- Email: [email protected]
or ICP-AES for NIST standards
Nuclear Instruments and Methods in Physics Research B 189 (2002) 86–93 www.elsevier.com/locate/nimb
Comparison of elemental quantity by PIXE and ICP-MS and/or ICP-AES for NIST standards K. Saitoh a
c
a,*
, K. Sera b, T. Gotoh c, M. Nakamura
c
Environmental Research and Information Center of Akita Prefecture (ERICA), 3-1-1 Sanno, Akita 010-8572, Japan b Cyclotron Research Center, Iwate Medical University, 348 Tomegamori, Takizawa, 020-0173, Japan Measurement and Analysis Division, Tohoku Afforestation and Environmental Protection Co., Ltd., 3-8-22, Sakuragi, Tagajo 985-0842, Japan
Abstract Urban particulate matter (SRM 1648), Buffalo River sediment (SRM 2704) and pine needle (SRM 1575) standard reference materials prepared by the National Institute of Standards and Technology (NIST, USA) were analyzed by three multi-element analysis methods, i.e., particle induced X-ray emission (PIXE), inductively coupled plasma-mass spectrometry (ICP-MS) and inductively coupled plasma-atomic emission spectrometry (ICP-AES); values determined by those analysis methods were compared with certified and/or non-certified values of NIST samples. Values determined by PIXE were 70–120% relative to certified and/or non-certified values of NIST samples except for Co in the urban particulate matter, for V and Co in Buffalo River sediment and for Ni and Br in the pine needles samples. In particular, Al, K, Ca, Cr, Mn, Fe, Cu, Zn and Pb were 85–110% in all samples. On the other hand, Na and Fe values determined by ICP-MS were very much different from the certified values in all samples, but the other elements were 70– 120%. As for ICP-AES, all elements except for Na were 80–100% in all samples. Comparing the values determined by PIXE and those determined by ICP-MS and/or ICP-AES, there was a slight difference between the samples, but the range was 75–120% except for Na, V, Fe and Co determined by ICP-MS and Na determined by ICP-AES, which was generally consistent with PIXE. Ó 2002 Elsevier Science B.V. All rights reserved. PACS: 32.30.Rj; 82.80.Ej Keywords: Multi-element analysis; Major-to-ultratrace elements; NIST sample; PIXE; ICP-MS; ICP-AES
1. Introduction Since particle induced X-ray emission (PIXE) allows analysis of minute samples smaller than a milligram without any complex chemical manipulation, and since it not only simultaneously detects
*
Corresponding author. Tel.: +81-18-860-4013; fax: +81-18860-4016. E-mail address: [email protected] (K. Saitoh).
elements from Na to U in a short time but also detects major-to-ultratrace elements at the concentration level of ppm to sub-ppb, it is being used in a variety of fields, including environmental research, medicine, geology and archeology, and PIXE analysis is fast becoming a universal method for highly sensitive analysis of multiple elements. Meanwhile, multi-element analysis by means of inductively coupled plasma-mass spectrometry (ICP-MS) and inductively coupled plasma-atomic emission spectrometry (ICP-AES) is
0168-583X/02/$ - see front matter Ó 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 5 8 3 X ( 0 1 ) 0 1 0 1 2 - 6
K. Saitoh et al. / Nucl. Instr. and Meth. in Phys. Res. B 189 (2002) 86–93
being performed on a daily basis by many research and analysis institutions. In the field of environmental research, ICP-MS and ICP-AES have become synonymous with multi-element analysis. Comparison of values determined by PIXE with those determined by ICP-MS and/or ICP-AES is important in the field of environmental research in terms of evaluating values determined by PIXE. Consequently, we compared elemental concentrations determined by PIXE with those determined by ICP-MS and/or ICP-AES using urban particulate matter, Buffalo River sediment and pine needles of standard reference material prepared by the National Institute of Standards and Technology (NIST, USA).
2. Experiment and method 2.1. Samples Samples were standard reference material (SRM) prepared by the NIST. The urban particulate matter sample was SRM 1648. The Buffalo River sediment sample was SRM 2704. The pine needle sample was SRM 1575. SRM samples were stored in a clean box until quantitative analysis by PIXE, ICP-MS and ICP-AES.
87
oven [3], where nearly 50 mg of the samples was put in a polytetrafluoroethylene (PTFE) vessel together with 1 ml of nitric acid, and Indium solution (1000 mg/l) was added as an internal standard in an amount of 1000 lg/g. This PTFE vessel was tightly sealed with a polypropylene jacket with a hand, and then it was heated for 2 min at low power corresponding to 200 W. Thin targets were prepared in a clean bench by taking 5 ll of it using a pipette and dropping it onto a 4 lm-thick polypropylene film, after which it was dried in clean air. Five such thin targets were prepared by decomposing 50 mg of each sample with nitric acid using a microwave oven in order to confirm reproducibility. Elemental concentrations in these thin targets were determined by PIXE at Nishina Memorial Cyclotron Center (NMCC), Japan Radioisotope Association, using 2.9 MeV protons from a baby cyclotron [4]. Beam currents, the accumulated charge and the typical measuring time were 20–40 nA, 11–19 lC, and 10–5 min, respectively. X-ray spectra were analyzed using the S A P I X program [4]. Quantitative analysis of elemental concentrations was performed based on a powdered internal standard method [1,2] for the urban particulate matter and the Buffalo River sediment samples, and on an internal standard method [3] for the pine needle sample.
2.2. Quantitative analysis by PIXE The urban particulate matter and Buffalo River sediment samples were prepared using the method developed by Sera and Futatsugawa [1,2], where 100 mg of sample was ground into fine powder in an agate mortar, and Palladium-carbon powder (5% Pd) was added as an internal standard in an amount of 10 mg, after which they were mixed uniformly in the agate mortar. Thin targets were prepared by taking 0.1 mg of the powder and putting it onto a 4 lm-thick polypropylene film, and dropping roughly 1 ll of 10% collodion solution diluted with ethyl alcohol onto it for fixing and smoothing. This process was carried out in a clean bench. Five such thin targets were prepared for each sample in order to confirm reproducibility. The pine needle samples thus prepared were decomposed with nitric acid using a microwave
2.3. Quantitative analysis by ICP-MS and ICPAES The urban particulate matter, Buffalo River sediment and pine needle samples were decomposed with nitric acid using a microwave oven, where nearly 50 mg of the sample was put in a dedicated PTFE vessel of the microwave oven together with 5 ml of nitric acid. This PTFE vessel was heated under the following conditions: 5 min (250 W), 5 min (400 W), 10 min (500 W) and 5 min (600 W), after which the PTFE vessel was reheated under the same conditions 15 min later. The decomposed solution was transferred through filter paper (5B, ADVANTEC Ltd., Japan) into a 100 ml calibrated flask, and the volume was adjusted to 100 ml with ultra-pure water. For ICPMS (Yokogawa Analytical Systems HP 4500,
88
K. Saitoh et al. / Nucl. Instr. and Meth. in Phys. Res. B 189 (2002) 86–93
Japan) measurement, this decomposed solution and solution diluted 10-fold with ultra-pure water were analyzed. For ICP-AES (Varian VISTA, USA) measurement, this decomposed solution only was analyzed. Five analysis solutions were prepared by decomposing 50 mg of each sample with nitric acid using a microwave oven in order to confirm reproducibility. The number of elements determined by ICP-MS was 28: Li, Be, Na, Mg, Al, Ca, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, As, Se, Rb, Sr, Ag, Cd, Cs, Ba, Tl, Pb, Bi, Th and U, and that determined by ICPAES was nine: Na, Mg, Al, K, Ca, Cr, Mn, Fe and Zn. The instrumental components and typical operating conditions for ICP-MS and ICP-AES are listed in Table 1. These operating conditions were selected after optimization of each parameter. Elemental concentrations were determined based on calibration curves generated from analysis of the continuing calibration standard solutions. The standard solution was multi-element 10 ppm (5% nitric acid base) mixed standard of SPEX Industries Co. The concentration of calibration standard solutions for ICP-MS was zero (ultra-pure water), 0.005, 0.02, 0.1, 0.5, 2, 10, 50 and 200 lg/l, and that for ICP-AES was zero (ultra-pure water), 0.005, 0.02, 0.1, 0.5, 1 and 2 mg/l. For generation of calibration curves, determination of each concentration was repeated three times.
