Other elements like V or W also show absorption in the observed
spectral region, but their concentration in the digested samples is too
low to achieve sufficiently high absorption values to be used as
internal spectroscopic standards.
Fig. 5 shows the effect of internal spectral standards on the same set
of calibration data: Monochromator drift correction enhances the
linearity of the calibration function as well as the precision of datasets.
The additional correction for monochromator resolution changes has
no significant effect on the linearity but enhances slightly the precision
of each data point for total boron concentrations N= 15 mmol L−1
. For
total boron concentrations less than 15 mmol L−1 the correction of
monochromator drift and resolution changes improves significantly
the calibration function as shown in Fig. 6: The variation of procedure
decreases due to two reasons: An increasing total boron concentration
leads to an improved Gauss fit and the use of internal spectroscopic
standards improves monochromator stability.
In order to show this, three calibration data sets (10 mmol L−1
,
15 mmol L−1 and 20 mmol L−1 B) were used and reprocessed for
correction of monochromator drift only and for correction of
monochromator drift and resolution changes, as shown in Fig. 6.
The coefficient of variation of procedure was calculated by
normalizing the residual standard deviation of the calibration on the
slope of the calibration function [18]. This results in the expression of
uncertainty in atom-%.
3.6. Reference materials
The effect of the use of internal spectroscopic standards on 3
reference steel samples and NIST SRM 952 is shown in Table 4 using a
linear calibration model. For all samples satisfactory agreement
with the certified value was attained by using drift and resolution
corrections.
For steel reference material S20 larger deviations from the certified
value were observed, when only the correction for monochromator
drift was applied or no internal spectroscopic standards were used.
We attribute this to resolution changes (15 mmol L−1
) and to
monochromator drift (20 mmol L−1). 4. Conclusion
In the present study we demonstrate that HR-CS-FAAS is suitable to
distinguish different enrichment levels of boron in steel samples.
Although this method cannot rival with mass spectrometric approaches
in terms of accuracy and precision, it is cheap, simple and
does not require the precise determination of the total boron content of the sample. Therefore this procedure is well suited for the identification
of 10B enriched scrap because less rigorous requirements have to
be met (±5% of the absolute 10B isotope content) for this purpose.
By introducing the concept of internal spectroscopic standards
small changes in the optical resolution of the monochromator as well
as small wavelength drifts could be compensated for.
The analyzed steel reference materials confirm the applicability of
this method.