Microplastic contaminants potentially distort our understanding of the ocean's carbon cycle.

Abstract

Direct observations confirm that admixtures of sedimentary organic matter (SOM) and microplastics (MPs) are fully oxidized during Elemental analysis (EA), with measured carbon yields, % carbon [%C], C:N ratios, stable- (δ13C) and radiocarbon (Δ14C) abundances consistent with predictions for SOM samples intentionally contaminated with plastic. As an example, MPs would comprise ~40% of all carbon atoms measured via EA in a 100 μg SOM sample (1% OC by mass) that has been contaminated with only 1 μg of polyethylene (PE = 77% C by mass). This MP contamination, amounting to just 1% of the total sample mass, would lower the sample's Δ14C by 258‰ to -622 ‰, lower the sample's δ13C by -3.65‰ to -25.22‰, and overestimate its conventional 14C age by ~4000 years. Moreover, this 1% MP contamination would imply a terrestrial source contribution of ~ 60% instead of the 20% for an uncontaminated SOM sample. Our results illustrate how these errors scale predictably with MP contamination level and dominant polymer types. While large errors might be recognized as outliers and scrutinized, even small levels of contamination (e.g., 0.1% by mass) can introduce significant but subtle errors that could go unnoticed (e.g., Δ¹⁴C error of -30‰). Most carbon biogeochemistry studies do not routinely recognize the presence of MPs in environmental samples, despite the ubiquity of MP in the ocean and their potential impact on measurements. Consequently, MP contamination either naturally-occurring in field samples or introduced while sampling and processing will necessarily lead to errors in organic matter characterization, source apportionment, and estimates of conventional 14C ages.