Best practice anti-contamination measures were taken in each laboratory. As the field laboratory was set up for the field campaign, it was thoroughly cleaned, and air conditioning remained off for the campaign. Footfall within the laboratory was restricted to the researcher, and cotton laboratory coats were worn. All instruments and containers were acid-washed before first use and were covered with aluminium foil during processing. At the British Antarctic Survey Laboratory - Cambridge, aluminium foil was used during filtering, and glass lids were used during the drying process. A bespoke Perspex shield was fitted around the FTIR stage to prevent airborne contamination during infrared analysis. The laboratory was thoroughly cleaned, and surfaces were wiped down with an ethanol dilution between each sample processing and analysis. Procedural blanks were taken to measure the introduction of contamination at any of these stages and recovery tests, to measure the loss of sample during processing.

A contamination library was also built. For this, fibre samples were collected from the garments worn by each person in the field and the laboratory, with particles from additional possible sources of plastic pollution also collected from the field camp. These samples were analysed using attenuated total reflection using a mobile Agilent FTIR. Spectra were collected and measured against the existing spectral reference library using the siMPle software. Hierarchical cluster analysis was carried out according to Primpke et al. 2018 using the Hellinger Distance for the calculation of the resemblance matrix (Primer 6 and Permanova, Primer-E) was performed to identify the primary polymer composition of these samples by their similarity to existing database entries and inform, post-hoc, on possible sources of contamination.

Procedural blanks were run in triplicate. Two separate sets of blanks were carried out. The first type, the ''''full procedural blank'''', replicated the complete procedure from collection to analysis, as per all other snow samples. A separate ECOtanka was used to collect 250ml of 0.2 micrometer filtered ''''clean water'''' and was processed in the same manner, i.e., filtering onto a silver filter in the field laboratory and subsequently transferred onto an Anodisc filter for FTIR analysis. Each blank was carried out on the same day the samples were processed in the field laboratory (n=3). In addition, a secondary set of procedural blanks was taken to inform about contamination introduced during the transference step from silver filter to Anodisc. These blanks, ''''laboratory blanks'''', replicate the procedure carried out in the Cambridge Laboratory, using 200 ml of 0.2 micrometer filtered Milli-Q water, filtered onto a silver filter and subsequently removed using 50ml of Milli-Q water onto an Anodisc and 10ml of 30 percent Ethanol to help remove any residual filtrate.

To determine the final ''''normalised concentration'''' presented in the results, the concentration of each polymer type was averaged from the three ''''full procedure'''' blanks and subtracted from each raw sample concentration. The concentration recorded in the lab blank was used only to inform what proportion may have been introduced in the laboratory. Furthermore, the fraction likely introduced during field processing is determined by subtracting the ''''lab result'''' from the ''''full result''''.

Recovery tests were designed to determine whether there was any substantial material loss during the transference step from silver filter to Anodisc. These recovery tests were carried out in triplicate, using Nylon fibres stained with Nile Red and had a uniform diameter of 16 microns, and cut to a length of 250 microns. A fluorescent microscope was used to count fibres once dry. Following this, the sample was enclosed in its case. It was shaken to mimic the disturbance of the sample within the case that was potentially encountered during transit. The sample was then removed, filtered onto the Anodisc filter, and counts made again. For both counts, photographs were also taken across the entire filter to validate counts. In addition, one of the recovery tests was subsequently analysed using FTIR to compare the manual counts with those counted by the software. Average laboratory blank nylon fibres (n=3) were subtracted from the final count.