Laser ablation inductively coupled plasma mass spectrometry for quantitative imaging of elements in ferromanganese nodule

Chemical Geology 427 (2016) 65–72

Contents lists available at ScienceDirect

Chemical Geology journal homepage: www.elsevier.com/locate/chemgeo

Laser ablation inductively coupled plasma mass spectrometry for quantitative imaging of elements in ferromanganese nodule Junichi Hirata a, Kazuya Takahashi b, Yu Vin Sahoo b, Miho Tanaka a,⁎ a b

Graduate School of Marine Science and Technology, Tokyo University of Marine Science and Technology, 4-5-7 Konan, Minato-ku, Tokyo 108-8477, Japan RIKEN, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan

a r t i c l e

i n f o

Article history: Received 24 July 2015 Received in revised form 14 February 2016 Accepted 15 February 2016 Available online 16 February 2016 Keywords: LA-ICP-MS Ferromanganese nodule Elemental imaging Determination

a b s t r a c t Laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) is used for the quantitative elemental imaging of ferromanganese nodules to obtain the distributions of both major and trace elements, owing to its high sensitivity and wide dynamic range. Methods for obtaining a calibration line, using Mg as an internal standard for correction, and data treatment were developed for the quantitative imaging of elements in ferromanganese nodules. The validity of the values determined by LA-ICP-MS was confirmed by comparison with elemental contents obtained by ICP-MS. A two-dimensional plotting system for LA-ICP-MS was established for expressing elemental contents as colors along with spatial information. The simple method describes the ease with which the colors can be changed to define the content ranges of elements, and the elemental distributions show the layered structure, clearly depicting the contrast. Reproducibility of these analytical processes was also confirmed by analyzing two ferromanganese nodules. The method is expected to be a powerful tool for investigating paleoenvironmental changes in the region surrounding a ferromanganese nodule and its formation processes. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Advances in techniques for analyzing elemental contents at trace levels are a boon in current research. Generally, these techniques examine the total metal content after the homogenization and digestion of samples but do not provide sufficient information on the spatial distribution of elements. Laser ablation inductively coupled plasma mass spectroscopy (LA-ICP-MS) is a popular technique of elemental imaging for solid samples (Barst et al., 2011; Becker et al., 2014; Konz et al., 2012, 2013; Lear et al., 2012; Moreno-Gordaliza et al., 2011; Urgast et al., 2012; Wang et al., 2013; Zhu et al., 2012). In contrast to destructive methods of analysis, LA-ICP-MS enables the spatial resolution of a small amount of a sample. The LA system provides a micro sampling method for solid samples as an alternative to the conventional “micro spatula” or a “micro drill”, whereby information at the micro sampling point can be obtained and the structure near the sampling point is preserved. Through observation by a CCD camera and micro sampling with a laser, the contents of elements in the sample can be obtained along with their spatial distribution. Therefore, elemental imaging by LAICP-MS provides important information on solid samples. In addition, the direct sample introduction system used in LA-ICP-MS prevents samples from contamination with reagents required for sample dissolution (Fernández et al., 2010; Shaheen et al., 2012). Other methods, such as ⁎ Corresponding author at: Tokyo University of Marine Science and Technology, 4-5-7 Konan, Minato-ku, Tokyo 108-8477, Japan. E-mail address: [email protected] (M. Tanaka).

http://dx.doi.org/10.1016/j.chemgeo.2016.02.017 0009-2541/© 2016 Elsevier B.V. All rights reserved.

electron probe micro analyzer (EPMA), scanning electron microscopeenergy dispersive X-ray spectroscopy (SEM–EDX), and secondary ion mass spectrometry (SIMS), can be used with solid samples, but a high-vacuum system is required, whereas the LA-ICP-MS analysis can be performed at atmospheric pressure. A high vacuum destroys the structure of samples with a high moisture content, such as oceanic sediment by dehydration. Thus, LA-ICP-MS enables the spatial resolution analysis using only a small amount sample introduction system, and LA-ICP-MS prevents samples from contamination and dehydration. Ferromanganese nodules are found on the seafloor. They have characteristic layered structures and scavenge trace elements in sea water, such as rare-earth elements (Duliu et al., 2009; Hein et al., 2013). These nodules are extremely good indicators of environmental changes, especially the elemental contents and distributions in their complicated layered structure. Previous studies on ferromanganese samples (ferromanganese nodules and crusts) include the bulk analysis of major and trace elements by instrumental neutron activation analysis (INAA), ICP-atomic emission spectrometry (ICP-AES), and ICP-MS (Aplin, 1984; Carlo and McMurtry, 1992; Glasby et al., 1987; Kunzendorf et al., 1993), the analysis of ferromanganese samples by a sequential leaching method to investigate how trace elements are scavenged by Mn and Fe oxides or hydroxides (Koschinsky and Halbach, 1995; Koschinsky and Hein, 2003; Mohwinkel et al., 2014), analysis of the distribution of major elements by EPMA, SEM–EDX, and micro area X-ray diffractometer (μXRD) (Gasparatos et al., 2005; Manceau et al., 2003; Palumbo et al., 2001; Wang et al., 2012), analysis of the chemical structure of Mo by X-ray absorption fine structure (XAFS) (Kashiwabara

