and corrects for matrix discrepancies by normal-
rithms based on intensities measured for pure ele-
izing to the Compton Kα (Comp. Kα) incoherent
ments, sensitivity coefficients, summations of ab-
radiation backscatter peak.*
sorption-enhancement terms in appropriate ma-
This study evaluates these two methods of rapid
trices, and correction factors for overlapping peaks
sample analysis by determining the concentrations
(Figura 1987). The X-Met 920 FP program also re-
of copper (Cu), zinc (Zn), arsenic (As), lead (Pb),
quired the analysis of a fully characterized stan-
chromium (Cr), cobalt (Co), nickel (Ni), mercury
dard material, which for this study was SRM 2710,
(Hg), thallium (Tl), selenium (Se), silver (Ag), an-
from the National Institute of Science and Tech-
timony (Sb), cadmium (Cd), tin (Sn), and barium
nology (NIST).
(Ba) in a variety of solid-particle matrices. All
To perform the RF/Comp. Kα normalization
analyses are performed with transportable high-
analysis, a standard must be available that con-
resolution XRF systems that can be configured for
tains the analytes of interest in a matrix that is
stand-alone operations. This combination of rapid
physically consistent (dry particles) with the
sample analysis and instrumental transportabil-
samples. When possible, the RF should be estab-
ity is well suited for screening purposes, thus a
lished for well resolved Kα, Kβ, Lα, or Lβ spectral
data quality objective of an accuracy of 50% and
lines (Table 1). Figure 1 is an example of an XRF
detection limits of less than 1000 g/g was used
spectrum obtained with the X-Met 920 showing
(Raab et al. 1987).
several characteristic peaks for metals along with
the incoherent (Compton) and coherent (Rayleigh)
sample matrix backscatter.
For this study a finely ground soil, SRM 2710
certified reference soil from NIST, was used to es-
Explanations of the principals of XRF analysis
tablish the analyte RFs for Cu, Zn, As, and Pb. Table
can be found elsewhere (Driscoll et al. 1991, Hewitt
2 is an example of some of the daily RFs estab-
1994a, b). The instruments used in this study were
lished for these four metals. For the determination
of Cr, the NIST SRM 2711 was spiked with 4000 g
the X-Met 920 (Metorex, Inc.; formerly Outokumpu
Electronics) and the Spectrace 9000 (Spectrace In-
struments) X-ray spectrophotometers. These two
Table 1. Primary sources and analyte lines for metals
field-portable systems are equipped with surface
of environmental concern that can be detected by X-
analysis probes allowing for either in situ or intru-
ray fluorescence spectrometry.
sive sample analysis and have software-supported
Emission lines (keV)
FP analysis capabilities. The Spectrace 9000 is
equipped with three primary radioactive sources,
Source
Metals
Kα
Kβ
Lα
Lβ
Fe-55, Cd-109, and Am-241, and has a HgI2 solid-
FE-55
Cr†
5.41
5.95
state detector. The X-Met 920 has two radioactive
sources, Cd-109 and Am-241, and has a Si(Li) de-
Cd-109
Cr†*
5.41
5.95
Cd-109
Mn†
5.89
6.49
tector that requires liquid N2 for operation. Table
Cd-109
Fe†*
6.40
7.06
1 lists some of the metals that can be determined
Cd-109
Co†
6.92
7.65
by these XRF systems. The table also shows those
Cd-109
Ni†
7.47
8.30
metals that could be quantitated by FP software
Cd-109
Cu†*
8.04
8.94
programs supplied by the respective manufactur-
Cd-109
Zn†*
8.63
9.61
As†*
10.5
11.8
ers.
Cd-109
Se†
11.2
12.6
Hg†
9.98
11.9
Cd-109
Tl
10.3
12.3
Pb†*
10.5
12.6
Am-241
Ag†
22.1
25.2
The FP software routines used in this study are
Am-241
Cd†
23.1
26.4
proprietary to the instrument manufacturer. In
Am-241
Sn†
25.2
28.8
general, these FP programs are a series of algo-
Am-241
Sb†
26.2
30.1
Am-241
Ba†
32.0
36.8
† Spectrace 9000 preprogrammed for fundamental parameter
* T.M. Spittler, U.S. Environmental Protection Agency,
analysis of this metal.
Environmental Services Division, Region 1, Lexington,
* X-Met 920 preprogrammed for fundamental parameter an-
alysis of this metal.
Massachusetts.
2
Previous Page
