Elsevier

Ultramicroscopy

Volume 157, October 2015, Pages 21-26
Ultramicroscopy

Energy dispersive X-ray analysis on an absolute scale in scanning transmission electron microscopy

https://doi.org/10.1016/j.ultramic.2015.05.010Get rights and content

Highlights

  • Scan-averaged energy dispersive X-ray (EDX) spectra are recorded in STEM.

  • Experimental and simulated EDX signals are shown to agree on an absolute scale.

  • Two independent methods are used for accurate thickness determination.

  • Probe channelling must be taken into account, even for very thin specimens.

  • Channelling effects are minimized for large probe-forming convergence angles.

Abstract

We demonstrate absolute scale agreement between the number of X-ray counts in energy dispersive X-ray spectroscopy using an atomic-scale coherent electron probe and first-principles simulations. Scan-averaged spectra were collected across a range of thicknesses with precisely determined and controlled microscope parameters. Ionization cross-sections were calculated using the quantum excitation of phonons model, incorporating dynamical (multiple) electron scattering, which is seen to be important even for very thin specimens.

Introduction

Analytical electron microscopy has long been able to determine elemental concentration ratios with a sensitivity of a few atomic percent in microanalysis via energy dispersive X-ray spectroscopy (EDX). This is done via EDX signal ratios and comparison with reference specimens to minimize the effect of uncertainties in factors such as thickness, ionization cross-section, fluorescence yield and detector geometry [1], [2]. However, in principle, absolute-scale comparison with ionization cross-section calculations is possible. Recent developments in aberration-corrected electron optics and X-ray detector design [3], [4] have facilitated scanning transmission electron microscopy (STEM) EDX mapping at atomic resolution [5], [6], [7], [8], [9], [10], [11], [12], and on this length scale absolute elemental concentrations will often be more useful than relative concentrations. For instance, in nanoparticles, nanoprecipitates, or dopant segregation at interfaces, the absolute number of atoms of a given species contains critical information about their structural arrangement and distribution.

In this paper, we show that absolute-scale agreement between experiment and simulations, which has recently been achieved in high-angle annular dark field (HAADF) imaging [13], [14], [15] and electron energy loss spectroscopy [16], [17], is also possible in STEM EDX. We emphasize that, in on-axis conditions, quantitative agreement requires accounting for dynamical electron scattering, also called “channelling”, even for very thin specimens.

Section snippets

Absolute-scale EDX

EDX measurements were taken using an aberration-corrected FEI Titan3 electron microscope operating at 302 kV. Experiments were carried out for a range of thicknesses in a SrTiO3 crystal viewed along the [001] zone axis. The sample was prepared by cross-sectional tripod/wedge polishing, with final thinning by Ar-ion milling [18]. Two different probe-forming aperture semi-angles were used, 15.2 and 21.5 mrad, both of which produce an atomically fine probe. However, the modest effective collection

Thickness determination

For absolute-scale comparison of the experimental STEM EDX data against simulation, independent measurement of sample thickness is necessary. Our primary method of thickness determination is position averaged convergent beam electron diffraction (PACBED). Some authors successfully apply PACBED with the same convergence angle used for imaging [37], but we obtain better thickness sensitivity for smaller probe-forming aperture angles [19], [38]. Therefore, PACBED patterns were recorded with a 9.2 

Effects of channelling

In Fig. 1 the effect of channelling, not hitherto included in quantitative STEM EDX measurements [2], is clearly large, even though the position-averaged signal approach is believed to reduce the severity of channelling effects [20], [41]. From Eq. (2), neglecting X-ray absorption for simplicity, the probe-position averaged signal is given byI¯AcI(R)dR=AP¯(r,t)Veff(r)dr,whereP¯(r,t)=Ac[0t|ψ(R,r,z)|2dz]dRgives the probability density of the fast electron projected across the specimen

Conclusion

We have demonstrated agreement between experiment and simulations incorporating dynamical electron scattering on an absolute-scale in STEM EDX imaging without any free adjustable parameters. Absolute chemical specific atom counting by STEM-EDX analysis with atomic resolution is now clearly in prospect.