3. Results and discussion 3.1. X-ray spectra analysis and reproducibility for PIXE Peaks of characteristic X-rays from seven elements (Na, Mg, Al, Si, P, S and Cl) were observed in a spectrum obtained with a Si (Li) X-ray detector without an absorber over the low X-ray energy range for all of the samples. In a spectrum from a Si (Li) detector with a 300 lm thick Mylar absorber, peaks of characteristic X-rays from thirteen elements (K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, Ga, Sr and Pb) could be observed for all of the samples. In addition to those element peaks, peaks of Co, As, Se, Br and Zr, peaks of Co, Rb and Zr and peaks of Br and Rb were observed for
Table 1 Operating conditions for ICP-MS and ICP-AES instrument ICP-MS
HP 4500 (Yokogawa Analytical Systems)
Plasma conditions Rf frequency Rf power
27.12 MHz 1.25 kW
Gas flow rate Carrier gas Auxiliary gas Coolant gas
Ar 1.0 l/min Ar 1.1 l/min Ar 15 l/min
Sampling conditions Sampling depth Sampling cone Skimmer cone Nebulizer Sampling uptake rate Data acquisition
7 mm from work coil Platinum, 1.0 mm orifice diameter Platinum, 0.4 mm orifice diameter Cross-flow type 0.4 ml/min
Data point Dwell time Integration Repetition
Multi-element mode by peak hopping 3 points/peak 20 ms/point 100 times 3 times
ICP-AES
VISTA (Varian)
Plasma conditions Rf frequency Rf power
40 MHz 1.0 kW
Gas flow rate Carrier gas Auxiliary gas Coolant gas
Ar 0.7 l/min Ar 1.5 l/min Ar 15 l/min
Sampling conditions Nebulizer Sampling uptake rate
Glass concentric type 1.0 ml/min
Spectrometer conditions Polychromator Focal length Grating Entrance slit width
Echelle polychromator 40 cm 94.74 grooves/mm 25 lm
Data acquisition Integration time Repetition
10 s 3 times
the urban particulate matter, the Buffalo River sediment and for pine needles, respectively. Errors for the analytical results are mainly from the spectrum fitting, the detection efficiency and values
K. Saitoh et al. / Nucl. Instr. and Meth. in Phys. Res. B 189 (2002) 86–93
of X-ray transmission through the absorber. Estimated relative error (100 error/count, in %) was smaller than 1% for three elements (Si, Ca and Zn), 1–5% for eight elements (Al, S, Ti, Cr, Mn, Fe, Cu and Pb), 5–10% for six elements (Mg, P, K, Co, Br and Sr) and 10–20% for three elements (Na, Cl and Ni). Repeated analysis of the same samples revealed that in urban particulate matter elements with a coefficient of variation (CV) of less than 10% were K, Ca, Fe, Cu, Se, Br, Sr and Pb, those with a CV of 10–15% were Na, Al, Cl, Mn, Co, Zn, Ga, As and Zr and those with a CV of 15–20% were Mg, Si and Ti. In Buffalo River sediment, elements with a CV of less than 10% were Mn, Fe, Zn and Pb, those with a CV of 10–15% were Na, Mg, P, S, V, Cr, Co, Ga, Rb and Sr and those with a CV of 15– 20% were Ca and Ti. In pine needles, elements with a CV of less than 10% were P, S, K, Ca, Mn, Fe and Sr and those with a CV of 10–15% were Mg, Zn and Pb. 3.2. Reproducibility for ICP-MS and ICP-AES The determination limits and concentration ranges of ICP-MS and ICP-AES were investigated from the reproducibility (relative standard deviation) of calibration standard solutions and linearity of calibration curves. Repeated analysis of the same samples by ICP-MS revealed that the CV of all elements was less than 10% except for Be, Ca, Ag, Bi and Th in urban particulate matter, Na and Fe in Buffalo River sediment and Na, V, Zn, Ag, Cd and Pb in pine needles. Elements with a CV of 10–15% were Be and Na in urban particulate matter and Fe in Buffalo River sediment. Elements with a CV of 15–20% were Ag and Th in urban particulate matter and Zn and Pb in pine needles. On ICP-AES, the CV of all elements in all samples was less than 10%. 3.3. Comparison of certified and/or non-certified values to values determined by PIXE, ICP-MS and ICP-AES Values determined by PIXE, ICP-MS and ICPAES and NIST certified and non-certified values
89
are shown in Tables 2–4. Elemental concentrations determined by PIXE were 70–120% relative to NIST certified and/or non-certified values except for Co in the urban particulate matter, V and Co in the Buffalo River sediment and Ni and Br in the pine needle samples. In particular, Mg, Al, K, Ca, Cr, Mn, Fe, Cu, Zn, Rb and Pb were 85–110% in all samples. In the case of PIXE, when determining V it is necessary to subtract the contribution of Ti Kb from the V Ka peak because V Ka overlaps with Ti Kb. The V Ka peak in urban particulate matter was clear and there was no major error when isolating it from Ti Kb. V was 76.4% relative to NIST certified and/or non-certified values. In contrast, in the case of Buffalo River sediment, the values determined by PIXE may have been considerably lower than the certified values because we overestimated the yield of the Ti Ka peak due to the influence of the escape peak of Fe, which overlaps with Ti Ka, resulting in overestimation of the contribution of Ti Ka from the V Ka peak. Regarding Co, because Co Ka closely resembles Fe Ka, when the Fe content is several thousand times greater than the Co content, as was the case with urban