66

J. Hirata et al. / Chemical Geology 427 (2016) 65–72

et al., 2009), and the determination of the age of a ferromanganese nodule by a dating method using 10Be/9Be (Frank et al., 1999; von Blanckenburg et al., 1996). In previous studies, various types of information have been accumulated on ferromanganese samples, such as classification of their origins, chemical compositions, and growth rate. However, the formation processes of ferromanganese nodules and paleo-environmental changes are yet to be studied. To elucidate this, not only averaged elemental content but also elemental distribution is required. Distribution of elemental contents is considered as significant basic information on ferromanganese nodules. For example, geochemical parameters can be calculated from elemental contents by preserving chronological order in the layered structure and the data obtained allows the analyst to decide exact position to be examined for isotope ratio analysis (Bollhöfer et al., 1996). Thus, although elemental distributions in the layered structure have been studied (Gasparatos et al., 2005; Manceau et al., 2003; Palumbo et al., 2001; Wang et al., 2012), the distribution of trace elements in the layered structure has not been investigated owing to analytical problems. Information on the scavenging mechanism of elements by ferromanganese nodules could throw light on the process, which could be artificially produced, thus applying ferromanganese nodules as scavengers of metals in waste water. To achieve this, first both the major and trace elements at a small point on the surface were analyzed by applying a laser shot. A spatial resolution analysis is very important at the borders of the layered structures to elucidate changes in the environment, which give the time of formation of each layer. Second, to investigate the environmental changes and formation processes of ferromanganese nodules, sample structures are preserved in the analytical procedures. In previous studies, the elemental imaging of major elements in ferromanganese samples by EPMA, SEM–EDX, and μXRD was reported (Gasparatos et al., 2005; Manceau et al., 2003; Palumbo et al., 2001; Wang et al., 2012) but, owing to the low sensitivity of these methods, it was difficult to analyze trace elements such as rare-earth elements. ICP-MS, which can analyze both major and trace elements owing to its high sensitivity and wide dynamic range, was applied for spatial resolution analysis in small spaces of a ferromanganese nodule. From the above considerations, LA-ICP-MS is the most suitable method for the elemental imaging of ferromanganese nodules, as a sample is ablated at atmospheric pressure then directly introduced into the MS detector. In previous studies on elemental imaging using LA-ICP-MS, the elemental imaging data were expressed as semiquantitative values, such as peak intensities and peak intensity ratios (Barst et al., 2011; Konz et al., 2013; Moreno-Gordaliza et al., 2011; Urgast et al., 2012; Wang et al., 2013; Zhu et al., 2012). To perform detailed data analysis, imaging data was expressed quantitatively. Quantitative elemental imaging by LA-ICP-MS has been reported, however, the studies have been limited to the field of biological samples (Becker et al., 2014; Konz et al., 2012; Lear et al., 2012). It is difficult to determine and image elements in geological samples like ferromanganese samples, because matrices in geological samples are complicated for each sample. In previous studies on a ferromanganese nodule, a powdered bulk sample was investigated, or spatial resolution analysis was performed in a straight line (in one dimension) across the layers by LA-ICP-MS (Axelsson et al., 2002; Garbe-Schönberg and McMurtry, 1994; Hlawatsch et al., 2002; Hirata et al., 2013; Hirata and Tanaka, 2014; Hoffmann et al., 1997; Zhi et al., 2007). The quantitative two-dimensional imaging of elements in a ferromanganese nodule by LA-ICP-MS has not been accomplished owing to technical problems. Using the LA system for micro sampling, not only ferromanganese nodules but also other geological samples can be imaged. However, measurements using an LA-ICP-MS system involve various problems yet to be solved. A sample is ablated by a laser, the generated nanoparticles are transported to the ICP torch by Ar or He gas, and elemental fractionation occurs through these processes (Koch and Günther, 2011; Russo et al., 2002; Shaheen et al., 2008). Therefore, to determine the elemental compositions in samples, matrix-matched standards

were prepared. In addition to the problem of elemental fractionation, the rate of introduced sample to ICP torch is unstable due to hardness of sample, defocusing of laser, and sampling position in the sample cell (Fricker et al., 2011; Koch and Günther, 2011; Russo et al., 2002; Shaheen et al., 2008). To obtain accurate elemental distributions and elemental contents, discrepancies in the sample introduction rate were corrected. The aim of this study is to establish a quantitative elemental imaging method of analysis using LA-ICP-MS. For this purpose, the preparation of standards, the correction of unstable sample introduction, validation of the reliability of determined values, and expression of the elemental content in two-dimensional imaging data were performed. Furthermore, these procedures were applied for two different ferromanganese nodules, to confirm the reproducibility of the analytical method. Thus, LA-ICP-MS for quantitative imaging of elements in ferromanganese nodules was established.

2. Experimental 2.1. Reagents JMn1 (Dulski, 2001; Terashima et al., 1995) was employed as a geological reference material for ferromanganese nodule and was provided by Geological Survey of Japan (GSJ). JMn1 and high-purity manganese dioxide (MnO2) powder (99.5%, Wako Chemical Co.) were used in the preparation of standards for LA-ICP-MS analysis. To dissolve a powdered ferromanganese nodule, HNO3, HCl, and HF (Kanto Chemical Co., ultra pure grade) were used. XSTC-1, XSTC-8, and XSTC-469 (SPEX Certi. Prep.) standard solutions were used for determining the elemental contents in the dissolved sample by ICP-MS. For the determination of Mg and Co, analytical-grade standard solutions (Kanto Chemical Co.) were used. The indium and bismuth solutions used for internal standard were analytical-grade (SPEX Certi. Prep. or Kanto Chemical Co.).

