ARTICLE pubs.acs.org/ac
Comparison of Multiple Prompt γ-Ray Analysis and Prompt γ-Ray Analysis for the Elemental Analysis of Geological and Cosmochemical Samples Mohammad Amirul Islam,† Mitsuru Ebihara,*,† Yosuke Toh,‡ and Hideo Harada‡ † ‡
Graduate School of Science and Engineering, Tokyo Metropolitan University, Hachioji, Tokyo 192-0397, Japan Research Group for Applied Nuclear Physics, Japan Atomic Energy Agency, Tokai, Ibaraki 319-1195, Japan ABSTRACT: Multiple prompt γ-ray analysis (MPGA) and conventional neutroninduced prompt γ-ray analysis (PGA) are nondestructive analytical methods for bulk chemical compositions, and their analytical capabilities were compared for elemental analyses of geological and cosmochemical samples. Detection sensitivities of PGA are often restricted by poor signal-to-noise ratios and interferences from different origins. MPGA can substantially reduce the background level, especially for hydrogenous samples, relative to PGA, which opens up a possibility to use lower energy prompt γ-rays of some trace elements. Although it is one of the major constituent elements of rock samples, Mg is hard to be determined by PGA. With MPGA, Mg contents could be determined with reasonable consistency with their corresponding recommended values in geological and cosmochemical samples by carefully selecting suitable coincident prompt γ-ray energy pairs without interference correction. MPGA was applied to a hydrogenous meteorite, Ivuna, which contains H at 2% mass level. MPGA detection limits for most of the elements studied can be reduced up to 1 order of magnitude when compared with PGA detection limits under the present experimental conditions.
eutron-induced prompt γ-ray analysis (PGA) is a highly sensitive, precise, multielemental, and nondestructive analytical method of elements and is used in various fields such as geology, archeology, cosmochemistry, industry, and environmental science. However, PGA has an inherently low signal-tonoise ratio primarily because of the large background associated with it.1 Therefore, difficulty of quantification arises when the γ-ray intensity from the trace element of interest is not strong enough in comparison with the background γ-ray from large amounts of other elements present in the sample. In such a situation, the multiple prompt γ-ray analysis (MPGA) method can be used to reduce background level.2,3 Since in MPGA only the γ γ coincident events from nuclei that simultaneously emit two or more cascade prompt γ-rays in de-excitation of the nuclei within a short interval of time are collected, the background caused by a single-γ-ray-emitting nuclide can be reduced. Among meteorites, Ivuna-type carbonaceous chondrites (CI) and some of Mighei-type chondrites (CM) are known to have high contents (about 2% in mass) of H. Elemental analysis of the CI group of meteorites is of special importance because they have essentially the same chemical composition as that of the sun except for a few extremely volatile elements.4 In PGA analyses of hydrogenous samples, the Compton-scattered γ-ray of H causes a considerable increase in the low-energy background of the spectrum.5 Therefore, MPGA can be used to analyze hydrogenous meteorites to reduce low-energy background caused by H in the sample. Hence, in MPGA, some trace elements could potentially be analyzed by using only the measurable lower energy prompt γ-ray which is not detectable in PGA. In analyzing geological and cosmochemical samples, X-ray fluorescence (XRF) and instrumental neutron activation analysis
N
r 2011 American Chemical Society
(INAA) are commonly used. Both XRF and INAA techniques are relatively simpler than PGA and MPGA. XRF is even simpler than INAA, but its application is limited to the surface analysis of solid samples because of the lower penetration ability of X-rays compared with neutrons and γ-rays. Therefore, XRF is not suitable for whole-rock analysis of heterogeneous solid samples. To overcome such a difficulty, compositionally homogeneous bead samples are to be prepared. Once the beads are prepared, the samples cannot be reused for other purposes. This can be a fatal defect in analyzing a limited amount of precious matter like meteorites. INAA can determine most major elements of silicate rocks, but Si, one of the most important constituent elements of silicate rock samples, is hardly determined. Another defect of INAA can be caused by the residual radioactivity in samples once irradiated with neutrons even in a short term. In PGA and MPGA, samples need not be physically decomposed before assaying and no significant amount of radioactivity remains in the samples after an appropriate cooling interval (normally few days). Although PGA is usable for determining most major elements simultaneously in rock samples, some elements cannot be always determined with high accuracy and precision due to known and unknown interferences from matrix elements and background sources.6 Magnesium (Mg) is one representative element for such a case. As, in MPGA, coincidence events caused by cascade prompt γ-rays from a nuclide are collected in a two-dimensional energy matrix, the spectral interferences from the other elements can essentially be avoided. In this study, a newly developed Received: July 2, 2011 Accepted: August 8, 2011 Published: August 08, 2011 7486
dx.doi.org/10.1021/ac201706g | Anal. Chem. 2011, 83, 7486–7491
Analytical Chemistry MPGA system at the Japan Atomic Energy Agency (JAEA) was used to analyze hydrogenous meteorites as well as rock samples and was characteristically compared with conventional PGA based on their analytical capabilities.
