A smoking gun...

A bit of history

The corrosive nature of human sweat has been known for nearly 100 years. Workers in the metal industry in the 1920s were called ‘rusters’ because their sweat readily corroded certain metals. By the 1970s, the dermatologist, John Burton had proposed a model for this corrosion based on the water in sweat forming an electrochemical cell: oxidation of metal at the anode and hydroxyl ion formation from water and dissolved oxygen at the cathode, with the resulting metal hydroxide undergoing hydrolysis to form metal oxide.⁴ The electrical circuit was completed by electron flow through the metal. This is a typical electrochemical corrosion reaction in which metal dissolution is accompanied by oxygen reduction.

We recently verified this model experimentally by demonstrating the effect that removing water from fingerprint sweat has on the electrochemical corrosion process. We placed fingerprint sweat deposits on semicircular brass disks with one half of each disk being left at room temperature in air and the other placed in a desiccator for 10 days. After this time, the semi-circular disks left in air showed varying fingerprint corrosion ridge detail.  In contrast, none of the disks placed in the desiccator showed any sign of corrosion.

However, while Burton and his co-workers thought that the role of sodium chloride in the sweat was merely to carry the electrical current within the drop of sweat, we now believe that its role, and, in particular, the chloride ion, is more complex. Research done by Andrew Meekins at the University of Sussex, UK,⁵ for example, has shown how the presence of chloride ions can lead to the formation of metal chloride ion complexes, with the end result being metal oxide formation. Additionally, Meekins’ work revealed that this process becomes more thermodynamically favourable as a result of hydrogen ion formation and consequent lowering of the pH of the fingerprint sweat.

Seeing the evidence

So how easily can fingerprint sweat corrosion of metal be visualised? Our early experiments, carried out by depositing fingerprints onto planar disks, revealed much variation in the ability of different people to produce corrosion that was visible to the naked eye. This variation was subsequently verified as being due to differences in the composition of sweat between individuals with, as suggested by Burton, statistical correlation between chloride ion concentration and visible corrosion of metals such as brass.

In our experiments with α-phase planar brass disks, over one third of samples gave fingerprint corrosion in which the corroded ridge detail was readily identifiable to the donor, and the fingerprints would have been of use to the police in a criminal investigation.⁶

These fingerprints were most easily visualised by optical interference and photographed obliquely under natural daylight. Daylight was found to give better results than laboratory light sources. But what of the fingerprints that are not so visible?

We decided to investigate whether the fingerprint deposit on the surface of the metal affects its electrical, as well as its optical, properties in the region of the metal/corrosion interface. We measured the electrical potential on the surface where there was corrosion and found it to be up to an order of magnitude less than that of the natural metal for, typically, an applied potential (V) of 2kV.

Initially, we attributed this potential difference (ΔV) to the insulating effects of the oxide layer, which was producing an ohmic (resistive) contact. However, a plot of ΔV against V showed an exponential relationship for V <~ 800V with a plateau for higher values of V: ΔV remained constant for V > ~ 800V.  This relationship is consistent with that measured for a commercial Schottky diode, which is a special form of diode formed between a metal in close contact with either a p-type or an n-type semiconductor,⁷ and discovered by the German physicist Walter Schottky in the 1930s. 

We concluded, therefore, that our corrosion products on brass must be semiconductors and therefore likely to be either copper (I) oxide (a p-type semiconductor) or zinc oxide (an n-type semiconductor). By simply reversing the polarity of the applied voltage, we could identify, using X-ray photoelectron spectroscopy, first copper (I) oxide and then zinc oxide as the corrosion products.

As a spin-off to this research, we have recently exploited this effect to develop a novel means of determining the degree of ‘de-zincification’, and therefore the condition of brass, using a portable hand held device.⁸ This device simply probes the surface of the brass and, by measuring the difference in ΔV when the polarity of the applied potential is reversed, we can estimate the amount of de-zincification at the surface.

