trace element analysis using edxrf with polarized optics

Copyright ©JCPDS-International Centre for Diffraction Data 2011 ISSN 1097-0002
TRACE ELEMENT ANALYSIS USING EDXRF
WITH POLARIZED OPTICS
Takao Moriyama1, Satoshi Ikeda1, Makoto Doi1 and Scott Fess2
1
Rigaku Corporation, Takatsuki, Osaka 569-1146, Japan
Applied Rigaku Technologies, Inc., Austin, TX 78717
2
ABSTRACT
We have developed an EDXRF spectrometer with polarized optics and new quantification
software which estimates non-measuring sample matrices using scattering intensities and full
profile fitting method combined with FP method. Accurate analysis down to ppm level can be
achieved even in complex sample composition with the quantification software. The
scattering FP method corrects for non-measuring components in samples such as coal fly ash,
soils and biological samples by using Compton and Thomson scattering intensities from a Mo
secondary target. Additionally, it is possible to analyze thickness and composition of
multi-layer thin films considering interlayer secondary excitation.
INTRODUCTION
Energy dispersive types of x-ray spectrometers are useful analytical tools for screening
analysis, such as for environmental applications, due to compact and easy sample handling.
However, the preparation of standard samples that match analyzing samples in applications
such as for industrial waste and recycling raw materials often requires special attention to
match sample types and preparation. This is because their sample matrices are complicated
and improved sensitivities for trace elements are demanded for these environmental
applications.
We have developed an EDXRF spectrometer with polarized optics and software using a
fundamental parameter method for accurate quantification combined with full profile fitting
method, including estimating non-measuring sample matrices using scattering intensities.
INSTRUMENT
The specifications of the new Rigaku EDXRF spectrometer NEX CG used in this study are as
follows:
X-ray tube : Pd target, air cooled
Tube power : 50W(50kV-2mA)
Secondary targets : 5 targets (max)
Detector : High performance SDD
(Silicon Drift Detector)
288
289
This document was presented at the Denver X-ray
Conference (DXC) on Applications of X-ray Analysis.
Sponsored by the International Centre for Diffraction Data (ICDD).
This document is provided by ICDD in cooperation with
the authors and presenters of the DXC for the express
purpose of educating the scientific community.
All copyrights for the document are retained by ICDD.
Usage is restricted for the purposes of education and
scientific research.
DXC Website
– www.dxcicdd.com
ICDD Website
- www.icdd.com
Copyright ©JCPDS-International Centre for Diffraction Data 2011 ISSN 1097-0002
289
290
Atmosphere : Vacuum, Air, He
Measuring area : 20mm diameter
METHODS
(1) Secondary targets with polarized optics optimized for low ppm applications
Secondary targets and polarized optics provide high P/B ratio spectra compared to direct
excitation optics in a wide energy range.
Blue and red lines in Fig. 1 show measurements of a multi-element oil sample by the
polarized optics and direct excitation optics, respectively. The polarized optics gives greatly
improved performance compared to the direct excitation optics for trace element analysis
1.E+05
Direct excitation optics
X-ray tube
Detector
Secondary
Targets
Intensity (Arb.Unit)
Sample
1.E+04
EDXL300
Direct
1.E+03
Polarized optics
1.E+02
4
6
8
10 12
keV
14
16
18
Fig. 1. Polarized optics and measurement comparison to direct excitation optics
of an oil sample
shows an example of a fitted and measured
profile using the standard soil sample
JSAC0466. The fitted spectrum consists of
As-K
Pb-L
10000
Measure
Total profile
As profile
Se profile
Pb profile
Hg profile
Se-K
Pb-L
8000
6000
4000
As-K
accurate quantification combined with full
profile fitting method. This method is useful
for samples with complex matrices. Fig.2
X-ray
intensity
(a.u.)
X-ray intensity
( a.u.)
(2) RPF-SQX (Rigaku Profile Fitting - Spectra Quant X) analysis
RPF-SQX is a standardless analysis
(Kataoka, 1989) software with a
fundamental parameters method for
2000
0
10
10
11
11 EnergykeV
( keV)
12
12
13
13
Fig. 2. Measured and fitted spectra of the soil
the sum of individually generated profiles
standard JSAC0466S
using the response function (Campbell and
Wang, 1992) for each element obtained by the FP method. Quantification results are obtained
by iteratively adjusting the fitted spectrum until it matches the measured spectrum. As shown
in Fig. 2, the calculated spectrum is in good agreement with the measured spectrum.