ARTICLE
Table 1. Some Characteristic Features of MPGA and PGA Systems at JAEAa MPGA
’ EXPERIMENTAL SECTION Samples and Standards. The Smithsonian Allende meteorite
powder sample (105.34 mg) (prepared by E. Jarosewich, the Smithsonian Institution; split/position = 22/6) and geological rock sample JB-1 (110.32 mg) (a basaltic standard rock sample issued by the Geological Survey of Japan (GSJ)) were analyzed repeatedly by MPGA in γ γ coincidence mode (4 MeV energy range). Analyses were repeated four times over half a year for JB1 and Allende with the same samples. The powder samples were sealed into thin FEP (fluorinated ethylene propylene) film bags. Analytical grade chemical reagents for all studied elements except Sm and Gd were used as reference standards for the comparison method. For Sm and Gd, recommended values of JB-1 from the literature, Imai et al.,7 were used. MPGA was also applied to some hydrogenous meteorites, Orgueil, Ivuna, Alais, Y-980115, B-7904, Y-793321, and Y-86720, as well as to a reference rock sample GSJ-JB-2 (mass range about 25 192 mg each). PGA was applied only to Orgueil, Ivuna, Allende, and GSJ-JB-2. MPGA. The MPGA irradiations of the samples were done for 2.5 4.4 h at a newly installed MPGA system by using guided cold neutrons (flux: 106 107 n cm 2 s 1) from the JRR-3 M research reactor of JAEA. The detailed descriptions on the developments of the neutron beamline and detector system for MPGA were reported elsewhere.8,9 The present detector system consists of eight clover Ge detectors with bismuth germanium oxide (BGO) Compton suppressors. The relative counting efficiency of each clover detector was approximately 120% relative to that of a 3 in. 3 in. NaI detector. The MPGA system at JAEA gains high γ-ray detection efficiency with the sophisticated clover Ge detector system in tight geometry and overcomes the weakness of the conventional coincidence technique with low detection efficiency. The energy calibration of the MPGA detector system was performed with the γ-rays emitted by 152Eu, 57Fe, and 66Cu. For dead time correction, the counts of the 81 and 356 keV coincident peak of 133Ba were used. The sample-to-detector distance was approximately 5 cm. Neutron beams were collimated to a size of 30 mm 20 mm at the position of sample irradiation, and the air in the beamline was replaced with helium/ carbon dioxide gas (3 L/min) to reduce the background γ-rays caused by neutron capture reactions with nitrogens. The neutron attenuators made up of acrylic plates were used to optimize the rate of true coincidence to the accidental coincidence by adjusting the neutron beam intensity. A blank sample (FEP film bag with sample holder) was also irradiated in every run for background correction. Events comprising a pair of coincident prompt γ-rays were collected in the experiment. The add-back sorting mode was used for offline sorting of the coincidence data. In MPGA, each clover detector is composed of four closely packed Ge crystals. In add-back mode, all output signals are summed up to obtain the total pulse height for an incident γ-ray that scatters from one crystal and is absorbed by one of the other three crystals, consequently increasing photo peak efficiency of a clover detector. With the sorted data, a histogram of the γ γ energy correlation, which was a two-dimensional γ-ray matrix spectrum, was constructed for elemental analysis.
