Spectroscopic Methods for Bloodstain Ageing

Oxyhaemoglobin has two characteristic absorption peaks in the visible region: at approximately 540 nm and 577 nm. These produce the Soret band visible to the eye as vivid red. Methaemoglobin, the dominant species in aged stains, has a characteristic absorption peak at around 630 nm (giving brown) and loses the sharp double-peak of oxyhaemoglobin. As haemichrome forms, absorbance spreads across the whole visible region, reducing reflectance broadly and giving the near-black appearance of very old stains.

Practical measurements use a fibre-optic reflectance probe or a handheld spectrometer pressed against the stain surface or, in some configurations, a hyperspectral imaging camera that records a full spectrum for every pixel of a scene photograph. The ratio of reflectance at 542 nm to reflectance at 630 nm (sometimes called the spectral age index or similar ratio metrics) declines monotonically as the stain ages under controlled conditions, which is exactly the directional relationship needed for a dating model.

The work of Edelman and colleagues, published between 2008 and 2012, established that UV-Vis reflectance spectra combined with PCA could separate bloodstains by age reliably in controlled laboratory conditions on a single substrate. The model performed well within its training conditions and was subsequently cited in several legal proceedings in the Netherlands. The critical question of cross-substrate and cross-environment generalisation was addressed in later studies, which found substantially larger errors when substrate and conditions varied, tempering the initial optimism.

NIR wavelengths (roughly 700-2500 nm) are longer than visible light and penetrate biological materials more deeply than UV-Vis. Rather than probing electronic transitions of chromophores, NIR measures overtones and combination bands of molecular bond vibrations. For a bloodstain this means the spectrum carries information about the protein environment, water content, and lipid composition of the stain, as well as haemoglobin-specific features.

Because NIR probes the protein matrix as a whole rather than a single chromophore, some researchers have proposed it as a substrate-independent approach. The argument is that while the colour signal is dominated by substrate effects, the NIR protein-matrix signal may be more consistent across surfaces. Published results have been mixed. On homogeneous substrates NIR shows age-related changes similar to UV-Vis, and chemometric models built on NIR data have achieved age prediction errors in the range of a few hours under controlled conditions. Cross-substrate studies show reduced accuracy, though the substrate effect appears less severe than in UV-Vis for some comparisons.

Raman spectroscopy is well established in analytical chemistry for structural identification because it is sensitive to the vibrational modes of specific molecular bonds. In haemoglobin, several marker bands in the region of 1300-1650 cm⁻¹ are diagnostic for the spin and oxidation state of the iron centre, the coordination chemistry of the haem group, and the conformation of the porphyrin ring. As haemoglobin converts from oxyhaemoglobin to methaemoglobin to haemichrome, these marker bands shift in frequency and relative intensity in ways that UV-Vis reflectance cannot resolve.

A key advantage of Raman in forensic context is that it is non-destructive and can in principle be performed without sampling the stain, either by pressing a probe to the exhibit packaging or by using a portable Raman instrument at the scene. This preserves the stain for subsequent DNA or immunological analysis. Several research groups have demonstrated that Raman spectra of bloodstains show statistically significant age-related changes and that PCA models trained on these data discriminate age classes correctly in controlled experiments, including Doty, McLaughlin and Lednev (2016), whose regression model for time since deposition achieved a cross-validated R2 of 0.97 for stains up to one week old.

The limitation is fluorescence interference. Many biological substrates and even some haemoglobin degradation products fluoresce when excited by the laser used for Raman, creating a large background signal that can overwhelm the Raman peaks. Surface-enhanced Raman spectroscopy (SERS) using nanoparticles can suppress fluorescence and boost sensitivity but introduces sample preparation steps that complicate the non-destructive advantage. Research into optimised excitation wavelengths and computational fluorescence-removal methods is ongoing.

A full reflectance or Raman spectrum is a vector of hundreds to thousands of intensity values. No human can interpret that directly. Chemometric methods, particularly PCA and partial least squares regression (PLS-R), compress the spectrum into a small number of latent variables that capture most of the age-related variance and then build a regression model linking those variables to stain age.

In a well-designed controlled study, PCA score plots show stains arranged along a time trajectory, with fresh stains clustered at one end and old stains at the other. Cross-validation within a controlled dataset can achieve prediction errors of a few hours to a fraction of a day for stains up to several weeks old. The critical question is what the independent validation error is on stains prepared under different conditions from the training set.

As of the mid-2020s, no spectroscopic bloodstain ageing method has been validated to the standard required for routine casework by the major forensic science regulatory bodies in any jurisdiction. The Forensic Science Regulator in England and Wales, the Organisation of Scientific Area Committees (OSAC) in the USA, and equivalent bodies elsewhere require an established error rate across realistic case conditions, inter-laboratory reproducibility data, and peer-reviewed validation independent of the method developers. None of the three spectroscopic approaches yet meet all of these criteria.

• The Dutch cases (2009-2014) in which UV-Vis reflectance evidence was admitted are the most frequently cited examples of spectroscopic ageing reaching court. Expert scrutiny of those admissions has been mixed, with some commentators arguing the stated confidence intervals understated true uncertainty.
• Raman spectroscopy has not to date been admitted as stain-age evidence in any published casework account, though non-destructive confirmatory identification of blood has been used.
• NIR ageing has been proposed as a screening tool rather than a point-estimate method, which is a more defensible framing given current data.
• The ENFSI and SWGMAT guidelines on bloodstain pattern analysis and trace evidence do not currently include approved protocols for spectroscopic stain age estimation.

None of this means the methods are scientifically uninteresting or that they will remain outside casework permanently. It means that, at the current state of validation, the appropriate professional position for a forensic scientist is to offer these methods as supporting intelligence-level information, not as a standalone dating opinion, and to be explicit about that framing in any report or testimony.

Hyperspectral imaging cameras collect a reflectance spectrum for every pixel in a scene. Applied to a bloodstain pattern, this generates a chemical map where the spectral signature of each pixel reflects the local haemoglobin oxidation state. Because the instrument images the whole scene non-destructively in a single acquisition, it can potentially reveal spatial age differences within a single stain, for example a rim of older dried material surrounding a more recent re-wetting, or confirm that stains of apparently similar colour have different spectral signatures suggesting different deposition times.

Research groups in the UK, USA, and Denmark have demonstrated hyperspectral imaging of bloodstained surfaces under laboratory conditions, with encouraging discrimination between stain ages. The barrier to wider adoption is practical: the cameras are expensive, the data volumes are large, and the processing pipeline requires specialist expertise. Portable systems have been developed for field use, but the same substrate and environment confounders that affect point-source spectroscopy apply equally to imaging versions. The method's greatest near-term value may be in scene documentation and triage rather than precise age estimation.