Copyright ©JCPDS-International Centre for Diffraction Data 2011 ISSN 1097-0002
290
291
(3) Scattering FP method for estimating non-measuring components
In the conventional FP method,
FP method
the information of all components
in a sample is required for
accurate analysis. The contents of
non-measuring elements can be
obtained when the elements and
Oxide
Non-measuring
balance
components
are known.
the composition are known. For
example, “CH2” is set in
polyethylene
as
the
Scattering FP method
Metal
Soil, Scale,
Waste oil
Liquid
Polymer
Non-measuring
balance components
are not known.
balance.
Fig. 3. Applicable samples of scattering FP method
However, when non-measuring
elements are not known, accurate calculation cannot be done using conventional fundamental
parameters (FP). The scattering FP method (Kataoka, et al., 2005) estimates the
non-measuring components from the scattering x-ray intensities of Compton and Thomson of
Mo-K line from the Mo secondary target as shown in Fig. 3. This method gives accurate
results for the applications having complex non-measuring components such a soil, scale and
so on.
X-ray Intensity
RESULTS
Four typical application results are described below. The measurement conditions are
summarized in Table 6.
(A) Coal fly ash and soil
The
hazardous
element
analyses of fly ash and soil are
current topics in XRF analysis.
The CRMs of coal fly ash and
soil analyzed were measured by
loose powder method. Two
grams of the sample was set in
polyethylene sample cups of
32mm diameter opening with 4
um Prolene film support. Fig. 4
shows the spectra of the
samples of fly ash and soil
samples of NIST CRMs. It
keV
7.0
8.0
9.0
10.0
11.0
12.0
13.0
14.0
Fig. 4. Measured spectra of coal fly ash(NIST1633a:
blue line) and San Joaquin soil (NIST2709: red line)
in the range of 7 to 14 keV using Mo secondary
target
shows the energy range for As,
Se and Pb and the peaks of the
trace elements are clearly detected. Table 1 and Table 2 list the analyzed results of hazardous
Copyright ©JCPDS-International Centre for Diffraction Data 2011 ISSN 1097-0002
291
292
elements and major elements, respectively. The RPF-SQX with the scattering FP method was
applied to the analysis of both samples. The analyzed results matched well with the certified
values even though the fly ash sample contains unburned carbon and the soil sample contains
organic matter.
Analyzed results of trace hazardous elements
Table 1
Samples
Unit : ppm
As
Cd
Cr
Hg
Pb
Se
17.9
n.d.
120.6
3.9
21.2
2.3
San Joaquin Soil
Analyzed
NIST2709
Std. Val.
17.7
0.4
130.0
1.4
18.9
1.6
Coal Fly Ash
Analyzed
149
n.d.
189
n.d.
75.8
12.6
NIST1633a
Std. Val.
145
1.0
196
0.16
72.4
10.3
Table 2
Unit : mass%
Analyzed results of major elements
Samples
Na
Mg
Al
Si
P
S
K
Ca
Ti
Fe
San Joaquin Soil
Analyzed
1.35
1.84
7.93
29.26
0.063
0.090
1.95
2.02
0.34
3.51
NIST2709
Std. Val.
1.16
1.51
7.50
29.66
0.062
0.089
2.03
1.89
0.34
3.50
Coal Fly Ash
Analyzed
0.31
0.405
14.4
23.3
0.23
0.19
1.91
1.06
0.31
9.0
NIST1633a
Std. Val.
0.17
0.455
14.3
22.8
-
-
1.88
1.11
-
9.4
(B) Thickness measurements of multilayer films on Si wafers
Multilayer thin film analysis is important in semi-conductor applications, as well as other
application such as photovoltaics. The multilayer thin film of Ti/Ni/Ag on Si wafers
illustrated in Fig. 5 were analyzed using the multilayer thin film FP software (Laguitton and
Mantler, 1977; Kataoka and Arai, 1990) and sensitivities of individual elements obtained by
measuring bulk pure metals are stored in the sensitivity library of the software for
standardless analysis. Table 3 lists the analyzed results of the two samples including repeat
measurement result to check the repeatability. The FP software in the NEX CG includes the
computation of the interlayer secondary excitation for accurate theoretical intensity
calculation. In this sample, secondary excitation calculation from Ni layer (Ni-K) to Ti
layer (Ti-K) is included. As shown in Table 3, good agreement between analyzed and
standard values was obtained without using standards with the same film structure.