2.2. Instrumentation High-resolution ICP-MS (ELEMENT XR, Thermo Fisher Sci. Co.) was used to analyze the dissolved ferromanganese nodule. For elemental imaging analysis, a laser ablation system (UP213, New Wave Research Co.) equipped with an ICP-MS system was used. The operating parameters for LA-ICP-MS are shown in Table 1.

Table 1 Operating parameters for LA-ICP-MS. Parameters UP213 (New Wave Research Co.) Laser Wavelength Spot diameter Repetition rate Fluence Laser energy Scan speed Carrier gas flow rate (Ar gas) ELEMENT XR (Thermo Fisher Sci.) Coolant gas flow rate Auxiliary gas flow rate RF power Mass window Time per pass Resolution a b

Nd; YAG 213 nm 100 μm 20 Hz 1.54–1.73a or 4.01–4.43b J/cm2 0.121–0.137a or 0.314–0.348b mJ 100a or 50b μm/s 1.23 L/min 16 L/min 0.80 L/min 1264 W 10% 2.0 s Medium resolution (Δm/m = 8000: FWHM)

Operating parameters for the analysis of Nodule-A. Operating parameters for Nodule-B.

J. Hirata et al. / Chemical Geology 427 (2016) 65–72

2.3. Sample preparation Ferromanganese nodules were provided by Japan Oil, Gas and Metals National Corporation (JOGMEC). The ferromanganese nodules in Fig. 1(a) and (b) are shown as “Nodule-A” and “Nodule-B”, respectively. In the case of discoidal ferromanganese nodules, glossy and less convex surface seem to be exposed to seawater, and more convex and rough surface seem to be buried in mud. According to visual observation of the samples in Fig. 1, the upper portion was exposed to seawater and lower portion was buried in mud. Nodule A and Nodule B were collected near Hawaii in the Pacific Ocean as shown in Fig. 2 (approximately 10°10′ N, 147°00′ W). The location shown in Fig. 2 is approximate, because precise sampling location was not published by JOGMEC. The size of Nodule-A was 6 cm × 5 cm × 3 cm and it was cut into half across the center (6 cm × 2.5 cm × 3 cm) as shown in Fig. 1(a)-(1). One half is analyzed by LA-ICP-MS with data reported in this paper. Whereas the cut obtained from second half has been published elsewhere (Hirata and Tanaka, 2014) which proves homogeneity of Mg in the sample. The size of Nodule-B was 6 cm × 5 cm × 4 cm and the half size of it after cutting was 6 cm × 2.5 cm × 4 cm (Fig. 1(b)-(1)). The circles in Fig. 1(a)-(2) and (b)-(2) indicate the area analyzed by LA-ICP-MS. 2.4. Determination of elements in the chip powder of ferromanganese nodules by ICP-MS To compare the elemental contents obtained by LA-ICP-MS, all of the chip powder from each sample (Fig. 1(a)-(1) and (b)-(1)) was collected, mixed, and an aliquot was decomposed in acid for ICP-MS analysis. These chip powders were appropriate for analysis, as they were considered to contain average contents of elements in the ferromanganese nodules over their cross section. The dissolved samples were diluted with ultrapure water (18.2 MΩ). Indium and bismuth solutions were added as internal standards, and the elemental contents in the chip powder samples were determined by ICP-MS. 2.5. Elemental imaging by LA-ICP-MS A 5 mm × 20 mm area on Nodule-A was analyzed by LA-ICP-MS, as shown in Fig. 1(a)-(2). In accordance with the conditions given in Table 1, analysis along a line of 20 mm length was performed on the cross section of the ferromanganese nodule. 100 peak intensities were obtained for each element, where the number of peak intensities was calculated from scan speed of LA (100 μm/s) and time per pass of mass scan (2 s). The line analysis was repeated 50 times, so that a total area of 5 mm × 20 mm was analyzed and a total of 5000 peak intensities were obtained for each element. The contents of elements were calculated from these peak intensities and converted into colors using a two-dimensional plotting system as discussed later.

