X-ray Imaging To The Nanoscale

Abstract The spatial resolutions of traditional computed tomography instruments and micro-tomography techniques are limited to a cubic millimetre and 100 nm respectively, contributing only to morphological studies of different specimens. However, applications in materials science and biomedical research such as the study of osteoporosis, require quantitative investigation methods with high precision of both morphological and density changes in specimens. Here we report quantitative 3D imaging of a mouse femur sample using a ptychographic coherent imaging approach. A 6.2 keV monochromised x-ray beam and a PILATUS 2M detector were used to collect, in a new approach, both transmitted and full diffracted beam intensities. In this way, we generated high contrast 3D electron density maps on the nanoscale using phase contrast information. In the present study images of the osteocyte lacunae and the inter-connective canalialar network were clearly resolved on the 100 nm length scale, and the average bone density was found to be 2.0 gcm-3 with a variation of about 5 mgcm-3. This is slightly higher than the tabulated value of 1.92 gcm-3; an average over various volumes, for human adult cortical bone. The high precision of this approach may contribute for the diagnosis of bone disease in the future . Summary Osteoporosis is among the most common disease in which bones become brittle and fragile due to a loss of density. It is currently and almost exclusively diagnosed by determining an overall reduction in bone density, giving little information about the local structure and bone density changes. Such information would require three-dimensional electron density maps to the nanoscale. Current CT scanners produce attenuation maps with resolutions below a cubic millimetre and other micro-tomography techniques produce 3D imaging at resolutions of 100 nm. However, these methods do not contribute to any quantitative assessment of density changes but to a rather morphological analysis. Three-dimensional structures of single cells can be observed due to a high penetration power of X-rays. This however is achieved with a low absorption contrast. Phase change on the other hand is quantitatively more noticeable than absorption variations but is harder to detect and quantify. The research group built on the principle of ptychographic X-ray computed tomography. This scanning method provided reliable quantitative phase contrast data of the wave field past the specimen with very high precision and sensitivity. This being done do with no assumptions of negligible absorption or small phase variations. The new method described, implements computer tomography techniques. However, measuring not only the beam intensity absorbed but also the full diffracted beam intensity by the specimen. In contrast to a weaker absorption variation with increasing specimen thickness, there is an increase in phase shift, of more than 2?, causing phase wrapping which was resolved by spatial correlations in the image. Experiments were carried out on a 25 µm in diameter and 35 µm in height mouse femora which was appropriately prepared to give a soft tissue and bone marrow free polymerized specimen. The specimen was exposed for a total of 36 h by a monochromatic 6.2 keV beam with a pinhole of diameter about 2.3 µm and X-ray attenuation data was collected with a PILATUS 2M detector which provided a low-noise pixel detection. The scan points located on concentric circles were spaced by 1.2 µm thus avoiding an uncertainty due to raster scanning and ensuring adequate overlap of the illuminated areas. 704 coherent diffraction patterns with 1 s exposure time each were taken for every angle over a range of 180 degrees to obtain high resolution projection images using a ptychographic algorithm developed by the research group. This new approach produced 3D density maps of the real part of the refractive index, ?, of the specimen. This reconstructed density provided quantitative information on the electron density distribution of the sample as this is directly proportional to ?. The mass density of the sample, which was found the electron density, was found to be around 2.0 gcm-3 with a variation of about 5 mgcm-3. These results were interpreted to be very significant due to their high precision and the team anticipates that their method should have applications in other fields. The authors of the Letter correctly conclude that a precision with variations of less than 0.2% would be valuable for the assessment of density changes in bone tissues as well as materials science where such high resolution and high electron density contrast would contribute in studies of density variations in materials. However, I believe there are currently a few limitations. A radiation dose of 2 MGy as implemented in the experiment would be fatal on patients, also although the time required to retrieve the data has been reduced to 8 h it would still not be practical for clinical implementation, especially, where no high brilliance synchrotron lights are available. However , this novel approach can be very valuable in materials science in its current stage and further development could prove to be valuable for clinical applications.