particulate matter and Buffalo River sediment, the Co Ka estimated relative error will be several 10 or several 100 percent higher even if the Fe Ka estimated relative error resulting from spectral analysis is 1%. The determined Co value may have been 10–20 times the certified value for this reason. Non-certified values for rare earth elements and heavy elements such as Cd and U are indicated for standard NIST samples, but these elements were not detected by PIXE. This may be because L-X rays of rare earth elements overlap with transition elements of K-X rays such as Fe and Cu or because the detection sensitivity toward rare earth elements and heavy elements is low under NMCC PIXE measurement conditions. On the other hand, in a comparison of values determined by ICP-MS and NIST certified and/or non-certified values, there were slight differences depending on the sample, but elements with a value 70–120% relative to certified and/or noncertified values were the 18 elements Mg, Al, Ca, V, Cr, Mn, Co, Ni, Cu, Zn, As, Rb, Sr, Ag, Cd, Cs, Ba and Pb. Values for Na and Fe were quite
90
K. Saitoh et al. / Nucl. Instr. and Meth. in Phys. Res. B 189 (2002) 86–93
Table 2 Elemental quantities by PIXE, ICP-MS and ICP-AES for urban particulate matter (SRM 1648) of NIST sample Determined value Element
PIXE (b)
ICP-MS (c) a
Mean SD Li (lg/g) Be (lg/g) Na (%) Mg (%) Al (%) Si (%) P (%) S (%) Cl (%) K (%) Ca (%) Sc (lg/g) Ti (%) V (lg/g) Cr (lg/g) Mn (lg/g) Fe (%) Co (lg/g) Ni (lg/g) Cu (lg/g) Zn (%) Ga (lg/g) As (lg/g) Se (lg/g) Br (lg/g) Rb (lg/g) Sr (lg/g) Zr (lg/g) Ag (lg/g) Cd (lg/g) In (lg/g) Sb (lg/g) I (lg/g) Cs (lg/g) Ba (lg/g) La (lg/g) Ce (lg/g) Sm (lg/g) Eu (lg/g) Hf (lg/g) W (lg/g) Tl (lg/g) Pb (%) Bi (lg/g) Th (lg/g) U (lg/g) a b
0.436 0.062 0.623 0.114 3.32 0.38 9.38 1.76 0.580 0.123 6.00 1.35 0.417 0.054 1.00 0.08 6.41 0.63 0.430 0.066 107 24 341 91 804 94 4.12 0.37 187 28 64.9 12.3 575 56 0.475 0.053 55.0 6.0 87.3 9.5 18.9 1.9 395 35
b/a (%)
102.6 77.9 97.1
Mean SD
10.2 0.2 0.90 0.13 0.134 0.008 0.631 0.067 2.76 0.06
c/a (%)
31.5 78.9 80.7
75.9 70.0 79.0
100.1
Mean and SD of five measurements. Indication of parentheses is non-certified value.
91.0 1.8 236 5 703 22 11.8 5.9 12.6 0.4 62.0 1.3 534 16.3 0.658 0.047 34.5 0.9 105 3 12.9 0.6
65.0 58.6 81.7 301.8 70.0 75.6 87.7 138.2 91.3 47.8
20.2 0.4 144 3
38.8
5.7 1.0 65.1 1.3
95.0 86.8
1.38 0.07 568 12
46.0 77.1
1.81 0.09 0.701 0.015 5.00 1.61 4.67 0.95 3.10 0.08
d/a (%)
0.183 0.016 0.723 0.067 2.76 0.27
43.1 90.4 80.7
0.425 0.002 (0.8)b 3.42 0.11
85.0
(5.0) (0.45) 1.05 0.01
a
0.893 0.103 5.77 0.52
12.6 1.4 107.5 76.4 84.6 93.5 105.4 1038.9 79.1 94.4 99.8
Mean SD
Certified value and/or non-certified value (a)
ICP-AES (d)
120.0 92.7 95.2
193 19 164 18
0.656 0.023
a
107.0 63.1 56.4
380 30 690 60 3.56 0.34
94.3 80.2 91.0
0.395 0.040
83.0
(7) (0.40) 140 3 403 12 (860) 3.91 0.10 (18) 82 3 609 27 0.476 0.014 115 10 27 1 (500) (52)
(6) 75 7 (1.0) (45) (20) (3) (737) (42) (55) (4.4) (0.8) (4.4) (4.8) 0.655 0.008 (7.4) 5.5 0.01
K. Saitoh et al. / Nucl. Instr. and Meth. in Phys. Res. B 189 (2002) 86–93
91
Table 3 Elemental quantities by PIXE, ICP-MS and ICP-AES for Buffalo River sediment (SRM 2704) of NIST sample Determined value Element
PIXE (b)
ICP-MS (c) a
Mean SD Li (lg/g) Be (lg/g) Na (%) Mg (%) Al (%) Si (%) P (%) S (%) Cl (%) K (%) Ca (%) Sc (lg/g) Ti (%) V (lg/g) Cr (lg/g) Mn (lg/g) Fe (%) Co (lg/g) Ni (lg/g) Cu (lg/g) Zn (lg/g) Ga (lg/g) As (lg/g) Se (lg/g) Br (lg/g) Rb (lg/g) Sr (lg/g) Y (lg/g) Zr (lg/g) Ag (lg/g) Cd (lg/g) Sn (lg/g) Sb (lg/g) I (lg/g) Cs (lg/g) Ba (lg/g) La (lg/g) Ce (lg/g) Sm (lg/g) Eu (lg/g) Dy (lg/g) Lu (lg/g) Hf (lg/g) Hg (lg/g) Tl (lg/g) Pb (lg/g) Bi (lg/g) Th (lg/g) U (lg/g) a b
b/a (%)
0.400 0.052 1.08 0.16 6.42 1.46 26.1 6.0 0.098 0.013 0.350 0.037 0.057 0.018 1.80 0.41 2.64 0.41
73.1 90.0 105.1 89.7 98.2 88.2
0.461 0.073 43.2 5.0 146 16 479 39 4.51 0.34 340 50 31.9 13.4 91.8 11.7 450 17 12.0 1.7
100.9 45.6 108.1 86.3 109.7 2428.6 72.3 93.1 102.7 80.0
90.0 101.5
198 22 110 13
198.0 84.6
305 111
101.7
192 16
119.2
a
Mean SD
ICP-AES (d) c/a (%)
a
Mean SD
d/a (%)
32.7 1.8 0.97 0.09 0.028 0.008 0.913 0.037 7.00 0.61
5.1 76.1 114.6
0.081 0.001 1.13 0.01 4.83 0.03
14.8 94.2 79.0
3.70 0.39
142.3
1.67 0.02 2.41 0.02
83.5 92.7
77.8 2.9 112 8 540 22 20.3 2.4 11.3 0.4 31.1 1.4 80.1 3.8 376 11 19.9 0.8 17.2 0.4 Not detected
81.9 83.0 97.3 493.9 80.7 70.5 81.2 85.8 132.7 73.5
120 3 550 7 4.01 0.03
88.9 99.0 97.6
Certified value and/or non-certified value (a) 47.5 4.1
68.8
440 10
100.4
0.547 0.014 1.20 0.02 6.11 0.16 29.08 0.13 0.0998 0.0028 0.397 0.004 (<0.01)b 2.00 0.04 2.60 0.03 (12) 0.457 0.018 95 4 135 5 555 19 4.11 0.10 14.0 0.6 44.1 3.0 98.6 5.0 438 12 (15) 23.4 0.8 1.12 0.05 (7) (100) (130) (2.8) (300)
93.1 1.8 70.7 1.1
93.1 54.4
0.4 0.1 2.15 0.09
62.3
4.79 0.07 275 8
79.8 66.4
0.45 0.03 138 13 Not detected 5.22 0.46 1.44 0.06
42.8 85.7
3.45 0.22 (9.5) 3.79 0.15 (2) (6) 414 12 (29) (72) (6.7) (1.3) (6) (0.6) (8) 1.47 0.07 1.06 0.07 161 17
56.7 46.0
(9.2) 3.13 0.13
Mean and SD of five measurements. Indication of parentheses is non-certified value.