67

To confirm the reproducibility, 6 mm × 20 mm area on Nodule-B was also analyzed, as shown in Fig. 1(b)-(2). Analysis along a line of 20 mm length was performed 60 times, in accordance with the conditions in Table 1. In total, 12,000 peak intensities were obtained for each element and the data was treated similarly as in Nodule-A. 2.6. Calibration lines and internal standard correction using Mg Elemental fractionation derived from laser ablation is serious problem for the determination analysis by LA-ICP-MS. Characteristics of elemental fractionation depend on sample matrix. In general, matrixmatched standard for the calibration of LA-ICP-MS was prepared to solve the problem of elemental fractionation (Koch and Günther, 2011; Russo et al., 2002). In previous studies, unknown samples were determined using just one matrix-matched calibration standard (Chen and Simonetti, 2013; Zhi et al., 2007) because it is difficult to prepare various contents of a matrix-matched standards. However, it is problematic for elemental imaging because the elemental contents to be determined vary over a wide range. When only one standard is used for calibration, reliability of the determined values is questionable. To obtain accurate determined values by LA-ICP-MS, especially in the case of elemental imaging, preparation of various content of matrixmatched standard is preferable. Preparation technique of calibration standards was developed, to perform as accurate as possible an analysis (Hirata et al., 2013). The reference material JMn1 was diluted with MnO2 in weight ratios of 25, 50, 75, and 100% by careful mixing with an agate mortar and pestle. These powder samples were hand-pressed into pellets used for infrared (IR) analysis, and the samples were used to derive matrix-matched standards for LA-ICP-MS. Calibration lines consisting of four standards were obtained by this method to achieve an accurate as possible determination. In general, because sample introduction by laser ablation is unstable, differences in sample introduction and ionization efficiency during measurements by LA-ICP-MS were corrected by internal standardization, (Becker et al., 2014; Konz et al., 2012). To improve the determination, internal standardization was also developed in our previous study (Hirata et al., 2013). An appropriate element to be used as an internal standard was considered and Mg was selected. Firstly, standard pellets were analyzed by LA-ICP-MS, where each pellet was analyzed using different LA conditions to check the availability of Mg as an internal standard. It was demonstrated by the experiment that an accurate calibration line can be obtained even under unstable LA settings owing to internal standard correction by Mg. These results demonstrate that Mg can be used to correct variations in ionization and sample introduction efficiency throughout measurements. To confirm above developed determination method by LA-ICP-MS, an unknown powdered ferromanganese nodule sample was analyzed by LA-ICP-MS. The results obtained by LA-ICP-MS were in good agreement with those by ICP-MS with acid decomposition procedure. For example, determined values of an unknown sample by ICP-MS were 53.0 g/kg for Fe, 11.7 g/kg for Cu, and 0.303 g/kg for Ce, respectively, whereas values by LA-ICP-MS using internal standard correction of Mg were 52.8 g/kg for Fe, 11.2 g/kg for Cu, and 0.296 g/kg for Ce, respectively. Thus, validity of a preparation procedure for the matrix matched standard and internal standard correction using Mg was confirmed in our previous study (Hirata et al., 2013). 2.7. Evaluation of the homogeneity of Mg in Nodule-A

Fig. 1. Photograph of ferromanganese nodules used in this study. (a) Photographs of Nodule-A. (b) Photographs of Nodule-B. All the chip powder from Nodule-A and B was collected for analysis by ICP-MS respectively. The circles indicate the area analyzed by LA-ICP-MS.

To apply internal standard correction for the spatial resolution analysis of a layered structure, the homogeneity of the required element is important. The elemental distributions from the second half of Nodule-A were examined by spatial resolution analysis using LA-ICPMS and ICP-MS to consider an appropriate element to be used as an internal standard in our previous study.

68

J. Hirata et al. / Chemical Geology 427 (2016) 65–72

Fig. 2. Map of the sampling location. The map was modified from google earth (http://earth.google.com). Sampling location data is approximate because precise sampling location was not published by JOGMEC.

To confirm homogeneity and applicability of Mg for internal standardization, elemental distributions of Nodule-A was examined in our previous study (Hirata and Tanaka, 2014). Overview of the experiments was shown in “Fig. 1 Preparation of the ferromanganese nodule sample” of this previous study. In the paper, Nodule-A (left side second half sample in Fig. 1(a)) was analyzed at 30 points by LA-ICP-MS. To compare the elemental contents obtained by LA-ICP-MS, scratch samples were obtained from five separated areas near the analyzed points of LA-ICPMS as shown in Fig. 1 of our previous study. Powdered samples were obtained by scratching 5 separated spots near the analyzed points of LAICP-MS and the obtained solid samples were determined by ICP-MS with acid decomposition procedure. Relative standard deviation (RSD%) of the peak intensities of Mg obtained by LA-ICP-MS at 30 points was checked. To calculate RSD%, data in “Fig. 2 Determination of Mg, Fe, and Mo in a ferromanganese nodule using LA-ICP-MS and ICP-MS” of our previous study (Hirata and Tanaka, 2014) was used. Although this figure indicates determined values by external calibration method, RSD% calculated from these data is the same meaning of RSD% from absolute peak intensities obtained by LA-ICP-MS. RSD% of Mg peak intensities was 25% and that of Fe peak intensities was 50%. The RSD% of Mg and Fe peak intensities are based on data derived from two types of information: one is the “real” distribution of the elemental content in Nodule-A, and the other is the variation of the rate of introduction during laser ablation. RSD% of 10–14% was obtained for the analysis of pellet sample by LA-ICP-MS (Hirata et al., 2013), the surface of which is much more smooth than that of Nodule-A. Data precision obtained by distribution analysis seemed to be worse than 10–14%. Taking into account above results, Mg is homogenously distributed in Nodule-A enough to apply for internal standardization. Furthermore, homogeneity of Mg was confirmed with the determined values obtained by ICP-MS. The contents of Mg by ICP-MS ranged from 17 to 24 g/kg as shown in “Table 3 Determined values of spots 1– 30 by LA-ICP-MS and Sample A-E by ICP-MS” of our previous study (Hirata and Tanaka, 2014). A Mg content range of 13 to 18 g/kg was also reported (Duliu et al., 2009) for a ferromanganese nodule collected from the Pacific Ocean (12–12°40′ N, 137°40–138°50 W), close to the location of the ferromanganese nodule in this study. These results support that Mg is “relatively homogenous” in our ferromanganese nodules.

