The smell of death

Volatile metabolites provide vital clues in murder investigation.

Studying the chemistry of decomposition is helping Belgian and US researchers improve the techniques available to law enforcement agencies searching for the victims of violent crime.

Bart Smedts of the Royal Military Academy’s chemistry department in Brussels analyses soil and vapour samples from the burial grounds of human and animal remains at various stages of decomposition. He has determined a series of chemical signatures that reflect a sample’s identity by measuring what he calls ’the traditional decomposition gases’ - including sulfur dioxide, methane, benzene derivatives and long chain hydrocarbons.

His main drive is to determine which compounds trigger which reactions in dogs specially trained to detect human remains - so-called cadaver dogs. The information could be used to design effective ’training mixes’ to teach dogs how to react, he says. Smedts tested a currently available commercial training mix. ’The compounds found in the mix are totally wrong,’ he told Chemistry World. ’They are never detected in volatile decomposition smell and they do not trigger the dog.’ This could explain variable success rates using the dogs, he says.

Smedts collaborated with forensic anthropologist Arpad Vass of the Oakridge National Laboratory in Tennessee. Vass works at the University of Tennessee Anthropological Research Facility (also known as the Body Farm) where hundreds of human cadavers - bequeathed by donors - are left to decompose in a variety of situations set to mimic recorded crime scenes. Vass plans to develop a device that can automatically detect and date cadavers.

Smedts’ human samples were collected from five real crime scenes in Belgium. He also had several animal species - rat, rabbit, muskrat, chicken, pig and dog - buried by the Disaster Victim Identification team of the Belgian Federal Police. These are the animals most likely to contaminate clandestine graves, he says.

Soil samples were removed, heated and the vapours collected for analysis. Vapour samples were collected at site by inserting a hollow probe into the soil and drawing in gases which were then trapped on solid sorbent tubes. Both vapour and soil samples were assayed by thermal desorption and GC-MS analysis. ’We know now what kind of compounds to look for in a body that still retains soft tissue,’ said Smedts, ’and we know the difference for the others that don’t (skeletons).’ His closely guarded data await publication, but Smedts will speak at the  RSC conference Forensic Analysis 2004 (University of Lincoln, UK, 20-22 June).

There is certainly room for improvement in cadaver dog training, says biological anthropologist Keith Jacobi of the University of Alabama, whose work in this area was published last year in the Journal of Forensic Sciences. ’We have found that it does make a difference what dogs are trained on,’ said Jacobi. ’If the dog is trained on pseudo compounds they may not do as well.’ Even the best-trained dogs, which according to his local sheriff are trained in Germany, don’t seem able to detect skeletonised remains, he notes.

Bea Perks