Accurate Determination of Chlorine, Bromine, and Iodine in

May 27, 2013 - Accurate Determination of Chlorine, Bromine, and Iodine in Sedimentary Rock Reference Samples by Radiochemical Neutron Activation ...
1 downloads 0 Views 599KB Size
Download PDF
Article pubs.acs.org/ac

Accurate Determination of Chlorine, Bromine, and Iodine in Sedimentary Rock Reference Samples by Radiochemical Neutron Activation Analysis and a Detailed Comparison with Inductively Coupled Plasma Mass Spectrometry Literature Data Shun Sekimoto*,† and Mitsuru Ebihara‡ †

Research Reactor Institute, Kyoto University, 2-1010 Asashiro-nishi, Kumatori, Sen-nan, Osaka 590-0494, Japan Department of Chemistry, Tokyo Metropolitan University, 1-1 minami-Osawa, Hachioji, Tokyo 192-0397, Japan

‡

S Supporting Information *

ABSTRACT: Trace amounts of three halogens (chlorine, bromine, and iodine) were determined using radiochemical neutron activation analysis (RNAA) for nine sedimentary rocks and three rhyolite samples. To obtain high-quality analytical data, the radiochemical procedure of RNAA was improved by lowering the background in gamma-ray spectrometry and completing the chemical procedure more rapidly than in conventional procedures. A comparison of the RNAA data of Br and I with corresponding inductively coupled plasma mass spectrometry (ICPMS) literature data revealed that the values obtained by ICPMS coupled with pyrohydrolysis preconcentration were systematically lower than the RNAA data for some reference samples, suggesting that the quantitative collection of Br and I cannot always be achieved by the pyrohydrolysis for some solid samples. The RNAA data of three halogens can classify sedimentary rock reference samples into two groups (the samples from inland water and those from seawater), implying the geochemical significance of halogen data.

A

ICPMS after preconcentration of the elements using pyrohydrolysis of rock powder.3,4 However, neither fluorine nor chlorine can be determined by ICPMS coupled with pyrohydrolysis. In principle, four halogens in rock samples can be determined using NAA. However, three halogens (chorine, bromine, and iodine) have been routinely determined by NAA with radiochemical purification.5−7 As for fluorine, the half-life of neutron-capturing nuclide 20F (11 s) is too short to be radiochemically purified in the RNAA procedure. Fluorine, however, can be determined by different activation analyses such as INAA with cyclic neutron irradiation8 and radiochemical photon activation analysis.9 In this study, trace three halogens (chlorine, bromine, and iodine) were determined by radiochemical NAA (RNAA) for 12 geological reference powder samples: 9 sedimentary rocks and 3 rhyolites. There are three major goals in this study. The first goal is to improve the radiochemical procedure for RNAA of three heavy halogens, aiming for increasing the analytical capability, and to apply the modified RNAA procedure to sedimentary rock reference samples, for which only preferable values are given. Three rhyolite samples were repeatedly analyzed throughout this study to monitor the quality of our halogen data. The second goal was to compare our RNAA values with the ICPMS values for the nine sedimentary rock samples in order to evaluate the consistency of the two data

ccurate and reliable data of halogen abundance have been rarely reported for terrestrial samples, such as crustal rock and mantle material. Since halogens differ in volatility from element to element, their content and relative abundance are highly informative when discussing the petrogenesis of samples.1 Among the halogens, iodine is the most informative element in discussion of the geochemical circulation of crustal material. In discussion of the geochemical circulation of iodine in the earth’s surface, oceanic crust, continental crust, and mantle, Deruelle et al.2 demonstrated a typical example. Halogens are also important in meteoritics, and in particular, iodine is of particular interest and of high importance in discussions of the cosmochemical behavior of its extinct nuclide 129 I (half-life of 15.7 million years) in the early solar system. Simply, the scarcity of reliable halogen data for meteorite samples is a problem. There is a shortage of accurate and reliable data of halogens in meteorites, as well as in terrestrial rock samples, as can be witnessed in the data library for geological rock samples prepared by the Geological Survey of Japan (GSJ) (at http:// riodb02.ibase.aist.go.jp/geostand/welcome.html). Most igneous and sedimentary rocks within the database are recorded with only preferable, not certified values, and for some rocks, no values are listed. This deficit must be largely related to difficulties in determining trace amounts of halogens within these samples. To determine trace halogens in rock samples, either inductively coupled plasma mass spectrometry (ICPMS) or neutron activation analysis (NAA) have commonly been utilized. Bromine and iodine are conventionally determined by © XXXX American Chemical Society