To confirm the validity of determined values by LA-ICP-MS using Mg as an internal standard, results between LA-ICP-MS and ICP-MS were compared. Variation of Mo determined by LA-ICP-MS and ICP-MS was shown in “Fig. 3(b) Determinations of Mg, Mo, and Fe by LA-ICP-MS using internal standard correction” of our pervious study (Hirata and Tanaka, 2014). It was found that the elemental content obtained by LA-ICP-MS was in good agreement with the results of ICP-MS. Thus, homogeneity of Mg and applicability of Mg as an internal standard were confirmed and ferromanganese nodule element contents were adequately determined as accurate as possible. 2.8. Two-dimensional plotting system The calculation to determine the elemental contents and the statistical calculations in this study were performed on EXCEL (Microsoft Co.). The determined values were imported into a two-dimensional plotting system, which was set up using Matplotlib, a Python two-dimensional plotting library (Hunter, 2007). Different ranges of elemental contents were expressed in different colors along with spatial information. Tones of colors can be selected from a wide variety of color tones in this two-dimensional plotting system. In this work, high contents were expressed in white, and low contents were expressed in black. Medium contents were expressed in blue, purple, and yellow in order of increasing content. 3. Results and discussion 3.1. Validity of calibration lines for determination of elemental contents Calibration lines were obtained for determining elemental contents in ferromanganese nodules by LA-ICP-MS. In this study, standards were prepared using JMn1, a reference material for the ferromanganese nodule. Under the LA conditions in Table 1, a laser was beamed on the standards by a raster scan for 60–80 s. This raster scan was repeated 4–6 times on each standard and the best data were selected to form calibration lines. Internal standard correction using Mg was applied for the calibration lines. From the calibration lines containing four standards, values of the square of the correlation coefficient (r2) are shown in Table 2 for

J. Hirata et al. / Chemical Geology 427 (2016) 65–72

69

Fig. 3. Imaging of elemental distributions in Nodule-A by LA-ICP-MS. (a) Images of elements in Nodule-A obtained from peak intensities. (b) Images of elemental contents in Nodule-A, where white and black colors represent the maximum and minimum determined values, respectively. (c) Images of elemental contents in Nodule-A, where white and black colors represent the μ + σ and the μ − σ, respectively.

Nodule-A. Values of r2 of at least 0.99 were obtained for Fe, Co, Cu, Mo, La, Ce, and Nd, and calibration lines with acceptable linearity were obtained, except for Mn and W. The calibration line for W was not linear, owing to the low peak intensity of W. The content of W was therefore

determined using only one standard consisting of 100% JMn1. In the case for Mn, MnO2 was used for the dilution of JMn1 in other standards and it was difficult to obtain an acceptable calibration line, pure JMn1 was used as the standard.

Table 2 Determination of elemental contents in Nodule-A by LA-ICP-MS and ICP-MS. LA-ICP-MS

Mn Fe Co Ni Cu Mo La Ce Nd W

Average (μ) g/kg

Maximum g/kg

Minimum g/kg

S. D. (σ)

2.7 × 102 45 1.9 7.9 8.6 0.55 6.9 × 10−2 0.24 8.2 × 10−2 5.6 × 10−2

5.8 × 102 2.0 × 102 7.3 20 26 5.5 0.49 2.0 0.57 0.64

49 0.20 0.37 0.34 1.4 0.10 1.6 × 10−2 5.6 × 10−2 1.4 × 10−2 1.1 × 10−2

62 32 0.66 2.2 2.1 0.18 3.6 × 10−2 0.14 4.2 × 10−2 1.9 × 10−2

μ: average values determined by LA-ICP-MS (n = 5000). [ICP-MS]: values of elements in chip powder shown in Fig. 1(a)-(1) determined by ICP-MS. Maximum and minimum: highest and lowest contents among 5000 values determined by LA-ICP-MS. σ: standard deviation (S. D.) calculated from 5000 values determined by LA-ICP-MS. r2: square of correlation coefficient calculated from calibration lines for LA-ICP-MS.

[ICP-MS] g/kg

[LA-ICP-MS]/ [ICP-MS]

