TEM tomography of pores with application to computational nanoscale flows in nanoporous silicon nitride (NPN)

Abstract

Silicon nanomembrane technologies (NPN, pnc-Si, and others) have been used commercially as electron microscopy (EM) substrates, and as filters with nanometer-resolution size cut-offs. Combined with EM, these materials provide a platform for catching or suspending nanoscale-size structures for analysis. Usefully, the nanomembrane itself can be manufactured to achieve a variety of nanopore topographies. The size, shapes, and surfaces of nanopores will influence transport, fouling, sieving, and electrical behavior. Electron tomography (ET) techniques used to recreate nanoscale-sized structures would provide an excellent way to capture this variation. Therefore, we modified a sample holder to accept our standardized 5.4 mm × 5.4 mm silicon nanomembrane chips and imaged NPN nanomembranes (50–100 nm thick, 10–100 nm nanopore diameters) using transmission electron microscopy (TEM). After imaging and ET reconstruction using a series of freely available tools (ImageJ, TomoJ, SEG3D2, Meshlab), we used COMSOL Multiphysics™ to simulate fluid flow inside a reconstructed nanopore. The results show flow profiles with significantly more complexity than a simple cylindrical model would predict, with regions of stagnation inside the nanopores. We expect that such tomographic reconstructions of ultrathin nanopores will be valuable in elucidating the physics that underlie the many applications of silicon nanomembranes.