X-ray Computed Tomography

X-ray computed tomography (XCT) produces a three-dimensional (3-D) computer model of the imaged volume of a sample. Contrast between different components is based on how strongly they absorb X-rays, which is typically a function of elemental composition and density. The technique is generally non-destructive and reveals the internal architecture of specimens, including pores, inclusions, and other internal features. XCT is available to the scientific community through the Environmental Molecular Sciences Laboratory (EMSL) User Program's open calls for proposals.

The versatile XCT capabilities available to EMSL users can be used for a variety of research experiments, including characterizing pore networks in soils or other geological samples, visualizing root structures of plants, and mapping the spatial arrangement of contaminant waste forms. Collaborative work with EMSL's visualization expertise enhances the 3-D information that can be obtained from XCT images. EMSL has also developed software for segmenting and extracting regions of interest from the 3-D volume, enabling quantitation of important parameters. For example, in root architecture studies, researchers can calculate the root surface area, length, average diameter, and more.

The XCT systems available (see below under Available instruments) enable research at multiple length scales, from core-scale for soils and sediments to cellular scale for biological systems.

Research application

• Supporting the Rhizosphere Function Integrated Research Platform, these resources provide image-based information and models of root/stem/leaf system growth and architecture that can be captured under different ecosystem conditions.
• Supporting the Biogeochemical Transformations Integrated Research Platform, these resources enable the 3-D characterization of soil/sediment microstructure, texture, and flow and transport properties to further research into the rhizosphere and nutrient cycling, as well as to bridge field-scale and pore-scale information in the subsurface and their influence on ecosystems.
• Supporting the Structural Biology Integrated Research Platform, these resources allow users to image individual cells, microbes, and fungi, even in their native environment, and understand how molecular information translates to organisms and how transformations observed on the micron scale are driven by organism on the nanoscale.

Available instruments

X-ray Computed Tomography, Nikon XTH 320/225: This instrument covers the roughly 1000 µm-10 µm resolution range and is well-suited for imaging soil and cores, soil-roots systems, and whole plants, up to about a foot in height and width at low resolution, and smaller, mm-sized samples down to 10 µm resolution. The uniqueness of this capability is its high resolution achievable on small specimens with micro-focused beam, while it can also accept large samples in its large cabinet with a heavy-duty sample manipulator (up to 40 kg sample weight). The Nikon system comes with three different X-ray options: a 225-kV fixed target, a 225-kV rotating target, and a 320-kV fixed target option. The X-ray wavelength can be varied using a multi-metal reflection target with Ag, Cu, Mo, and W surfaces available to provide characteristic wavelengths. The large sample chamber and the various X-ray targets available allow considerable flexibility, including in situ experiments of fluid displacement in pore networks and examination of root architecture of living plants. Sample dimensions between ca. 1 and 20 cm are easily imaged, with resolution approximately 1/1000 of the largest sample dimension.

X-ray nano-tomography microscope, Faast Nano XRM by Sigray: The nano-CT system uses tender X-rays to scan thin samples (60 µm maximum thickness) on a rotating platform that also has x-, y-, and z-axis adjustments. There are two X-ray sources: a rhodium (1.7 keV) and a silicon carbide (2.7 keV) source. The samples are mounted to 3-mm diameter electron microscopy grids, and the grids are held by custom, 3-D-printed titanium positioning pins. The tender X-rays allow for 3-D imaging of biological samples such as cells and fungi down to 30-nm resolution. Imaging can be performed on resin embedded and stained samples using absorption contrast or on cryogenically preserved samples using phase contrast. There are two options for field of view: large-scale imaging at 150 x 150 µm and small-scale at 30 x 30 µm.

