X CT: A Modern Approach to Biological and Material Imaging

X-ray computed tomography at the micro and nano scale has become one of the most versatile imaging tools in modern science, capable of revealing the three-dimensional internal structure of everything from mouse embryos to 3D-printed titanium parts without cutting into them. The technique works by collecting hundreds or thousands of X-ray “shadow” images from different angles around a sample and then computationally reconstructing a full 3D volume, where every tiny cube of space gets assigned a brightness value corresponding to how much X-ray energy it absorbed. What makes this approach so powerful is its range: depending on the setup, researchers can resolve features smaller than a micrometer in samples ranging from a few millimeters to several centimeters across, and the list of fields relying on it keeps growing.

How X-ray Micro-CT Produces a 3D Image

A micro-CT scanner generates X-rays using a micro-focus tube, where electrons are accelerated by a voltage (often up to 240 kV or more) and slammed into a metal target, typically tungsten. The X-rays pass through the sample and land on a flat detector, producing a two-dimensional shadow image for each angular position. The sample sits on a rotation stage that turns it through 180° or 360°, and the scanner captures many hundreds or thousands of these projection images during the rotation. Afterward, software uses filtered back-projection algorithms to combine all those flat images into a three-dimensional dataset where each volumetric pixel, or voxel, represents the actual X-ray density at that point in the sample.¹

The result is a complete digital replica of the object’s interior. You can slice through it in any plane, measure internal distances, segment individual structures, and quantify features like porosity or tissue boundaries. Compared with optical and electron microscopy methods that build 3D volumes by physically sectioning a sample or by limited depth-of-focus tricks, micro-CT can image the internal structure of opaque, macroscopic samples at sub-micron resolutions without destroying them.²

Making Soft Tissue Visible

Micro-CT was originally the domain of hard, mineralized materials like bone, teeth, and rock, because those absorb X-rays strongly and show up with high contrast. Soft biological tissues are a different story. Made largely of water and light elements like carbon, hydrogen, nitrogen, and oxygen, they absorb X-rays weakly and look nearly transparent in a standard scan. This was the central barrier to using micro-CT in much of biology for years.

The workaround that opened the floodgates is chemical staining. Soaking a tissue sample in an X-ray-dense contrast agent before scanning dramatically boosts the contrast between different soft tissue types. Two stains in particular have become workhorses: iodine-based solutions and phosphotungstic acid (PTA). Both are relatively cheap, easy to handle, and far less toxic than the osmium tetroxide that electron microscopists traditionally relied on.³

Iodine staining has proven especially useful because it does not distribute evenly across tissues. In heart muscle, for example, iodine preferentially accumulates in muscular fibers rather than in connective tissue. This selectivity means that a stained micro-CT scan of a canine heart can clearly distinguish muscle from connective tissue and even resolve the orientation of individual atrial fibers, which is valuable for building computational models of how electrical signals travel through the heart.⁴ The same principle has been applied to peripheral nerves. Using Lugol’s solution (a simple mixture of iodine and potassium iodide in water), researchers have visualized regenerating nerves inside a silk fibroin tube in full 3D, something that would previously have required serial histological sectioning.⁵

Developmental Biology and Whole-Organism Imaging

One of the areas where staining-enhanced micro-CT has had the biggest impact is developmental biology. When you want to understand how an embryo’s organs form, you need to see the whole organism in three dimensions at fine resolution, ideally without destroying it. Micro-CT fills that niche. Staining protocols tuned for embryos allow researchers to acquire high-contrast, high-resolution datasets of whole embryos and their specific organ systems, and the technique has become a major tool for phenotyping mouse development, both qualitatively and quantitatively.⁶

The method is not limited to mammals. Researchers have used micro-CT to track the entire developmental trajectory of the African clawed frog, from pre-metamorphosis tadpoles through to fully mature adults, generating 3D models that capture changes in body shape, skeletal formation, teeth, brain, and gut at each stage.⁷ In marine biology, synchrotron-based micro-CT with phosphotungstic acid staining has been applied to map the embryonic development of the small-spotted catshark, spanning from gastrulation through late organogenesis. The resolution was fine enough to segment sensory organs, fins, muscles, and skeletal elements as they formed across 36 whole-embryo scans.⁸

Inspecting Manufactured Parts

In materials science, micro-CT has carved out a critical role in quality control and failure analysis, particularly for additively manufactured (3D-printed) metal components. When a laser melts metal powder layer by layer, tiny pores can form inside the finished part, and those pores affect how the material performs under stress. Micro-CT lets engineers map exactly where those pores are, how large they are, and what shape they take, all without sawing the part open.