r2 values of calibration lines

2.8 × 102 53 2.2 16 12 0.53 9.3 × 10−2 0.30 0.11 5.8 × 10−2

0.96 0.85 0.86 0.49 0.72 1.0 0.74 0.80 0.74 0.97

– 0.990 1.00 0.987 0.991 0.998 0.994 0.996 0.996 –

70

J. Hirata et al. / Chemical Geology 427 (2016) 65–72

3.2. Validity of elemental contents in ferromanganese nodule obtained by LA-ICP-MS The peak intensities of Nodule-A obtained by LA-ICP-MS were examined using both the calibration line and the internal standard methods to obtain the contents of each element. To correct the discrepancy between the contents of Mg in the standard and in the ferromanganese nodule, a factor (Hirata et al., 2013) was calculated using the reference value of JMn1 and the value obtained from the chip powder by ICP-MS after acid decomposition. For evaluating the accuracy of the elemental contents obtained by LA-ICP-MS, the values in Nodule-A determined by LA-ICP-MS were compared with those of the chip powder decomposed with acid then subjected to ICP-MS measurements (Fig. 1(a)-(1)). The selected area on Nodule-A was analyzed by LA-ICP-MS under the conditions mentioned earlier. The peak intensities of the target elements were corrected using the peak intensity of Mg as an internal standard. From the corrected peak intensity and the calibration lines, 5000 values of contents were determined for each target element. The calculation was performed using EXCEL and the data were converted to CSV files. The average values (μ) and standard deviations (σ) of elemental contents in Nodule-A obtained by imaging analysis using LA-ICP-MS are shown in Table 2. To evaluate the validity of the values determined by LA-ICP-MS, the chip powder in Fig. 1(a)-(1) was analyzed by ICP-MS and the results are also shown in Table 2. For ease of comparison, the ratios between the values determined by LA-ICP-MS and by ICP-MS are shown as [LA-ICP-MS]/[ICP-MS]. As can be seen in the table, the ratios for Fe, Mo, and La are 0.85, 1.0, and 0.74, respectively. To check the consistency, determined values for each element by LA-ICP-MS were plotted against those by ICP-MS with acid decomposition. Spearman's rank coefficient of correlations from the plots was 1.00. Thus, the values determined by LA-ICP-MS were consistent with those determined by ICP-MS for 9 elements except for Ni. 3.3. Consideration of peak intensities of Mg obtained by imaging analysis As mentioned in the Experimental section, Mg was considered to be a reasonable internal standard for the ferromanganese nodule. That is, effectiveness of Mg as an internal standard was considered (Hirata et al., 2013), and the homogeneity of Mg in second half of Nodule-A was examined by spatial resolution analysis using LA-ICP-MS and ICPMS (Hirata and Tanaka, 2014). Although applicability of Mg was confirmed in our previous studies, it is important to consider the peak intensities of Mg obtained by imaging analysis using LA-ICP-MS. The elemental distribution of major elements were examined using the peak intensities, as shown in Fig. 3(a), and a photograph of the target area is also shown on the right of the figure. The distributions of Mg, Mn, and Fe in Fig. 3(a) are based on data derived from two types of information: one is the “real” distribution of the elemental content in the ferromanganese nodule, and the other is the variation of the rate of introduction during laser ablation. To apply the internal standardization for imaging data, one data value each for, Mg, Mn and Fe was selected and corrected for the sample introduction rate. Moreover, to evaluate the original distributions of elements in Nodule-A, the peak intensity ratios between the highest peak and the lowest peak for each element were used. The peak intensity ratios were 42 for Mg, 57 for Mn, and 150 for Fe. As the peak intensity ratio of Mg was lowest, Mg appears to be relatively homogeneously distributed in Nodule-A justifying the use of Mg as an internal standard. The reasons for Mg peak intensity ratios to vary by a factor of 42, could be attributed to absolute count rates affected by ablation yield because of the following two factors. Firstly, sample introduction efficiency varies depending on the ablation position in the sample cell of LA system, because velocity of carrier gas is not constant in a sample cell. From the simulation by Koch and Günther (2011), carrier gas velocity in the sample cell approximately varied from 0.1 to 10 m/s. To elucidate

elemental distributions of the layered structures in ferromanganese nodules, larger area was analyzed in our study as compared to previous study (Becker et al., 2014). Larger analytical area increases variation of the range of Mg count, because carrier gas velocity is varied depending on the position in sample cell of LA system. Secondly, generation of outlier cannot be avoided throughout the imaging analysis. It is difficult to obtain flat and smooth surface for imaging analysis, because ferromanganese nodule is a porous sample. The asperity on the sample surface could generate outlier owing to defocusing of the laser ablation. Mg count rates become larger when outliers are obtained. Considering these problems, Mg peak intensity ratios vary by a factor of 42 which in LA-ICP-MS is acceptable. 3.4. Quantitative imaging of elements in ferromanganese nodule by LA-ICP-MS The data obtained from the quantitative elemental imaging of Nodule-A by LA-ICP-MS using Mg as an internal standard are shown in Fig. 2(b). Here, the highest content of an element obtained from 5000 data points is shown in white and the lowest content is shown in black. The maximum and minimum contents for each element are shown in Table 2 and Fig. 3(b). From the elemental images in Fig. 3(b), the following observations were made. (1) The contents of Fe and rare-earth elements are enriched near the core of the ferromanganese nodule. This result could imply elemental content in the core fragment inside of the Nodule-A (Rizescu et al., 2001). (2) The distribution of Cu was different from that of Fe. (3) The elemental distributions of Mn and Fe were concentrated at the border of the layered structure. (4) Low homogeneity was observed for the distributions of Mo and W as compared with that of Fe. The elemental distributions along the layered structure and the difference in the distributions among the elements were elucidated by LA-ICP-MS. Here, the imaging of an element requires 5000 data points, among which several data points may be out of range. For instance, in the imaging of Mo and W by LA-ICP-MS, detailed layered structures are not visible. To improve the contrast in elemental images, the average values (μ) and standard deviations (σ) were calculated (Table 2). The content ranges used to define colors in the plotting system were limited to within “μ ± σ” for the 5000 determined values, and the corrected data are shown in Fig. 3(c). As can be seen in the images, the distributions of Mo and W show the layered structures with clearly depicted contrast. The simple method describes the ease with which the colors can be changed to define the content range of elements. Estimation of data precision is also significant however imaging analysis essentially requires a very large amount of data. Thus to obtain a fair precision, as an example, four of Fe contents were extracted from the core section indicated by an arrow in the data for Fe in Fig. 3(c). The nodules at these four points are expected to be formed approximately in the same period and contain similar Fe contents. Therefore, these four determined value of Fe contents were used as an approximate estimation for a precision obtained by imaging analysis. Average content of these four Fe contents was 97 g/kg and standard deviation of those was 9.4. Although this estimation of precision does not indicate the “real” precision, the results indicate elemental contents by LA-ICP-MS were applicable for geochemical discussions. Thus, quantitative imaging of elements in Nodule-A using LA-ICP-MS was performed as accurate as possible. 3.5. Reproducibility of the quantitative elemental imaging of ferromanganese nodule To establish a quantitative elemental imaging for ferromanganese nodules using LA-ICP-MS, reproducibility of the method was performed by measuring Mn, Fe, Cu, La, Nd, and Ce in Nodule-B (Fig. 1(b)-(2)) following the same process as in Nodule-A. That is, almost the same conditions of LA-ICP-MS (Table 1), calibration line and Mg as internal