92
K. Saitoh et al. / Nucl. Instr. and Meth. in Phys. Res. B 189 (2002) 86–93
Table 4 Elemental quantities by PIXE, ICP-MS and ICP-AES for pine needles (SRM 1575) of NIST sample Determined value Element
PIXE (b)
ICP-MS (c) a
Mean SD Li (lg/g) Be (lg/g) Na (%) Mg (%) Al (lg/g) Si (%) P (%) S (%) Cl (lg/g) K (%) Ca (%) Sc (lg/g) Ti (%) V (lg/g) Cr (lg/g) Mn (lg/g) Fe (lg/g) Co (lg/g) Ni (lg/g) Cu (lg/g) Zn (lg/g) Ga (lg/g) As (lg/g) Se (lg/g) Br (lg/g) Rb (lg/g) Sr (lg/g) Ag (lg/g) Cd (lg/g) Sb (lg/g) Cs (lg/g) Ba (lg/g) La (lg/g) Ce (lg/g) Eu (lg/g) Hg (lg/g) Tl (lg/g) Pb (lg/g) Bi (lg/g) Th (lg/g) U (lg/g) a b
0.018 0.006 0.074 0.009 509 111 0.115 0.048 0.089 0.004 0.099 0.004 30.1 9.8 0.370 0.021 0.375 0.011 13.2 3.6 2.33 0.62 2.58 0.83 618 20 175 12 2.12 0.52 2.98 0.74 61.2 6.9 3.73 1.74
1.87 0.61 11.0 3.1 4.16 0.17
b/a (%)
93.4
a
Mean SD
Not detected Not detected 0.015 0.004 0.122 0.002 378 3
Mean SD
d/a (%)
Certified value and/or non-certified value (a)
0.015 0.001 0.100 0.001 434 8
79.8
545 30
ICP-AES (d) c/a (%)
69.4
a
0.12 0.02
74.2
100.0 91.5
99.2 91.6 87.5 60.6 99.3
20.8 94.0 86.7
0.346 0.013
0.19 0.08 1.8 0.1 704 8 120 5 Not detected 2.18 0.21 3.31 0.35 28.7 5.5 0.22 0.06 Not detected Not detected 9.34 0.22 3.34 0.05 3.0 1.2 0.08 0.05
84.4
69.2 104.3 60.0
0.342 0.003 0.393 0.007
92.4 95.8
<10 610 4 160 6
90.4 80.0
62.3 110.3
0.37 0.02 0.41 0.02 (0.03)b
2.6 0.2 675 15 200 10 (0.1) (3.5) 3.0 0.3
50 3 0.21 0.04
79.8 69.6
(9) 11.7 0.1 4.8 0.2 (<0.5) (0.2)
Not detected 5.06 0.04
9.39 1.22
86.9
Not detected 9.8 1.8 Not detected Not detected Not detected
90.7
(0.2) (0.4) (0.006) 0.15 0.05 (0.05) 10.8 0.5 0.037 0.003 0.020 0.004
Mean and SD of five measurements. Indication of parentheses is non-certified value.
different from the certified values in urban particulate matter and Buffalo River sediment in which these elements account for a percentage of the content. In pine needles, in which Na accounts
for mg/g of the content and Fe accounts for lg/g of the content, the Fe value was 60% relative to the certified value, and in the case of Na, for which certified and non-certified values are not
K. Saitoh et al. / Nucl. Instr. and Meth. in Phys. Res. B 189 (2002) 86–93
indicated, the value was not much different from the value determined using PIXE. As this demonstrates, the compatibility of Na and Fe values with certified values differed greatly depending on the Na and Fe concentrations in samples. Values determined by ICP-AES were 80–100% of certified and/or non-certified values for all elements except Na in all samples. Na values tended to be the same as those determined by ICP-MS. 3.4. Comparison of PIXE to ICP-MS and ICPAES for elemental quantity Comparing values determined by PIXE with those determined by ICP-MS, elements for which PIXE values were not within the range of 80–120% relative to ICP-MS values were Na, Cr, Fe, Co, Zn and Se in urban particulate matter and Na, Ca, V, Fe, Co, Ga and Sr in Buffalo River sediment. Of these elements, PIXE values for V and Co and ICP-MS values for other elements were very different from NIST certified and/or non-certified values. In a comparison of values determined by PIXE and those determined by ICP-AES, ICPAES values for all elements except for Na in urban particulate matter and Buffalo River sediment were 75–120% relative to PIXE values. There was a large discrepancy between ICP-MS and ICPAES Fe values.
93
4. Conclusion We compared values determined by PIXE, ICPMS and ICP-AES using standard NIST urban particulate matter (SRM 1648), Buffalo River sediment (SRM 2704) and pine needle (SRM 1575) samples. We found that although reproducibility of PIXE was somewhat lower than that of ICPMS and ICP-AES, except for V and Co, values determined by PIXE were consistent with NIST certified and/or non-certified values compared to those determined by ICP-MS and ICP-AES. V and Co values determined by PIXE were quite different from the certified and/or non-certified values because of spectral analysis, so improving the spectral analysis program may solve this problem. Compared to ICP-MS and ICP-AES, PIXE can be used to precisely and efficiently analyze many elements, so it may become the most effective means of analysis in environmental studies in the near future.
References [1] K. Sera, S. Futatsugawa, Int. J. PIXE 8 (1998) 185. [2] K. Sera, S. Futatsugawa, D. Ishiyama, Int. J. PIXE 9 (1999) 63. [3] S. Futatsugawa, S. Hatakeyama, S. Saitou, K. Sera, Int. J. PIXE 3 (1993) 319. [4] K. Sera, T. Yanagisawa, H. Tsunoda, S. Hutatukawa, Y. Saitoh, S. Suzuki, H. Orihara, Int. J. PIXE 2 (1992) 325.
