Virtual histology
Histology answers questions no other technique can, and it destroys the sample to do it. Sections are thin, sparse, and cut in one plane chosen before you know what you are looking for; three-dimensional architecture has to be inferred.
Synchrotron phase-contrast tomography gives the same tissue in three dimensions, intact, at a resolution that resolves individual structures, with contrast that comes from electron density rather than from a stain that may or may not have taken. We work on lung — where the disease processes we care about are fundamentally architectural — on cardiac muscle fibre architecture, and, in the DREAMLAND consortium, on human fetal tissue imaged jointly with MRI.
The bottleneck is no longer acquisition. A single campaign produces terabytes, and the useful output is a segmentation, not a stack of slices. Building segmentation pipelines that a whole consortium can reuse is as much of the work as the imaging.
This axis is where the group connects to the rest of the 3T team and to the clinical questions that motivate it.
Heritage science
The same physics, applied to objects that cannot be opened. Carbonised papyri that would crumble if unrolled; bound manuscripts with text concealed in the binding; printing ink whose carbon black is nearly transparent to X-rays but scatters distinctively.
These projects are opportunistic — they run when a beamtime allocation and an object line up — but they are not a hobby. They are the most demanding possible test of weak-contrast imaging, and they force us to work at the limit of what the method can detect.
To add: images from the lung and cardiac campaigns, and a paragraph on the ink-detection feasibility study.