Aberration corrected 3-dimensional (3D) scanning transmission electron microscopy (STEM) is capable of high-resolution 3D imaging of specimens without a tilt stage. In a manner similar to confocal light microscopy, the sample is scanned layer by layer by changing the objective lens focus so that a focal series is recorded. Optimized 3D STEM is expected to exhibit significant advantages over tilt-series transmission electron microscopy (TEM) for conventional thin sections, such as 1) better resolution, 2) absence of mechanical tilt, 3) the capability of imaging large area thin sections, and 4) faster 3D data collection. Recently, we have obtained 3D images of conventional thin sections containing 3T3 cells showing a lateral resolution of 0.6 nm and an axial resolution of 60 nm. Our calculations predict that the axial resolution on conventional thin sections could be improved to 7 nm. Our goals are to fully understand the image formation mechanisms, to evaluate the feasibility of 3D STEM with respect to radiation damage, and to optimize the performance of 3D STEM for the imaging of thin sections of biomedical relevance.
The specific aims are to: 1) Determine the resolution of 3D STEM of test specimens. The factors determining the resolution of 3D STEM are not yet well understood. An important aspect that needs to be investigated is the influence of electron scattering by the embedding medium on the resolution. We will examine the lateral and axial resolution for test specimens of realistic dimensions in the absence of radiation damage and extreme sample heterogeneity. A Monte Carlo model of the 3D STEM will be calibrated with these experiments and used to examine 3D STEM image formation and beam-sample interactions in further detail. 2) Determine resolution of 3D STEM on conventional thin sections. We will optimize the sample parameters and the microscope settings to obtain a lateral resolution of 1 nm and an axial resolution of 7 nm for a conventional thin section containing 3T3 cells. In addition to the factors investigated in Aim 1 the resolution obtained on biological specimens is influenced by the size of the grains of the stain, by local variations of the stain density, and by radiation damage. We will develop a theoretical model of the electron dose limited resolution to predict the optimal microscope settings for each sample thickness. We will then apply 3D STEM to image several samples of biomedical relevance. 3) Improve resolution by deconvolution. The images recorded with 3D STEM will contain both the in-focus information from the focal plane as well as out-of-focus contributions. This is an effect commonly found in wide-field focal series from optical microscopy. We will develop an iterative deconvolution strategy for 3D STEM including constrains based on knowledge of electron scattering in the sample.
We aim to improve the resolution of the datasets obtained under Aim 2 by a factor of 2-3. 3D STEM can be used to study the complex organization of the nanometer-sized assemblies of macromolecules and compartments (e.g., ribosomes, proteasomes, Golgi apparatus, and mitochondria) within Eukaryotic cells. Understanding how these structures are organized, and thereby function, within the crowded 3D volume of the cell can be applied, for example, to aid the development of new therapeutics, or improve existing ones.
Showing the most recent 10 out of 15 publications