Comparison of elemental quantity by PIXE and ICP-MS and/or ICP-AES for NIST standards K. Saitoh a
c
a,*
, K. Sera b, T. Gotoh c, M. Nakamura
c
Environmental Research and Information Center of Akita Prefecture (ERICA), 3-1-1 Sanno, Akita 010-8572, Japan b Cyclotron Research Center, Iwate Medical University, 348 Tomegamori, Takizawa, 020-0173, Japan Measurement and Analysis Division, Tohoku Afforestation and Environmental Protection Co., Ltd., 3-8-22, Sakuragi, Tagajo 985-0842, Japan
Abstract Urban particulate matter (SRM 1648), Buffalo River sediment (SRM 2704) and pine needle (SRM 1575) standard reference materials prepared by the National Institute of Standards and Technology (NIST, USA) were analyzed by three multi-element analysis methods, i.e., particle induced X-ray emission (PIXE), inductively coupled plasma-mass spectrometry (ICP-MS) and inductively coupled plasma-atomic emission spectrometry (ICP-AES); values determined by those analysis methods were compared with certified and/or non-certified values of NIST samples. Values determined by PIXE were 70–120% relative to certified and/or non-certified values of NIST samples except for Co in the urban particulate matter, for V and Co in Buffalo River sediment and for Ni and Br in the pine needles samples. In particular, Al, K, Ca, Cr, Mn, Fe, Cu, Zn and Pb were 85–110% in all samples. On the other hand, Na and Fe values determined by ICP-MS were very much different from the certified values in all samples, but the other elements were 70– 120%. As for ICP-AES, all elements except for Na were 80–100% in all samples. Comparing the values determined by PIXE and those determined by ICP-MS and/or ICP-AES, there was a slight difference between the samples, but the range was 75–120% except for Na, V, Fe and Co determined by ICP-MS and Na determined by ICP-AES, which was generally consistent with PIXE. Ó 2002 Elsevier Science B.V. All rights reserved. PACS: 32.30.Rj; 82.80.Ej Keywords: Multi-element analysis; Major-to-ultratrace elements; NIST sample; PIXE; ICP-MS; ICP-AES
1. Introduction Since particle induced X-ray emission (PIXE) allows analysis of minute samples smaller than a milligram without any complex chemical manipulation, and since it not only simultaneously detects
*
Corresponding author. Tel.: +81-18-860-4013; fax: +81-18860-4016. E-mail address: [email protected] (K. Saitoh).
elements from Na to U in a short time but also detects major-to-ultratrace elements at the concentration level of ppm to sub-ppb, it is being used in a variety of fields, including environmental research, medicine, geology and archeology, and PIXE analysis is fast becoming a universal method for highly sensitive analysis of multiple elements. Meanwhile, multi-element analysis by means of inductively coupled plasma-mass spectrometry (ICP-MS) and inductively coupled plasma-atomic emission spectrometry (ICP-AES) is
0168-583X/02/$ - see front matter Ó 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 5 8 3 X ( 0 1 ) 0 1 0 1 2 - 6
K. Saitoh et al. / Nucl. Instr. and Meth. in Phys. Res. B 189 (2002) 86–93
being performed on a daily basis by many research and analysis institutions. In the field of environmental research, ICP-MS and ICP-AES have become synonymous with multi-element analysis. Comparison of values determined by PIXE with those determined by ICP-MS and/or ICP-AES is important in the field of environmental research in terms of evaluating values determined by PIXE. Consequently, we compared elemental concentrations determined by PIXE with those determined by ICP-MS and/or ICP-AES using urban particulate matter, Buffalo River sediment and pine needles of standard reference material prepared by the National Institute of Standards and Technology (NIST, USA).
2. Experiment and method 2.1. Samples Samples were standard reference material (SRM) prepared by the NIST. The urban particulate matter sample was SRM 1648. The Buffalo River sediment sample was SRM 2704. The pine needle sample was SRM 1575. SRM samples were stored in a clean box until quantitative analysis by PIXE, ICP-MS and ICP-AES.
87
oven [3], where nearly 50 mg of the samples was put in a polytetrafluoroethylene (PTFE) vessel together with 1 ml of nitric acid, and Indium solution (1000 mg/l) was added as an internal standard in an amount of 1000 lg/g. This PTFE vessel was tightly sealed with a polypropylene jacket with a hand, and then it was heated for 2 min at low power corresponding to 200 W. Thin targets were prepared in a clean bench by taking 5 ll of it using a pipette and dropping it onto a 4 lm-thick polypropylene film, after which it was dried in clean air. Five such thin targets were prepared by decomposing 50 mg of each sample with nitric acid using a microwave oven in order to confirm reproducibility. Elemental concentrations in these thin targets were determined by PIXE at Nishina Memorial Cyclotron Center (NMCC), Japan Radioisotope Association, using 2.9 MeV protons from a baby cyclotron [4]. Beam currents, the accumulated charge and the typical measuring time were 20–40 nA, 11–19 lC, and 10–5 min, respectively. X-ray spectra were analyzed using the S A P I X program [4]. Quantitative analysis of elemental concentrations was performed based on a powdered internal standard method [1,2] for the urban particulate matter and the Buffalo River sediment samples, and on an internal standard method [3] for the pine needle sample.
2.2. Quantitative analysis by PIXE The urban particulate matter and Buffalo River sediment samples were prepared using the method developed by Sera and Futatsugawa [1,2], where 100 mg of sample was ground into fine powder in an agate mortar, and Palladium-carbon powder (5% Pd) was added as an internal standard in an amount of 10 mg, after which they were mixed uniformly in the agate mortar. Thin targets were prepared by taking 0.1 mg of the powder and putting it onto a 4 lm-thick polypropylene film, and dropping roughly 1 ll of 10% collodion solution diluted with ethyl alcohol onto it for fixing and smoothing. This process was carried out in a clean bench. Five such thin targets were prepared for each sample in order to confirm reproducibility. The pine needle samples thus prepared were decomposed with nitric acid using a microwave
2.3. Quantitative analysis by ICP-MS and ICPAES The urban particulate matter, Buffalo River sediment and pine needle samples were decomposed with nitric acid using a microwave oven, where nearly 50 mg of the sample was put in a dedicated PTFE vessel of the microwave oven together with 5 ml of nitric acid. This PTFE vessel was heated under the following conditions: 5 min (250 W), 5 min (400 W), 10 min (500 W) and 5 min (600 W), after which the PTFE vessel was reheated under the same conditions 15 min later. The decomposed solution was transferred through filter paper (5B, ADVANTEC Ltd., Japan) into a 100 ml calibrated flask, and the volume was adjusted to 100 ml with ultra-pure water. For ICPMS (Yokogawa Analytical Systems HP 4500,
88
K. Saitoh et al. / Nucl. Instr. and Meth. in Phys. Res. B 189 (2002) 86–93
Japan) measurement, this decomposed solution and solution diluted 10-fold with ultra-pure water were analyzed. For ICP-AES (Varian VISTA, USA) measurement, this decomposed solution only was analyzed. Five analysis solutions were prepared by decomposing 50 mg of each sample with nitric acid using a microwave oven in order to confirm reproducibility. The number of elements determined by ICP-MS was 28: Li, Be, Na, Mg, Al, Ca, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, As, Se, Rb, Sr, Ag, Cd, Cs, Ba, Tl, Pb, Bi, Th and U, and that determined by ICPAES was nine: Na, Mg, Al, K, Ca, Cr, Mn, Fe and Zn. The instrumental components and typical operating conditions for ICP-MS and ICP-AES are listed in Table 1. These operating conditions were selected after optimization of each parameter. Elemental concentrations were determined based on calibration curves generated from analysis of the continuing calibration standard solutions. The standard solution was multi-element 10 ppm (5% nitric acid base) mixed standard of SPEX Industries Co. The concentration of calibration standard solutions for ICP-MS was zero (ultra-pure water), 0.005, 0.02, 0.1, 0.5, 2, 10, 50 and 200 lg/l, and that for ICP-AES was zero (ultra-pure water), 0.005, 0.02, 0.1, 0.5, 1 and 2 mg/l. For generation of calibration curves, determination of each concentration was repeated three times.
