Computed Tomography Applied to Geosciences
by Allen H. Reed and Yoko Furukawa
Computed tomography (CT) provides an exciting capability to geosciences by capturing the geometry of objects bounded by optically opaque materials. The geometric information provides guidance, constraints and boundaries within which to model processes, systems, transitions, and changes. The development of CT has progressed dramatically in the 30 years since it was first developed. In the beginning, CT was slow and cumbersome, but it demonstrated a clear ability to resolve a material embedded within another material. Hounsfield (1973) performed the first scans, one of which was of a pig's pancreas. The scan took the entire day and by the end of the day gas bubbles had developed as artifacts within the decaying tissues, yet the ability to resolve the pancreas in 3D had been achieved. Soon afterwards, the medical profession adapted this technology and pioneered many changes to increase speed and utility and to enhance image quality. Seeing an opportunity to advance understanding in fluid mechanics, groups at Shell and Chevron applied medical CT systems to evaluate multiphase flow in core samples (2-3). Currently, medical scanners have improved image-resolution (to ~250-400 um from 1-2 um; Note: image resolution is equivalent to the size of the voxels, a 3D pixels). As industry and security needs have emerged, high-resolution industrial scanners (to <10 um resolution) have enabled scanning of smaller objects, such as pores and grains in a sediment or rock assemblage and the fluid and gas phases trapped in the pore space. With this advance, the ability to visualize and research multiphase fluid flow, grain interactions, pore clogging, gas formation, and a host of other geological processes is possible.
Computed tomography enables volumetric imaging of objects by emitting high-energy photons, (i.e., gamma rays or x-rays) through an object. These photons, produced either by gamma emitters, such as Ce-137 or by conventional electrical sources, are emitted from an x-ray tube through the sample material and to an x-ray detector. Once the x-rays reach the detector they are converted, through a series of processes, to an image. In the case of conventional systems, x-rays are produced electronically by illuminating a cathode to produce photons within an x-ray tube. The photons are accelerated through a magnetic field to an anode which produces x-rays when contacted by the photons. X-rays that pass through the focal spot, adjacent to the anode and at the front of the X-ray tube, are projected towards the sample material. The X-rays that pass through the sample and reach the X-ray detector, the image intensifier, are converted to an electric signal. This signal is then converted to an array of gridded gray-scale values or the voxels that comprise the image (2).
The CT system at the Naval Research Laboratory has a stationary source (x-ray tube) and receiver (image intensifier) and the sample is rotated and translated between the source and receiver. The distance between the source and receiver can be adjusted to accommodate large diameter samples (up to 11 cm) and small diameter samples (~3-5 um). Sample diameter, as well as density and atomic number, play the key roles in determining sample resolution. See references 2 and 3 for more information on the specific determinants of x-ray attenuation and image resolution.
Samples scanned at NRL range from 6 um to 11 cm and it is common at NRL to scan small cores of sand to capture the pore- and grain-scale geometry and mud cores to capture features (e.g., gas bubbles, burrows) that persist in these cores. The small diameter sand cores, (6 to 8-um) are imaged with 10 um resolution. The objects in mud cores, 6 to 11 cm diameter) are often larger. The 6-cm diameter cores of mud provide resolution of 120 um.
The capabilities of the microfocus x-ray CT system to resolve a variety of materials is presented in a series of figures and movies. The examples displayed below pertain to research in the geosciences, including geochemistry, marine geology, and geomechanics; they demonstrate the capabilities of this system. Specifically, these examples assist in the evaluation of structural changes in mud due to biogeochemical alteration of organic matter or gas bubble production. It turns out that assigning this problem to one sector of the geosciences would be too limited. As research in bubble production demonstrates, the mechanics of bubble growth differs in natural sediments from reconstituted sediments due to the difference in mechanical characteristics between natural and reconstituted mud (Boudreau et al., 2005; Figure 2, Figure 3; Movie 1:Gas bubbles in natural sediments (38.9MB MPEG Movie)).
Geosciences may require resolution of fine-scale geometrical details. The ability to resolve fine-scale details of small objects is demonstrated in the partial skeleton of the Mosquito fish and the inner ear of an extinct dolphin (Figure 4). The fish skeleton, although slightly off the topic, puts the resolution capabilities of NRL' CT into perspective that may be easier to grasp. The image of the fish, a 1.5-cm long section scanned that was 0.5 um thick, displays the ability to resolve small-scale features.