For 3D-printed titanium alloy parts, micro-CT imaging has shown that the porosity, pore size, and pore location depend on the build direction, which in turn leads to different mechanical properties.⁹ In copper-chromium alloys, high-resolution scans have revealed that samples made from virgin powder develop keyhole pores and deep pit cavities, while samples made from surface-modified powder show mostly small, spherical, evenly distributed defects, a finding that directly informs which powder preparation strategy produces more reliable parts.¹⁰

Micro-CT is also being used to study fiber-reinforced composite materials. When microvascular channels are embedded in glass-fiber composites (channels that could carry healing agents or coolants in service), the fibers around them get pushed out of alignment. Micro-CT image analysis has quantified that misalignment at up to about 6° out of plane, with the channels reducing local stiffness by up to 19%.¹¹

Geology and Digital Rock Physics

Underground rock formations contain networks of pores and channels that determine how fluids like water, oil, and gas move through them. Drilling a core sample and measuring permeability in a lab is expensive and slow. Micro-CT offers an alternative: scan the rock, reconstruct its pore network digitally, and simulate fluid flow on a computer. This approach, known as Digital Rock Physics, has become a dominant technique for characterizing fluid flow in rock.¹²

The power of this approach extends to multiphase flow experiments, where researchers combine actual flooding experiments with high-resolution X-ray imaging to measure not just the pore space but also the fluid configurations and local contact angles within the rock.¹³ That said, the method has known limitations. Imaging constraints including limited resolution, segmentation biases in coarser scans, and the finite field of view can lead to systematic errors: porosity can be underestimated by as much as half, and permeability overestimated by up to a factor of ten, compared to physical measurements.¹⁴ Researchers working in this space spend considerable effort calibrating their digital measurements against traditional lab tests for exactly this reason.

Watching Damage Happen in Real Time

One of the more striking recent developments is “4D” micro-CT, where the fourth dimension is time. Instead of scanning a sample once and looking at a frozen snapshot, researchers scan it repeatedly while applying a load, heating it, or exposing it to some other stimulus, effectively making a movie of the internal structure as it changes.

In metals research, this means loading a 3D-printed stainless steel sample inside the scanner while pulling it apart and capturing CT volumes at each stage. This in-situ approach lets researchers watch microscale voids initiate, grow, and merge with neighboring voids under stress, processes that were previously only inferred from post-mortem examination of fractured samples.¹⁵ The same principle has been applied to recycled aggregate concrete, where 4D CT scanning under compression reveals how cracks form and evolve around recycled aggregate particles and reinforcing fibers.¹⁶

Ceramic matrix composites used in aerospace are another application. In silicon carbide fiber composites, 4D in-situ micro-CT combined with deep-learning segmentation has been used to automatically identify cracks as they open under tension, tracking their initiation, propagation, and the progressive opening of individual matrix cracks in full 3D.¹⁷

Phase-Contrast and Spectral Imaging

Standard micro-CT measures how much the sample absorbs X-rays. But X-rays also undergo a phase shift when passing through matter, and that phase shift is much more sensitive to light elements than absorption is. For elements like hydrogen, carbon, nitrogen, and oxygen, the phase shift cross section is roughly a thousand times larger than the absorption cross section. Phase-contrast X-ray CT exploits this difference and can image soft biological tissues without any staining and without heavy radiation exposure. Early work demonstrated that phase-contrast X-ray CT could clearly distinguish cancerous lesions from normal tissue and depict fine structures related to cancerous degeneration.¹⁸

In clinical-scale applications, propagation-based phase-contrast CT of the human knee has shown that it can visualize altered cartilage regions that are invisible in absorption CT, with surface detail comparable to dedicated 3-Tesla MRI sequences. It also picks up cartilage and meniscal calcifications that neither conventional CT nor MRI can detect.¹⁹ In zoology, the same technique has been used to image mollusks with their shells still intact, revealing soft-tissue features and shell growth lines that are nearly invisible in conventional absorption-based micro-CT.²⁰

A separate frontier is spectral CT, which uses photon-counting detectors or multiple X-ray energies to distinguish materials based on how they absorb X-rays at different energies. Dual-energy micro-CT, for instance, can quantify the effective atomic number and electron density of each voxel with relative errors within about 5%, while also strongly reducing beam-hardening artifacts and improving contrast compared to single-energy scans.²¹ When combined with phase retrieval, spectral CT can decompose a sample into its constituent materials, such as calcium, iodine, and water, with relative quantification errors within 10%.²²

Synchrotron Versus Lab Sources

Most of the applications described so far can be done with a laboratory micro-CT scanner, which sits on a benchtop and is available around the clock. But some experiments demand the extreme brightness of a synchrotron, a particle accelerator that produces an extraordinarily intense, highly focused X-ray beam. The practical upshot is speed: synchrotron sources can cut scan times from hours down to minutes, which reduces the chance that the sample moves during the scan and yields sharper final images.²³ That speed advantage also matters for in-situ experiments where you need to capture fast processes.

