A method for thickness determination of thin films of amalgamable metals by total-reflection X-ray fluorescence

https://doi.org/10.1016/j.nimb.2009.04.018Get rights and content

Abstract

A method for thickness determination of thin amalgamable metallic films by total-reflection X-ray fluorescence (TXRF) is presented. The peak’s intensity in TXRF spectra are directly related to the surface density of the sample, i.e. to its thickness in a homogeneous film. Performing a traditional TXRF analysis on a thin film of an amalgamated metal, and determining the relative peak intensity of a specific metal line, the layer thickness can be precisely obtained. In the case of gold thickness determination, mercury and gold peaks overlap, hence we have developed a general data processing scheme to achieve the most precise results.

Introduction

Several methods for thickness determination have been developed up to now. “In situ” methods, such as those based on microbalances [1] or crystal deposition [2] have shown to be very useful to sputtering deposition and multi-layer growth. Although “in situ” and non-destructive, they have a low accuracy, mainly due to the fact that the measured film is the one deposited on the monitor, not the actual one. Other thickness determination methods range from the direct observation with a SEM [3] to more indirect methods such as those based on interferometric techniques, gamma-ray transmission [4] or PIXE [5]. Several proposals for thickness measurement using TXRF have been previously made [6]. TXRF is a not a widely known technique. It has been developed as an efficient, fast method of elemental analysis for both the quantitative and qualitative determination of elements with Z > 14 [7]. The method is advantageous because it is multielemental, has low detection limits (on the order of tens of nanograms per gram), and requires only minute specimen quantities. Furthermore, the usual practice of adding an internal standard to affect quantification may often be avoided in the analysis of trace elements in organic matrix materials because most of the constituent of the sample are low Z elements [8]. In this work TXRF is not used as a standard method for chemical characterization but rather as an alternative way to determine a physical property of the sample: its thickness. This is done by modifying the chemical composition of the sample in a known precise way. A mass of mercury is added to the surface of the sample. If the relative sensitivities of detection of the elements involved are known, the thickness of the sample can be determined. The method is applicable for thickness determination of thin-film amalgamable metals, such as gold or platinum, by TXRF. The proposed method is very simple, because it uses previously established infrastructure of hardware and software, originally designed to determine chemical composition.

Section snippets

Theoretical

The basic equation for TXRF for a thin film is given by [9]Ii=KiI0N0σiωimiAi,where the intensity of the line originated in the element i, Ii, are written in terms of the following parameters: Ki which depends on experimental geometry and energy detection efficiency; I0 that stands for the intensity of the excitation source; Avogadro’s Number N0; mi which represents the mass surface density of the element i with atomic number Z = i and atomic mass Ai and finally the element i collision cross

X-ray spectrometer

The TXRF analytical system employed in this study is a modified unit designed at the Atominstitut [13] which consists of a high voltage generator (Siemens Kristaloflex 710 H), operated at 40 kV and 20 × 10−3 A. The generator supplies a line, fine-focus, molybdenum-anode X-ray tube (Philips model 2215/20) so that a fine-line X-ray source beam enters the attached TXRF module. The module contains three stages. In the collimator the divergent beam falls on tungsten slit parallel to the X-ray tube line

Results

Each sample spectrum was obtained as described in items 4 and 5 of Experimental Procedure. Each spectrum was acquired four times, and then those were compared to check the stability of the data. Later, all the spectra were summed and the area was normalized in the ROI.

The values of α with their relative errors were determined for each sample solving Eq. (17), using the ROI between channels 530 and 720 (as is shown in Fig. 2, for the amalgam shown in Fig. 1). In Fig. 3, Fig. 4 we present the α

Limitations of the technique

In the light of the TXRF characteristics [7], the main and unique limitation of the proposed determination is produced because of the attenuation inside of the thin metallic film of: (1) The X-ray exciting beam. As the thickness increases, the inner layers become less studied; so information about those atoms cannot be appropriately obtained. (2) The attenuation of the fluorescent emerging spectrum produced inside the sample, in its travel outside the sample. The attenuation of the emerging

Conclusions

A very simple method and data treatment for thickness determination of thin films of amalgamable metals by TXRF has been developed. The method was tested in gold and platinum films, but it can be used with any metal that can be amalgamated, like, Cu, Pd, Ir, Ag, etc. If the metallic film is multi-elemental, the same procedure of data analysis can be generalized for all elements which form the film.

For high exactitude in the ionic solution of mercury we get high exactitude for thickness

Acknowledgements

We are most grateful to J. Gomez for his continuous interest and cooperation on this work. This work was carried out in part trough Diuc Project No. 206.011.043-1.

References (20)

  • R.A. Barrea et al.

    Nucl. Instr. and Meth. B

    (1998)
  • E.D. Greaves et al.

    Spectrochim. Acta B

    (1997)
  • L. Bennun et al.

    Spectrochim. Acta B

    (1999)
  • P. Wobrauschek et al.

    Spectrochim. Acta B

    (1989)
  • L. Bennun et al.

    Spectrochim. Acta B

    (1997)
  • P.D. Swift et al.

    Thin Solid Films

    (2003)
  • S. Martin

    Closing remarks

    Faraday Discuss.

    (1997)
  • T.W. Barbee

    Opt. Eng.

    (1987)
  • V. Sammelseg et al.

    J. Anal. Atom. Spectrom.

    (1999)
  • J.C. Martinez-Antón

    Meas. Sci. Technol.

    (2000)
There are more references available in the full text version of this article.

Cited by (5)

  • Robust and fast absolute quantification of a colored wood surface coating by scanning densitometry

    2020, Applied Surface Science
    Citation Excerpt :

    Infrared spectroscopy is an excellent method to confirm the presence of the coating if no quantitative data is needed, while X-ray fluorescence (XRF) gives accurate data on the coating composition, thickness and homogeneity without destroying the samples. However, XRF is very sensitive to the background and limited to the analysis of coatings containing transition metals or elements with an atomic number of Z ≥ 14 [2], and is therefore used mostly for thin inorganic coatings, such as silicon, nickel, silver or gold or metal-based paints [1–4]. Measuring the difference of color by spectrophotometry (e.g. with the CIELAB color system) is also a technique widely used in industry to record a change of color.

  • Atomic spectrometry update-X-ray fluorescence spectrometry

    2010, Journal of Analytical Atomic Spectrometry
View full text