3. Results and discussion 3.1. X-ray spectra analysis and reproducibility for PIXE Peaks of characteristic X-rays from seven elements (Na, Mg, Al, Si, P, S and Cl) were observed in a spectrum obtained with a Si (Li) X-ray detector without an absorber over the low X-ray energy range for all of the samples. In a spectrum from a Si (Li) detector with a 300 lm thick Mylar absorber, peaks of characteristic X-rays from thirteen elements (K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, Ga, Sr and Pb) could be observed for all of the samples. In addition to those element peaks, peaks of Co, As, Se, Br and Zr, peaks of Co, Rb and Zr and peaks of Br and Rb were observed for
Table 1 Operating conditions for ICP-MS and ICP-AES instrument ICP-MS
HP 4500 (Yokogawa Analytical Systems)
Plasma conditions Rf frequency Rf power
27.12 MHz 1.25 kW
Gas flow rate Carrier gas Auxiliary gas Coolant gas
Ar 1.0 l/min Ar 1.1 l/min Ar 15 l/min
Sampling conditions Sampling depth Sampling cone Skimmer cone Nebulizer Sampling uptake rate Data acquisition
7 mm from work coil Platinum, 1.0 mm orifice diameter Platinum, 0.4 mm orifice diameter Cross-flow type 0.4 ml/min
Data point Dwell time Integration Repetition
Multi-element mode by peak hopping 3 points/peak 20 ms/point 100 times 3 times
ICP-AES
VISTA (Varian)
Plasma conditions Rf frequency Rf power
40 MHz 1.0 kW
Gas flow rate Carrier gas Auxiliary gas Coolant gas
Ar 0.7 l/min Ar 1.5 l/min Ar 15 l/min
Sampling conditions Nebulizer Sampling uptake rate
Glass concentric type 1.0 ml/min
Spectrometer conditions Polychromator Focal length Grating Entrance slit width
Echelle polychromator 40 cm 94.74 grooves/mm 25 lm
Data acquisition Integration time Repetition
10 s 3 times
the urban particulate matter, the Buffalo River sediment and for pine needles, respectively. Errors for the analytical results are mainly from the spectrum fitting, the detection efficiency and values
K. Saitoh et al. / Nucl. Instr. and Meth. in Phys. Res. B 189 (2002) 86–93
of X-ray transmission through the absorber. Estimated relative error (100 error/count, in %) was smaller than 1% for three elements (Si, Ca and Zn), 1–5% for eight elements (Al, S, Ti, Cr, Mn, Fe, Cu and Pb), 5–10% for six elements (Mg, P, K, Co, Br and Sr) and 10–20% for three elements (Na, Cl and Ni). Repeated analysis of the same samples revealed that in urban particulate matter elements with a coefficient of variation (CV) of less than 10% were K, Ca, Fe, Cu, Se, Br, Sr and Pb, those with a CV of 10–15% were Na, Al, Cl, Mn, Co, Zn, Ga, As and Zr and those with a CV of 15–20% were Mg, Si and Ti. In Buffalo River sediment, elements with a CV of less than 10% were Mn, Fe, Zn and Pb, those with a CV of 10–15% were Na, Mg, P, S, V, Cr, Co, Ga, Rb and Sr and those with a CV of 15– 20% were Ca and Ti. In pine needles, elements with a CV of less than 10% were P, S, K, Ca, Mn, Fe and Sr and those with a CV of 10–15% were Mg, Zn and Pb. 3.2. Reproducibility for ICP-MS and ICP-AES The determination limits and concentration ranges of ICP-MS and ICP-AES were investigated from the reproducibility (relative standard deviation) of calibration standard solutions and linearity of calibration curves. Repeated analysis of the same samples by ICP-MS revealed that the CV of all elements was less than 10% except for Be, Ca, Ag, Bi and Th in urban particulate matter, Na and Fe in Buffalo River sediment and Na, V, Zn, Ag, Cd and Pb in pine needles. Elements with a CV of 10–15% were Be and Na in urban particulate matter and Fe in Buffalo River sediment. Elements with a CV of 15–20% were Ag and Th in urban particulate matter and Zn and Pb in pine needles. On ICP-AES, the CV of all elements in all samples was less than 10%. 3.3. Comparison of certified and/or non-certified values to values determined by PIXE, ICP-MS and ICP-AES Values determined by PIXE, ICP-MS and ICPAES and NIST certified and non-certified values
89
are shown in Tables 2–4. Elemental concentrations determined by PIXE were 70–120% relative to NIST certified and/or non-certified values except for Co in the urban particulate matter, V and Co in the Buffalo River sediment and Ni and Br in the pine needle samples. In particular, Mg, Al, K, Ca, Cr, Mn, Fe, Cu, Zn, Rb and Pb were 85–110% in all samples. In the case of PIXE, when determining V it is necessary to subtract the contribution of Ti Kb from the V Ka peak because V Ka overlaps with Ti Kb. The V Ka peak in urban particulate matter was clear and there was no major error when isolating it from Ti Kb. V was 76.4% relative to NIST certified and/or non-certified values. In contrast, in the case of Buffalo River sediment, the values determined by PIXE may have been considerably lower than the certified values because we overestimated the yield of the Ti Ka peak due to the influence of the escape peak of Fe, which overlaps with Ti Ka, resulting in overestimation of the contribution of Ti Ka from the V Ka peak. Regarding Co, because Co Ka closely resembles Fe Ka, when the Fe content is several thousand times greater than the Co content, as was the case with urban