator or laboratory scientists with a view to obtaining information to support ideas and opinions established during the scene investigation.
Although this is a justifiable procedure, it should be remembered that any results obtained from the testing do not necessarily prove the validity of the scene diagnosis. For example, positive results from tests designed to investigate a diagnosed ignition cause do not prove that this was the actual cause. These results merely confirm that such a theoretical cause is a practical concept. Furthermore, any negative results do not prove that the diagnosed cause could not have occurred. In these cases, testing would have to be repeated to ascertain the likelihood that a fire would commence in this way; that is, an investigation would determine the statistical improbability of this particular ignition source. So long as there is a theoretical possibility that a fire can start in a particular way, negative testing will only provide evidence as to how unlikely this is, not that it is impossible.
A classical example of this philosophical viewpoint in the interpretation of test results is fires in bedding that are caused by smoldering cigarettes. Simple tests using lighted cigarettes and cotton sheeting show that it is difficult (or a statistical improbability) to start fires in this way. However, it is well established that there are many deaths caused through smoking in bed. Therefore, the testing proves that, for every death that occurs, many people survive because cigarettes usually burn out in this situation.
It is obvious from the preceding discussion, that the quality and quantity of items and samples have the potential to impact profoundly on laboratory results and, therefore, on the fire-cause diagnosis. This article addresses the scientific considerations for sampling, including the necessary control and reference samples. Furthermore, because samples have such scientific significance, they can have evidential value in any subsequent criminal or civil hearings. Therefore, legal considerations, such as item security and item continuity, must be addressed in order to satisfy jurisdictional and legislative requirements.
Sampling from fire scenes also serves other purposes. Important examples include research and education into product performance under actual fire exposure conditions. The results from these observations can lead to academic research and investigations into product performance, safety and improvement.
The recording of the performance of materials subjected to fire under different conditions assists the fire investigator to ‘read’ the fire’s progress and therefore to diagnose a fire origin or starting point. However, this product performance can also provide the necessary evidence required for a product re-call and also lead to recommendations for changes and improvements in national standards for materials used in building, furnishing, clothing etc.
Sample Classification
Before considering the scientific aspects of items and samples, it is important to classify samples according to their intended functions. This will allow identification of sampling requirements and of the limitations on information that these samples can provide. The following definitions apply and can, in fact, apply to items collected from scenes of incidents other than fire.
Test sample
A test sample is taken to provide specific information that will assist in reaching conclusions or establishing facts concerning the incident. Examples include clothing with biological stains, vehicle panels with paint smears and floor coverings containing flammable liquids. A feature of test samples is that they usually consist of a substrate containing a foreign material.
Control sample
A control sample is taken for comparative purposes and provides information that is often necessary to determine the significance of the results produced from test samples. This is achieved by sampling and analyzing the material that comprises the substrate of the test sample, and eliminating these results from the test sample results. Examples of control samples include carpet, floorboards, soil and other fire debris. The extraction of natural products such as lanolin from woollen carpet underlay, and terpenes from timber when collected as control samples, allows the scientist to account for these materials when interpreting the analytical results from test samples. Because these samples are basically used for elimination purposes, an important feature of control samples is that they can fulfill their purpose without the necessity of a complete scientific identification of their composition or of substances that occur naturally within them. Control samples are not always required but should always be considered.
Reference samples
A reference sample allows an absolute scientific or technical identification to be made of a test or control sample, or of an extract from a test or control sample. Because a reference sample is used for comparison purposes, it might also serve as a control sample under certain circumstances. For any sample to serve as a reference sample it must satisfy one important criterion, that is, the material must have scientific and/or technical evidence of its identity and composition. The most common reference samples used by the laboratory in fire investigations are fuels and solvents provided by oil refineries. However, there might also be the need to examine reference samples of carpet when reporting and explaining the significance of burn patterns or the identity of pyrolysis products. The manufacturer would normally supply such samples but, providing comprehensive technical information is made available, a carpet supplier could provide the necessary samples.
Samples from Fire Scenes
As previously discussed, samples and items are collected for various purposes. However, by far the most common reason for collecting samples from fire scenes is to provide information concerning the ignition of the fire. Furthermore, the majority of these samples consist of fire debris suspected of containing flammable liquid residues.
Selection of sampling sites
The sampling sites for fire debris samples are usually selected after an interpretation of the ‘fire travel indicators’. These allow the fire investigator to interpret the fire travel direction and intensity and, therefore, determine the fire origin or starting point. Whether the ignition of the fire was assisted by the presence of flammable liquids is decided after a consideration of various observations and other empirical evidence (see below).