An important factor determining oxidation of organic matter and mineralogical changes in buried sediments is bottom water oxygen. In sediments, infauna rework sediments and create burrows, which serve as conduits for air in terrestrial environments or for oxygen rich bottom waters in aqueous environments. Burrow geometry is presented in bioturbated sediments (Briggs et al., 2005; Figure 5; Movie 2: Bioturbation Example 1 (4.6MB MPEG Movie) and Movie 3: Bioturbation Example 2 (4.3MB MPEG Movie)) from which one can determine the magnitude of oxygen penetration.
The composition and interrelationship of sand grains in many settings determines fluid flow, diffusion, dispersion and chemical precipitation. Modeling these processes in marine systems, aquifers, and the vadose zone has been facilitated by the ability to resolve sand grains in high detail (Thompson et al., 2006; Figure 6; Movie 4: Microfossils (34.9MB MPEG Movie)).
As demonstrated in these examples, resolving the geometry of objects provides means to analyze geological problems, geochemical processes and geomechanical constraints. Numerous other objects and possibilities for research exist. It is our hope that this presentation will entice your imagination, foster new ideas and promote research options.
The wide array of materials that have been imaged at NRL demonstrate great potential for evaluating problems that exist within three-dimensional geometrical boundaries. These geometries provide the boundaries and constraints within which chemical and physicochemical processes can be analyzed. In addition to static three-dimensional problems we are currently evaluating dynamic processes in four-dimensions by including a time-component. These evaluations include rates of bioturbation, chemical transfer across burrow walls, single and multi-phase fluid flow, and bubble formation and ebullition; many additional possibilities exist.
Interested parties may contact the authors for further information on collaboration, research ideas and projects, and access to the CT facility. Post-doctoral candidates are sought to perform research on related projects.
Movie 1: Gas bubbles in natural sediments. Gas bubbles in natural sediments collected from the same area as the sample in figure 3.
Movie 2: Bioturbation Example 1 and Movie 3: Bioturbation Example 2. These movies depict sandy sediment systems which have inclusions of shells (white) and burrows (brown). The sand has been made transparent. Cores are 6-cm in diameter. See Briggs et al. (2006) for more information.
Movie 4: Microfossils. An assemblage of sand-sized microfossils from Mihlos, Crete is presented in a slice progression movie. The sand sized materials have a median grain size of ~470 micrometers. The image resolution is 11 um.
Allen H. Reed
Marine Geosciences Division
Naval Research Laboratory
Stennis Space Center, MS 39529
We are grateful for the support of NRL, which provided the funds under a CPP to acquire the CT and to Drs. Eric Hartwig, Herbert Eppert, Philip Valent, Michael Richardson, and Kevin Briggs. Funding for these projects was provided by provided by NRL core funds and ONR's offices of Coastal Geosciences, Acoustics, and Optics. Periotic membrane provided courtesy of Calvert Marine Museum, Solomons MD. We thank Kevin Wilson, an undergraduate student, for producing the movies, Karsten Thompson for sand images and Chris Algar/Mark Barry for image of injected bubble.
- Hounsfield G. N., 'Computerized transverse axial scanning (tomography). 1. Description of system', Br. J. radiology, 46(552) 1016-1022, 1973.
- Wellington, S.L., Vinegar, H.J., X-ray computerized tomography. Journal of Petroleum Technology 39 (8), 885-898, 1987.
- Flannery, B.P., Deckman, H.W., Roberge, W.G., D'Amico, K.L., 'Three- dimensional X-ray microtomography', Science 237, 1439-1444, 1987.
- Boudreau B. P. C., Algar C, B. D. Johnson , I. Croudace , A. H. Reed, Furukawa Y, Dorgan KM, Jumars PA, Grader AS, Gardiner BS, 'Bubble growth and rise in soft sediments,' Geology (6), 517-520, 2005.
- Briggs, K. B., A. H. Reed, and M. D. Richardson, 'Using CT to image storm-generated stratigraphy in sandy sediment off Fort Walton Beach, Florida, USA', Proceedings of the International Conference 'Underwater Acoustic Measurements: Technologies & Results,' Crete, Greece, 2005.
- Thompson, K. E., C.S. Willson, and W. Zhang, 'Quantitative computer reconstruction of particulate materials from microtomography images,' Powder Technology 163, 169-182, 2006.
- Reed, A. H., K. E. Thompson, K. B. Briggs, and M. D. Richardson, 'Quantification of pore and grain properties of quartz sands from microfocus computed tomography images,' in prep for IEEE-Oceanic Engineering 2007.