particulate matter and Buffalo River sediment, the Co Ka estimated relative error will be several 10 or several 100 percent higher even if the Fe Ka estimated relative error resulting from spectral analysis is 1%. The determined Co value may have been 10–20 times the certified value for this reason. Non-certified values for rare earth elements and heavy elements such as Cd and U are indicated for standard NIST samples, but these elements were not detected by PIXE. This may be because L-X rays of rare earth elements overlap with transition elements of K-X rays such as Fe and Cu or because the detection sensitivity toward rare earth elements and heavy elements is low under NMCC PIXE measurement conditions. On the other hand, in a comparison of values determined by ICP-MS and NIST certified and/or non-certified values, there were slight differences depending on the sample, but elements with a value 70–120% relative to certified and/or noncertified values were the 18 elements Mg, Al, Ca, V, Cr, Mn, Co, Ni, Cu, Zn, As, Rb, Sr, Ag, Cd, Cs, Ba and Pb. Values for Na and Fe were quite
90
K. Saitoh et al. / Nucl. Instr. and Meth. in Phys. Res. B 189 (2002) 86–93
Table 2 Elemental quantities by PIXE, ICP-MS and ICP-AES for urban particulate matter (SRM 1648) of NIST sample Determined value Element
PIXE (b)
ICP-MS (c) a
Mean SD Li (lg/g) Be (lg/g) Na (%) Mg (%) Al (%) Si (%) P (%) S (%) Cl (%) K (%) Ca (%) Sc (lg/g) Ti (%) V (lg/g) Cr (lg/g) Mn (lg/g) Fe (%) Co (lg/g) Ni (lg/g) Cu (lg/g) Zn (%) Ga (lg/g) As (lg/g) Se (lg/g) Br (lg/g) Rb (lg/g) Sr (lg/g) Zr (lg/g) Ag (lg/g) Cd (lg/g) In (lg/g) Sb (lg/g) I (lg/g) Cs (lg/g) Ba (lg/g) La (lg/g) Ce (lg/g) Sm (lg/g) Eu (lg/g) Hf (lg/g) W (lg/g) Tl (lg/g) Pb (%) Bi (lg/g) Th (lg/g) U (lg/g) a b
0.436 0.062 0.623 0.114 3.32 0.38 9.38 1.76 0.580 0.123 6.00 1.35 0.417 0.054 1.00 0.08 6.41 0.63 0.430 0.066 107 24 341 91 804 94 4.12 0.37 187 28 64.9 12.3 575 56 0.475 0.053 55.0 6.0 87.3 9.5 18.9 1.9 395 35
b/a (%)
102.6 77.9 97.1
Mean SD
10.2 0.2 0.90 0.13 0.134 0.008 0.631 0.067 2.76 0.06
c/a (%)
31.5 78.9 80.7
75.9 70.0 79.0
100.1
Mean and SD of five measurements. Indication of parentheses is non-certified value.
91.0 1.8 236 5 703 22 11.8 5.9 12.6 0.4 62.0 1.3 534 16.3 0.658 0.047 34.5 0.9 105 3 12.9 0.6
65.0 58.6 81.7 301.8 70.0 75.6 87.7 138.2 91.3 47.8
20.2 0.4 144 3
38.8
5.7 1.0 65.1 1.3
95.0 86.8
1.38 0.07 568 12
46.0 77.1
1.81 0.09 0.701 0.015 5.00 1.61 4.67 0.95 3.10 0.08
d/a (%)
0.183 0.016 0.723 0.067 2.76 0.27
43.1 90.4 80.7
0.425 0.002 (0.8)b 3.42 0.11
85.0
(5.0) (0.45) 1.05 0.01
a
0.893 0.103 5.77 0.52
12.6 1.4 107.5 76.4 84.6 93.5 105.4 1038.9 79.1 94.4 99.8
Mean SD
Certified value and/or non-certified value (a)
ICP-AES (d)
120.0 92.7 95.2
193 19 164 18
0.656 0.023
a
107.0 63.1 56.4
380 30 690 60 3.56 0.34
94.3 80.2 91.0
0.395 0.040
83.0
(7) (0.40) 140 3 403 12 (860) 3.91 0.10 (18) 82 3 609 27 0.476 0.014 115 10 27 1 (500) (52)
(6) 75 7 (1.0) (45) (20) (3) (737) (42) (55) (4.4) (0.8) (4.4) (4.8) 0.655 0.008 (7.4) 5.5 0.01
K. Saitoh et al. / Nucl. Instr. and Meth. in Phys. Res. B 189 (2002) 86–93
91
Table 3 Elemental quantities by PIXE, ICP-MS and ICP-AES for Buffalo River sediment (SRM 2704) of NIST sample Determined value Element
PIXE (b)
ICP-MS (c) a
Mean SD Li (lg/g) Be (lg/g) Na (%) Mg (%) Al (%) Si (%) P (%) S (%) Cl (%) K (%) Ca (%) Sc (lg/g) Ti (%) V (lg/g) Cr (lg/g) Mn (lg/g) Fe (%) Co (lg/g) Ni (lg/g) Cu (lg/g) Zn (lg/g) Ga (lg/g) As (lg/g) Se (lg/g) Br (lg/g) Rb (lg/g) Sr (lg/g) Y (lg/g) Zr (lg/g) Ag (lg/g) Cd (lg/g) Sn (lg/g) Sb (lg/g) I (lg/g) Cs (lg/g) Ba (lg/g) La (lg/g) Ce (lg/g) Sm (lg/g) Eu (lg/g) Dy (lg/g) Lu (lg/g) Hf (lg/g) Hg (lg/g) Tl (lg/g) Pb (lg/g) Bi (lg/g) Th (lg/g) U (lg/g) a b
b/a (%)
0.400 0.052 1.08 0.16 6.42 1.46 26.1 6.0 0.098 0.013 0.350 0.037 0.057 0.018 1.80 0.41 2.64 0.41
73.1 90.0 105.1 89.7 98.2 88.2
0.461 0.073 43.2 5.0 146 16 479 39 4.51 0.34 340 50 31.9 13.4 91.8 11.7 450 17 12.0 1.7
100.9 45.6 108.1 86.3 109.7 2428.6 72.3 93.1 102.7 80.0
90.0 101.5
198 22 110 13
198.0 84.6
305 111
101.7
192 16
119.2
a
Mean SD
ICP-AES (d) c/a (%)
a
Mean SD
d/a (%)
32.7 1.8 0.97 0.09 0.028 0.008 0.913 0.037 7.00 0.61
5.1 76.1 114.6
0.081 0.001 1.13 0.01 4.83 0.03
14.8 94.2 79.0
3.70 0.39
142.3
1.67 0.02 2.41 0.02
83.5 92.7
77.8 2.9 112 8 540 22 20.3 2.4 11.3 0.4 31.1 1.4 80.1 3.8 376 11 19.9 0.8 17.2 0.4 Not detected
81.9 83.0 97.3 493.9 80.7 70.5 81.2 85.8 132.7 73.5
120 3 550 7 4.01 0.03
88.9 99.0 97.6
Certified value and/or non-certified value (a) 47.5 4.1
68.8
440 10
100.4
0.547 0.014 1.20 0.02 6.11 0.16 29.08 0.13 0.0998 0.0028 0.397 0.004 (<0.01)b 2.00 0.04 2.60 0.03 (12) 0.457 0.018 95 4 135 5 555 19 4.11 0.10 14.0 0.6 44.1 3.0 98.6 5.0 438 12 (15) 23.4 0.8 1.12 0.05 (7) (100) (130) (2.8) (300)
93.1 1.8 70.7 1.1
93.1 54.4
0.4 0.1 2.15 0.09
62.3
4.79 0.07 275 8
79.8 66.4
0.45 0.03 138 13 Not detected 5.22 0.46 1.44 0.06
42.8 85.7
3.45 0.22 (9.5) 3.79 0.15 (2) (6) 414 12 (29) (72) (6.7) (1.3) (6) (0.6) (8) 1.47 0.07 1.06 0.07 161 17
56.7 46.0
(9.2) 3.13 0.13
Mean and SD of five measurements. Indication of parentheses is non-certified value.