PGA
neutron energy
cold
neutron flux (at sample position)
106 107 n cm
thermal
detector system
eight clover (Ge)
2
s
1
2.4 107 n cm
2
s
1
single (Ge) detector
detectors Compton suppression
BGO
BGO
detector efficiency (relative) 120%
24.5%
system
a
sample-to-detector distance
5 cm
24.5 cm
beam size (at sample box)
30 mm 20 mm
20 mm 20 mm
Refs 8 and 10.
PGA. The same samples and reagent standards measured in MPGA were used for the PGA irradiation. Irradiations were done for 2.0 4.4 h by using guided thermal neutrons (flux: 2.4 107 n cm 2 s 1) of JRR-3 M at JAEA. The powder/chip samples were sealed into thin FEP film bags. Neutron beams were collimated to size of 20 mm 20 mm at the entrance of the sample box, which was filled with He gas. The sample-to-detector distance was 24.5 cm. Prompt γ-ray was measured by a Ge detector (relative counting efficiency 23.5%) surrounded by a BGO Compton suppressor and coupled with a 16K channel pulseheight analyzer.10 Prompt γ-ray intensities measured on different days were normalized to an average count rate of the 341.7 and 1381.7 keV γ-rays of Ti, which was routinely measured with the PGA system to monitor the neutron flux fluctuation. Some basic characteristic features of the MPGA and PGA systems at JRR-3 M are tabulated in Table 1.
’ RESULTS AND DISCUSSION Compton Background Reduction in Hydrogenous Meteorites. MPGA and PGA were applied to the same specimen
of a hydrogenous meteorite, Ivuna. The measurement times (real time) for the same specimen in both methods were set to be the same (4.4 h). Figure 1 compares the prompt γ-ray spectra obtained by PGA and MPGA methods for the Ivuna CI meteorite. A conventional γ-ray spectrum is shown for PGA (Figure 1a), whereas a gated spectrum for the Cd 558.5 keV prompt γ-ray is shown for MPGA (Figure 1b). Usually, a twodimensional γ-ray spectrum is used for MPGA analysis. For the purpose of comparison with the PGA spectrum, Figure 1b shows the projection made by gating on the Cd 558.5 keV prompt γ-ray. As Ivuna contains about 2 mass % of H, its PGA spectrum (Figure 1a) shows a strong peak of H at 2223.3 keV energy. Compton scattering of this high-energy prompt γ-ray of H leads to the increase of Compton continuum in the spectrum, especially in the low-energy region. Although the BGO Compton suppression system was used in the conventional PGA system, the background level at the low-energy region is relatively higher than the level in the MPGA spectrum gated for the Cd 558.5 keV γ-ray. The Ivuna meteorite contains 0.683 μg/g of Cd.11 Due to the high Compton background and its close proximity to the Co peak at 555.9 keV, the Cd peak at 558.5 keV was not visible in the conventional PGA spectrum (Figure 1a). Since the γ-ray multiplicity of H is unity, the H peak as well as its influence in the 7487
dx.doi.org/10.1021/ac201706g |Anal. Chem. 2011, 83, 7486–7491
Analytical Chemistry
ARTICLE
Figure 1. Partial spectra of conventional PGA (a) and MPGA (b) for the Ivuna CI meteorite. The MPGA spectrum is a gated one for the Cd 558.5 keV prompt γ-ray. Both spectra were obtained using the same specimen of 191.7 mg and counting time of 4.4 h. In the expanded spectrum of MPGA, a Cd peak at 558.5 keV became clearly visible mainly because the background level is reduced in MPGA compared with that in PGA.