Received: February 28, 2013 Accepted: May 27, 2013

A

dx.doi.org/10.1021/ac400637d | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

samples were cooled for a few minutes to enable the decay of Al and were then subjected to radiochemical separation of neutron-capturing radionuclides of halogens (38Cl, 82Br, and 128 I). The neutron irradiation was performed by using the same reactor for the determination of chemical yields of halogens for their radiochemical separation. The detailed description is given below. Radiochemical Purification Procedure in RNAA of Chlorine, Bromine, and Iodine. Although we followed the radiochemical procedure for RNAA of three heavy halogens described by Ozaki and Ebihara7 and others cited therein, several modifications were introduced in our analytical scheme in accordance with the aim of our study. Here, the (improved) radiochemical purification procedure is outlined. The schematic analytical procedure is shown in the Supporting Information. (1) A known amount of each halogen carrier (10 mg for chlorine, 20 mg for bromine and 30 mg for iodine) together with proper amounts of saturated NaOH solution and a manganese carrier (1 mg of manganese) are put into a nickel crucible and gently heated on a hot plate until dryness. (2) An irradiated rock powder is transferred to the crucible and fused with 1 g of NaOH. The crucible is heated mildly for first 3 min and then strongly for 5 min. (3) After fusion, the fused cake is dissolved in water and the hydroxide precipitate is separated from the supernatant by centrifugation. A few mg of sodium sulfite (solid) is then added to the supernatant solution containing the halogens. (4) The solution is made slightly acidic by adding 6 M HNO3. Iodine remains as iodide due to the presence of sodium sulfite, even with an addition of HNO3. (5) By adding a proper amount of Pd(NO3)2 solution into the supernatant, a PdI2 precipitate forms and is separated from the supernatant by centrifugation. The PdI2 precipitate is then washed with a sufficient amount of 0.2 M HNO3. The PdI2 precipitate is finally collected onto a disk filter paper (25 mm diameter), using filter holders (17 mm diameter), and dried under a heat lamp. The PdI2 sample thus prepared is sandwiched by tape (Scotch Book Tape) and immediately subjected to gamma-ray counting with a Ge semiconductor detector. (6) After the preparation of the PdI2 sample, a proper amount of AgNO3 is added to the supernatant containing chlorine and bromine so that the precipitate of the AgCl + AgBr mixture forms. The precipitate is washed with a sufficient amount of 0.2 M HNO3 and is prepared for the gamma-ray counting in the same manner as applied for the PdI2 precipitate. Gamma Ray Measurement and Yield Determination. The radioactivity of 128I and 38Cl was measured by using a Ge semiconductor detector for 300−1000 s and 500−1000 s, respectively. For the determination of bromine, either 80Br or 82 Br can be used. As discussed in Shinonaga et al.,5 the measurement of 82Br is more advantageous and, hence, the activity of 82Br was measured for 50 000−100 000 s on the

sets. However, because there was no chlorine data available from ICPMS, only bromine and iodine data could be compared. The final goal was to search for any geochemical implications based on the data obtained from the analysis of the nine sedimentary rocks. However, this goal was not entirely realized due to the limited number of sediment samples in this study and will be pursued in future studies.

