FIB Tomography

Table of Contents

Overview #

FIB tomography is a destructive technique that enables inspection of the volume of a sample and creation of a 3D model of the sample. During a tomogram acquisition, the sample is slice through using the ion beam and then the revealed section is imaged. The slicing and imaging is repeated, forming a stack of image slices. The slice thickness can be varied from as small as 3 nm to several microns, depending on the sample feature size. It is common to match the slice thickness with the pixel size of the image, which results in square voxels when the dataset is complete. After acquisition of the full image stack, a high performance PC can be used to reconstruct the full 3D model of the sample (Figure 1). 

The acquisition time for FIB tomgraphy can be lengthy, anywhere from several hours to days or weeks, depending on the various parameters. One must balance the need for high quality and high resolution data with the need to complete the acquisition in a timely fashion. Post processing steps of the tomogram also presents challenges to obtain precise data: image alignment, noise and artifact reduction, and segmentation.

Figure 1. Schematic of the process of FIB tomogram acquisition: repeated sample slicing and imaging, followed by reconstruction of the image stack, and finally full 3D model of the sample. The example on this slide shows a BSE image stack of a magnetotactic bacteria, however SE, EDS or EBSD measurements can be recorded as well, allowing full scale analytical capabilities not just in 2D but in 3D.

FIB Tomogram Acquisition Steps #

Sample Preparation #

Preparing a sample for a FIB tomography acquisition begins with many steps used to prepare for cross-sectioning, but with additional considerations. Since the tomogram acquisition may take several days, good preparation before the acquisition is necessary for success. The first step is to adhere the sample to the sample holder in a secure fashion; gluing the sample to a stub is preferred over using double sided tape which may relax over the course of the acquisition.

Protection Layer Deposition #

After the region of interest is identified in the sample, the next step is the deposition of a protection layer to the surface of the area of interest. The surface of the area of interest will see a moderate amount of ion-beam imaging to confirm the position of the sample fiducials, so a robust protection layer is necessary. If the surface is of great importance to the work, deposition of an initial thin EBID layer should be performed, followed by a thicker IBID layer.

For high accuracy of measurement of slice thickness, many researchers include a pattern of lines within the protection layer of known geometry in a material with contrast from the main protection layer material. For example, if Pt is used as the main protection layer, C is used as a suitable material since C will appear darker than the surrounding Pt in electron beam images. The five lines consist of a group of three evenly spaced lines to allow measurement of a known distance in the center of the protection layer, and two lines angled from the center three at known angle (for example 15°) to allow measurement of the slice thickness (Figure 2).

Figure 2. Schematic of pattern deposited in protection layer for calculation of FIB slice thickness (left), and a schematic of the calculation thickness, t, of a slice from two consecutive images.

To calculate the thickness of a slice, the

Exposing the Region of Interest #

Once the area of interest is protected, the exterior of the region of interest is exposed from the bulk material. This is done in a similar fashion as cross-sectioning: a rough milling pattern is typically performed to expose the front of the region of interest quickly, followed by a polishing step to give a clear image at the front of the region of interest. Additionally, many times the sides of the region of interest are also exposed to provide a pathway for redeposition material to escape (Figure 3). If the side reliefs are not cut, the redeposition may deposit in front of the ongoing milling, partially obscuring the field of view in the tomogram.

Figure 3. Schematics of a sample prepared for FIB tomography.
Deposited Pt is in gray, C in gray, while the substrate is shown in yellow.
Fiducials are shown on the left side of the deposit for both electron and ion beam directions.

Fiducial Deposition #

After the sample is defined with milling, the last step of the sample preparation is the deposition and milling of fiducial markers for the electron and ion beams images. The fiducial markers are unique patterns that are added to the sample to allow precise alignment of the beams to the samples. The ion beam fiducial mark is critical for accurate placement of the ion beam patterns and is placed in an area close to the region of interest on the top surface, but out of the way of the tomogram (The X marker in Figure 3). A fiducial mark for the electron beam is optional, but can be advantageous for more reproducible electron beam images by enabling closer registration of the electron beam images.

Tomogram Acquisition #

Slicing with the Ion Beam #

The first step in removing a small slice of material from the sample is to image the sample with the ion beam to find the fiducial mark. The dwell time per pixel for the image should be as short as possible to get a clear image of the fiducial to minimize the sputtering of the protection layer on top of the region of interest. After the fiducial is located, a small box cut is placed at the leading edge of the sample and milled, leaving a freshly exposed surface.

The parameters of the cut should be optimized to the sample, minimizing the milling time and artifacts to leave a clean surface for electron beam imaging. The depth of the cut into the sample can also vary, but many practitioners try to match the depth of the cut to the pixel size of the electron beam image; i.e if the electron beam image has a pixel size of 10 nm per side, then the depth of the cut should also be 10 nm. The result is that the voxels after reconstruction are square. Cut of less than 10 nm in thickness also tend to be near or below the spot size of the ion beam, which can result in more time than necessary for acquiring data because of repeated milling on the same area.

Imaging with the Electron Beam #

After a ion beam cut is complete, the electron beam is the used to image the area of interest. If a fiducial mark is used for the electron beam alignment, then a quick lower resolution image is taken to place the mark. Images may be captured with any detector on the FIB-SEM, including EDS and EBSD detectors. In some cases, the sample may need to be tilted to a different angle than that used for milling, in which case electron beam fiducials are very important.

In most FIB-SEM systems, the region of interest will be tilted relative to the electron beam and will result in a distorted image. Many microscopes can correct for the sample tilt in the image during acquisition, but applying the correct homography correction to the image can also be done during the post-processing steps.

After the imaging is complete, the ion beam milling step is performed again.

Stack Reconstruction and 3D Modeling #

After the full region of interest is milled and imaged, the stack of 2D images are then reconstructed into a 3D stack. The reconstruction process can have many different steps, depending on the need for subsequent analysis. See the tomography post-processing page for more details.

What are your Feelings