Figure 2. Part of the PGA spectra for GSJ-JB-1 (a) and Allende (b). The energy range from 2820 to 2840 keV is shown. The PGA peak at 2828.2 keV, which is normally used for the Mg determination, is interfered by peaks of Fe (2832.5 keV) and N (2830.8 keV).
Table 2. Average Mg Contents (in %) for JB-1 and Allende Obtained by Seven Intense Coincidence Prompt γ-Ray Pairs JB-1 (n = 4) coincidence prompt γ-ray in keVa
a c
average (%)
Allende (n = 4)
relative to recommended value
average (%)
relative to recommended value
2828.2(75.7) 585.1(100)
4.64 ( 0.27
1.00 ( 0.06
13.8 ( 0.3
0.93 ( 0.02
1129.6(46.0) 1808.7(93.0)
4.29 ( 0.64
0.92 ( 0.14
13.4 ( 0.8
0.91 ( 0.05
389.7(13.6) 585.1(74.8)
4.69 ( 0.54
1.01 ( 0.12
13.0 ( 0.8
0.88 ( 0.05
3054(19.6) 389.7(13.6)
4.17 ( 0.42
0.90 ( 0.09
12.4 ( 0.9
0.84 ( 0.06
1003.3(8.2) 1129.6(46.0)
5.37 ( 0.96
1.16 ( 0.21
13.4 ( 1.2
0.91 ( 0.08
2438.5(10.6) 389.7(13.6)
3.98 ( 0.56
0.86 ( 0.12
13.6 ( 0.8
0.92 ( 0.05
1003.3(8.2) 1808.7(93.0)
3.56 ( 1.08
0.76 ( 0.23
13.8 ( 0.4
0.93 ( 0.03
recommended values
4.65b
14.8c
Nuclear data (values within parentheses are γ-ray intensities, %) from the Web sites listed in refs 12 and 13. b Recommended value for JB-1 (ref 7). Recommended value for Allende (ref 14).
background continuum was reduced in MPGA as noticed in Figure 1b. The removal of such spectral interference mainly
caused by H in MPGA opens up a possibility of determining some elements emitting prompt γ-rays in the low-energy region 7488
dx.doi.org/10.1021/ac201706g |Anal. Chem. 2011, 83, 7486–7491
Analytical Chemistry
ARTICLE
of the spectrum such as Cd (558.5 keV), Co (277.1 and 555.9 keV), and Sm (333.9 keV). As shown in Figure 1b, a trace Cd peak is clearly visible in the MPGA spectrum gated by the Cd 558.5 keV γ-ray for the Ivuna meteorite sample. Selection of the Coincidence Energy Pair in MPGA. In PGA, intense prompt γ-ray peaks of elements of interest are sometimes interfered by γ-rays from matrix elements or background sources. For such a case, any interference can principally be eliminated if proper correction is performed. However, even with careful interference correction, high accuracy and precision cannot be always guaranteed for some elements. In contrast, MPGA can avoid any spectral interference in principle because coincidence events caused by the cascade prompt γ-ray from each element are collected in a two-dimensional energy matrix and can be completely separated from each other. As a typical example, Mg data for geological and cosmochemical samples are testified for PGA and MPGA here. In this study, the consistency of Mg data using MPGA for different combinations of coincidence energy pairs is evaluated by repeated analysis of JB-1 and Allende powder. For the quantification of Mg by PGA, the most intense peak of 585.1 keV is commonly used. This peak is interfered by the 583.6 keV peak of F, which is a major constituent element of the structure material of the PGA system and sample holder at JAEA. Magnesium has another intense peak of 2828.2 keV, which is also interfered by Fe (2832.5 keV) and N (2830.8 keV) as shown in Figure 2. Although He gas flow was performed to replace air in the sample box, a trace N could always be measured. In MPGA, Mg can be determined by using several coincidence prompt γ-ray energy pairs (with decay time being less than 100 ns). Table 2 lists such pairs of prompt γ-rays with relatively higher emission probabilities. The average Mg contents and standard deviations