28

■

EXPERIMENTAL METHODS Sample Preparation. Nine sedimentary rock samples analyzed are listed in Table 1, together with additional Table 1. Geological Reference Samples of Sedimentary Rocks Analyzed in This Study sample name

sample type

JLk-1

lake sediment

JLs-1 JDo-1 JSl-1 JSl-2 JSd-1

limestone dolomite slate slate stream sediment stream sediment stream sediment chert

JSd-2 JSd-3 JCh-1

sampling location fresh water lake sediment, Lake Biwa, Shiga, Japan Garo limestone, Kamiiso, Hokkaido, Japan Kuzuu dolomite, Kuzuu, Tochigi, Japan Toyoma clay slate, Toyoma, Miyagi, Japan Toyoma clay slate, Okatsu, Miyagi, Japan composite sample of northern region, Ibaraki, Japan composite sample of eastern region, Ibaraki, Japan composite sample of central region, Ibaraki, Japan Ashio chert, Ashikaga, Tochigi, Japan

information. The samples were prepared by GSJ and were commercially distributed;10 three rhyolite reference samples (JR-1, JR-2, and JR-3) were also analyzed. All the samples were in powder and were not subjected to any additional treatment such as drying. Approximately 100 mg of each powder sample was weighed, inserted into a clean, small plastic vial that was sealed, and then resealed inside a clean polyethylene bag. Chemical standard solutions of the three halogens were prepared for their quantifications by NAA. An appropriate amount of each halogen solution (containing 90 μg for chlorine, 50 μg for bromine, and 10 μg for iodine) was dropped onto a paper disk (17 mm diameter), weighed, dried under a heat lamp, and doubly sealed into polyethylene bags. Extreme care was taken when preparing the iodine reference sample.11,12 Neutron Irradiation. Two rock samples, together with a set of three reference halogen samples, were irradiated for 10 min with a thermal neutron flux of 3.3 × 1012 cm−2 s−1 at Kyoto University Research Reactor Institute (KURRI). When a high flux reactor is used, the samples need to be cooled to prevent iodine loss.7 However, no such cooling was required with the neutron flux used in this study. After irradiation, the rock

Table 2. Nuclear Data of Radioactive Nuclides Concerned in This Study element

target nuclides

17Cl

37

35Br 53I 13Na 25Mn

81

Cl Br 127 I 23 Na 55 Mn

isotopic abundance (%)

cross sectiona (barn)

24.24 49.31 100 100 100

0.43 2.64 6.2 0.53 13.3

produced nuclideb 38

Cl Br 128 I 24 Na 56 Mn 82

half life

γ-ray energyc (keV)

37.18 m 35.34 h 25.0 m 14.96 h 2.58 h

1642, 2168 776, 554 443 1369, 2754 847, 1811

a

Neutron capture cross section for thermal neutron. bProduced by neutron capture reaction. cUsed in gamma-ray spectrometry for quantification of elements. B

dx.doi.org/10.1021/ac400637d | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

following few days after the RNAA procedure. The reference samples of three halogens were measured individually after the completion of the measurements for rock samples for 200−500 s for Cl and I and 1000−2000 s for Br. Chemical yields of three halogens were determined by the reactivation method. After the completion of all gamma-ray measurements, a set of 4−6 samples (either of PdI2 or AgBr + AgCl mixture), along with reference samples for the corresponding halogens (of roughly the same amounts as used for their carriers for radiochemical separation), were irradiated for 10 s with a thermal neutron flux of 3.3 × 1012 cm−2 s−1 at KURRI and measured for radioactivity of 38Cl, 82Br, and 128I for 100 s for each nuclide. The nuclear data related to the NAA of halogens concerned in this study are summarized in Table 2.

