banner image for surface and microanalysis science division with links to chemical science and technology laboratory and www.nist.gov and return to division home page Chemical Science and Technology Laboratory label with link to www.cstl.nist.gov Surface and Microanalysis Science Division label with link to home.html NIST logo label with link to www.nist.gov

Analyzing the Tough Ones: Quantitative X-ray Microanalysis of Extreme Topography

Dale E. Newbury

Objective: Evaluate the accuracy of the peak-to-background method for quantitative electron beam x-ray microanalysis of specimens with extreme topography.

Problem: Conventional quantitative electron probe microanalysis is restricted to surfaces that are flat to a relief of 100 nm amplitude or less. Otherwise, geometric effects on electron scattering and x-ray absorption violate the basic assumption of matrix correction calculations that composition is the only reason x-ray intensities differ between the unknown and the standard(s) (e.g., pure elements). Specimens examined in the scanning electron microscope (SEM) with semiconductor energy dispersive x-ray spectrometry (Si-EDS) are often characterized by extreme topography. Conventional quantitative x-ray microanalysis of rough specimens can be severely compromised. As shown in the table, the conventional analysis of IN100 alloy in the form of a sphere in which the analyzed location is on the backside of the sphere away from the Si-EDS results in large errors, especially for characteristic peaks with low photon energies below 4 keV.

Table of Relative Errors for IN-100 Alloy Sphere (122 micrometer diameter; backside from EDS)
Element
C (wt. frac.)
Conv. (raw) error
Conv. (norm.) error
P/B error
Al-K
0.0603
-77%
-71%
-18%
Mo-L
0.0353 + 6
-62
-51
+ 6
Ti-K
0.0519
-16
+7
+19
Cr-K
0.0965
-23
- 3
+12
Co-K
0.155
-18
+5
- 0.3
Ni-K
0.601
-14
+9
- 1

Approach: To overcome the limitations of conventional analysis in such cases, an additional source of information must be supplied that can compensate for geometric factors. For point beam analysis, the peak-to-local background (P/B) method, which incorporates measures of both the characteristic and continuum (bremsstrahlung) x-rays at the same energy, can provide the necessary information for correction. While the P/B method has been available for many years, there exist no systematic studies of its efficacy for rough specimens. NIST-NIH Desktop Spectrum Analyzer (DTSA) provides an ideal x-ray spectrum-processing platform to study each stage of P/B corrections.

Results and Future Plans: The impact of the P/B corrections upon the accuracy of the analysis is shown in the table, where the relative errors are seen to be significantly reduced after P/B corrections. This on-going measurement program will be extended to other alloy systems as well as non-conducting targets to develop sufficient data to fully characterize the error distribution of the peak-to-background correction method.

Publications:
DE Newbury, Microbeam Analysis 2000, Inst. Physics Conf. Series #165, 427-428, 2000.
DE Newbury, Microscopy & Microanalysis, 6, suppl. 2, 928-929, 2000.

Last Updated March 5, 2002

Web Contact micro@nist.gov