EDS Analysis of Light Elements with Silicon Drift Detectors: Quantification of Boron

International Union of Microbeam Analysis Societies V (IUMAS), Seoul - Korea 2011

J. Berlin, R. Terborg, T. Salge, M. Falke & S. SchellerBruker Nano GmbH, Schwarzschildstr. 12, 12489 Berlin, Germany

IntroductionThe analysis of light elements (Be, B, C, N, O and F) represents a challenge for energy dispersive X-ray spectrometry (EDS). Some of the problems are technical in nature, relating to instrument design and measurement procedure, while others are due to inherent physical effects (e.g., low fluorescence yields). With recent advances in EDS techno-logy, dramatic progress in the detection of light elements has been made. Ultra-thin polymer-Al windows have made the detection of B-K (183 eV) and even Be-K (108 eV) possible (Fig. 1).

Conclusions

EDS

Fig. 1. Efficiency curves for an SDD with a super light element window (SLEW) compared to an SDD with conventional Be window.

Figure 3 shows spectra of B4C (purple), LaB6 (green), BN (blue), Ni3B (red), a pure boron (black) and a pure carbon standard (gray) at 5 kV accelerating voltage.

Furthermore, with the introduc-tion of silicon drift detectors (SDD) and their high count rate capabilities, a rapid evolution in pulse processing was spurred, which in turn led to a dramatic improvement of the peak resolution – also for light elements (Fig. 2).

Fig. 2. Energy resolution for different elements (e.g., C-K 46 eV, Mn-K

123 eV) measured with a Bruker XFlash® 5010 SDD on a standard sample (EDS-TM001), specifically designed by the German Federal Institute for Materials Research and Testing (BAM) for testing the resolution of EDS detectors [1].

Results

Fig. 5. Spectra of Ni3B obtained at 2, 5, 10 and 15 kV accelerating voltage. At 10 and 15 kV, the boron peak is barely visible. To ensure a sufficient peak to back-ground ratio for the quantification of minor boron concentrations, an accelerating voltage 5 kV needs to be used. Sample: courtesy of U. Burkhardt, MPI Dresden.

BC O

Ni (L)

Ni (K )

Ni (K )

2 kV5 kV

10 kV15 kV

B

C

Ni (L)

O

Here, we show specific examples for the analysis of boron and discuss remaining challenges for light element analysis with EDS.

B C

N

O

La (L)

Ni (L) BoronCarbonB4CLaB6BNNi3B

Fig. 3. Spectra of B4C (purple), LaB6 (green), BN (blue) and Ni3B (red) compared to spectra of a pure boron (black) and a pure carbon standard (gray). Analytical conditions: 5 kV accelerating voltage, 0.4 nA beam current, 3.5-7 kcps input count rate,XFlash® 5030 silicon drift detector, 60 s measurement time.

Table 1. Quantification results (standard-based, PhiRhoZ) for Ni3B. Average of 5 measurements for each condition.

74.625.494.15.9Ni3B (2 kV)

75.025.094.25.8given value60.1*39.989.110.9Ni3B (15 kV) 71.528.593.26.8Ni3B (10 kV) 75.224.894.25.7Ni3B (5 kV)

Ni (atom%)B (atom%)Ni (wt%)B (wt%)

*quantified using Ni-K�

Advances in detector techno-logy have triggered dramatic progress in the analysis and quantification of light elements with EDS!

0,00

0,01

0,02

0,03

0,04

0,05

0,06

0,07

0,08

0,09

0,10

0 5 10 15 20 25

Accelerating voltage (kV)

k-ra

tio B

-K�

in N

i B

predictedmeasured

3

Fig. 6. K-ratios for B-K in Ni3B as a function of accelerating voltage. The solid curve indicates the prediction of the PhiRhoZ correction calculation. Dots represent our measurements.

When comparing the Ni3B spectrum with the spectrum of the carbon standard (which does not contain any boron), the boron peak of Ni3B (with 5.8 wt% B) is small but clearly visible. It is also possible to map such minor boron concentrations. Figure 4 shows an element map of boron for an area within a weld seam (base metal was a Cr-Ni stainless steel). The sample contains phases with 3 different boron contents as indicated by the color scale: red = highest B-content (Cr-boride), green = intermediate B-content (Cr,Ni-boride) and blue = lowest B-content (Ni3B). Figure 5 shows spectra of a Ni3B (reference) sample at 2, 5, 10 and 15 kV accelerating voltage. Corresponding quantification results (average of 5 measurements for each condition) are given in Table 1. References:

[1] http://www.rm-certificates.bam.de/de/rm-certificates_ media/rm_cert_layer_and_surface/eds_tm001de.pdf

[2] Bastin G. F. and Heijligers H. J. M. (1997) Quantitative electron probe microanalysis of boron in binary borides (2nd rev. Ed.). Technische Universiteit Eindhoven, ISBN 90-3860-898-5.

Quantification was carried out standard-based (with pure element standards for B and Ni) using the XPP-optimized PhiRhoZ model. Figure 6 shows the quantification results in terms of intensity ratios as a function of accelerating voltage.

Fig. 4. Element map of boron (obtained in 30 min at 5 kV; 150 kcps input count rate) for an area within a weld seam and corresponding SE image. The sample contains phases with 3 different B-contents as indicated by the color scale: red = highest (Cr-boride), green = intermediate (Cr-Ni-boride) and blue = lowest (Ni3B).

Our results underline the importance of using low accele-rating voltages ( 5 kV) to ensure a sufficient peak to background ratio for the quantification of light elements. This, however, complicates the analysis, because K lines of many other elements, which may be present in the sample, are not excited when optimum overvoltages for the light elements are attained. Uncer-tainties in mass absorption coefficients [2] and peak overlaps with L, M and N lines of heavier elements represent further challenges that need to be considered.

Discussion