■

RESULTS AND DISCUSSION Improvement of the Analytical Sensitivity. Accuracy and sensitivity are two major concerns for analytical chemists. It is well acknowledged that analytical data obtained by RNAA are generally very high in accuracy because an analyte (a radionuclide produced from a target nuclide) is radiochemically purified after the neutron activation of samples and any loss of the analyte during the purification procedure can be corrected. In this study, we aimed to improve the analytical sensitivity for three halogens, especially for iodine. Two approaches were tried for this purpose: decreasing the background in gamma-ray spectrometry and shortening the time for radiochemical purification of radionuclides concerned. Under the present irradiation condition, rock samples including sedimentary rocks and rhyolites yield extremely high radioactivity of 24Na (half-life, 14.96 h) and 56Mn (halflife, 2.58 h). Manganese can be largely separated from halogens in the first step, where halogens stay in the supernatant, whereas manganese is coprecipitated with iron as hydroxides. To enhance the coprecipitation with iron, about 1 mg of manganese was added to the irradiated rock sample before the NaOH fusion so as to act as a holdback carrier. Nevertheless, a certain amount of 56Mn activity could be detected in the supernatant. To reduce the amount of 56Mn along with that of 24 Na, which was almost totally present in the supernatant, the PdI2 precipitate was washed two times with 30 and 5 mL of 0.2 M HNO3 and then prepared in a disk sample for gamma-ray spectrometry. Figure 1a,b compares the gamma-ray spectra for the conventional procedure and the improved procedure including 56Mn removal, for the PdI2 sample obtained from the rhyolite reference sample JR-3. As seen in Figure 1b, peaks intensities of 56Mn and, particularly, of 24Na were lowered by more than factors of 2 and 10, respectively. As a result, counting statistics for the 128I peak at 443 keV was improved by a factor of 2, and the uncertainty due to counting statistics were reduced by 33%. A similar washing was also applied to the AgCl and AgBr precipitate. Another approach for increasing the analytical sensitivity was shortening the time for radiochemical purification of radionuclides to be measured. Among the three heavy halogens, iodine has two disadvantages for its determination in rock samples by RNAA; the lowest abundance in most rock samples, being about 3 to 1 orders of magnitude lower than those of chlorine and bromine, and the shortest half-life of the radionuclide (128I with a half-life of 25.0 min) used for the determination (Table 2). In our RNAA procedure, after alkaline fusion of rock powders with NaOH and dissolution

Figure 1. (a) γ-ray spectrum of the PdI2 precipitate without HNO3washing for JR-3. (b) γ-ray spectrum of the PdI2 precipitate with HNO3-washing for JR-3.

of fusion cakes in H2O, the supernatant was neutralized and made slightly acidic by adding 6 M HNO3. During this procedure, iodide ion (I−) in the supernatant is partly oxidized to iodine (I2). To prevent any loss of iodine, conventionally, the sodium sulfite solution (6%) was added to the slightly acidified supernatant as a reducing agent.5−7 This step was replaced by the addition of a few milligrams of Na2SO3 to the supernatant before the addition of HNO3. This modification not only prevents the loss of iodine but also facilitates the neutralization of the supernatant. Coupling with other modifications for speeding up the chemical procedure, a measurement of 128I was enabled within two half-lives (about 50 min) after the end of neutron irradiation, with about 10 min being shortened. Cl, Br, and I Contents in Reference Rock Samples of Sediments and Rhyolites. Three halogens (chlorine, bromine, and iodine) were determined for nine sedimentary rocks and three rhyolites. Each sample was analyzed three to five times (individual values are tabulated in the Supporting Information). Relative reproducibilities (1σ) of repetitive analytical results are 1.2−14%, 2.0−19%, and 3.6−17% for chlorine, bromine, and iodine, respectively. In general, a large uncertainty is to be obtained from a set of values with poor counting statistics, corresponding to a low elemental content. Nevertheless, the variation of relative reproducibilities stays similar for the three halogens. Mean values with 1 sigma of standard deviation for all the samples analyzed in this study are summarized in Table 3. The chlorine and bromine contents vary from 4.76 mg kg−1 for JCh-1 (chert) to 982 mg kg−1 for JR-1 (rhyolite) and from 0.027 mg kg−1 for JCh-1 (chert) to C

dx.doi.org/10.1021/ac400637d | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

