Speaker
Mr
Dmitry Sorokoletov
(Budker Institute of Nuclear Physics SB RAS)
Description
To investigate the possibility of increasing the spatial resolution of the installation "SR micro-XRF" [1], we conducted earlier an experiment on the elemental mapping of one-dimensional test sample (a fragment of an RF bipolar transistor chip, containing a number of gold stripes), as well as processing of some of the experimental results (one mapping line) [2-3] using the classical Tikhonov regularization algorithm [4, p. 73]. In this work we present the results of the processing of all data of this experiment (by two variational realizations [4, p. 119] of the Tikhonov algorithm) and a detailed analysis of the results. For example, we performed a numerical simulation of the direct and inverse deconvolution problems. We did that for cases with different resolutions, to study the plausibility of the results of solving the deconvolution problem. We also analyzed the influence of certain effects that cause systematic errors in the calculations and the resulting solution.
As a result of this work, we found out that it was very difficult to practically increase the spatial resolution 3 times (which corresponds to the distance between the gold stripes of the sample and their width of 5 μm at an FWHM of the lens instrumental function of 15 μm), while it was theoretically possible, on the "SR micro-XRF" installation with the available intensities of fluorescent signals and noise levels. A quantitative analysis of the reasons for this is discussed in this work and illustrated in detail in a number of model cases. At the same time, in certain experimental mapping regimes, the achievable resolution can be increased 1.5-2 times (provided there are no significant systematic errors in the calculations).
The results of the research will help to plan subsequent mapping experiments (they showed that the data of the mapping conducted earlier were ill-suited to processing because of non-optimal experimental regime, which led to a substantial systematic error with the desired 3-fold resolution increase, because of the edge effects and not too good signal-to-noise ratio).
This work was supported by the RFBR Grants № 14-02-00631, 16-32-00705. The work by D.S. Sorokoletov was supported by a scholarship of the President of the Russian Federation (SP-2761.2016.2).
References
[1] http://ssrc.inp.nsk.su/CKP/stations/passport/3/
[2] Sorokoletov D.S., Rakshun Ya.V. "Some aspects of quantitative micro-XRF based on the use of polycapillary lenses" // Book of Abstracts. The International Joint School «Smart Nanomaterials and X-Ray Optics 2015: Modeling, Synthesis and diagnostics» for young researchers. Rostov-on-Don, 26-30, September, 2015. (ISBN 978-5-98615-164-9)
[3] Sorokoletov D.S., Rakshun Ya.V. "An Tikhonov regularizing method in micro-XRF inverse problem on example of the studying of a test sample - a fragment of an RF bipolar transistor chip (in Russian)" // Book of Abstracts. The Young Scientist School «Applications of synchrotron and
terahertz radiation for studies of high energy materials». Biysk, 15-20, September, 2015.
[4] Leonov A.S. Solving ill-posed inverse problems: essay on the theory, practical algorithms, and demonstrations in Matlab (in Russian). Second edition. Moscow, "Publishing House "Lybrokom"". 2013. 336 p.
Primary author
Mr
Dmitry Sorokoletov
(Budker Institute of Nuclear Physics SB RAS)
Co-authors
Ms
Arina Lobova
(Budker Institute of Nuclear Physics SB RAS)
Mr
Fedor Darin
(Budker Institute of Nuclear Physics SB RAS)
Dr
Yakov Rakshun
(Budker Institute of Nuclear Physics SB RAS)