(C) Biological samples
Analysis of mineral elements such as Mg, Ca and K and hazardous heavy elements are
important in biological samples. The CRMs of biological samples analyzed were prepared by
weighing 2 grams of the sample and making a hydraulically pressed pellet in 32mm of
diameter using 10 tons of pressure for 30 seconds. Fig. 6 shows the spectra of hazardous
heavy elements obtained by measuring various kinds of biological samples. The peaks of
several ppm of hazardous elements can be clearly seen. Fig. 7 shows the spectra of mineral
elements in plant and food samples. High P/B ratio could be obtained for the peaks of Na and
Copyright ©JCPDS-International Centre for Diffraction Data 2011 ISSN 1097-0002
292
293
Mg by using a special secondary target for light element analysis equipped in the NEX CG
without interference of higher energy peaks as shown in Fig. 7 (c).
Table 4 shows the analyzed results of the aforementioned biological samples using the
RPF-SQX software with the scattering FP method. The influence of non-measuring elements
highly contained in these biological samples such as N, O, C and H, were effectively
corrected.
Table 3 Analyzed results of Ti/Ni/Ag film on Si Wafer
Ti/Ni/Ag/Si Wafer
n=1
2
3
4
5
6
7
8
9
10
Ti
976
978
983
970
976
968
970
990
973
968
Unit:Angstrom
Ni
Ag
3917 1166
3923 1177
3921 1167
3923 1185
3921 1196
3923 1169
3921 1212
3917 1169
3925 1201
3921 1172
Average
975
3921 1181
Std. dev.
RSD%
7.1
0.7
2.6
0.1
Ti (1000A)
Ni (4000,2000A)
Ag (1200A)
Si Wafer
Fig. 5. Structure of
multilayer film
Sample B
Sample A
16.4
1.4
n=1
2
3
4
5
6
7
8
9
10
Average
Std. dev.
RSD%
Ti
Unit:Angstrom
Ni
Ag
1066
1052
1056
1066
1066
1065
1065
1078
1064
1056
1063
7.3
0.7
2016
2013
2008
2013
2013
2020
2015
2012
2013
2015
2014
3.1
0.2
1177
1187
1170
1162
1201
1178
1192
1145
1171
1180
1176
15.8
1.3
Copyright ©JCPDS-International Centre for Diffraction Data 2011 ISSN 1097-0002
293
294
Heavy elements
Hg-L
(b)
X-ray Intensity
X-ray Intensity
Hg-L
As-K
(a)
21.6 ppm
18.0 ppm
Se
16.6 ppm
6.06 ppm
5.63 ppm
4.64 ppm
2.14 ppm
0.27 ppm
1.40 ppm
Ag-K
9.6
14.0keV
(d)
X-ray Intensity
(c)
X-ray Intensity
12.0
26.8 ppm
20.8 ppm
21.0
22.0
23.0
10.4
10.8 keV
34.7 ppm
0.77 ppm
0.043 ppm
20.0
10.0
Cr-K
10.0
Cd-K
8.0
6.0
5.0
24.0 keV
5.8
5.4
6.2
keV
Fig. 6. Spectra of marine organism standard samples of National Research Council Canada
Red lines : Lobster Hepatopancreas (TORT-2), Blue lines : Dogfish Muscle (DORM-2), Green
lines :Dogfish Liver (DOLT-2) (a) Mo secondary target, (b) magnified spectrum of (a) in
region of Hg, (c) Al secondary target, (d) Cu secondary target
1.4
1.8
2.2
2.6
1.56
2.43
mass% mass%
keV
RX9 secondary target
Secondary target for light elements
(c)
0.8
0.9
1.0
1.1
3.2
3.6
4.0
4.4 keV
Fig. 7. Spectra of Peach Leaves (NIST1547)
Mg-K
0.7
2.8
and Non-fat Milk Powder (NIST1549)
NIST1547:(a) RX9 secondary target, (b) Cu
secondary target
1200 ppm
NIST1549:(c)
Comparison
of
spectra
compared between RX9 secondary target and
special secondary target for light elements
Ká
á
Na-K
Cl-K escape-peak
X-ray Intensity
NaNa
-K
4970 ppm
Ca-K
360
ppm
K-K
2000
ppm
(b)
X-ray Intensity
Cl-K
1370
ppm
S-K
Si-K
X-ray Intensity
Al-K
(a)
P-K
Light elements
1.2
1.3
keV
Copyright ©JCPDS-International Centre for Diffraction Data 2011 ISSN 1097-0002
294
295
Table 4 Analyzed results of various biological samples
Na
Samples
Mg
Si
P
Cl
S
K
Ca
Mn
Fe
Co
mass% mass% mass% mass% mass% mass% mass% mass% ppm
ppm
ppm
n.d.