Over the past 20 years, some fire investigators have, to a certain extent, avoided the orthodox scene inspection procedure and have resorted to the use of hydrocarbon vapor detectors known as ‘sniffers’ to ‘screen’ the scene for the presence of fuel and solvent vapors.
‘Sniffers’ have undergone development from the initial simple instruments based on a color change in crystals exposed to hydrocarbons, to sophisticated portable gas chromatographs. Fire investigators must be familiar with the specificity and sensitivity specifications of an instrument before using it in the field and herein lies the dilemma.
Many thermal breakdown products, known as pyrolysis products, are present in fire debris as a result of the thermal decomposition of natural and synthetic products. Many of these products are very similar in their chemical structure to aliphatic and aromatic compounds occurring in petroleum fuels and solvents. Therefore, a positive reading from a ‘sniffer’ should not be regarded as confirmation of petroleum compounds but only as an indicator of their possible presence.
Also, petroleum products often undergo severe evaporation when exposed to an intense fire. This evaporation will continue after the fire has been extinguished and before sampling occurs. In fact, it is not uncommon to recover residues of fuels and solvents that only consist of the highest boiling point components of the original mixture. These chemical compounds can exert such low vapor pressures that the ‘sniffer’ cannot detect them.
As a consequence of these limitations of ‘sniffers’, the fire investigator must make a decision based on experience as to whether a sample should be taken from a promising site, regardless of a negative response from the ‘sniffer’, and whether a sample should not be taken from an unlikely site regardless of a positive response from the ‘sniffer’.
The question then arises whether the ‘sniffer’ is serving any purpose at all and, if not, why use it at all?
The situation is somewhat different when the ‘sniffer’ is a portable gas chromatograph. If the use of this instrument produces a chromatogram that is definitive evidence of a flammable liquid, then it is providing valuable investigative information that has the potential to become evidence in any subsequent court hearing.
There is no substitute for a competent and logical scene investigation when diagnosing fire origins. The presence of flammable liquids at a fire scene has some significance but the presence of a flammable liquid at the diagnosed fire origin is of greater significance.
Fire debris samples
Samples of fire debris usually consist of floor coverings, floorboards and soil but burnt clothing, curtains, soft toys, bedding, etc. are also collected and submitted for examination. If maximum information is to be gained from these items, it is essential to have at least a basic knowledge of the chemical structure of these materials and their burning characteristics. For the more common sampling media, the following information applies.
Floor coverings When these materials burn with the assistance of flammable liquids, there are usually characteristic ‘pooling patterns’ produced that are characterized by a sharp demarcation between burned and unburned material. However, it should be remembered that similar patterns can be produced in other ways, for example, falling drapes, burning wallpaper and protection provided by furniture and fittings. In most cases, an experienced fire investigator will be able to distinguish the actual cause of any burning pattern and will sample accordingly.
If it is decided that sampling is warranted, the sample must include both burned and unburned carpet (that is, the sample should be taken at the edge of the burn) and any underlay must be included. Normally an area of at least 1 m2 should be taken and this is best achieved using a large heavy-duty knife.
Floor boards So long as holes in flooring are not too extensive, the fire investigator should be able to distinguish the cause of these holes. Burning debris usually causes holes burnt from the upper surface, whereas holes burnt from the lower surface can often be associated with a flammable liquid. In the latter case, if the burning is not too extensive, the characteristic ‘tongue and groove’ burning patterns are often present. This pattern arises when the liquid seeps through the cracks in adjoining floorboards and burns up through this seam. This can result in a series of holes in the flooring that follow the joins of the floorboards. The pattern can occur without the involvement of flammable liquids, but its presence always demands a close inspection.
If it is decided to take a sample, the sample must include the unburned material on each floorboard in addition to the charred areas. These samples should always be accompanied by control samples because wooden flooring usually contains natural products that have the potential to complicate any laboratory analysis. Sampling is best achieved using an axe or a saw. If a power saw is used then serious contamination considerations arise if the saw is fueled by a petroleum product.
If the flooring has been burned to such an extent that the holes are almost joined, it is extremely unlikely that any flammable liquid residues will remain and sampling is probably worthless.
Soil In many instances where the involvement of a flammable liquid is suspected, but floor coverings and flooring are too severely charred to warrant sampling, consideration should be given to sampling the soil underneath the floor. In these cases, the top few centimeters of soil should be removed for testing using a clean spade or trowel. Again, a control sample should be taken, although it will probably not be used because extracts from soil are often relatively free of contamination.