Tips for success

• Scanning times: The length of time it takes to scan samples varies depending on the scientific objectives, the geometrical and material characteristics of the samples scanned, and the CT system used. On the Nikon CT, scan times generally range between 1 to 3 hours. In general, if the details of interest have a large density contrast to the surrounding matrix, or if they are fairly large in proportion to the overall volume imaged, scan times are shorter. However, if the features being investigated are small or consist of subtle density variations, longer scan times should be expected to distinguish them with sufficient clarity. Denser objects tend to take longer to scan than less-dense objects, and very small objects often require longer scan times to achieve suitable signal-to-noise ratio in the images.
• Expected resolution: The achievable resolution largely depends on sample characteristics—size being the most critical. Note: We have to differentiate spatial resolution from phase resolution/contrast. Spatial resolution refers to the smallest size or feature we can resolve, while phase contrast can be thought of as our ability to resolve or distinguish different components from each other (density contrast). For spatial resolution, a good rule of thumb is to expect resolution to be approximately 1/1000 of the largest sample dimension. Specifically, the images collected on the Nikon CT are at most 2000 × 2000 pixels, and the entire sample or the volume of interest should fit within the field of view. Since it takes a few pixels on a computer image to distinguish a feature, the maximum resolution is correspondingly a few 2000ths of the maximum dimension of the object in the scan plane. This estimation can be applied to data from other scanners, in general, if one takes the corresponding detector and sample sizes. The phase contrast/resolution depends on both sample and X-ray beam characteristics. For components that are difficult to distinguish, one can experiment with different X-ray power settings (voltage for changing the acceleration of the X-ray photons to achieve different penetration thicknesses, current for changing the photon flux, that is, the number of photons per unit area), different X-ray targets to generate different wavelength X-rays, and different X-ray filters. A general guiding principle is that high-density objects typically require higher X-ray voltages for transmission and medium-to-low beam currents (photon flux), while low-density materials tend to require lower accelerating voltages and higher beam currents. Note that the X-ray power settings can also limit beam spot size and, thus, the spatial resolution.
• Sample size: The largest field of view when scanning using our Nikon system is about 40 cm. The smallest sample that was imaged using this scanner was smaller than 1 mm x 1 mm (at about 10 µm resolution).
• Sample preparation: It depends on the sample scanned. In many cases, no preparation needs to be done at all. In general, the optimal sample geometry for scanning efficiency and image quality is a cylinder. If this geometry is not possible, it is still important to have the sample as uniform in size as possible in the scan plane. One frequent consideration is whether the sample can be made smaller (subsampled) to fit it in a smaller field-of-view for higher resolution images. Samples are often packed in a cylindrical container surrounded by air or another material of low density.
• Data or file format produced: The native data format from the Nikon scanner is Tag Image File format (TIFF) files in 32-bit gray scale (i.e., intensity values range from 0 to 255). The original TIFFs (radiographs) are reconstructed into a .vol volume file to create 3-D data (same 32-bit depth). The data can be converted to 16-bit or 8-bit formats, which entails some loss of image information, but in most cases it is not significant enough to be noticeable to the naked eye, and most software packages and computer memories for image processing or analysis tend to work better/faster with 8-bit images. Users can get their data in different formats: as volume data in .vol, .raw, TIFF stacks, .stl, etc., or they can get representative slice or 3-D rendered images. Additional analysis (volumetric analysis, particle/pore size distribution, etc.) results in data in spreadsheets and/or graphs. All of the data acquired is uploaded to a server the user can get access to, as well as archived.

Contributing teams and resources

EMSL develops and deploys capabilities for the user program by conducting original research independently or in partnership with others and by adapting/advancing science and technologies developed outside of EMSL. In some instances, EMSL directly deploys mature capabilities developed by others where there is value for the EMSL user community. The following grants/activities, PI’s and teams contributed to the development of this capability: 

• Nancy Hess, Emily Graham; 1000 Soils Research Campaign, DE-AC05-76RL01830, Environmental Molecular Sciences Laboratory
• Pubudu Handakumbura, DE-AC05-76RL01830, Environmental Molecular Sciences Laboratory
• Steven Luksic, DE-AC05-76RL01830, Pacific Northwest National Laboratory

Related publications

Bowman, MM, et al. "One thousand soils for molecular understanding of belowground carbon cycling." Frontiers in Soil Science. Vol. 3, 1120425. (2023) [DOI: 10.3389/fsoil.2023.1120425]

Handakumbura, PP, et al. "Visualizing the Hidden Half: Plant-Microbe Interactions in the Rhizosphere." mSystems. Vol. 6, No. 5. e00765-21. (2021) [DOI: 10.1128/mSystems.00765-21]

Luksic, SA, et al. "In situ characterization of foam morphology during melting of simulated waste glass using x-ray computed tomography." Ceramics International. Vol. 46, 17176-17185. (2020) [DOI: 10.1016/j.ceramint.2020.02.215]