The spatial resolution of x-ray microanalysis in thin foils

Author(s):  
D. B. Williams ◽  
J. R. Michael ◽  
J. I. Goldstein ◽  
A. D. Romig

The spatial resolution of x-ray microanalysis in a thin foil is determined by the size of the beam-specimen interaction volume. This volume is a combination of the incident electron beam diameter (d) and the beam broadening (b) due to elastic scatter within the specimen. Definitions of spatial resolution have already been proposed on this basis but all present a worst case value for the resolution based on the dimensions of the beam emerging from the exit face of the foil.

Author(s):  
J. R. Michael ◽  
A. D. Romig

There have been many experimental efforts to measure the spatial resolution for x-ray microanalysis in the analytical electron microscope (AEM). There have been three commonly utilized specimen geometries in these experiments: 1) segregant at a grain boundary, 2) interphase boundaries oriented parallel to the electron beam, and most recently 3) spherical particles embedded at various depths in thin foils. The results of many of these experiments have been analyzed with a number of models for the broadening of the electron beam as it traverses the thin foil. These models are typically based on incoherent electron scattering, typical of Monte Carlo simulations. A vast majority of the published spatial resolution data support the incoherent scattering models as the best simulation of spatial resolution for x-ray microanalysis in the AEM. Recent experimental work using embedded particles to measure beam broadening has been used to support the coherent scattering model of beam broadening.


Author(s):  
D. Imeson ◽  
J. B. Vander Sande

It is well established that when an electron beam is incident upon a thin foil in the form of a focused probe, as in STEM, multiple scattering events in the sample cause considerable lateral spreading of the electron beam. The volume of material excited by the electron beam is therefore much greater than that volume defined by simple projection of the incident beam through the sample. In the application of the techniques of quantitative X-ray analysis in STEM to regions of composition variation with small spatial extent this point becomes of crucial importance, being the main determinant of the ability to map such composition changes, or even to detect them. It is the view of the authors that there exists some confusion over the nature of the beam spreading phenomenon in thin foils of crystalline material and the concept of spatial resolution of composition determination by characteristic X-ray emission. We intend here to clarify these concepts by discussing the meaning and use of the term “beam broadening” in analytical transmission electron microscopy.


Author(s):  
E. A. Kenik ◽  
J. Bentley

Cliff and Lorimer (1) have proposed a simple approach to thin foil x-ray analy sis based on the ratio of x-ray peak intensities. However, there are several experimental pitfalls which must be recognized in obtaining the desired x-ray intensities. Undesirable x-ray induced fluorescence of the specimen can result from various mechanisms and leads to x-ray intensities not characteristic of electron excitation and further results in incorrect intensity ratios.In measuring the x-ray intensity ratio for NiAl as a function of foil thickness, Zaluzec and Fraser (2) found the ratio was not constant for thicknesses where absorption could be neglected. They demonstrated that this effect originated from x-ray induced fluorescence by blocking the beam with lead foil. The primary x-rays arise in the illumination system and result in varying intensity ratios and a finite x-ray spectrum even when the specimen is not intercepting the electron beam, an ‘in-hole’ spectrum. We have developed a second technique for detecting x-ray induced fluorescence based on the magnitude of the ‘in-hole’ spectrum with different filament emission currents and condenser apertures.


Author(s):  
R. Hutchings ◽  
I.P. Jones ◽  
M.H. Loretto ◽  
R.E. Smallman

There is increasing interest in X-ray microanalysis of thin specimens and the present paper attempts to define some of the factors which govern the spatial resolution of this type of microanalysis. One of these factors is the spreading of the electron probe as it is transmitted through the specimen. There will always be some beam-spreading with small electron probes, because of the inevitable beam divergence associated with small, high current probes; a lower limit to the spatial resolution is thus 2αst where 2αs is the beam divergence and t the specimen thickness.In addition there will of course be beam spreading caused by elastic and inelastic interaction between the electron beam and the specimen. The angle through which electrons are scattered by the various scattering processes can vary from zero to 180° and it is clearly a very complex calculation to determine the effective size of the beam as it propagates through the specimen.