(1σ) for these seven coincidence pairs were calculated and summarized in Table 2. One intense coincidence energy peak of 3054 and 585.1 keV of Mg is not considered here as it is overlapped with the 3051.4 583.6 keV intense peak of F. As indicated by ratios close to unity for relative values to a recommended one in Table 2, Mg can be reliably determined by MPGA within 10% deviation from the recommended value for JB-1 by using any coincidence prompt γ-ray pairs except the 1003.3 1129.6, 2438.5 389.7, and 1003.3 1808.7 keV energy pairs. Among the remaining four coincidence pairs, the 2828.2 585.1 keV pair yields the highest consistency with the recommended value of Mg and relatively small standard deviation (6%; 1σ). For the Allende meteorite, which contains a few times higher Mg content compared with crustal rock samples like JB-1, most of the coincidence pairs chosen in Table 2 yielded reasonable values of Mg except for the 3054 389.7 keV energy pair. Among the remaining six coincidence energy pairs, the 2828.2 585.1 and 1003.3 1808.7 keV pairs give more consistent values with a recommended value of Mg for Allende than the other pairs, with the 2828.2 585.1 keV pair having higher intensity and lower standard deviation (2%; 1σ). Thus, it is proposed for the 2828.2 585.1 keV coincidence pair to be used for the determination of Mg by MPGA for geological and cosmochemical rock samples. Mg Contents of Some Hydrogenous Meteorites and a Rock Sample. Mg contents of samples of four CI, three CM chondrites, and one basaltic rock were determined by MPGA. Among the CI meteorites, two meteorites, Orgueil and Ivuna, have been extensively analyzed by many researchers by using different techniques. The Mg contents of the samples determined with the proposed 2828.2 585.1 keV coincidence peak by MPGA along with available literature data are given in Table 3. All the Mg contents determined by MPGA are in good agreement with literature data. Although all coincidence energy pairs give systematically lower values of Mg relative to the recommended value for Allende as mentioned above (Table 2), no such systematically lower values of Mg were obtained for CI and CM chondrites by MPGA as shown in Table 3. The Allende powder may have inhomogeneity in Mg-bearing minerals, or a Mg value for Allende may need to be revised. Comparison of Some Trace Element Data between MPGA and PGA. PGA can determine most major and some minor elements in geological and cosmochemical samples, but it is not so suitable for the determination of trace elements in rock samples as INAA. This is partly due to poor signal-to-noise ratios for PGA. As signal-to-noise ratios can principally be enhanced in MPGA compared with PGA, analytical capability was tested by applying MPGA and PGA to some hydrogenous meteorites and a geological rock sample, and analytical results for several trace elements (Cl, Sm, and Gd) are compared in Table 4,
Table 3. Mg Contents of CI and CM Meteorites and a Rock Sample Determined by MPGA Using a Coincidence Prompt γ-Ray Pair of 2828.2-585.1 keVa sample names
type
this work (%)
literature values (%)
N
Orgueil
CI1
9.01 ( 0.26
9.59 ( 0.44b
17
Ivuna
CI1
8.74 ( 0.19
9.71 ( 0.065b
3
Alais
CI1
9.54 ( 0.24
9.42 ( 0.33b
Y-980115
CI1
10.5 ( 0.3
10.87c
B-7904
CM2
13.2 ( 0.4
13.6d
Y-793321
CM2
12.5 ( 0.4
12.25c
Y-86720
CM2
12.8 ( 0.4
13.9c
GSJ-JB-2
basaltic rock
2.87 ( 0.05
3 15
2.79e
a
Uncertainties with this work’s values are due to counting statistics (1σ); N stands for the number of a literature data. b Lodders (ref 11). c Yanai and Kojima (ref 15). d Choe et al. (ref 16). e Imai et al. (ref 7).
