Just like visible light, X-rays also experience refraction. The refraction is caused by what is called phase shift through the material the radiation passes. Even though this effect is much weaker for X-rays than for visible light, it can be used to create contrast for imaging. Phase-contrast imaging requires special techniques to be detected but can be very beneficial. For materials that have low absorption, such as soft biological matter, polymers and many other organic compounds, the phase can give more than 1000 times stronger contrast than absorption. Without going into the detailed physics behind phase contrast, we can still understand the methods that are used to exploit it.
To make phase-contrast imaging, the phase shift induced by the object must be measured. The unfortunate catch is that X-ray detectors don’t measure phase, but only intensity. The phase shift is instead converted to intensity variations, that can be detected. There are several methods to do this, which are all good in different ways.
Compared to absorption-contrast imaging, the requirements on the equipment are in general higher for phase-contrast imaging. To be able to obtain contrast, an X-ray source with high transverse coherence is needed. This means that the source either has to have a small emission spot or has te be moved far away from the object. Both options are traditionally related to very low flux, which leads to very long exposure time. To keep the exposure time short, an X-ray source with a small emission spot that can still maintain a high flux is very beneficial. Another commonly used metric is the source brightness. The source brightness is a measure that includes both coherence and flux and is important to have high in many cases!
Phase-contrast imaging was originally developed for biomedical applications, where it is very beneficial for imaging of soft tissues. Recently, it has seen a growing interest also in materials science, engineering and industrial non-destructive testing.
Simple example of how the phase shift through an object perturbs the wavefront. The wavefront cannot be directly measured but is here converted to intensity variations at a distance. The contrast arises from both absorption and phase. The edge enhancements introduced by the phase shift is apparent along all contours.
Propagation-Based Phase-Contrast Imaging (PBI)
PBI has the simplest measurement geometry, similar to conventional attenuation-contrast imaging and requires no optical elements. The only difference is that free-space distances are kept between the object, sample and detector. This leads to that the phase can generate intensity variations, visible as edge enhancements. The image represents a mixing of the absorption-contrast signal and the phase-contrast signal. Implementing PBI in a compact laboratory-scale setup puts high requirements on the spatial coherence of the source, detector resolution and long distance between the sample and the detector. Both the MetalJet and NanoTube sources meet the requirements for doing laboratory-based high-resolution PBI. Furthermore, the high brightness helps to keep the exposure time down, while still enabling high resolution.
To obtain an image that represents the object thickness, the phase-contrast image must be treated with a process called phase retrieval. There are many methods for phase retrieval, most of them based on convolutions. If using a single image, assumptions on the object must be made, such as assuming a single material object. With multiple images acquired at different propagation distances, this assumption can be circumvented. On the other hand, multiple images introduce a higher complexity in the experiments.
Application Examples
High resolution propagation-based imaging system for in-vivo dynamic computed tomography of lungs in small animals
By operating a MetalJet D2+ at 250 W and 15 µm spot size, it has been shown that phase-contrast tomography can be used for dynamic imaging of live mice. In research work done in Australia, time-resolved computed tomography was performed to image the ventilation in the lungs of a mouse. A flat-panel detector acquired projections with only 18 ms exposure time, allowing a full tomography in 32 s. These very short exposure time and controlled breathing, allowed small airways down to 55-60 µm in diameter to be imaged dynamically. This high quality dynamic imaging of lungs enables determination of the lung function, even down on a regional level. Furthermore, high quality dynamic CT potatially has many other applications.
Time-resolved computed tomography of a live mouse (A). Close-up region (B) shows the anatomical features. The method demonstrates the differences in volume of air in the lungs after 0 hours mechanical ventilation (C)-(E), and after 2 hours (F)-(H). Image reprinted from M. Preissner et al., “High resolution propagation-based imaging system for in vivo dynamic computed tomography of lungs in small animals”, Phys. Med. Biol. (2018).
The Nanotube N1 60 kV is based on advanced electron optics refined from the 
The PBI tomographic axial (a) and sagittal (b) slices and the zoom-in ROI (c) as the black box in (b).
3-D rendering of the PBI tomogram of the soft tissue in the thorax.
PBI tomography slice along transverse (a) and another along longitundnal (b) directions; automatic volume rendering of the reconstructed volume (c) and the zoomed-in cellular segmentation (d). The molecular layer (ML), granular layer (GL), white matter (WM) and Purkinje cell layer (PCL) of the cerebellar vermis at cellular resolution.
Single grating imaging with MetalJet source and a high-resolution detector. A nylon fiber is imaged and a single image captures(a) attenuation(b) phase and(c) dark field. Scalebar is 260 µm.
The radiographs of a dried spider in attenuation-contrast, dark-field contrast and phase-shift along transverse and longitudinal directions (b). (a) is the photo of the sample and one raw speckle image.
