Transparent ceramics have recently acquired a high degree of interest and notoriety, the basic applications being high energy lasers, transparent armor windows, nosecones for heat seeking missiles, radiation detectors for non-destructive testing, high energy physics, space exploration, security and medical imaging applications. In one way or another ALEM is involved in all these aspects, being a developer of optical ceramics for the past 10 years. However our prime interest is in the last (medical) applications.

Diagnostic, imaging devices using ionizing radiation (x or gamma rays), need to convert the absorption patterns of this radiation into visible images, as is commonly done in static x-rays (chest, bone damage, dental etc) as well as time dependent imaging such as fluoroscopy, CT, SPECT, various gamma cameras and PET. The converter, commonly known as scintillator, converts the high-energy photons into easily detectable visible photons. In general, the successful materials have to be highly absorbing for the ionizing radiation, efficient generators of suitable wavelength visible photons and accomplishing the conversion in minimum time, for some applications on the order of nanoseconds.

These characteristics place extraordinary demands on material science and can in most cases be accomplished by inorganic crystals of various configurations ranging from microns (powders) to objects on the inch scale. The objective is forever a moving target because an improvement in one property, invariably leads to a compromise in some other. Yet the imaging qualities place an ever-increasing demand.

Research on scintillators started in the late 19 century with Roentgens discoveries, has lasted through all of the 20 century and goes on unabated at present. However the law of diminishing returns is beginning to make itself felt, posing strategic planning demands on new ways of accomplishing the tasks. Basically there are two directions, one being the direct conversion of ionizing radiation to electrical current (bypassing the visible photon creation) or entirely new materials. The first is very attractive but encountering many fundamental limitations. ALEM is a partisan of the second approach but essentially leaving the discovery of new and better single crystal scintillators to those possessing better ideas and better equipment and concentrating on ceramic. It is now a well known fact that some highly desirable materials cannot be grown as single crystals either because they are not congruently melting, decompose below the melting point or are simply too refractory to be grown economically. Yet poly crystalline ceramic formation is possible at considerably lower temperature. One excellent example is the scintillator GOS, (gadolinium oxy sulfide), impossible to grow as single crystal, but in translucent ceramic form used in various imaging modalities. Others are ceramics developed by GE and Eu doped lutetium oxide (Eu:Lu₂O₃), developed by ALEM. The later two have the great advantage of cubic structure, greatly facilitating high transparency, but limited by speed of response.

Basically the search for ceramic scintillators offers possibilities, which may not be achievable in other ways. The main problem is that only isotropic (cubic) ceramics can be obtained in a transparent state and this greatly restricts the number of choices. In applications involving relatively low energy x-rays (mammography, dental, CT), full transparency is not essential because the ionizing photons are easily stopped and the generated visible photons only have to traverse a millimeter or less before encountering the PMT or photodiode detector. Here the translucent GOS is sometimes useful. On the other hand in Positron Emission Tomography (PET), the distances become on the order of centimeters. If the ceramic is anisotropic, multiple scattering by randomly oriented, grains can increase this distance by orders of magnitude. Since there are no materials totally free of residual absorption (impurities, tails of UV bands, etc), this introduces a loss, which lowers the efficiency of the scintillator. Hence, the key problems of developing anisotropic ceramic scintillators are purity and reduction of scattering. The first may be a difficult but generally well understandable area. The second represents much more than an enigma because both experimental and theoretical knowledge of multiple scattering in ceramics is practically non-existent. This then we regard as one of the main fields of ALEM�s activity.

While developing one or a family of materials we try to change (improve) some property by introducing variations in the methods of preparation of the material. It is therefore essential to have a numerical representation of the property, in order to monitor the progress. Clearly we need a measure of transparency, expressed numerically on some suitable scale � for instance as a percentage. And there comes the rub, because such a measure does not exist. Matters are even worse because you will not find a definition of transparency as a measurable physical quantity. The situation is similar to the encounter of the Supreme Court with pornography, " we cannot define it but we know it when we see it". Dictionaries are not much help either. So for instance Webster�s dictionary provides a correct but still qualitative definition of transparency as "having the property of transmitting light without appreciable scattering, so that bodies lying beyond (my underlying) are entirely visible".

Probably Dr. Craig Bohren expressed it best in a letter to the author: "There is indeed often confusion between transparency and translucency but, if you'll forgive the pun, the difference between them should be transparent: Images can be transmitted by transPARENT materials whereas light can be transmitted by transLUCENT materials. A transparent material is necessarily translucent but not vice versa."

J. G. J. Peelen made a valiant but unfinished attempt at coming to grips with the problem. It is a common test for distinguishing between translucency and transparency of a ceramic plate, by putting it in contact with a pattern (maybe simply print) on top of a light box. Countless pictures of this kind litter the ceramic literature. Often it is quite possible to see the pattern and read the print, if the plate is in direct contact with the image. However if one lifts the plate a few millimeters above the pattern, it becomes totally obliterated. Transparency of the plate is therefore related to the distance separating it from the pattern. For an obviously transparent material like glass this distance being arbitrarily large ceases to play any role. An often used variant of this test is to hold the tested plate at arm�s length from the eye and try to determine the visibility of a well lit edge, such for instance as the edge of a fluorescent light fixture. If this is possible we declare the plate as "somewhat" transparent.

Fig 1. Scatterometer measurements of specimens with various levels of transparency, as indicated by intensity of central peak (in-line transmission). Note variation in shape of large-angle background, related to microstructure of ceramic. (Horizontal axis is scaled as square root of angle for clarity of presentation.)

Because of the obvious relationship between transparency and scattering we have used a Stover scatterometer [], consisting essentially of a laser source and a goniometor arm carrying a detector, to map the angular dependence of the light intensity in a plane parallel to the incoming laser beam. The instrument has an incredible dynamic range of some 12 orders of magnitude and a resolution better than 0.1 degree. This allows for wide angle or narrow angle, high resolution scans. Some typical traces are reproduced on Fig. 1. Note that with the exception of a ground glass, which is definitely not transparent, all of the other materials are characterized by the presence of a narrow forward peak and do indeed pass the viewing test described in the previous paragraph.

From what we said and observed it is obvious that the existence of the forward peak appears to be a necessary condition for some degree of transparency. However what feature of the peak (sometimes referred to as "inline transmission"), such as its magnitude, functional representation or FWHA, is important we still do not know.

Another obvious conclusion is that transparency, ultimately insuring perfect transmission of images through plates of transparent materials, must be related to the range of spatial frequencies characterizing the material. Thus it is evident that there must be a close, but yet not fully exploited correspondence between scattering profiles of Fig 1 and Modulation Transfer Functions (MTF) characterizing the materials.