Table 4. Analytical Results of Some Trace Element Contents Using MPGA and PGA Cl (μg/g)
Sm (μg/g)
Gd (μg/g)
sample name MPGA (786.3 1164.9)a PGA (1164.9) lit. value MPGA (333.9 439.4) PGA (333.9) lit. value MPGA (181.8 781.6) PGA (181.8) lit. value
a c
Allendeb
290 ( 10
336 ( 17
316
0.319 ( 0.034
c
0.34
0.405 ( 0.071
0.416 ( 0.065
0.42
Orgueild
781 ( 25
778 ( 18
700
0.133 ( 0.039
c
0.145
0.190 ( 0.082
0.204 ( 0.022
0.201
Ivunad
625 ( 19
784 ( 26
724
0.138 ( 0.024
c
0.143
0.160 ( 0.050
0.248 ( 0.052
0.187
JB-2e
284 ( 7
305 ( 20
281
2.43 ( 0.10
2.33 ( 0.084
2.31
3.31 ( 0.28
3.07 ( 0.07
3.28
Values within parentheses are the prompt γ-ray energies in keV; uncertainties are due to counting statistics. b Lit. values: Jarosewich et al. (ref 14). No meaningful value was obtained. d Lit. values: Lodders (ref 11). e Lit. values: Imai (ref 7). 7489
dx.doi.org/10.1021/ac201706g |Anal. Chem. 2011, 83, 7486–7491
Analytical Chemistry
ARTICLE
where literature data of these elements are also included. The uncertainties associated are due to counting statistics from the single measurement. In PGA, Sm and Gd contents are determined by using low-energy γ-rays 333.9 and 181.8 keV, respectively. No meaningful peaks were observed for lower contents of Sm in meteorites (0.143 0.34 μg/g) by PGA, whereas Sm contents in the same specimens could be determined by MPGA. Both MPGA and PGA show similar accuracy for Cl and Gd determinations in meteorites as well as in the geological rock sample. Background Counts and Detection Limits in PGA and MPGA. In PGA, background counts associated with peaks of target elements mainly depend on chemical compositions of coexisting elements in the sample. Other sources of background include γ-rays from the natural background, surrounding materials including the sample holder, and detector neutron activation. In MPGA, noncoincident signals from background can be
Figure 3. Partial two-dimensional matrix of MPGA of the Ivuna meteorite around the coincident peak 898 352 keV of 57Fe. Peak area and different background areas are indicated.
eliminated in principle. However, even in MPGA, in addition to the true coincidence event from cascade prompt γ-rays of a nuclide of interest, several chance coincident events contribute to background counts. The random coincidence events caused by two γ-rays from independent events or coincident events of Compton-scattered prompt γ-rays from matrix elements cause background counts, which may be named simple background. There is another type of background called line background, which is constituted by coincidence events of any one of the prompt γ-rays of a target nuclide of interest and a Comptonscattered higher energy cascade γ-ray of the same nuclide. It is socalled as appearing in a horizontal or vertical line over the coincidence peak. Figure 3 shows a partial two-dimensional spectrum of MPGA for the Ivuna meteorite sample, where two types of backgrounds along with a coincident peak of 57Fe at 352.4 and 898.6 keV can be observed. In the case of 57Fe, another intense prompt γ-ray of 1260.6 keV is also emitted as a cascade prompt γ-ray. The line background nos. 2 and 4 shown in Figure 3 are caused from coincident events of 352.4 keV and Compton-scattered γ-rays of 898.6 and 1260.6 keV prompt γrays. Background line nos. 1 and 3 can also be caused when one γ-ray of 898.6 keV and a Compton-scattered γ-ray from 352.4 and 1260.6 keV prompt γ-rays are coincidentally counted. Detection limits of several elements for both MPGA and PGA were calculated for the Ivuna meteorite sample (191.7 mg) under the present experimental conditions and compared in Table 5. The measuring time was set as 4.4 h for both PGA and MPGA. Here, the detection limit was defined as a value corresponding to 3σ of the background counts for the peak area. In multiparameter coincident spectrometry, the detection limit of an element is sometimes calculated from the area of the simple background18 as shown in Figure 3. In this study, however, the detection limit for MPGA was defined in the same as that for PGA; the detection limit was calculated from line background on the two-dimensional spectrum. The calculated detection limits are compared in Table 5. For most elements, it was confirmed that detection limits by MPGA are 1 order of magnitude lower than those by PGA.
