PERFORMANCE OF TXRF and GI-XRF ANALYSES - Total-Reflection X-Ray Fluorescence Analysis and Related Methods - page 266

Figure4.26. An ion source of the Kaufman type, suitable for planar sputter-etching of extended
areas of a sample. (a) Sectional view; (b) the complete device with load-lock, sample chamber, ion-
gun, ion-pump, and high-vacuum valve. Figure from Ref. [77], reproduced with permission from
Springer-Verlag. Copyright2005, Springer-Verlag.
microbalance (UMT2, Mettler Toledo, Greifensee, Switzerland). Special care
has to be taken: (1) the microbalance should be located in a calm, air-
conditioned room free from vibrations; (2) weighing should not start before
a waiting period of 5 min; and (3) weighing has to be repeated three times and
averaged. The mass of about 400 mg can be determined with a precision of
0.1 μ g so that the mass difference of only 2-8 μ g for the sputtered sublayer can
be given with an RSD of only 5-10%. The total loop is repeated sublayer by
sublayer, and sputtering is reiterated until the implanted layer is completely
removed from the surface.
The thickness of the total layer is determined afterward by an interfer-
ometer. For this purpose, a small droplet of a correction fluid (0.1 μ l) is
dripped onto the wafer piece and dried before the repetitions start. A tiny
circular spot of < 1mm 2 of the sample is masked by a thin film (about 10 μ m),
which remains unaffected during sputtering. Such a tiny spot influences the
results of repeated ion sputtering and TXRF analyses by < 0.3%, conse-
quently it can be neglected. After the layer is completely etched and the
iteration process is finished, the correction fluid is removed by a solution of
acetone. A truncated cone with a flat plateau is left, which is observed now,
for example, by a simple Tolansky device attached to the mentioned
microscope. For this purpose, the sample is coated first by a thin layer of
gold with 1-2 nm thickness and illuminated then with monochromatic light,
for example, of Na-D at 589 nm wavelength. Parallel fringes appear with a
shift at the border of the mentioned cone. The total sputter depth can be
determined in the range of 100-200 nm with an RSD < 2%. A mean thick-
ness of a single sublayer follows by division with the number of repeated
steps or sublayers.