because these rock samples have been the most frequently analyzed among Japanese geological reference samples for halogens. As seen in Table S-1 in the Supporting Information and Table 3, our halogen data for the three rhyolites are highly reproducible from run to run of this study and are substantially consistent with those of Shinonaga et al.5 and Ozaki and Ebihara.7 Thus, we may conclude that the halogen data obtained in this study are reproducible and further that the RNAA procedure applied and developed in this study is effective in determining halogens in geological rock samples. The nine sedimentary rock samples analyzed in this study were relatively newly issued reference rock samples by GSJ. The data usable for the comparison of our data are still limited in number, as evident in Table 3, where only a few data are given for chlorine and bromine (Imai et al.13 and from the GSJ web page). Our data are consistent with the data compiled by Imai et al.13 within uncertainties of 30%, except for chlorine in JSL-1, JSL-2, JSd-3, and JCh-1. Considering the data quality of our values as discussed above, we envisage that our data would be reflected in any future compilation in the establishment of a database of GSJ reference rock samples. A comparison of our data with the ICPMS data reported by Chai and Muramatsu14 is discussed in the following section. Comparison of RNAA Data with ICPMS Data for Sedimentary Rock Reference Samples. The nine sedimentary rock reference samples analyzed using RNAA in this study were also analyzed using ICPMS,14 by which, however, only bromine and iodine were determined. In the procedure of Chai and Muramatsu,14 bromine and iodine were preconcentrated using pyrohydrolysis, in which the rock samples are heated with vanadium pentoxide under an oxygen flow in an electric furnace and the expelled halogens are then trapped into an alkaline solution. In terrestrial rock samples, halogens are supposed to be present in the form of a monovalent anion, such as F−, Cl−, Br−, and I−. In pyrohydrolysis, some halogens (at least, bromine and iodine) are supposed to be mostly oxidized to gaseous halogen molecules (dimer) like Br2 and I2, which are trapped in alkaline solution, where they are probably in the form of a monovalent anion. The solution is then introduced into the ICPMS instrument. As already mentioned, the sedimentary samples were analyzed three to four times and the values obtained were reasonably consistent with one another for each halogen. Our RNAA data are mostly consistent with the ICPMS data of Chai and Muramatsu,14 even in bromine in JCh-1 (chert), where the ICPMS value is 2 times higher than the RNAA value (0.060 ± 0.040 mg kg−1 vs 0.027 ± 0.005 mg kg−1) but is in agreement with the RNAA value within uncertainties. The bromine content of this sample is extremely low compared with those in other sedimentary rocks. The next lowest value (approximately 0.060 mg kg−1) was obtained for JSL-2 (slate), for which RNAA and ICPMS values were consistent with each other, although a slightly lower value was obtained using ICPMS. Our RNAA values are compared with the ICPMS values of Chai and Muramatsu14 for all nine sedimentary rocks in Table 4, where concentration ratios of ICPMS values to RNAA values are given. These ratios are plotted in Figure 2, where ratios for concentrations of bromine (Figure 2a) and iodine (Figure 2b) are compared separately. In calculating the mean ratio of bromine, a value for JCh-1 was ruled out using the outlier test under the 2σ-criterion. An uncertainty quoted for each ratio is simply the value calculated from two individual uncertainty values accompanied by RNAA and ICPMS data. Please note

Table 3. Cl, Br, and I Contents in GSJ Reference Samples Analyzed by RNAA in This Studya and from the Literature sample Sedimentary Rocks JLk-1 RNAA (n = 3) compiled13 ICPMS14 JLs-1 RNAA (n = 3) ICPMS14 JDo-1 RNAA (n = 3) compiled13 ICPMS14 JSL-1 RNAA (n = 4) compiled13 ICPMS14 JSL-2 RNAA (n = 3) compiled13 ICPMS14 JSd-1 RNAA (n = 4) compiled13 ICPMS14 JSd-2 RNAA (n = 3) compiled13 ICPMS14 JSd-3 RNAA (n = 3) compiled13 ICPMS14 JCh-1 RNAA (n = 3) compiled13 ICPMS14 Rhyorites JR-1 RNAA (n = 4) RNAA5 RNAA7 JR-2 RNAA (n = 5) RNAA5 RNAA7 JR-3 RNAA (n = 3) RNAA7 a

Cl (mg kg−1) 59.1 −b −b 16.4 −b 35.9 −b −b 13.6 21.5 −b 7.56 18.5 −b 64.0 67.5 −b 22.7 28 −b 25.8 39.0 −b 4.76 14 −b

± 1.8

± 1.4 ± 5.0

± 1.0

± 0.75

± 6.5

± 1.5

± 0.3

± 0.24

982 ± 71 992 ± 90 977 ± 58 789 ± 39 1000 ± 10 867 ± 36 134 ± 12 129 ± 5

Br (mg kg−1)

I (μg kg−1)

