Does microplastic analysis method affect our understanding of microplastics in the environment?

Two analytical methods - both in active use at different laboratories - were tested and compared against each other to investigate how the procedure influences microplastic (MP) detection with micro Fourier Transform Infrared Spectroscopy (μFTIR) imaging. A representative composite water sample collected from the Danube River was divided into 12 subsamples, and processed following two different methods, which differed in MP isolation procedures, the optical substrate utilized for the chemical imaging, and the detection limit of the spectroscopic instruments. The first instrument had a nominal pixel resolution of 5.5 μm, while the second had a nominal resolution of 25 μm. These two methods led to different MP abundance, MP mass estimates, but not MP characteristics. Only looking at MPs > 50 μm, the first method showed a higher MP abundance, namely 418-2571 MP m^-3 with MP mass estimates of 703-1900 μg m^-3, while the second method yielded 16.7-72.1 MP m^-3 with mass estimates of 222-439 μg m^-3. Looking deeper into the steps of the methods showed that the MP isolation procedure contributed slightly to the difference in the result. However, the variability between individual samples was larger than the difference caused by the methods. Somewhat sample-dependent, the use of two different substrates (zinc selenide windows versus Anodisc filters) caused a substantial difference between results. This was due to a higher tendency for particles to agglomerate on the Anodisc filters, and an 'IR-halo' around particles on ZnSe windows when scanning with μFTIR. Finally, the μFTIR settings and nominal resolution caused significant differences in identifying MP size and mass estimate, which showed that the smaller the pixel size, the more accurately the particle boundary can be defined. These findings contributed to explaining disagreements between studies and addressed the importance of harmonization of methods.