Implications of method- and instrument-based size detection limits in μFTIR-based microplastic analysis

Spectroscopic detection methods are considered the gold standard in microplastic research, but they are not yet fully standardized. Results, such as numerical concentrations from different studies, are often reported and compared without considering all the relevant factors during the analysis process that may influence these concentrations. For the first time, we quantify the influence of both the sieve mesh sizes used during laboratory microplastic extraction and the choice of objective (4×, 15×, or 25× magnification) in μFTIR analysis on microplastics number and mass concentrations in suspended solids samples from two Dutch rivers. A high-resolution μFTIR objective and thus lower size detection limit enhances the detection of particles just above this limit, thereby increasing the total number of detected particles. Size distributions captured with a 4× objective (20.6 μm pixel resolution) differed significantly from those with a 15× (5.5 μm) and 25× objective (3.3 μm), while the results for the 15× and 25× objectives were more comparable. Using the 15× objective appears to be a reasonable choice, as lowering the detection limit further would only marginally improve results while significantly increasing workload. During extraction, sieves with mesh sizes of 20 μm and 5 μm did not produce sharp size cut-offs. By applying cascade filtration, it was possible to determine whether the actual size boundaries of the retained particles matched these nominal limits. Our findings indicate that 14-38 % of the particles are nominally either too large or too small to be retained by a given sieve mesh size. Due to negligible differences for polymer type and size distributions found on the 20 and 5 μm sieves, the 20 μm sieve is recommended for an improved time efficiency in microplastic analysis. To further improve comparability across studies, we recommend expanding datasets as measured here so that read-across techniques can be developed, allowing conversion between the diverse resolutions and mesh sizes applied.