Estimating the uncertainties in combined GIXRF-XRR for the characterization of innovative materials - CEA - Commissariat à l’énergie atomique et aux énergies alternatives Accéder directement au contenu
Poster De Conférence Année : 2022

Estimating the uncertainties in combined GIXRF-XRR for the characterization of innovative materials

Résumé

The combination of XRR and GIXRF is used for the characterization of thin films and multilayered materials, since both techniques use similar measuring and data analysis procedures. Since XRR is more sensitive to the electronic density, it is used to determine the thickness and roughness of thin layers, and GIXRF, being more sensitive to the elemental density, provides information on the depth distribution of the elements. Therefore, combining these two techniques leads to more accurate characterization than with a single technique [1]. However, there is a major difficulty to determine the associated uncertainties, due to the large number of parameters and more specifically, to the combination of the two methods with dramatically different dynamic ranges. The objective in this work is to quantify the uncertainties on the parameters of the optimized structure (thicknesses, roughness, density, etc.). Since repeating the combined measurements is a time-consuming process, we propose a recursive method for the estimation of the uncertainties of the data from the GIXRF-XRR analysis, based on the Bootstrap statistical method [2]. This method allows the generation of random uniform weights, between 0 and 1, that are multiplied by the absolute difference between the experimental and theoretical values. The application of these weights results in new sets of calculated values and the uncertainties on the optimized parameters can be derived. This method was applied to combined GIXRF-XRR measurements which where carried out in the goniometer CASTOR (Chamber of Spectrometric Analysis in Transmission or in Reflection) at the hard X-ray Metrology beamline of the synchrotron SOLEIL facility [3]. One of the studied samples is an amorphous chalcogenide GeSbTe (germanium, antimony and tellurium) thin film. In order to excite the Kα-line of germanium and the Lα-lines of tellurium and antimony, the GIXRF data were acquired at an excitation energy of 11.5 keV, while the XRR measurements were carried out at an excitation energy of 8 keV, in order to benefit from the highest photon flux. First, XRR data were analyzed with IMD [4] to obtain an estimation of the structural composition and then GIXRF spectra were fitted with COLEGRAM [5] to derive the intensity of the fluorescence X-ray lines. Finally, starting with the model obtained with IMD and the fitted fluorescence data, the GIXRF-XRR analysis was performed using in-house software to derive the sample structure and associated uncertainties.
Fichier principal
Vignette du fichier
StephanieMelhem_Presentation_Quantitative Analysis without Reference by Grazing Incidence X-ray Fluorescence _EXRS.pptx (3.71 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

cea-04174135 , version 1 (31-07-2023)

Identifiants

  • HAL Id : cea-04174135 , version 1

Citer

Stéphanie Melhem, Yves Ménesguen, Marie-Christine Lépy. Estimating the uncertainties in combined GIXRF-XRR for the characterization of innovative materials. EXRS-2022 (European Conference on X-ray Spectrometry 2022), Jun 2022, Bruges, Belgium. 2022. ⟨cea-04174135⟩
14 Consultations
10 Téléchargements

Partager

Gmail Facebook X LinkedIn More