Table 5. Comparison of Detection Limits (DL) of Some Elements for the Ivuna Meteorite (191.7 mg) in PGA and MPGA with 4.4 h of Irradiation element
unit
PGA energy (keV)a
PGA DL
MPGA coincidence energy pair (keV)a
MPGA DL
PGA/MPGA
Ti
%
1381.7(85.5)
0.0121
341.7(24.8) 1381.7(85.5)
0.00138
8.8
K
%
770.3(43.0)
0.0364
770.3(43.0) 1158.9(7.80)
0.00911
4.0
S
%
841(68.0)
0.0180
6.1
Mn
%
314.4(9.39)
0.0357
271.2(5.70) 104.6(8.40)
0.00908
3.9
Ca
%
1942.6(88.5)
0.179
2001.6(18.9) 1942.6(88.5)
0.0376
Mg
%
2828.2(56.7)
1.12
2828.2(56.7) 585.1(74.8)
0.0606
Ni Fe
% %
464.9(27.4) 1725.3(6.30)
0.059 0.397
2842.1(1.64) 339.4(5.37) 1260.6(2.45) 352.4(9.50)
0.097 0.0746
0.6 5.3
Na
%
472.2(90.4)
0.105
0.119
0.9
Si
%
3539(70.2)
0.376
Cl
μg/g
1164.9(25.5)
0.109
2379.7(0.45) 841(68.0)
781.4(3.17) 91(44.5)
4.8 18
2092.9(19.5) 1273.3(16.9)
0.196
1.9
786.3(9.59) 1164.9(25.5)
5.67
7.8
44.2
Cr
μg/g
749(84.8)
Cob
μg/g
277.1(19.9)
Smb
μg/g
333.9(98.9)
0.144
439.4(55.5) 333.9(98.9)
0.0546
2.6
Cdb Gd
μg/g μg/g
558.5(72.7) 181.8(0.39)
0.522 0.109
651.3(13.9) 558.5(72.7) 181.8(0.39) 781.6(0.97)
0.127 0.118
4.1 0.9
951
1784.7(6.07) 834.1(0.33)
73.3
497.3(4.43) 229.7(25.9)
186 23.0
5.1 3.2
a Nuclear data from the Web sites listed in refs 12 and 13 (values within parentheses are γ-ray intensities, %). b γ-Ray intensity (%) data from the Web site listed in ref 17.
7490
dx.doi.org/10.1021/ac201706g |Anal. Chem. 2011, 83, 7486–7491
Analytical Chemistry However, detection limits of MPGA for Ni, Na, and Gd yield a nearly similar level to or higher than those of PGA. For these elements, intensities (emission probabilities) of prompt γ-rays for the most intense coincidence peaks are lower than those for a single γ-ray usable for PGA. Since the analytical sensitivity in MPGA depends on the intensity of paired cascade prompt γ-rays of the peak concerned, lower intensity of either γ-ray of the MPGA coincident peak results in higher detection limit.
’ CONCLUSIONS MPGA can reduce the background level substantially especially for hydrogenous samples relative to PGA, which opens up a possibility to use some low-energy prompt γ-rays emitted by constituent elements and enhances the analytical capability of PGA. By carefully selecting suitable coincident prompt γ-ray pairs, MPGA can determine Mg contents in geological and cosmochemical samples with reasonable consistency with their corresponding recommended values. MPGA can reduce detection limits for most of the elements determined in the Ivuna meteorite sample by 1 order of magnitude when compared with PGA detection limits under the present experimental conditions.
