Transport of silver nanoparticles in the soil-water nexus

Mahdi, Karrar Najah


Nanomaterials are materials with particle sizes ranging from 1 to 100 nm in at least one dimension and have properties that are different from their bulk material. Engineered nanomaterials (ENMs) are a promising field of study for science and engineering since it offers many prospects for new products. Silver nanoparticles (AgNPs) have the widest range of applications due to their antimicrobial, catalytic, optical, electronic and magnetic properties. The increasing production of AgNPs will, without a doubt, lead to the emission of these materials to the environment. AgNPs can enter the soil from different pathways. Their available concentration in soil is expected to be low which makes their detection beyond the ability of many current measuring techniques.

Moreover, extracting AgNPs from soil samples is a troublesome process due to the difficulties posed by the soil matrix and the high adsorption affinity between soil particles and AgNPs. The possible risks and impacts of AgNPs on soils and plants have been confirmed by many studies. However, knowledge about the transport mechanisms of AgNPs through soil and over soil surfaces as a result of leaching, runoff and erosion is still limited. These soil processes can facilitate AgNP migration to deeper soil layers and downhill areas and to ground and surface water systems. This thesis presents results for AgNP detection in soil as well as measurements and modelling results of its fate and transport.

In chapter 2, this thesis describes the development and validation of an extraction method for AgNP quantification and particle size determination in soil samples. The developed method is simple and fast thus reducing analysing time and costs. AgNPs recoveries were calculated by comparing AgNP concentration measured in the soil samples to the initial (spiked) AgNPs concentration. Two AgNP-spiking suspensions were prepared with concentrations of 0.5 and 5 (mg L-1) using commercially obtained 60 and 100 nm stock suspensions. The extraction method was developed in multiple steps but the final method consisted of pre-wetting the soil sample, exposing it to sonication and then extracting from the aqueous solution. Pre-wetting was done with the aim of re-suspending the adsorbed AgNPs, which facilitated their extraction. All of the samples were analysed using the principle of single particle inductively coupled plasma mass spectrometry (spICP-MS). Validation of the method showed that the recovery rate of AgNPs was 44% for sandy soil and 42% for clayey soil. The repeatability and reproducibility values for the concentrations were 1.2 %, which is within the limits of the Horwitz ratio (0.5-2). This makes the method suitable for its purpose and effective in environmental, toxicological and pollution studies since the method limit of detection for concentrations (LOD c) is 5 µg kg‑1soil.

In chapter 3, AgNP transport through the soil profile was monitored using a series of polyethylene (PE) hydraulic soil columns with a diameter of 12 cm and a depth of 25.5 cm. Three soil types were used: loam soil with high organic matter (3.4±0.46) (LSH), loam sand with low organic matter (1.8±0.11) (LSH) and sand (zero organic matter). Soils were put into soil columns to a depth equal to ±16 cm in 4 equal layers. The column was filled with 3 soil layers without AgNPs. A top layer that contained AgNPs was then added. The results showed that in the LSH columns, 6.7% of the AgNPs applied to the top soil layer (layer 1) was transported to the 3 other soil layers in the soil column, 3.4% was leached with effluent water and the rest (89.9%) stayed in the top soil layer. In the LSL columns, 8.9% of the AgNPs applied to the top layer was transported to the 3 other soil layers and 4.3 % was leached with effluent, leaving 86.8% in the top layer. In the sandy soil columns, 24.6% of the amount applied was transported to the 3 other soil layers and 13.9% was detected in the effluent. AgNP size decreased during transport through soil layers due to the soil cation exchange capacity (CEC). The decrease in AgNP size increased as the AgNPs moved deeper though soil column layers. Thus, the size of the AgNPs in layer 1 was > in layer 2, and so on. In addition to that, AgNP size in the columns packed with LSL and LSH decreased more than that in the columns packed with sand because of the limited value of CEC. There was less of a reduction in AgNP size in the effluent water for all soil columns as compared to AgNPs that were retained in the soil layers. Similar to what was seen in the soil, the effluent from the sand columns had the least amount of reduction in their particle size. This work highlights the ability to track the concentrations and particle sizes of low concentration applied AgNPs as they leach through soil into groundwater, which increases our knowledge of the patterns of AgNPs during their transport in soil.

In chapter 4, AgNP transport via runoff and erosion processes were analysed using 2 different soil flumes slopes (10% and 20%). Each soil flume had the same dimensions 100x50x30 cm (length x width x depth). These flumes were divided into 5 segments of 20 cm each and a bucket was placed at the flume outlet to collect runoff water (RW) and sediment (RS). Rainfall application was constant (60 mm h-1) during the whole experiment. AgNPs mixed with soil were applied to the top (first) flume segment with the target (initial) concentration of AgNPs in the soil at 50 μg kg−1. Due to runoff and erosion processes which occurred after applying the rain, two main soil clusters with different characteristics appeared on the flume surface and were categorized as background soil (BS) and surface sedimentation (SS). Therefore, the AgNPs were measured in these different soil clusters on flume as well as in the eroded sediments and runoff water. BS samples contained a higher amount of silt while SS samples had a higher sand content. Organic matter (OM) and cation exchange capacity (CEC) in the BS samples were about three times higher than those measured in the SS samples, while the pH was found to be similar for both. Thus, AgNP content was higher in the BS than in the SS. The measured AgNP content in the BS samples in the S20 experiments were higher than in the S10 experiments however, the SS distribution area on the soil flume surface in S10 (measured by picture analysis) was larger than in S20. This could raise concerns for AgNPs pollution on the soil surface where the slope is less steep. The amounts of RW collected were comparable for the 10% and 20% slopes, while the amounts of RS collected after each event were higher for the S20 than for the S10. With each subsequent rain event, the AgNP content increased in the RS samples and decreased in the RW samples. Overall, the total AgNP content measured was higher in S20 than in S10 since 7.3% of the total applied AgNPs was transported in the S20 flume, while only 4.3% was transported in the S10 flume. This study also confirmed that the highest amount of AgNPs was transported with RW, which enhances our understanding of the potential river water pollution by AgNPs due to the runoff process.

In chapter 5, a hydrological soil erosion model (LISEM) was combined with a recently developed particle-transport model (PestPost) in order to simulate the transport of AgNPs caused by runoff and erosion processes. AgNPs were mixed with soil and applied to the upper part of the soil as described in the previous chapter (Chapter 4). The models were calibrated separately using the results of the flume experiment conducted in the laboratory. The aim of this study was to simulate AgNP transport over the soil surface with runoff water and sediment samples for each time-step during runoff and erosion processes. This was achieved using the output of the LISEM of runoff and erosion that migrated in the PestPost model and using this as data base for measuring AgNP transport over soil surface. The calibration of the LISEM showed high model efficiency (MEF) for both sediment =0.81 and runoff =0.79. For the PestPost model, a simple optimisation procedure was used to find the simulated results that were closest to the measured ones. The results of the AgNP simulation in runoff water were better than those in the eroded sediment for both flumes (Δpl for runoff <0.01 and for sediment =0.06 in S10 and 0.02 in S20). It was shown that the processes affecting the transport of AgNPs on the soil surface, mainly runoff and erosion due to rainwater, can be predicted using the combination of the LISEM and PestPost model.

Overall, the outcome of this study shows the possibilities of AgNP detection and transport in the soil, which will increase the knowledge and awareness of AgNPs in our daily-life applications.