Access is the main drawback. Synchrotron beamtime is allocated through competitive proposals, and demand far outstrips supply. For routine characterization or industrial quality control, lab-based systems remain the practical choice. Recent developments in lab-based nano-CT, which uses zone-plate focusing optics to achieve even finer resolution than standard micro-CT, have brought high-throughput characterization of tiny samples like individual battery electrode particles into the realm of minutes per scan rather than hours or days.²⁴

Plant Roots, Soil, and Agriculture

Plant scientists face a peculiar imaging challenge: roots grow in soil, an opaque medium, and digging them up destroys the very structure you want to study. Micro-CT sidesteps this by imaging roots within their natural soil environment non-invasively. Automated software tools like RooTrak treat the CT volume as a sequence of cross-sectional images and use visual tracking algorithms to follow root objects through the stack, recovering the full 3D root architecture.²⁵

Beyond mapping root shape, high-resolution CT has revealed details about how roots physically change the soil around them. It turns out that as a root pushes through packed soil, it increases porosity right at the root-soil interface, creating a thin zone of loosened structure. This finding contradicts an older assumption that growing roots compress and compact the surrounding soil. The zone of structural influence appears to be more localized than previously thought, operating at the micrometer scale rather than the millimeter scale.²⁶

Reading Ancient Texts Without Touching Them

Perhaps the most unexpected application of X-ray CT is in archaeology. The Herculaneum papyri, ancient scrolls carbonized by the eruption of Mount Vesuvius in 79 AD, are far too brittle to unroll physically. Researchers have developed a pipeline that uses volumetric X-ray CT images to scan intact, rolled scrolls and then “virtually unwrap” them using a combination of geometric modeling and machine learning, revealing hidden text without ever touching the fragile material.²⁷

A similar approach was used on the En-Gedi scroll, the oldest Pentateuchal scroll in Hebrew found outside the Dead Sea Scrolls. The scroll was so badly damaged that its sheets were disintegrating and untouchable. Virtual unwrapping of X-ray CT data produced readable columns of text at high enough quality to support serious scholarly analysis of the writing.²⁸ Paleontology benefits too: X-ray micro-CT has been used to non-destructively image the interiors of rare silicified fossil plants, revealing fine cracks and small cavities that help researchers understand preservation conditions and internal anatomy.²⁹

Artifacts, AI, and the Limits of the Image

No imaging technique is artifact-free, and micro-CT has its own well-known headaches. Beam hardening is the most pervasive: because a typical lab X-ray source produces a spectrum of energies rather than a single energy, lower-energy photons are absorbed more readily as they pass through the outer layers of a dense sample. By the time the beam reaches the center, its average energy has shifted upward, creating artificial brightness variations in the reconstructed image, notably “cupping” artifacts where the center of a uniform object appears darker than the edges, and ring artifacts. Deep-learning approaches using convolutional neural networks have shown promise at suppressing these artifacts by learning the mapping between artifact-affected and artifact-free images.³⁰

AI is also being applied to a more fundamental limitation: radiation dose. Scanning at full dose gives clean images but can damage biological samples or simply takes too long for high-throughput workflows. Reducing the dose introduces noise. Neural networks trained on paired low-dose and high-dose scans can restore image quality from noisy low-dose data, outperforming traditional denoising methods in terms of noise suppression, structural preservation, and contrast.³¹ For spectral CT, dictionary-learning-based approaches tackle a different problem: the instability of decomposing multi-energy data into material maps, where noise in the input images gets amplified by the decomposition math. These methods improve the accuracy of material identification from spectral CT scans.³²

The broader lesson is that a micro-CT image is not a photograph. Every voxel’s brightness is a reconstruction influenced by the source spectrum, the detector response, the number of projections taken, and the algorithms used to turn projections into a volume. Careful attention to scan parameters and calibration against physical measurements remains essential, whether you are quantifying pores in a rock or tracing nerve fibers in a regenerating limb.