Terrestrial nanoplastics (NPs) pose a serious threat to agricultural food production systems due to the potential harm of soil-born micro- and macroorganisms that promote soil fertility and ability of NPs to adsorb onto and penetrate into vegetables and other crops. Very little is known about the dispersion, fate and transport of NPs in soils. This is because of the challenges of analyzing terrestrial NPs by conventional microscopic techniques due to the low concentrations of NPs and absence of optical transparency in these systems. Herein, we investigate the potential utility of small-angle neutron scattering (SANS) and Ultra SANS (USANS) to probe the agglomeration behavior of NPs prepared from polybutyrate adipate terephthalate, a prominent biodegradable plastic used in agricultural mulching, in the presence of vermiculite, an artificial soil. SANS with the contrast matching technique was used to study the aggregation of NPs co-dispersed with vermiculite in aqueous media. We determined the contrast match point for vermiculite was 66% D 2 O / 33% H 2 O. At this condition, the signal for vermiculite was ~50-100%-fold lower that obtained using neat H 2 O or D 2 O as solvent. According to SANS and USANS, smaller-sized NPs (50 nm) remained dispersed in water and did not undergo size reduction or self-agglomeration, nor form agglomerates with vermiculite. Larger-sized NPs (300-1000 nm) formed self-agglomerates and agglomerates with vermiculite, demonstrating their significant adhesion with soil. However, employment of convective transport (simulated by ex situ stirring of the slurries prior to SANS and USANS analyses) reduced the self-agglomeration, demonstrating weak NP-NP interactions. Convective transport also led to size reduction of the larger-sized NPs. Therefore, this study demonstrates the potential utility of SANS and USANS with contrast matching technique for investigating behavior of terrestrial NPs in complex soil systems.

For Figures 1 (and S2, which is a replotting of Fig 1 with error bars), the data contained were collected at Oak Ridge National Laboratory (ORNL), after data reduction as described in the main paper. The "info" tab provides information on the SANS and USANS instrumentation and when the data was collected. For Figure 2, the same comments as given for Fig. 1 apply. The power law fits were obtained via standard regression (using the Linest function of Excel. Analyses are not shown.) For Figure 3, the same comments as given for Fig. 1 apply. The experimental data was obtained by sbtracting the power law fits from the SANS and USANS data, both of which are given in Fig. 2. The model fits were obtained from performing a form factor model using modeling software developed by the National Institute of Standards, National Center for Neutron Research (NIST-NCNR) with IGOR-Pro serving as the platform.