Acknowledgements

The authors gratefully acknowledge Dr. Y. Zhu for the sample and Drs. N.R. Lugg and J.M. LeBeau for helpful discussions. We are especially grateful to Dr. N.J. Zaluzec (Argonne National Laboratory) and Alan Sandborg (EDAX Inc.) for assistance with the geometry of our EDX detector. We thank FEI for the measurements of the Be double tilt holder. This research was supported under the Australian Research Councils Discovery Projects funding scheme (Projects no. DP110102228 and DP140102538), its

References (47)

  • J.I. Goldstein et al.

    Principles of Analytical Electron Microscopy

    (1986)
  • M. Watanabe et al.

    The quantitative analysis of thin specimensa review of progress from the Cliff–Lorimer to the new ζ-factor methods

    J. Microsc.

    (2006)
  • T. Hashimoto et al.

    Development of two steradian EDX system for the HD-2700 FE-STEM equipped with dual X-MaxN 100 TLE large area windowless SDDs

    Microsc. Microanal.

    (2014)
  • P.J. Phillips et al.

    A new silicon drift detector for high spatial resolution STEM-XEDS: performance and applications

    Microsc. Microanal.

    (2014)
  • A.J. D'Alfonso et al.

    Atomic-resolution chemical mapping using energy-dispersive X-ray spectroscopy

    Phys. Rev. B

    (2010)
  • M. Chu et al.

    Emergent chemical mapping at atomic-column resolution by energy-dispersive X-ray spectroscopy in an aberration-corrected electron microscope

    Phys. Rev. Lett.

    (2010)
  • D.O. Klenov et al.

    Structure of the InAlAs/InP interface by atomically resolved energy dispersive spectroscopy

    Appl. Phys. Lett.

    (2011)
  • L.J. Allen et al.

    Chemical mapping at atomic resolution using energy-dispersive X-ray spectroscopy

    MRS Bull.

    (2012)
  • B.D. Forbes et al.

    Contribution of thermally scattered electrons to atomic resolution elemental maps

    Phys. Rev. B

    (2012)
  • P.G. Kotula et al.

    Challenges to quantitative multivariate statistic analysis of atomic-resolution X-ray spectral

    Microsc. Microanal.

    (2012)
  • P. Lu et al.

    Atomic-scale chemical quantification of oxide interfaces using energy-dispersive X-ray spectroscopy

    Appl. Phys. Lett.

    (2013)
  • G. Kothleitner et al.

    Quantitative elemental mapping at atomic resolution using X-ray spectroscopy

    Phys. Rev. Lett.

    (2014)
  • J.M. LeBeau et al.

    Quantitative atomic resolution scanning transmission electron microscopy

    Phys. Rev. Lett.

    (2008)
  • Cited by (36)

    • A deep convolutional neural network to analyze position averaged convergent beam electron diffraction patterns

      2018, Ultramicroscopy
      Citation Excerpt :

      To reduce human error and enhance the repeatability of the measurements, a semi-automated approach is usually employed. To this end, least square fitting (LSF) has been the primary tool [3,4,8–10]. The parameter of interest, e.g. thickness, tilt, polarity, etc., is found by searching for the best fit amongst a library of patterns.

    • A menu of electron probes for optimising information from scanning transmission electron microscopy

      2018, Ultramicroscopy
      Citation Excerpt :

      This may be advantageous for specimens where the structure at a particular depth or location is to be targeted. On the other hand, if used in combination with a through-focal series, a more even signal contribution with depth can be obtained, with potential applications in EDX [45] and achieving high resolution 3D ADF-STEM tomography of specimens with structures that are inhomogeneous with respect to depth [18]. For vortex probes using smaller convergence angles, the intensity minima at the centre of the probe is sufficiently large to open the possibility of ‘reverse’ EELS or EDX where there can be confidence that the column beneath the incident probe is making a small contribution to the final ADF, EDX or EELS signal.

    View all citing articles on Scopus
    1

    These authors contributed equally to this work.

    View full text