Three NIST databases are provided on the Physical Reference Data Web page from which values of x-ray cross sections can be obtained: the X-Ray Attenuation and Absorption for Materials of Dosimetric Interest (XAAMDI) database, the XCOM: Photon Cross Sections Database, and the X-Ray Form Factor, Attenuation and Scattering Tables (FFAST) database. The first two databases were developed for different purposes than the third. The XAAMDI (1 keV to 20 MeV) and XCOM (1 keV to 100 GeV) databases were developed for radiological physics and dosimetry and are based on the same theoretical calculations, which will be referred to as the XCOM data set. The FFAST database was produced for x-ray diffraction, interferometry, crystallography, and related areas and covers energies from threshold to 433 keV.
The values in the FFAST data set are calculated by different methods than the XCOM data set and may produce different results. Disagreements in the total attenuation cross sections are mostly less than 5 %, but can be larger in some cases, especially near absorption edges. Comparisons with experimental data [1,2] do not allow us to choose between the theoretical methods due to the scatter in the values of different experimental data sets.
Some general guidance can be given for the interpretation of the data contained in these two data sets. Both use approximations to carry out the theoretical calculations and cannot be expected to give exact results. For energies below 200 eV to 500 eV (the higher value for higher atomic numbers Z), both calculations are expected to give results with large uncertainties. (See estimated uncertainties in Table II of the FFAST paper in JPCRD 2000.) In fact, the XCOM data set presents results only for energies above 1 keV. The current FFAST data set has resolved previous difficulties near edges for some elements with Z above 30 at energies between 1 keV and 3 keV. Above 300 keV there is minor disagreement between the FFAST and XCOM data sets in their photoabsorption components; however, because the attenuation is dominated by scattering at these energies, the total attenuation coefficients tend to be less affected. The single electron models used for both data sets produce photoabsorption values which can be unreliable for highly correlated atomic systems, such as multi-electron atoms with low atomic number, especially Z = 2 (helium).
Figure 1 summarizes the differences in the values of the total attenuation coefficient given by these data sets. FFAST provides internally consistent values of form factors f1 and f2 and photoelectric absorption coefficients, while the XCOM data set provides only the attenuation coefficient (whose photoelectric component is related to f2). FFAST and XCOM provide various components of the total attenuation cross section such as the photoelectric and scattering contributions. Figure 2 and figure 3 compare the photoelectric and scattering components, respectively, of FFAST and XCOM. The large differences (>50 %) between FFAST and XCOM seen near absorption edges (i.e., where the functions are rapidly varying) in the graphical comparisons of the photoelectric cross section and the total absorption coefficient can be attributed primarily to two factors: 1) minor differences in the function shapes, and 2) a difference in energy resolution, which can cause the comparison of data points which are on opposite sides of an absorption edge.