Comparison of Phosphor Scintillators with
Scintillating Fibre Optic Plates

The availability of terbium activated scintillating fibre optic plates has been known for some years. One of the main suppliers, Schott Fibre Optics (USA), supplied samples of LG-9 material some 5 years ago for evaluation by Photonic Science Ltd.

We found that at energies up to 50 keV, the scintillating fibre optic plates were much less efficient than conventional phosphor screens. At 200 keV and above, a definite advantage can be obtained from the fibre optic plates due to the long absorption depths possible. We have used this material for many high energy applications.

Recently (1996) claims have been made as to the superiority of Schott scintillating fibre optics over ‘conventional’ phosphors at low energies. This being directly opposite to our experience, an experimentally rigorous test was carried out on the latest batch of material from Schott to compare the performance of current material.
We had in stock plates of LG-9 material with a diameter of 75mm and a thickness of 15mm, and these were used for the tests. They were compared with an optimised Gadolinium Oxysulphide scintillator of our own manufacture.
The tests were made on a XIOS camera. This camera, a fibre-optic input slow scan imager, is routinely supplied for crystallography. We could also have used an intensified CCD (such as our ‘X-ray Crystallographer’ or ‘LA’ version) cameras, which are optimised for time-resolved studies.
The scintillating fibre optic was aluminised on the input face to virtually double the light yield compared to an uncoated plate. The phosphor was of our own production, as used on both the XIOS camera and the large area intensified cameras made for high time resolution. The physical thickness of the phosphor was 40 microns, this being considerably smaller than the optical pixel size of both imagers, so producing a very small reduction in resolution. The yield was measured in terms of analogue to digital converter unit per mA second of x-ray dose. The source was a 2kW copper target tube running at 15kVp. The vast majority was at 8 keV being the K a lines of the copper target. The aluminised scintillating fibre optic plate gave 0.34 ADU/mAsec and the XIOS-type phosphor give 3.5 ADU/mAsec.

Conclusion

The ‘conventional’ phosphor produced by ourselves produced OVER AN ORDER OF MAGNITUDE MORE LIGHT than the scintillating fibre optic plate at 8 keV.
This confirms our earlier test and shows that, particularly where a taper is used behind the scintillator, the phosphor based scintillator should be chosen to overcome the deadly ‘quantum sink’ effect, in which DQE is lost due to less than one electron (on average) being produced either in the CCD (XIOS) or from the photocathode (high speed cameras). Clearly no great advances have been made in improving the output of the Schott material, and this situation is confirmed by Schott themselves.

Discussion

It is not clear where the ideas of the superiority of scintillating fibre optic plates over optimised phosphors originated. However, it has been suggested that the measurements made by others may have used a ‘commercial’ phosphor screen as used in mammograghy or medical radiography. This material has its phosphor layer embedded in a plastic matrix which greatly reduces the light output compared to a silicate-bonded polycrystalline layer such as ours, where the light transfer out of the material is considerably enhanced by selection of particle size, size distribution, particle density and binder technique. This severe limitation to medical ‘intensifying’ screens is a result of the needs of their intended use - they are not designed as low energy x-ray detectors. Further, most medical screens are at least 80 microns thick, and since at crystallography wave length most of the absorption is in the outer surface of such a layer, considerable self-absorption will occur within the screen. Use of such a simple, non-optimised structure may well be the explanation of the origin of these technically unfounded claims. We concluded that for low energy applications our phosphor design is approximately 10 times better than LG-9 (and similar products, e.g. Collimated Holes). We will of course continue to use the scintillating fibre optic plates for the high energy applications for which they were designed and where a real advantage is to be gained.
© Photonic Science Limited 1998