A critical view on sample preparation for the analysis of microplastics from stormwater retention ponds

Accurate detection of microplastics (MPs) in complex environmental matrices is often compromised by analytical biases introduced during sample preparation. This study systematically demonstrates how the selection of matrix digestion agents (NaOCl vs H₂O₂) and density separation media (zinc chloride (ZnCl₂) vs. sodium polytungstate (SPT)) leads to significant discrepancies in MP loads determined by laser direct infrared (LDIR) imaging in stormwater samples. The results reveal a pronounced underestimation when using ZnCl₂ compared to SPT, with total MP concentrations in all density fractions measured at 66 particles/L and 125 particles/L, respectively (tire and road wear particles excluded). Critically, the analysis reveals that hetero-aggregation of low-density polymers, such as polypropylene, with mineral matter causes a substantial amount of MPs to be lost in high-density fractions (e.g., ρ > 1.7 g/mL). This phenomenon, which is more pronounced with ZnCl₂ solution, indicates that the analysis of only the buoyant fraction - a common practice - results in a severe underestimation of the true MP load. Applying the SPT-based protocol confirms the high particle retention efficiency of a stormwater retention system with active treatment using in situ microflotation technology. Outflow concentrations (17 ± 11 particles/L) were substantially lower than inflow concentrations (79 ± 22 particles/L) for the combined fractions ρ < 2.0 g/mL (tire and road wear particles excluded). This work underscores the need to revise current analytical practices to prevent the systematic underestimation of environmental MP loads.