J. Hirata et al. / Chemical Geology 427 (2016) 65–72

71

Table 3 Determination of elemental contents in Nodule-B by LA-ICP-MS and ICP-MS. LA-ICP-MS

Mn Fe Cu La Ce Nd

Average (μ) g/kg

Maximum g/kg

Minimum g/kg

S. D. (σ)

3.3 × 102 65 13 9.7 × 10−2 0.36 0.13

6.4 × 102 2.2 × 102 33 0.59 1.5 0.83

66 8.3 2.1 1.7 × 10−2 4.0 × 10−2 2.5 × 10−2

78 37 3.2 5.9 × 10−2 0.22 6.9 × 10−2

[ICP-MS] g/kg

[LA-ICP-MS]/ [ICP-MS]

r2 values of calibration lines

3.1 × 102 54 11 9.5 × 10−2 0.33 0.12

1.0 1.2 1.1 1.0 1.1 1.1

– 0.998 1.00 0.998 0.999 0.991

μ: average values determined by LA-ICP-MS (n = 12,000). [ICP-MS]: values of elements in chip powder shown in Fig. 1(b)-(1) determined by ICP-MS. Maximum and minimum: highest and lowest contents among 12,000 values determined by LA-ICP-MS. σ: standard deviation (S. D.) calculated from 12,000 values determined by LA-ICP-MS. r2: square of correlation coefficient calculated from calibration lines for LA-ICP-MS.

standard were applied for determination of elemental content in Nodule-B, and the elemental imaging was performed by the twodimensional plotting system. To confirm the validity of elemental contents in Nodule-B determined by LA-ICP-MS, the data were compared with those obtained by ICP-MS with acid decomposition procedure. The data obtained is shown in Table 3. As can be seen in the table, the values of [LA-ICPMS]/[ICP-MS] for Mn, Fe, Cu, La, Nd, and Ce are 1.0, 1.2, 1.1, 1.0, 1.1, and 1.1 respectively. To check the consistency, determined value for each element by LA-ICP-MS was plotted against those by ICP-MS with acid decomposition and the correlations from the plots were 1.00. Although Nodule-B was a different sample from Nodule-A, accurate elemental contents could be obtained by LA-ICP-MS, as seen by the consistent data by ICP-MS. Elemental imaging of Nodule-B was also performed using the method mentioned earlier. From the elemental images in Fig. 4, the following observations were made. (1) The content of Fe and rare-earth elements are enriched near the core of the ferromanganese nodule. This result could imply elemental content in the core fragment inside of the Nodule-B (Rizescu et al., 2001). (2) The distribution of Cu was different

from that of Fe. (3) The elemental distribution of Fe and Mn were concentrated at the borders of the layered structure. These observations of Nodule-B were in good agreement with those of Nodule-A. The reproducibility implies usefulness of Mg as internal standard and the determination methods for ferromanganese samples. Thus, the quantitative elemental imaging for ferromanganese nodules using LA-ICP-MS was established. It was also confirmed that contents of both major and trace elements could be determined. The contents and distributions of elements are visually depicted, allowing us to elucidate the formation processes of the layered structure. This method of imaging and determining the contents of both major and trace elements is a powerful tool for obtaining information on ferromanganese nodules. The distribution of elemental contents gives us significant basic information on ferromanganese nodules, from which geochemical parameters could be established in the layered structure. For example, growth rate can be calculated from elemental contents (Frank et al., 1999; Lyle, 1982) and Ce/La (Ce anomaly) is a significant indicator of the paleo-environments (Aplin, 1984). The detailed spatial distributions and content images of both major and trace elements in ferromanganese nodules will provide imaging of geochemical parameters.

Fig. 4. Imaging of elemental distributions in Nodule-B by LA-ICP-MS.

