Tracking Microplastics with Laser-Based Infrared Imaging: The Potential and Challenges of LDIR
Spotlight on Olga Novillo Sanjuan
Detecting microplastics in environmental samples is not simply a matter of finding small particles. Researchers also need to determine what those particles are made of, how large they are, what shape they have and how many are present.
At the InPlasTwin Summer School 2026, Olga Novillo Sanjuan from DTU Sustain, Technical University of Denmark, presented the possibilities and challenges of Laser Direct Infrared Imaging (LDIR) for microplastic analysis in her presentation “Tracking Microplastics with Laser-based Infrared Imaging: LDIR-QCL.”
Her presentation showed why LDIR is becoming a valuable tool for microplastic research, while also emphasising an important point: automation can make analysis faster, but it does not remove the need for careful scientific judgement.
What is LDIR?
LDIR stands for Laser Direct Infrared Imaging spectroscopy. It combines infrared spectroscopy with high-resolution imaging and uses a Quantum Cascade Laser (QCL) as its infrared source to rapidly detect, image and identify microscopic particles. The technique analyses the characteristic infrared response of particles and compares their spectra with reference libraries, helping researchers identify the materials present in a sample. One of its major advantages is its ability to combine chemical identification with information about individual particles. The system presented during the Summer School offers high spatial resolution, can target specific wavelengths and operates in the micrometre size range.
Why is LDIR interesting for microplastic research?
Microplastic analysis can involve hundreds or even thousands of particles, making speed and automation particularly valuable. Among the potential benefits of LDIR highlighted in the presentation were high throughput, speed of analysis, ease of use and automation. Another advantage is that it does not face the fluorescence problems that can complicate the analysis of environmental samples with some other spectroscopic techniques. In automatic mode, LDIR can analyse particles approximately 20-500 µm in size and provide information that can subsequently be used to report particle counts, concentrations and size distributions. But getting a result from the instrument is only part of the analytical process.
Sample preparation still matters
As with other approaches to microplastic analysis, the quality of the final result depends strongly on what happens before the sample reaches the instrument. Sample preparation depends on the complexity of the matrix and can include homogenisation, pre-filtration, digestion, further treatment and vacuum filtration, ideally while applying measures to minimise contamination. One useful feature of the LDIR workflow is the possibility of reducing intermediate handling between filtration and polymer analysis. This can simplify the analytical process and potentially reduce opportunities for contamination. However, small sample volumes and small analytical surfaces mean that subsampling is often necessary. The physical characteristics of the prepared sample also matter: particle height and surface flatness can affect analysis.
From a spectrum to a polymer identification
LDIR generates infrared spectra that can be compared with spectral libraries to identify particles. In principle, this sounds straightforward. In practice, environmental samples are rarely simple. Olga demonstrated examples involving PET fibres and particles, polypropylene, rubber, cellulosic material and polyamide. Some spectra can be difficult to distinguish, particularly when different materials produce similar signals. One example from the presentation illustrates this challenge particularly well: a particle produced a library match of 0.820 for rubber and 0.802 for cellulose. Such a small difference demonstrates why the highest automated library match should not always be accepted without further examination.
Automation does not eliminate uncertainty
Environmental microplastic samples are heterogeneous, and several sources of uncertainty can affect the final results. These include subsampling error, sample heterogeneity, particle-detection thresholds and the aggregation of fibres and particles during image analysis. Bundles and agglomerated particles can further complicate automated particle recognition and classification. This means that researchers need to understand not only the analytical instrument but also how software settings influence what is detected and how particles are classified. This is particularly important because LDIR results can be strongly affected by settings such as sensitivity and the selected detection area.
Good quality control remains essential
Fast analytical technology cannot replace good laboratory practice. Among the best practices highlighted during the presentation were:
- homogenising the sample;
- performing and reporting replicates, blanks and recovery tests;
- reporting particle size ranges and detection limits;
- ensuring sample flatness as much as possible;
- applying strict anti-contamination measures;
- using glassware and filtered reagents where possible; and
- manually double-checking selected spectra.
These steps are essential if results are to be reliable, reproducible and comparable between studies.
No single analytical method is perfect
LDIR is only one part of the broader microplastic analytical landscape. Olga compared it with µFTIR, Raman spectroscopy and Py-GC/MS, showing that each technique has particular strengths and limitations. LDIR provides particle-based information and high throughput; µFTIR is robust and standardised but slower; Raman can analyse smaller particles but may face fluorescence issues; and Py-GC/MS provides quantitative mass-based information but does not provide individual particle data. The message from this comparison is simple but important: No single method is perfect. The appropriate analytical approach therefore depends on the research question, sample matrix, particle size of interest and type of information required.
Where do we go from here?
LDIR has considerable potential to make microplastic analysis faster and more automated. However, several challenges remain. Future work should address the limitations associated with small sample volumes, improve differentiation between similar polymer types and develop simpler sample-preparation procedures that reduce contamination while improving detection. Machine learning could also contribute to better particle identification. Inter-laboratory studies and greater transparency will be particularly important for improving comparability and confidence in microplastic measurements. Ultimately, LDIR demonstrates how advanced analytical technology can help researchers process complex microplastic samples more efficiently. But the technology works best when automation is combined with careful sample preparation, quality control and expert interpretation.
About the expert
Olga Novillo Sanjuan is a postdoc researcher affiliated with DTU Sustain, Environmental Contamination and Chemicals, at the Technical University of Denmark. Her background is in microplastics quantification in environmental matries, ranging from wastewater to wildlife. At the InPlasTwin Summer School 2026, she presented “Tracking Microplastics with Laser-based Infrared Imaging: LDIR-QCL,” introducing participants to the principles, applications, strengths and analytical challenges of LDIR for microplastic research. Within InPlasTwin, she has been supporting microplastic identification in soil with the LDIR technique.