For all the above cases, the taking of control samples is a consideration. They should be collected from areas that are confidently diagnosed as not being involved in the starting of the fire. In many instances they will not require subsequent scientific examination, but their existence can sometimes be of great assistance to the laboratory.
There may also be the opportunity at the scene to collect important reference samples that will assist the laboratory to identify extracts from the test samples. Importantly, these samples might also provide important incriminating evidence in investigations and any subsequent criminal proceedings.
Sample size
Sample size is obviously determined by individual circumstances. If a section of carpet or other fire debris is saturated in flammable liquid, then the sample size required by the laboratory will obviously be significantly less than for a sample that has no odor and a diffuse burning pattern or is extensively charred. Sample sizes for control and reference samples will not normally be as large as those for test samples where the amount of sample extract is the major consideration.
As a rule, samples cannot be too large but they can be too small. If a sample is inadequate, it may not be possible to return to the scene and re-sample. Often, the laboratory analysis is not conducted for some time after sample submission and the fire scene could be cleared during this interval. Also, even if the scene does remain, it is unlikely that there will have been any scene security since the original inspection and the integrity of any further samples is therefore compromised.
Packaging
Over the years, many different types of packaging and containers have been used to collect and store samples that contain fire debris and flammable liquid residues. Basically, these containers fall into the following classifications.
Metal containers Drums and cans have been used for many years. They are light, strong, cheap, effectively retain flammable liquid vapors and can be used during the analysis of their contents. However, they are bulky and, when the varnish coating on the interior is scratched, they are quickly susceptible to rusting. Unfortunately, fire debris samples usually contain nails, broken glass, rocks, etc; all of which easily damage the varnish layer. Also, there are usually copious amounts of water in these samples from the fire fighting operations and so corrosion commences immediately. Therefore, any samples collected and stored using metal cans should be analyzed as soon as possible by the laboratory.
Bottles Glass bottles have also been used for many years with reasonable success. Depending on the type of lid seal used, they are generally effective in retaining flammable liquid vapors for long periods. They can also be washed and reused, although this is not advisable and unnecessary given the cheap cost. However, glass bottles are heavy and fragile and additional care must be used in their handling.
Plastic bottles have been used to overcome the deficiencies of glass bottles. They can be effective so long as the plastic is impervious to hydrocarbon vapors.
A major disadvantage of all bottles is the restrictions placed on the sample size by the size of the bottle neck and lid.
Plastic bags Paper bags are obviously totally ineffective for collecting samples. However, the fact that samples are still occasionally submitted using these bags suggests that this is worth noting here.
‘Polythene’ (polyethylene) bags have been used with limited success. These bags are cheap, light, are available in a variety of sizes, are durable and can be easily sealed in the field by tying at the neck. Unfortunately, they are not impervious to hydrocarbon vapors and can be affected by some solvents and chemicals. The use of these bags should be regarded as temporary packaging at the best.
The shortcomings of these bags can be overcome by the use of polyvinylidine bags (available as ‘oven’ or ‘freezer’ bags). These bags have been shown to be impervious to hydrocarbon vapors and will contain flammable liquid vapors for periods in excess of a year under reasonable storage conditions. However, a major disadvantage with these bags is that, if the bag is pierced, a tear usually results very quickly.
Because all these packaging types have certain advantages and disadvantages, no particular method is recommended. Rather, a combination of methods should be considered depending on the circumstances and availability of the containers. For example, samples could be placed into ‘polythene’ bags to take advantage of their resilience and then placed into cans or polyvinylidine bags to retain any flammable liquid vapors.
Sample hygiene
The recovery of samples from areas of interest often requires tools and equipment that are not normally present in a laboratory. These tools include pinch bars, large knives, shovels, axes and saws. Many fire investigators also include a small chain saw with their equipment.
This equipment must be thoroughly cleaned after each sample has been taken, to prevent contamination of any subsequent samples. This can be done using detergent or some other cleaning agent that does not contain substances with a petroleum or ethanol base.
For similar reasons, if a chain saw is included in the equipment, extreme care must be taken to avoid contamination of the samples by the fuel for the chainsaw.