7.82 ± 0.53 8.7 8.0 ± 0.3 0.105 ± 0.012 0.068 ± 0.007 0.622 ± 0.051 0.790 0.530 ± 0.020 0.123 ± 0.014 −b 0.110 ± 0.010 0.060 ± 0.006 −b 0.054 ± 0.008 1.84 ± 0.10 1.65 1.65 ± 0.07 1.13 ± 0.04 −b 1.18 ± 0.07 3.92 ± 0.08 3.9 4.2 ± 0.5 0.027 ± 0.005 −b 0.060 ± 0.040

9050 ± 620 −b 9400 ± 100 318 ± 28 260 ± 20 789 ± 39 −b 710 ± 60 107 ± 8 −b 86 ± 6 97 ± 9 −b 101 ± 9 1100 ± 90 −b 1040 ± 40 675 ± 63 −b 733 ± 30 4230 ± 330 −b 4200 ± 100 115 ± 13 −b 97 ± 4

2.07 ± 0.06 2.3 ± 0.1 2.12 ± 0.13 1.64 ± 0.12 2.2 ± 0.3 1.65 ± 0.07 0.577 ± 0.045 0.533 ± 0.016

84 ± 3 75 ± 13 98 ± 7 86 ± 15 72 ± 7 62 ± 8 482 ± 37 462 ± 27

Mean values followed by standard deviations (1σ) . bNot reported.

7.82 mg kg−1 for JLk-1 (lake sediment), respectively. The iodine content varies from 84 μg kg−1 for JR-1 (rhyolite) to 9050 μg kg−1 for JLk-1 (lake sediment). The variation is thus over 2 orders of magnitude for the three halogens, for which the RNAA procedure applied in this study can work. Methodologically, it has a much larger dynamic range in determining halogens in rock samples. In Table 3, our data are compared with the values in the literature.5,7,13,14 Halogen data reported by Shinonaga et al.5 and Ozaki and Ebihara7 were obtained by RNAA, whereas those by Chai and Muramatsu14 were obtained by ICPMS with pyrohydrolysis concentration. The values presented by Ozaki and Ebihara7 are mean values of replicate analysis results. Imai and co-workers,13 who were in charge of preparing a series of geological reference samples at GSJ, evaluated all the available data, including those obtained from personal communication, and presented the recommended and preferable values for some halogens in selected reference samples, which are listed in Table 3. The three rhyolites (JR-1, JR-2, and JR-3) were repeatedly analyzed as control samples during the course of this study D

dx.doi.org/10.1021/ac400637d | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

relatively high compared with those in the other seven samples. In the sample containing more than a few mg kg−1 bromine and/or with more than 1000 μg kg−1 iodine, consistent data between RNAA and ICPMS are likely to be obtained. Although inconsistency between RNAA data and ICPMS data cannot necessarily be seen for a sample such as JSL-2 where bromine and iodine contents are relatively low, ICPMS data of both bromine and iodine are evidently lower than RNAA data for the limestone sample (JLs-1), which consists mainly of calcium carbonate and is decomposed into refractory calcium oxide (having a high melting point of 2927 K) and carbon dioxide on heating. Presumably, the quantitative collection of bromine and iodine cannot always have been achieved in the procedure used by Chai and Muramatsu.14 This suggests the limitation of the pyrohydrolysis procedure in the determination of halogens (at least bromine and iodine) by ICPMS. It may be noted, however, that RNAA and ICPMS data agree within 10% for both bromine and iodine, as shown by the mean values of individual ratios, and further that ICPMS coupled with pyrohydrolysis can work well for such studies where only rough abundances of halogens are required. However, if accurate values of halogens are essential for the study, which corresponds to the analysis of reference samples, RNAA is the best choice of analytical procedure. There is also an advantage of choosing RNAA over ICPMS when the sample mass used for analysis is limited and duplicate analyses are not possible or desirable, as is the case when analyzing extraterrestrial samples such as meteorites and lunar samples. The applicability of our proposed RNAA procedure is described in the Supporting Information in terms of detection limits for three halogens (chlorine, bromine, and iodine). The suitability of the RNAA procedure for other types of samples than geological samples is also briefly described there. Comparison of Halogen Contents between InlandWater Sedimentary Rocks and Seawater Sedimentary Rocks. Halogen abundances for the nine sedimentary rocks analyzed in this study are compared in Figure 3, where abundances are normalized to CI chondrite values.15 Halogen abundances for present bulk continental crust, lower con-