2000 ◽  
Vol 6 (4) ◽  
pp. 307-316 ◽  
Author(s):  
E.D. Boyes

AbstractThe current status and general applicability of scanning electron microscopy (SEM) at low voltages is reviewed for both imaging (low voltage scanning electron microscopy, LVSEM) and chemical microanalysis (low voltage energy-dispersive X-ray spectrometry, LVEDX). With improved instrument performance low beam energies continue to have the expected advantages for the secondary electron imaging of low atomic number (Z) and electrically non-conducting samples. They also provide general improvements in the veracity of surface topographic analysis with conducting samples of all Z and at both low and high magnifications. In new experiments the backscattered electron (BSE) signal retains monotonic Z dependence to low voltages (<1 kV). This is contrary to long standing results in the prior literature and opens up fast chemical mapping with low dose and very high (nm-scale) spatial resolution. Similarly, energy-dispersive X-ray chemical microanalysis of bulk samples is extended to submicron, and in some cases to <0.1 μm, spatial resolution in three dimensions at voltages <5 kV. In favorable cases, such as the analysis of carbon overlayers at 1.5 kV, the thickness sensitivity for surface layers is extended to <2 nm, but the integrity of the sample surface is then of concern. At low beam energies (E0) the penetration range into the sample, and hence the X-ray escape path length out of it, is systematically restricted (R = F(E05/3)), with advantages for the accuracy or elimination of complex analysis-by-analysis matrix corrections for absorption (A) and fluorescence (F). The Z terms become more sensitive to E0 but they require only one-time calibrations for each element. The new approach is to make the physics of the beam–specimen interactions the primary factor and to design enabling instrumentation accordingly.


2000 ◽  
Vol 6 (4) ◽  
pp. 307-316
Author(s):  
E.D. Boyes

Abstract The current status and general applicability of scanning electron microscopy (SEM) at low voltages is reviewed for both imaging (low voltage scanning electron microscopy, LVSEM) and chemical microanalysis (low voltage energy-dispersive X-ray spectrometry, LVEDX). With improved instrument performance low beam energies continue to have the expected advantages for the secondary electron imaging of low atomic number (Z) and electrically non-conducting samples. They also provide general improvements in the veracity of surface topographic analysis with conducting samples of all Z and at both low and high magnifications. In new experiments the backscattered electron (BSE) signal retains monotonic Z dependence to low voltages (<1 kV). This is contrary to long standing results in the prior literature and opens up fast chemical mapping with low dose and very high (nm-scale) spatial resolution. Similarly, energy-dispersive X-ray chemical microanalysis of bulk samples is extended to submicron, and in some cases to <0.1 μm, spatial resolution in three dimensions at voltages <5 kV. In favorable cases, such as the analysis of carbon overlayers at 1.5 kV, the thickness sensitivity for surface layers is extended to <2 nm, but the integrity of the sample surface is then of concern. At low beam energies (E0) the penetration range into the sample, and hence the X-ray escape path length out of it, is systematically restricted (R = F(E05/3)), with advantages for the accuracy or elimination of complex analysis-by-analysis matrix corrections for absorption (A) and fluorescence (F). The Z terms become more sensitive to E0 but they require only one-time calibrations for each element. The new approach is to make the physics of the beam–specimen interactions the primary factor and to design enabling instrumentation accordingly.


2012 ◽  
Vol 108 (7) ◽  
Author(s):  
Michael Schnell ◽  
Alexander Sävert ◽  
Björn Landgraf ◽  
Maria Reuter ◽  
Maria Nicolai ◽  
...  

2000 ◽  
Vol 6 (S2) ◽  
pp. 112-113
Author(s):  
D. B. Williams ◽  
M. Watanabe

Commercial TEM/STEMs are ill designed for quantitative X-ray microanalysis of thin foils. There have been no fundamental advances in their design since the first instruments appeared in the mid- 1970s. These instruments had thermionic sources with useful probe sizes of ∼10 nm, small (∼ 0.1 sr) detector collection angles, illumination systems giving serious levels of stray radiation, substantial hydrocarbon and water-vapor contamination due to the poor vacuum (∼10-5Pa) and sliding o-ring seals on unstable side-entry goniometer stages. The instruments lacked a direct measure of the probe current at the specimen. Often a 60 s accumulation time resulted in enough drift or contamination that the microanalysis was not trustworthy and spatial resolution was compromised. Today's modern TEM/STEMs, apart from replacement of thermionic source with a FEG, are little better.Dedicated STEMs however, have invariably offered better X-ray performance. The first VG HB 5 DSTEM became available, also in the 1970s, and the X-ray performance improved through to the latest design (the VG HB 603) in the mid-1990s.


Author(s):  
J.M. Brown ◽  
H.L. Fraser

Quantitative X-ray microanalysis of thin foils may be achieved either by calculation or by making use of standards. This paper, although advocating the use of pure elemental standards for accurate quantification, points out some of the problems involved in the choice of good standards. Specifically, the presence of surface films on as-prepared thin foil standards must be removed or accounted for during data analysis.


Sign in / Sign up

Export Citation Format

Share Document