NIST1570a
Analyzed
1.83
0.85
0.11
0.59
0.50
0.65
2.79
1.47
66.5
271
Spinach Leaves
Std. Val.
1.82
0.89
-
0.52
0.46
-
2.90
1.53
-
-
-
NIES CRM No.1
Analyzed
n.d.
0.38
0.27
0.14
0.26
0.39
1.58
1.45
2021
215
15.6
Pepperbush
Std. Val.
-
0.41
-
0.11
-
-
1.51
1.38
2030
205
23.0
NRCC DOLT-2
Analyzed
0.83
0.12
n.d.
1.08
1.29
0.83
0.87
0.06
5.7
1124
n.d.
Dogfish Liver
Std. Val.
-
-
-
-
-
-
-
-
6.9
1103
-
Ni
ppm
Cu
ppm
Zn
ppm
As
ppm
Se
ppm
Br
ppm
Rb
ppm
Sr
ppm
Cd
ppm
Ba
ppm
Hg
ppm
Samples
NIST1570a
Spinach Leaves
Analyzed
Std. Val.
2.2
14.0
80.0
n.d.
n.d.
34.5
13.0
52.7
n.d.
n.d.
n.d
2.1
12.2
82.0
-
-
-
13.0
55.6
-
-
-
NIES CRM No.1
Pepperbush
Analyzed
Std. Val.
10.3
12.9
347.5
n.d.
n.d.
0.8
75.5
35.0
5.7
12.0
340.0
-
-
-
75.0
36.0
6.7
165.5
165.0
n.d.
8.7
NRCC DOLT-2
Dogfish Liver
Analyzed
n.d.
27.7
90.8
14.2
5.7
23.7
2.0
2.7
21.1
n.d.
1.6
Std. Val.
-
25.8
85.8
16.6
6.1
-
-
-
20.8
-
-
(D) Copper Alloy
Copper alloy samples were analyzed as examples of trace element analysis in metal samples.
The copper alloy samples were polished using a lathe to make flat surface. Fig. 8 shows the
spectra obtained by measuring a naval brass standard. A Pb peak of 0.069mass% can be seen
with high P/B ratio due to the high speed detector with pile-up rejection and Mo secondary
target. Fig. 9 shows the spectra in the light element range obtained by measuring three copper
alloys containing phosphorous. Monochromatic Pd-L line effectively excites the P-K line
so that the peak of 90 ppm phosphorous could be seen in the spectrum. The detection of
P-K is difficult in WDXRF due to the interference by the 4-th order of Cu-K line. Table 5
exhibits the analyzed results of various kinds of copper alloys and the results match well with
the standard values from low level to high contents.
10.0
6.0
8.0
10.0
Pb-L
Alターゲット
(b)
11.0
12.0
12.0
13.0
14.0 keV
X-ray Intensity
X-ray Intensity
Pb-L
(a)
20.0
24.0
28.0
32.0
36.0 keV
Fig. 8. Spectra of naval brass standard (NIST C1108) (a) Mo secondary target, (b) Al
secondary target
X-ray Intensity
Copyright ©JCPDS-International Centre for Diffraction Data 2011 ISSN 1097-0002
295
296
0.13mass%
0.07mass%
0.009mass%
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4 keV
Fig. 9. Spectra of light element range including phosphorus of copper alloys
Blue: NIST C1118, red:NIST C1119, green:NIST C1114
Table 5 Analysis results of copper alloy samples
Samples
Cu
Zn
Pb
Fe
Sn
Ni
Al
Sb
As
Mn
P
mass%
mass%
mass%
mass%
mass%
mass%
mass%
mass%
mass%
mass%
mass%
NIST
C1103
Free-Cutting
Brass
Analyzed
59.1
35.6
3.88
0.25
0.869
0.17
n.d.
n.d.
n.d.
n.d.
0.0045
Std.
