EFFICIENCY AND EVALUATION - Total-Reflection X-Ray Fluorescence Analysis and Related Methods - page 384

6.1ANALYTICALCONSIDERATIONS
The efficiency of TXRF is mainly a result of the fact that it is an energy-
dispersive method of X-ray fluorescence analysis. Specimens need be depos-
ited on totally reflecting carriers in only small amounts. This capability for
microanalyses is the second feature that determines the efficiency of TXRF.
The benefits of this efficiency affect cost, detection power, reliability, and
applicability.
6.1.1GeneralCostsofInstallationandUpkeep
The simplicity of the TXRF instrumentation described in Chapter 3 ought to be
reemphasized. Complete instruments as well as individual supplements or
components are commercially available. At present five different manufactur-
ers offer a certain selection of devices. The basic purchase costs amount to
about $180 000 for a complete instrument including a high-voltage (HV)
generator, X-ray tubes, reflectors, sample changer, detector, multichannel
analyzer, computer, and software. 1 An instrument additionally offering an
angle-scan of the sample and therefore suited for surface and thin-layer
analyses (GI-XRF) may cost approximately $20 000 more. The complete
instruments are already adjusted, mechanically stable, and nearly mainte-
nance-free. Operational problems are very infrequent, but sometimes a leak-
age of the thin detector window does occur, which then has to be replaced by
the manufacturer.
For installation, a space of only about 4 m 2 is required and the room should
be air-conditioned. A power supply of 3 kW is needed, and a coolant system
with a flow rate of 5 l water/min must be connected. Low-power X-ray tubes do
not need water cooling because they are air-cooled but they are less effective.
The older detectors, such as Si(Li) and Ge(Li) detectors, need to be cooled with
liquid nitrogen, and an accessory Dewar of 10 l volume has to be filled at least
once but not more than twice a week. However, the modern silicon drift
detectors (SDDs) run without liquid nitrogen because they are thermoelectri-
cally cooled (Peltier cooling). In general, vacuum will not be needed if one does
not intend to detect light elements with low atomic numbers ( Z < 11). For light-
element detection a fine vacuum (1 hPa) is sufficient; furthermore, a detector
with a special diamond-like window or a thin polypropylene foil or even a
windowless detector will have to be utilized, the latter being more susceptible to
trouble.
Sample preparation before analysis takes the bulk of the time and effort,
especially when samples cannot be analyzed directly but have to be digested
prior to analysis. However, a lot of preparation techniques are well-tried
and tested for approved methods like electrothermal atomic absorption
spectrometry (ET-AAS) or inductively coupled plasma optical emission
1 All costs are given in US currency here.