Identifying Nontoxic Gunshot Residue Using LIBS and Machine Learning

A recent study published in Talanta reports a promising analytical strategy for identifying gunshot residue from nontoxic ammunition, which is an important challenge for forensic science as heavy-metal-free alternatives replace conventional lead-based rounds (1). This study, which was conducted by lead authors Cicero Cena, Matheus Cicero Ribeiro, and Bruno S. Marangoni at the Universidade Federal de Mato Grosso do Sul in Campo Grande, Brazil, demonstrated that combining laser-induced breakdown spectroscopy (LIBS) with machine learning (ML) can successfully discriminate between shooters and non-shooters, even when traditional elemental markers are absent (1).

Gunshot residue consists of microscopic particles generated when a firearm is discharged. (1). These particles originate from multiple sources, including the primer, cartridge case, bullet, and internal firearm components (1,2). Analyzing gunshot residue is a specialized area in forensics, which focuses on analyzing trace elements left behind in a shooting (2). The most common technique often employed for this type of analysis scanning electron microscopy coupled with energy dispersive spectroscopy (SEM-EDS) because of its microanalysis capabilities (2,3). Other techniques used in the past include atomic absorbance spectroscopy (AAS) and inductively coupled plasma–mass spectrometry (ICP-MS), but these techniques were inferior to SEM-EDS for several reasons. For AAS, it is an expensive and destructive technique to use (2). For ICP-MS, although it performed elemental analysis better than AAS, it also is an expensive technique to use (2).

In the firearm industry, there has been an ongoing trend of adopting heavy-metal-free (HMF) or nontoxic ammunition (NTA). This change is mostly because of growing health and environmental concerns (1). The issue, though, is that these newer formulations are typically free of lead (Pb) and often lack barium (Ba) and antimony (Sb) as well, leaving forensic analysts without the traditional elemental signatures used in SEM-EDS (1). As a result, identifying gunshot residue from nontoxic ammunition (GSR-NTA) has emerged as a critical gap in current forensic capabilities (1).

As a result, the researchers investigated whether LIBS can be an alternative analytical technique. LIBS works by focusing a high-energy laser pulse onto a sample surface, generating a microplasma and emitting light characteristic of the elements present (1). Unlike SEM-EDS, LIBS can rapidly detect a broad range of elements, including lighter elements such as hydrogen, carbon, nitrogen, and oxygen, alongside metals (1).

The researchers obtained GSR-NTA particles directly from the hands of individuals who had fired two consecutive shots, as well as from non-shooters. Then, the LIBS spectra were acquired across two spectral ranges—186–1042 nm and 186–570 nm (1). These two ranges allowed the researchers to collect as much elemental information as possible. The elements detected included hydrogen (H), nitrogen (N), oxygen (O), carbon (C), titanium (Ti), zinc (Zn), copper (Cu), Ba, strontium (Sr), iron (Fe), magnesium (Mg), and aluminum (Al) (1).

The team also used multivariate analysis and supervised machine learning (ML) algorithms to the LIBS data. The data set was divided into training and external validation sets to test the robustness of the classification models. Among the algorithms evaluated, linear discriminant analysis (LDA) stood out, achieving 100% classification accuracy in distinguishing shooter samples from non-shooter samples (1).

According to the authors, the success of the approach lies in recognizing elemental patterns rather than relying on a single characteristic marker. Spectral analysis revealed that zinc, titanium, copper, and iron were the primary contributors to sample differentiation, with minor contributions from barium and strontium (1). This finding aligns with current ASTM E1588/20 guidelines, which classify GSR-NTA particles as characteristic when combinations such as Gd-Ti-Zn or Ga-Cu-Sn are present, and as consistent when Ti-Zn-Sr is detected (1).

Although it is still unclear whether LIBS can effectively replace SEM-EDS, the study’s findings do show that LIBS can serve as a rapid screening and complementary technique, particularly in cases involving modern ammunition formulations.

“The combination of LIBS and ML shows potential as a forensic tool for identifying GSR-NTA particles on the hands of individuals who have, or have not, discharged a firearm,” the authors wrote in their study (1).

References

1. Wenceslau, R.; Cabral, J. S.; da Silva Souza, G. et al. Analysis of Gunshot Residue from Nontoxic Ammunition by Laser-induced Breakdown Spectroscopy and Machine Learning Algorithms. Talanta 2026, 296, 128483. DOI: 10.1016/j.talanta.2025.128483
2. Nanoscience Instruments, Automated Gunshot Residue Analysis Using Scanning Electron Microscopy. Nanoscience.com. Available at: https://www.nanoscience.com/applications/automated-gunshot-residue-analysis-using-scanning-electron-microscopy/#:~:text=Gunshot%20residue%20analysis%20(GSR)%20is,originate%20primarily%20from%20the%20primer. (accessed 2026-01-30).
3. Serol, M.; Ahmad, S. M.; Quintas, A. et al. Chemical Analysis of Gunpowder and Gunshot Residues. Molecules 2023, 28 (14), 5550. DOI: 10.3390/molecules28145550