59.19
35.7
3.81
0.26
0.88
0.16
-
-
-
-
0.003
NIST
C1108
Naval Brass
Analyzed
65.0
34.4
0.069
0.048
0.397
0.031
n.d.
n.d.
n.d.
0.029
0.003
Std.
64.95
34.42
0.063
0.05
0.39
0.033
-
-
-
0.025
-
NIST
C1115
Commercial
Bronze
Analyzed
88.1
11.6
0.011
0.127
0.102
0.070
n.d.
n.d.
n.d.
0.004
0.003
Std.
87.96
11.73
0.013
0.13
0.1
0.074
-
-
-
-
0.005
NIST
C1118
Aluminum
Brass
Analyzed
75.3
21.8
0.022
0.060
0.0047
n.d.
2.88
n.d.
0.009
0.002
0.133
Std.
75.1
21.9
0.025
0.065
-
-
2.8
-
0.007
-
0.13
NIST
C1119
Aluminum
Brass
Analyzed
77.2
20.2
0.053
0.028
n.d.
n.d.
2.09
0.049
0.043
0.005
0.062
Std.
77.1
20.4
0.05
0.03
-
-
2.14
0.05
0.04
-
0.07
Table 6 Measurement conditions
(A) Coal fly ash and soil
Atmosphere: helium
Secondary target
Light element
target
RX9
Cu
Mo
Al
kV-mA
25-auto *)
25-auto
50-auto
50-auto
50-auto
Energy ranges(KeV)
1.0-1.3
1.3-2.8
2.8-8.0
8.0-15.0
15.0-40.0
Counting time (s)
200
100
100
100
200
)
* auto: current is adjusted such that dead time is at a predetermined level.
(B) Multilayer film
Atmosphere: vacuum
Secondary target
Cu
Mo
Al
kV-mA
50-auto
50-auto
50-auto
Energy ranges(KeV)
2.8-8.0
8.0-15.0
15.0-40.0
Counting time (s)
100
100
100
Copyright ©JCPDS-International Centre for Diffraction Data 2011 ISSN 1097-0002
(C) Biological samples
296
297
Atmosphere: vacuum
Secondary target
Light element
target
RX9
Cu
Mo
Al
kV-mA
25-auto
25-auto
50-auto
50-auto
50-auto
Energy ranges(KeV)
1-1.3
1.3-2.8
2.8-8.0
8.0-15.0
15.0-40.0
Counting time (s)
600
300
300
300
600
(D) Copper alloy
Atmosphere: vacuum
Secondary target
RX9
Cu
Mo
Al
kV-mA
25-auto
50-auto
50-auto
50-auto
Energy ranges(KeV)
1.0-2.8
2.8-8.0
8.0-15.0
15.0-40.0
Counting time (s)
100
50
50
50
CONCLUSION
The spectrometer with secondary targets, polarized optics, and high speed detector with
pile-up rejection demonstrated extremely high P/B ratios from light to heavy elements
resulting in low LLD results. Accurate analyses down to ppm level could be achieved even in
complex sample composition for the quantification software which combined fundamental
parameter method and full profile fitting. The scattering FP method corrected for
non-measuring components in samples such as soils, biological samples by using Compton
and Thomson scattering intensities from Mo secondary target.
Additionally, good results of multilayer thin films could be obtained by the thin film FP
software considering interlayer secondary excitation.
REFERENCES
Campbell, J. L. and Wang, J.-X. (1992), “Lorentzian contributions to x-ray lineshapes in
Si(Li) spectroscopy”, X-ray Spectrometry, 21, 223-227.
Kataoka, Y. (1989), “Standardless x-ray fluorescence spectrometry (Fundamental parameter
method using sensitivity library), The Rigaku Journal, 6, 33-39.
Kataoka, Y. and Arai, T. (1990), “Basic studies of multi-layer thin film analysis using
fundamental parameter method”, Advances in X-ray Analysis, 33, 213-223.
Kataoka, Y., Kawahara, N., Hara, S., Yamada, Y., Matsuo, T. and Mantler, M. (2005),
“Fundamental parameter method using scattering x-rays in x-ray fluorescence analysis”,
Advances in X-ray Analysis, 40, 255-260.
Laguitton, D. and Mantler, M. (1977), “LAMA I-A general fortran program for quantitative
Copyright ©JCPDS-International Centre for Diffraction Data 2011 ISSN 1097-0002
x-ray fluorescence analysis”, Advances in X-ray Analysis, 20, 515-528 .
297
298