ARTICLE
(11) Lodders, K. Astrophys. J. 2003, 591, 1220–1247. (12) Chu, S. Y.; Nordberg, H.; Firestone, R. B.; Ekstrom, L. P. Isotope Explorer 2.23; U.S. Department of Energy, U.S.A., 1999. http://dbserv. pnpi.spb.ru/elbib/tablisot/toi98/www/isoexpl/isoexpl.htm (accessed on July 25, 2011). (13) National Nuclear Data Center, Brookhaven National Laboratory, Upton, NY 11973-5000, U.S.A. http://www.nndc.bnl.gov/capgam/index. html (accessed on February 15, 2011). (14) Jarosewich, E.; Clarke, R. S.; Barrows, J. Smithson. Contrib. Earth Sci. 1987, 27, 1–49. (15) Yanai, K.; Kojima, H. Catalog of the Antarctic Meteorites; National Institute of Polar Research (NIPR): Tokyo, Japan,1995. (16) Choe, W. H.; Huber, H.; Rubin, A. E.; Kallemeyn, G. W.; Wasson, J. T. Meteorit. Planet. Sci. 2010, 45, 531–554. (17) Oak Ridge Associated Universities, U.S.A. http://www.orau. org/ptp/PTP%20Library/library/DOE/lbl/1PROMPTNEUTRON1. PDF (accessed on December 20, 2010). (18) Kimura, A.; Toh, Y.; Oshima, M.; Hatsukawa, Y. J. Radioanal. Nucl. Chem. 2008, 278, 521–525.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected]. Fax: 81-0426-77-2525.
’ ACKNOWLEDGMENT MPGA and PGA were made possible by an interuniversity cooperative research program for the use of JAEA facilities supported by the University of Tokyo. M. A. Islam thanks JAEA for giving him a chance to join the Summer Term Research Program of JAEA for graduate students. We are grateful to H. Kojima of the National Institute of Polar Research, Japan for loaning us meteorite samples. Other members of the Research Group for Applied Nuclear Physics, JAEA are acknowledged for their assistance in MPGA experiments. This work is financially supported in part by a Grant-in-Aid defrayed by the Ministry of Education, Science and Culture of Japan (No. 22224010). ’ REFERENCES (1) Gardner, R. P.; Mayo, C. W.; El-sayyed, E. S.; Metwally, W. A.; Zheng, Y.; Poezart, M. Appl. Radiat. Isot. 2000, 53, 515–526. (2) Wang, J.; Calderon, A.; Peeples, C. R.; Ai, X.; Gardner, R. P. Nucl. Instrum. Methods Phys. Res., Sect. A [Online early access]. Doi: 10.1016/ j.nima.2010.08.011. Published Online: Aug 26, 2010. http://www. sciencedirect.com/science/article/pii/S0168900210017353. (3) Amber, P. P.; Belgya, T.; Molnar, G. L. Appl. Radiat. Isot. 2002, 56, 535–541. (4) Anders, E.; Grevesse, N. Geochim. Cosmochim. Acta 1989, 53, 197–214. (5) Toh, Y.; Oshima, M.; Kimura, A.; Koizumi, M.; Furutaka, K.; Hatsukawa, Y.; Goto, J. J. Radioanal. Nucl. Chem. 2008, 278, 685–689. (6) Karouji, Y.; Ebihara, M. Anal. Sci. 2008, 24, 659–663. (7) Imai, N.; Terashima, S.; Itoh, S.; Ando, A. Geochem. J. 1995, 29, 91–95. (8) Toh, Y.; Oshima, M.; Furutaka, K.; Kimura, A.; Koizumi, M.; Hatsukawa, Y.; Goto, J. J. Radioanal. Nucl. Chem. 2008, 278, 703–706. (9) Toh, Y.; Oshima, M.; Koizumi, M.; Kimura, A.; Hatsukawa, Y. J. Radioanal. Nucl. Chem. 2008, 276, 217–220. (10) Yonezawa, C.; Wood, A. K. H.; Hoshi, M.; Ito, Y.; Tachikawa, E. Nucl. Instrum. Methods Phys. Res., Sect. A 1993, 329, 207–216. 7491
dx.doi.org/10.1021/ac201706g |Anal. Chem. 2011, 83, 7486–7491