72

J. Hirata et al. / Chemical Geology 427 (2016) 65–72

4. Conclusion The quantitative imaging of elements in ferromanganese nodules by LA-ICP-MS was first accomplished by applying a calibration line and internal standard method to obtain elemental contents and distributions in ferromanganese nodules. The contents of elements determined by LA-ICP-MS, such as Mn, Fe, Mo, and rare-earth elements, were in good agreement with values obtained by ICP-MS. A two-dimensional plotting system for elemental imaging by LAICP-MS was established using Matplotlib, which is a Python twodimensional plotting library. The distributions of elements in a ferromanganese nodule were imaged by colors. To obtain better contrast in elemental imaging, the colors corresponding to elemental content ranges can be easily changed by the plotting system. Furthermore, reproducibility of this analytical method was confirmed through the analysis of two different ferromanganese nodule samples. Although LA-ICPMS is widely applied in biological samples, this is an initial step toward the imaging of multi-elements in geological materials. It is expected that by applying this analytical method to a number of ferromanganese nodules and imaging of geochemical parameters, new information on the formation mechanism of ferromanganese nodules will be obtained. Acknowledgments We would like to thank Japan Oil, Gas and Metals National Corporation (JOGMEC) for providing the ferromanganese nodule. We also greatly appreciate help from Yamamoto K. to develop the two-dimensional plotting system, and Iwata S. to calculate statistical work. Appendix A. Supplementary data Supplementary data associated with this article can be found in the online version, at doi: http://dx.doi.org/10.1016/j.chemgeo.2016.02. 017. These data include the Google map of the most important areas described in this article. References Aplin, A.C., 1984. Rare earth element geochemistry of Central Pacific ferromanganese encrustations. Earth Planet. Sci. Lett. 71, 13–22. Axelsson, M.D., Rodushkin, I., Baxter, D.C., Ingri, J., Öhlander, B., 2002. High spatial resolution analysis of ferromanganese concretions by LA-ICP-MS. Geochem. Trans. 3, 40–47. Barst, B.D., Gevertz, A.K., Chumchal, M.M., Smith, J.D., Rainwater, T.R., Drevnick, P.E., Hudelson, K.E., Hart, A., Verbeck, G.F., Roberts, A.P., 2011. Laser ablation ICP-MS Colocalization of mercury and immune response in fish. Environ. Sci. Technol. 45, 8982–8988. Becker, J.S., Matusch, A., Wu, B., 2014. Bioimaging mass spectrometry of trace elements — recent advance and applications of LA-ICP-MS: a review. Anal. Chim. Acta 835, 1–18. Bollhöfer, A., Eisenhauer, A., Frank, N., Pech, D., Mangini, A., 1996. Thorium and uranium isotopes in a manganese nodule from the Peru basin determined by alpha spectrometry and thermal ionization mass spectrometry (TIMS): are manganese supply and growth related to climate? Geol. Rundsch. 85, 577–585. Carlo, D.H.E., McMurtry, M.G., 1992. Rare-earth element geochemistry of ferromanganese crusts from the Hawaiian Archipelago, central Pacific. Chem. Geol. 95, 235–250. Chen, W., Simonetti, A., 2013. In-situ determination of major and trace elements in calcite and apatite, and U–Pb ages of apatite from the Oka carbonatite complex: insights into a complex crystallization history. Chem. Geol. 353, 151–172. Duliu, O.G., Alexe, V., Moutte, J., Szobotca, S.A., 2009. Major and trace element distributions in manganese nodules and micronodules as well as abyssal clay from the Clarion-Clipperton abyssal plain, Northeast Pacific. Geo-Mar. Lett. 29, 71–83. Dulski, P., 2001. Reference materials for geochemical studies: new analytical data by ICPMS and critical discussion of reference values. Geostand. Newslett. 25, 87–125. Fernández, B., Costa, J.M., Pereiro, R., Sanz-Medel, A., 2010. Inorganic mass spectrometry as a tool for characterization at the nanoscale. Anal. Bioanal. Chem. 396, 15–29. Frank, M., O'Nions, R.K., Hein, J.R., Banakar, V.K., 1999. 60 Myr records of major elements and Pb–Nd isotopes from hydrogenous ferromanganese crusts: reconstruction of seawater paleochemistry. Geochim. Cosmochim. Acta 63, 1689–1708. Fricker, M.B., Kutscher, D., Aeschlimann, B., Frommer, J., Dietiker, R., Bettmer, J., Günther, D., 2011. High spatial resolution trace element analysis by LA-ICP-MS using a novel ablation cell for multiple or large samples. Int. J. Mass Spectrom. 307, 39–45. Garbe-Schönberg, C.D., McMurtry, G.M., 1994. In-situ micro-analysis of platinum and rare earths in ferromanganese crusts by laser ablation-ICP-MS (LAICPMS). Fresenius J. Anal. Chem. 350, 264–271.

Gasparatos, D., Tarenidis, D., Haidouti, C., Oikonomou, G., 2005. Microscopic structure of soil Fe-Mn nodules: environmental implication. Environ. Chem. Lett. 2, 175–178. Glasby, G.P., Gwozdz, R., Kunzendorf, H., Friedrich, G., Thijssen, T., 1987. The distribution of rare earth and minor elements in manganese nodules and sediments from the equatorial and S.W. Pacific. Lithos 20, 97–113. Hein, J.R., Mizell, K., Koschinsky, A., Conrad, T.A., 2013. Deep-ocean mineral deposits as a source of critical metals for high- and green-technology applications: comparison with land-based resources. Ore Geol. Rev. 51, 1–14. Hirata, J., Tanaka, M., 2014. Determination of trace elements across annual ring structure of ferromanganese nodule by laser ablation-ICP-MS. Bunseki Kagaku 63, 763–773. Hirata, J., Takahashi, K., Tanaka, M., 2013. Determination method of multi elements in ferromanganese samples by LA-ICP-MS. Anal. Sci. 29, 151–155. Hlawatsch, S., