Catalina David & Emmanuel Rinnert discuss morphological and chemical characterisation of microplastics using particle image analysis combined with Raman spectroscopy
Microplastics (MPs) are tiny pieces of plastics ranging from 5mm down to 100nm in diameter. They represent a global anthropogenic pollution that is expected to last for many generations and become a geological marker of our time. Scientists have found MPs in the air, on mountain summits, on polar ice caps, at the bottom of the deepest abyssal trenches, as well as in virtually all marine lifeforms – from zooplankton to whales. Now, the attention of the community has turned towards the impact of MPs on human health as recent studies revealed their presence in bottled water and even in human stools and placenta.
Since regulation and standardisation agencies only recently started to look at the problem, much effort is required to improve existing methodologies, ensure reproducibility, comparability and representativeness of the analysis, and put in place validated standard operating protocols (SOPs). In the end, the quality of the data produced should meet predefined performance criteria for wide acceptance.
While MPs sorting and counting is still often done by visual inspection or under a dissection microscope, chemical characterisation is becoming essential to differentiate MP-like polymers from other organic and inorganic materials and avoid misidentification when analysing samples. Another key aspect is the need for developing faster procedures and imaging technologies, as mapping large surfaces of filters used to collect MPs can be extremely time consuming, especially when a spatial resolution of the order of micrometres is wanted.
The French Research Institute for Exploitation of the Sea (Ifremer) is currently working on an efficient and time-effective method for MPs monitoring in marine environment, by using a Horiba LabRam Raman microspectrometer in conjunction with the ParticleFinder module proposed with the LabSpec 6 software suite.[1]
Particles collected at sea with a Manta trawl are first deposited on a gold-coated microscope slide to avoid parasitic signal from the borosilicate glass. A large, high-resolution, centimetre-scale optical image of the sample is recorded by the microscope camera, using the integrated video montage/mosaic tool. Image analysis algorithms are applied to perform a binary segmentation based on the optical contrast and to quickly locate all the particles, highlighting their positions. The analysis of this binary image provides at the same time various morphological descriptors (diameter, area, perimeter, major/minor axis, circularity, as well as brightness). The particles can then be filtered according to the statistics results, allowing only particles that correspond to a specific size/shape to be analysed. In the present study, MPs were classified into three groups: 0.335 -1mm (335 µm is the mesh size of the Manta trawl), 1-2mm and 2-5mm.
The next step consists of the collection of one Raman spectrum at the centre of each particle, based on its previously determined position, with two accumulations of an integration time of 10s. The particles are illuminated with a 785nm wavelength laser source for optimum signal/fluorescence ratio, using a 10x magnification objective lens. The Raman spectra cover the 200-1,700cm−1 range with a spectral resolution of about 4cm−1, using a grating with 300 grooves per mm. Care is taken to maintain the laser power below a certain threshold to avoid structural or chemical medication of the polymers.
Chemical identification is finally obtained from commercial Raman spectra libraries (KnowItAll Informatics Systems, Bio-Rad, Raman ID Expert). Out of the 962 particles collected in the surface water, 75% were chemically characterised. Microplastics of different polymer nature (polyethylene 48%, polypropylene 12%, polystyrene 11%) accounted for 71% of the whole sample. Other identified particles corresponded to inorganic minerals, mainly calcium carbonate (2%) and quartz (2%). The non-identified particles exhibited either Raman spectra with no correspondence with databases, saturated spectra (due to strong fluorescence background) or an absence of signal.
This workflow, based on selective and targeted Raman analysis, providing automated localisation, counting and characterisation of environmental microplastic particles, hugely reduces the total analysis time compared to a full point-by-point raster scanning approach. As the proposed method is time effective, a large sample size can be analysed in a reasonable amount of time, which prevents the downsizing to subsamples as is done in most studies.
Reference
[1] Frere L., Paul-Pont I., Moreau J., Soudant P., Lambert C., Huvet A., Rinnert E. (2016). A semi-automated Raman micro-spectroscopy method for morphological and chemical characterisations of microplastic litter. Marine Pollution Bulletin, 113(1-2)
Catalina David is with Horiba and Emmanuel Rinnert is with Ifremer
