From Pore to Core: Do Engineered Nanoparticles Violate Upscaling Assumptions? A Microtomographic Investigation

Abstract:

The recent success in using Synchrotron X-ray Computed Microtomography (SXCMT) for the quantification of nanoparticle concentrations within real, three-dimensional pore networks [1] has opened up new opportunities for collecting experimental data of pore-scale flow and transport processes. One opportunity is coupling SXCMT with nanoparticle/soil transport experiments to provide unique insights into how pore-scale processes influence transport at larger scales. Understanding these processes is a key step in accurately upscaling micron-scale phenomena to the continuum-scale.

Upscaling phenomena from the micron-scale to the continuum-scale typically involves the assumption that the pore space is well mixed. Using this ‘well mixed assumption’ it is implicitly assumed that the distribution of nanoparticles within the pore does not affect its retention by soil grains. This assumption enables the use of volume-averaged parameters in calculating transport and retention rates. However, in some scenarios, the well mixed assumption will likely be violated by processes such as deposition and diffusion. These processes can alter the distribution of the nanoparticles in the pore space and impact retention behaviour, leading to discrepancies between theoretical predictions and experimental observations.

This work investigates the well mixed assumption by employing SXCMT to experimentally examine pore-scale mixing of silver nanoparticles during transport through sand packed columns. Silver nanoparticles were flushed through three different sands to examine the impact of grain distribution and nanoparticle retention rates on mixing: uniform silica (low retention), well graded silica sand (low retention) and uniform iron oxide coated silica sand (high retention). The SXCMT data identified diffusion-limited retention as responsible for violations of the well mixed assumption. A mathematical description of the diffusion-limited retention process was created and compared to the experimental data at the pore and column-scale. The mathematical description accurately predicted trends observed within the SXCMT-datasets such as concentration gradients away from grain surfaces and also accurately predicted total retention of nanoparticles at the column scale.

1. ES&T 2014, 48, (2), 1114-1122.