92
K. Saitoh et al. / Nucl. Instr. and Meth. in Phys. Res. B 189 (2002) 86–93
Table 4 Elemental quantities by PIXE, ICP-MS and ICP-AES for pine needles (SRM 1575) of NIST sample Determined value Element
PIXE (b)
ICP-MS (c) a
Mean SD Li (lg/g) Be (lg/g) Na (%) Mg (%) Al (lg/g) Si (%) P (%) S (%) Cl (lg/g) K (%) Ca (%) Sc (lg/g) Ti (%) V (lg/g) Cr (lg/g) Mn (lg/g) Fe (lg/g) Co (lg/g) Ni (lg/g) Cu (lg/g) Zn (lg/g) Ga (lg/g) As (lg/g) Se (lg/g) Br (lg/g) Rb (lg/g) Sr (lg/g) Ag (lg/g) Cd (lg/g) Sb (lg/g) Cs (lg/g) Ba (lg/g) La (lg/g) Ce (lg/g) Eu (lg/g) Hg (lg/g) Tl (lg/g) Pb (lg/g) Bi (lg/g) Th (lg/g) U (lg/g) a b
0.018 0.006 0.074 0.009 509 111 0.115 0.048 0.089 0.004 0.099 0.004 30.1 9.8 0.370 0.021 0.375 0.011 13.2 3.6 2.33 0.62 2.58 0.83 618 20 175 12 2.12 0.52 2.98 0.74 61.2 6.9 3.73 1.74
1.87 0.61 11.0 3.1 4.16 0.17
b/a (%)
93.4
a
Mean SD
Not detected Not detected 0.015 0.004 0.122 0.002 378 3
Mean SD
d/a (%)
Certified value and/or non-certified value (a)
0.015 0.001 0.100 0.001 434 8
79.8
545 30
ICP-AES (d) c/a (%)
69.4
a
0.12 0.02
74.2
100.0 91.5
99.2 91.6 87.5 60.6 99.3
20.8 94.0 86.7
0.346 0.013
0.19 0.08 1.8 0.1 704 8 120 5 Not detected 2.18 0.21 3.31 0.35 28.7 5.5 0.22 0.06 Not detected Not detected 9.34 0.22 3.34 0.05 3.0 1.2 0.08 0.05
84.4
69.2 104.3 60.0
0.342 0.003 0.393 0.007
92.4 95.8
<10 610 4 160 6
90.4 80.0
62.3 110.3
0.37 0.02 0.41 0.02 (0.03)b
2.6 0.2 675 15 200 10 (0.1) (3.5) 3.0 0.3
50 3 0.21 0.04
79.8 69.6
(9) 11.7 0.1 4.8 0.2 (<0.5) (0.2)
Not detected 5.06 0.04
9.39 1.22
86.9
Not detected 9.8 1.8 Not detected Not detected Not detected
90.7
(0.2) (0.4) (0.006) 0.15 0.05 (0.05) 10.8 0.5 0.037 0.003 0.020 0.004
Mean and SD of five measurements. Indication of parentheses is non-certified value.
different from the certified values in urban particulate matter and Buffalo River sediment in which these elements account for a percentage of the content. In pine needles, in which Na accounts
for mg/g of the content and Fe accounts for lg/g of the content, the Fe value was 60% relative to the certified value, and in the case of Na, for which certified and non-certified values are not
K. Saitoh et al. / Nucl. Instr. and Meth. in Phys. Res. B 189 (2002) 86–93
indicated, the value was not much different from the value determined using PIXE. As this demonstrates, the compatibility of Na and Fe values with certified values differed greatly depending on the Na and Fe concentrations in samples. Values determined by ICP-AES were 80–100% of certified and/or non-certified values for all elements except Na in all samples. Na values tended to be the same as those determined by ICP-MS. 3.4. Comparison of PIXE to ICP-MS and ICPAES for elemental quantity Comparing values determined by PIXE with those determined by ICP-MS, elements for which PIXE values were not within the range of 80–120% relative to ICP-MS values were Na, Cr, Fe, Co, Zn and Se in urban particulate matter and Na, Ca, V, Fe, Co, Ga and Sr in Buffalo River sediment. Of these elements, PIXE values for V and Co and ICP-MS values for other elements were very different from NIST certified and/or non-certified values. In a comparison of values determined by PIXE and those determined by ICP-AES, ICPAES values for all elements except for Na in urban particulate matter and Buffalo River sediment were 75–120% relative to PIXE values. There was a large discrepancy between ICP-MS and ICPAES Fe values.
93
4. Conclusion We compared values determined by PIXE, ICPMS and ICP-AES using standard NIST urban particulate matter (SRM 1648), Buffalo River sediment (SRM 2704) and pine needle (SRM 1575) samples. We found that although reproducibility of PIXE was somewhat lower than that of ICPMS and ICP-AES, except for V and Co, values determined by PIXE were consistent with NIST certified and/or non-certified values compared to those determined by ICP-MS and ICP-AES. V and Co values determined by PIXE were quite different from the certified and/or non-certified values because of spectral analysis, so improving the spectral analysis program may solve this problem. Compared to ICP-MS and ICP-AES, PIXE can be used to precisely and efficiently analyze many elements, so it may become the most effective means of analysis in environmental studies in the near future.
References [1] K. Sera, S. Futatsugawa, Int. J. PIXE 8 (1998) 185. [2] K. Sera, S. Futatsugawa, D. Ishiyama, Int. J. PIXE 9 (1999) 63. [3] S. Futatsugawa, S. Hatakeyama, S. Saitou, K. Sera, Int. J. PIXE 3 (1993) 319. [4] K. Sera, T. Yanagisawa, H. Tsunoda, S. Hutatukawa, Y. Saitoh, S. Suzuki, H. Orihara, Int. J. PIXE 2 (1992) 325.