With metallic and semiconducting samples, the penetration of macroscopic layers requires photon energies in excess of 10 keV. This precludes the use of optical approaches for imaging. Only projection radiography can be employed in this energy range to see through hot and molten metals. Projection radiography uses the divergence of the beam from a small source. The ultimate resolution is limited by the diameter of the source. Hence, x-ray projection radiography requires micro-focus x-ray tubes. Our X-ray Transmission Microscope (XTM) utilizes one of the smallest micro-focus sources commercially available. The spot size (electron beam focus on the target) is set down to around a one micrometer diameter.
 |
This figure schematically indicates the major components of the system and their placement. A metal sample (one mm thickness) is contained in a specially designed, high transmittance crucible. A furnace on a translation stage imposes a temperature gradient onto the sample (one end molten, the other solid). The solid-liquid interface is positioned in close proximity to the focal spot of a micro-focus x-ray source. The diverging x-ray beam permeates the sample and the resulting shadow falls on an x-ray image intensifier. The resulting visible image is converted to a digital image by a CCD camera and stored in a computer. This image is displayed on a high resolution monitor, either in real time or after further processing (contrast enhancement, filtering, etc.). |
| For bigger images, click on the diagrams. | |
 |
This diagram shows how the x-ray shadow of an object is magnified by projection. The rays emanate from the irradiated spot such that each portion of the spot casts its own shadow. The merged shadows from all areas of the spot form the final projected image. The diagram shows the broadening of features from the specimen plane onto the projection plane. The finite source spot size limits the resolution of the image. Magnification is established by the ratio of a + b to a. This method of microscopy produces infinite depth of field since no focusing is required. |
| For bigger images, click on the diagrams. | |
 |
The limit to image contrast (equally important to resolution) is examined in this figure. Here, the x-ray shadow (image) is formed from a specimen of different sized features which (for simplicity) are fully absorbing (or black). The shadow on the detector is converted to a visible light image which shows the features with differing amounts of darkness depending on their size. The problem is that very small small features can disappear because their signal is too low (below noise level). This is the contrast transfer function which limits the ability to image small features. This limitation is established by the image conversion and detection processes. Technology improvements to the XTM are best concentrated in this area. |
At typical solidification rates, motion-induced blurring limits the exposure time to less than a few seconds. With state-of-the-art x-ray image intensifier/camera combinations, which have a spatial resolution of order 100 µm (100 micrometers), a magnification on the intensifier input window of some 20X is required to obtain a spatial resolution of 10 µm. Such resolution is needed to see the dendritic structures formed in some solidifying metals. (See Related Web Sites for info regarding dendrite structures.) Since magnification is the direct ratio of detector-source to specimen-source distances, magnifications of 20X or more will require the (heated) specimen to be less than 1 cm from the housing of the x-ray tube. This leaves little room for the crucible, insulation and cooled housing creating challenging design problems for the x-ray furnace.
Of course, such observations require sufficient contrast (difference in absorptance) between features to be resolved and the retention of this contrast by the imaging devices ( image intensifier, camera, recording device). In monocomponent metallic systems, contrast between solid and melt is determined by the (electron cloud) density of the two phases resulting in less than 2% radiographic (image) contrast. [This low contrast is the main limit to the more universal application of this form of study.] In alloy systems, solute segregation will lead to further contrast enhancement. The magnitude of contrast is proportional to the difference in atomic number of the components and their concentration.
How much of the original image contrast is retained, depends on the dynamic range of the detector (imaging train) and the size of the features in the object. For small length scales, the contrast retained by the imaging train becomes much smaller than the original image contrast. (This phenomenon was described above.) This can only be (partly) compensated for if the dynamic range of the imaging train is high enough and that the lowest intensities of interest remain above the noise of the system.
Last Updated July 05, 2000