Number of samples
As discussed previously, control and reference samples should be taken liberally although, in many cases, they will not require examination. The number of test samples collected will depend on the perceived potential of the samples and the individual requirements of the jurisdiction. For example, many jurisdictions are served by fire investigators with little practical or theoretical training but who have extensive experience. When these investigators are unsure as to the fire origin and fire cause, they often take many samples with the hope that the laboratory will detect a flammable liquid residue in one or more of these samples and so identify the fire cause and origin. This approach is actually the reverse of accepted fire investigation procedures that require the origin to be diagnosed before sampling commences to investigate the cause. Also, as a consequence, large amounts of laboratory time are used to examine samples taken with no logical basis and the results obtained are open to various interpretations.
If it is obvious from burn patterns and odors that flammable liquids have been present in a number of rooms in a building, it should not be necessary to sample from all of these sites. Rather, a high quality sample should be taken that will undoubtedly confirm the presence and identity of any flammable liquids, and accurate notes taken of the odors and burn patterns in the other rooms.
Other items and samples
With the vast number of diagnosed ignition sources, there are a corresponding variety of items that require collection and examination. In many cases, test samples will not be taken because the fire has destroyed all the physical evidence in the vicinity of the fire origin. However, in many of these cases, the collection of control samples will be vital if any information is to be salvaged. An example of just such a situation is a fire that is ignited in bedding, either deliberately or by a smoldering cigarette. Often in these cases, no bedding remains to be examined and the actual ignition source is completely destroyed. However, vital information will still become available if identical bedding is available for laboratory testing. In these situations, reference samples of the bedding might also be required, especially if the burning characteristics of the bedding become the major issue.
When the ignition cause is diagnosed as a piece of machinery or electrical component, a subsequent laboratory examination can often provide supporting information. In these instances, it can be advantageous if similar components that have also been exposed to the fire can be examined in order that the significance of the test sample results can be determined. Effectively, the additional components act as control samples.
Electrical fuses, fused electrical wiring and other electrical components can provide much information during an examination at the scene and, in fact, may critically influence the scene examination and final fire cause diagnosis. If it is considered that further information could become available through a more detailed examination in the laboratory, control samples should also be collected. In the case of fused wires and ‘blown’ fuses, samples of unaffected and fire-affected components should be submitted as control samples.
There are occasions when no test samples are available and no control or reference samples will assist the investigator. A classical case is a bush fire (wild fire) caused by a lightning strike. In these cases, even if the particular tree struckby lightning survived the ensuing fire, nothing will be achieved by examining the tree in the laboratory that could not be achieved by examining the tree in situ. Photographs recording the condition of the tree will be required, as a formality, but the actual fire cause diagnosis will depend more on prevailing weather conditions and burning patterns in the immediate area.
Sample Labeling
All samples must carry appropriate identification through adequate labeling. Agency and legislative requirements can influence what is entered on labels but, as a general rule, the following information should be included.
The label must include details of who tookthe sample (including a signature) and the time and date the sample was collected. Ideally, the label will also include a basic identification of the sample and the site from which it was taken, for example: ‘Sample of carpet from near the settee in the north east corner of the loungeroom’, Sampled by N. Collins, 1400 hrs, 31/11/98, signature.’
This labeling fulfills two functions: (1) it will allow the results to be interpreted by the fire investigator, magistrate, coroner, jury, etc.; and (2) it will allow a positive identification to be made of the sample at a later date. This latter point is significant when it is realized that any subsequent court proceedings may be years after the incident.
Basic labeling can be applied at the scene and consolidated when the items are submitted to the laboratory at a later date. Obviously, when a sample is labeled, the site of the sampling must be entered onto the scene diagram in the scene case notes.
Sample continuity
Adequate labeling can also assist in the establishment and maintenance of the records necessary to prove sample continuity. If a court subsequently requires any item, or result obtained from an item, collected at a scene, continuity must be strictly maintained and proven. In effect, sample continuity or integrity means that the presence of the sample must be accountable at all times. This must be documented and, if necessary, established through written or oral evidence. Evidence bags are available with sealing and labeling facilities that allow all persons who have access to the bag to record the fact on the bag. However, statements will still be required if it is intended that these items become exhibits.
Summary
The recovery of evidence from fire scenes should be based on observations, interpretation and an appreciation of the laboratory’s requirements and capabilities. As laboratory examinations are time-consuming and expensive, it is inefficient, wasteful and potentially damaging to a case to use the laboratory for the primary diagnosis of a fire cause. The laboratory can, however, provide valuable supporting information for a diagnosed ignition source if the necessary test, control and reference samples are collected, preserved, labeled and submitted.