We then applied a fine conducting powder (particle size < 5μm) to the metal, and found that it preferentially adhered to areas of lower potential, that is areas coincident with the fingerprint ridge detail, enabling the fingerprint to be visualised. This was achievable even after the metal had been washed thoroughly in warm soapy water and when the corrosion was not readily visible to the naked eye.

Unlike conventional fingerprint enhancement techniques, this method is reproducible. This is possible because, unlike the sweat deposit that adheres to the surface of the metal, the metal oxide forms within the crystal lattice and thus requires abrasive cleaning to remove it.

Scene of crime

Of course, the ‘real life’ application of this research is to be able to identify fingerprints from spent brass cartridge cases. So, once the techniques had been perfected on planar disks, we then tried to reproduce these results on brass cartridge cases where the fingerprint sweat had been deposited pre-firing, a situation that would be more advantageous to the police than finding a fingerprint deposited after the casing had been fired from the weapon. 

Applying the powder to a cylinder of brass – with a typical diameter of 9mm or less – proved more troublesome than on a planar disk. The angle of contact between the powder and casing was found to be critical to obtain powder adherence to the corrosion.⁹

We developed equipment for shell casings that mimicked the delivery of the powder to a flat disk and, as a result, many law enforcement agencies across the world use this technology to look for fingerprint corrosion, typically after conventional techniques to enhance sweat deposits have been tried and have failed.

Moreover, the age of the specimen has not been found to be a barrier, with ‘cold case’ homicides committed over 30 years ago producing identifiable fingerprint corrosion. This has been possible not least because it is the police’s responsibility to retain all evidence but also because the fingerprint doesn’t disappear unless abrasively cleaned, and, once the water in the sweat has evaporated, the electrochemical corrosion process ceases – so we really do have fingerprints forever. Many cases, in which spent cartridge cases have been subjected to these techniques, are high profile homicides and have featured on the America’s Most Wanted TV programme.¹⁰

The future

Although fingerprint corrosion on metal (particularly brass) has enjoyed some success in producing identifiable fingerprints after conventional techniques to enhance sweat deposits have failed, there remains much to understand regarding the mechanisms controlling the corrosion process. Not least of these is the desire for a greater understanding of what components of sweat, other than sodium chloride, influence the formation of corrosion and its enhancement and visualisation.

In addition, we continue to research ‘non invasive’ techniques to aid the visualisation of fingerprint corrosion. Such techniques are favoured by law enforcement agencies because they do not compromise the evidence, either from being able to use other enhancement techniques subsequently or from innocent contamination.

Our latest work has focused on the use of long wave (8–15μm) infrared radiation. The physics of why this might be of interest comes down to the emissivity (ε) of a material. This is a measure of a material’s relative ability to emit radiation, where 0≤ ε ≤1 and so generally reflective surfaces have lower emissivity. The value of ε for brass is ~ 0.06, much less than the values for copper oxide (~0.7) and zinc oxide (~0.25), so we hypothesised that the variation in emitted power at infrared wavelengths might enable fingerprint corrosion, which appears faint at visible wavelengths, to be visualised.

Our initial results are encouraging but there remains much more work to be done, as well as in other areas of fingerprint corrosion visualisation. Current research at the University of Leicester, UK, in these other areas includes trying to re-initiate the corrosion process after the water has evaporated from the sweat and investigating the potential of fingerprint sweat to corrode other common metals.

Evidence in court

On 13 September 2008, businessman Tong Van Le was murdered outside his home in Novato, California, US. He had been shot and, typically for homicides involving firearms, the police found spent brass shell casings at the murder scene.

Having been treated for latent fingerprint deposits without success, in April 2010, Inspector Sara Delaney from the District Attorney’s office in Marin County, California, travelled to the UK with the shell casings to see if fingerprint corrosion techniques could reveal any fingerprints. Two years later, in April 2012, John Bond, who examined the shell casings, gave evidence in the Marin County Superior Court of California – the first time that evidence related to fingerprint corrosion has been heard in a court of law.