Table 4. Ratios of ICPMS Values to RNAA Values ICPMS/RNAAa sample JLk-1 JLs-1 JDo-1 JSL-1 JSL-2 JSd-1 JSd-2 JSd-3 JCh-1 mean

bromine 1.02 0.65 0.85 0.89 0.90 0.90 1.04 1.07 2.22 0.92

± ± ± ± ± ± ± ± ± ±

0.08 0.10 0.08 0.13 0.16 0.06 0.07 0.13 1.54 0.19b

iodine 1.04 0.82 0.90 0.80 1.04 0.95 1.09 0.99 0.84 0.94

± ± ± ± ± ± ± ± ± ±

0.07 0.10 0.09 0.08 0.13 0.09 0.11 0.08 0.10 0.10

a

RNAA and ICPMS values from Table 3. bA ratio for JCh-1 is ruled out (see text for details). An uncertainty means a standard deviation (1σ).

Figure 2. (a) Concentration ratios of Br between RNAA values and ICPMS values. A bar indicates a range of 1σ deviation for replicate analyses. (b) Concentration ratios of I between RNAA values and ICPMS values. A bar indicates a range of 1σ deviation for replicate analyses.

here that the uncertainty of RNAA is a standard deviation (1σ) of three or four separately determined values, whereas the uncertainty of ICPMS is a standard deviation of several (normally three) consecutive measurements of one solution sample. Although the two sets of data are fairly consistent between RNAA and ICPMS, it can also be seen that the ICPMS data are systematically lower than the RNAA data. In particular, ICPMS data for bromine for three samples (JLs-1, JDo-1, and JSd-1) are lower than the corresponding RNAA data, even with the uncertainties being concerned. A similar situation can be seen for iodine in four samples (JLs-1, JDo-1, JSL-1, and JCh-1). However, the ICPMS data for bromine and iodine are very consistent with the RNAA data for two samples (JLk-1 and JSd-3), where bromine and iodine contents are

Figure 3. Relative abundances of Cl, Br, and I in sedimentary rocks, crustal samples, and seawater normalized to primitive meteorite values. E

dx.doi.org/10.1021/ac400637d | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

tinental crust, upper continental crust,16 oceanic crust,2,17 and seawater18 are also shown for comparison. The nine sedimentary reference rock samples are classified into two groups: sedimentary rocks in inland water (JLk-1, JSd-1, JSd-2, and JSd-3) and those in seawater (JLs-1, JDo-1, JCh-1, JSL-1, and JSL-2). There appear to be several systematic differences in absolute and relative abundances of three heavy halogens between these two groups. First, sedimentary rocks in seawater have a lower abundance of halogens than those in inland water. However, JDo-1 (dolomite) has exceptionally high halogen (especially, bromine) abundance in the sedimentary rocks from seawater. The difference in the abundance between the two groups is largest for bromine and smallest for chlorine, with iodine in the middle. The second difference can be seen in the relative abundance of the three halogens between the inlandwater sedimentary rocks and the seawater sedimentary rocks; the latter samples have relatively high abundances of chlorine compared with those of the former samples. The high abundance of chlorine in sediment rocks from seawater may be related to the high content of chlorine in seawater. As seen in Figure 3, halogen abundance of bromine and iodine are similar in bulk and upper continental crusts16 and inland-water sedimentary rocks, especially JSd-1 and JSd-2, in terms of relative as well as absolute abundance. However, there appears to be a large difference in chlorine abundance; the crustal abundance is roughly an order of magnitude higher than in inland-water sedimentary rocks. It is likely that the crustal material is also affected by seawater, similarly to the sediment rocks from seawater, and that its effect is higher in the continental crusts than in seawater sedimentary rocks.

the three halogens are apparently different between marine sediment rocks and river sediment rocks. It is thus expected that scientifically meaningful implications can be extracted from halogen data obtained by the RNAA procedure described in this paper.

■

ASSOCIATED CONTENT

S Supporting Information *

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Fax: +81-72-451-2632. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors express their gratitude to the members of the research reactor group in KURRI for the preparation and operation in the neutron irradiation. This study was supported by Kyoto University Global COE Program “International Center for Integrated Research and Advanced Education in Material Science” (to S.S.) and by a grant-in-aid for scientific research by MEXT (to M.E.).

