Analysis of 3D Elemental Mapping Artifacts in Biological Specimens using Monte Carlo Simulation

We performed Monte Carlo simulation to demonstrate the feasibility of using the focused ion beam based X-ray microanalysis technique (FIB-EDS) for the 3D elemental analysis of biological samples. We used a marine diatom as our model organism and NISTMonte for the place Monte Carlo simulations. We explored several beam energies commonly used for the X-ray microanalysis to examine their effects on the resulting 3D elemental volume of the model organism. We also performed a preliminary study on the sensitivity of X-ray analysis for detecting nanoparticles in the model. For the conditions considered in this work, we show that the X-ray mapping performed using the 5 keV beam energy results in 3D elemental distributions that closely reflect the elemental distributions in the original model.

Intended impact

Development of a 3D characterization technique for the detection and quantification of intra- and inter-cellular distributions of engineered nanoparticles and biologically relevant elements in biological matrices.

Objective

Development of 3D elemental mapping technique for biological matrices using Focused Ion Beam Scanning Electron Microscopy.

Goals

Develop Monte Carlo models for 3D x-ray microanalysis of biological matrices and evaluate the technique under various beam conditions.

Technical approach

Although the beam-sample interaction is a well known problem in the X-ray microanalysis field, its effects have not been analyzed carefully for the purpose of 3D volumetric analysis of biological specimens. We studied the effects of different beam energies for generating 3D X-ray maps of biologically relevant specimens and evaluated the detection and resolution limitations of the FIB-EDS technique for this type of sample. Based on our analysis, for the 3D elemental analysis of resin-embedded bulk biological samples and the conditions specified in this study, 5 keV beam energy is likely to be the maximum usable X-ray mapping beam energy. Even at 5 keV, 3D reconstruction suffers from noticeable distortions in the features due to the sub-surface beam-sample interactions. However, the general shape and the size of the features are reproduced reasonably well at 5 keV beam energy. Maps generated using beam energies higher than 5 keV produce unrecognizable 3D cellular features.

The 3D volumes generated from the simulated Si, Mg, and S X-ray map stacks are shown in the figure below. The effects of increasing beam-sample interaction volume are pronounced here. Fig. A is the schematic of the original model of the diatom. The 3D volume reconstructed from the 5 keV X-ray maps (Fig. B) show some distortions in the diatom shell and the organelles as well as general broadening of internal features. However, the major cellular components are identifiable and appear distinct from each other. In the 10 keV case (Fig. C), the distortions due to the increased beam-sample interaction volume become severe enough that the organelle shapes are completely lost, although regions of high Mg or Si concentration are recognizable. Finally in the 20 keV case (Fig. D), all elemental volumes overlap each other and none of the organelles are recognizable.By improving the milling resolution and by supplementing the lower resolution X-ray maps with the corresponding high resolution structural data from SEM images, FIB-EDS can provide more detailed elemental and structural information than existing methods. However, the proper acquisition and the interpretation of X-ray mapping data depend on many factors such as the beam parameters, detector settings, sample composition, fixing and staining methods, and embedding material. Models used in this work are relatively simple representations of a diatom. In addition, our simulations are based on a single organism and we cannot generalize our results to all possible biological specimens. Work is still needed in generating accurate and diverse biological models as well as simulating experimental conditions relevant to these samples. However, based on these simulations, we have been able to establish baseline X-ray mapping conditions that can provide reasonable results for bulk biological samples such as diatoms.

• Completed Monte Carlo simulation of 3D x-ray mapping of nanoparticles in diatoms using NISTMonte

• Identified and evaluated key experimental parameters for the FIB based x-ray microanalysis of resin embedded biological specimens.

• Developed biologically relevant material models for Monte Carlo simulations.