Differentiation of Body Fluid Stains Using a Portable, Low-Cost Ion Mobility Spectrometry Device—A Pilot Study

• Published in: Molecules (Basel, Switzerland) Switzerland), 2023-09, Vol.28 (18), p.6533
• Main Authors: Heaton, Cameron, Clement, Simon, Kelly, Paul F., King, Roberto S. P., Reynolds, James C.
• Format: Article
• Language: English
• Subjects: Analysis; biofluids; blood; Chemical tests and reagents; Crime scenes; Criminal investigations; Drugs; Evidence; Explosives; forensics; Gas flow; IMS; ion mobility spectrometry; Ions; Machine learning; Mass spectrometry; Reagents; Reynolds, George J; Scientific imaging; Spectrum analysis; urine; VOCs; Volatile organic compounds
• ISSN: 1420-3049; 1420-3049
• DOI: 10.3390/molecules28186533

The identification and recovery of suspected human biofluid evidence can present a bottleneck in the crime scene investigation workflow. Crime Scene Investigators typically deploy one of a number of presumptive enhancement reagents, depending on what they perceive an analyte to be; the selection of this reagent is largely based on the context of suspected evidence and their professional experience. Positively identified samples are then recovered to a forensic laboratory where confirmatory testing is carried out by large lab-based instruments, such as through mass-spectrometry-based techniques. This work proposes a proof-of-concept study into the use of a small, robust and portable ion mobility spectrometry device that can analyse samples in situ, detecting, identifying and discriminating commonly encountered body fluids from interferences. This analysis exploits the detection and identification of characteristic volatile organic compounds generated by gentle heating, at ambient temperature and pressure, and categorises samples using machine learning, providing investigators with instant identification. The device is shown to be capable of producing characteristic mobility spectra using a dual micro disc pump configuration which separates blood and urine from three visually similar interferences using an unsupervised PCA model with no misclassified samples. The device has the potential to reduce the need for potentially contaminating and destructive presumptive tests, and address the bottleneck created by the time-consuming and laborious detection, recovery and analysis workflow currently employed.