■

REFERENCES

(1) Unni, C. K.; Schilling, J.-G. Nature 1978, 272, 19−23. (2) Deruelle, B.; Dreibus, G.; Jambon, A. Earth Planet. Sci. Lett. 1992, 108, 217−227. (3) Schnetger, B.; Muramatsu, Y.; Yoshida, S. Geostand. Newslett. 1998, 22, 181−186. (4) Schnetger, B.; Muramatsu, Y. Analyst 1998, 121, 1627−1631. (5) Shinonaga, T.; Ebihara, M.; Nakahara, H.; Tomura, K. Chem. Geol. 1994, 115, 213−225. (6) Ebihara, M.; Ozaki, H.; Kato, F.; Nakahara, H. J. Radioanal. Nucl. Chem. 1997, 216, 107−112. (7) Ozaki, H.; Ebihara, M. Anal. Chim. Acta 2007, 583, 384−391. (8) Parry, S. J.; Benzing, R.; Bolstad, K. L.; Steinnes, E. J. Radioanal. Nucl. Chem. 2000, 244, 67−72. (9) Nakamoto, T.; Oura, Y.; Ebihara, M. Anal. Sci. 2007, 23, 1113− 1119. (10) Terashima, S.; Ando, A.; Okai, T.; Kanai, Y.; Taniguchi, M.; Takizawa, F.; Itoh, S. Geostand. Newslett. 1990, 14, 1−5. (11) Ebihara, M.; Saito, N.; Akaiwa, H. J. Radioanal. Nucl. Chem. 1986, 108, 241−255. (12) Ebihara, M.; Saito, N.; Akaiwa, H.; Tomura, K. Radioisotopes 1991, 40, 433−439. (13) Imai, N.; Terashima, S.; Itoh, S.; Ando, A. Geostand. Newslett 1996, 20, 165−216. (14) Chai, J. Y.; Muramatsu, Y. Geostand. Geoanal. Res. 2007, 31, 143−150. (15) Anders, E.; Grevesse, N. Geochim. Cosmochim. Acta 1989, 53, 197−214. (16) Wedepohl, K. H. Geochim. Cosmochim. Acta 1995, 59, 1217− 1232. (17) Jambon, A.; Déruelle, B.; Dreibus, G.; Pineau, F. Chem. Geol. 1995, 126, 101−117. (18) Wu, S. F.; You, C. F.; Valsami-Jones, E.; Baltatzis, E.; Shen, M. L. Mar. Chem. 2012, 140−141, 33−43.

■

CONCLUSION The radiochemical procedure used in RNAA for determining trace amounts of chlorine, bromine, and iodine in silicate rock samples was improved by lowering the background level in gamma-ray spectrometry and shortening the procedural time for radiochemical purification of neutron-capturing halogen nuclides. Being applied to geological reference samples, the procedure was proved to be more effective and reliable than the procedure reported previously. The obtained data of chlorine, bromine, and iodine can contribute in establishing their certified values. The bromine and iodine data determined by RNAA were compared with the corresponding data in literature which were obtained by ICPMS with pyrohydrolysis for the extraction of the two halogens from rock samples. It was confirmed that the ICPMS data were systematically smaller than the RNAA data for both bromine and iodine although the difference is less than 10%. Presumably, bromine and iodine cannot be quantitatively extracted from silicate rock samples, even though the samples are in powder and the extraction is performed in an electric furnace with the presence of oxidizing agent and oxygen gas. The use of RNAA with an improved radiochemical procedure can be very favorable when accurate values of trace amounts of chlorine, bromine, and iodine are to be obtained. This method must be the most preferable for geological reference samples and extraterrestrial samples, such as meteorites and lunar samples. The effectiveness of the improved RNAA procedure was illustrated by analyzing sedimentary rock samples collected from the ocean and the river. These rock samples can be classified into two groups based on their chlorine, bromine, and iodine abundances. Absolute as well as relative abundances of F

dx.doi.org/10.1021/ac400637d | Anal. Chem. XXXX, XXX, XXX−XXX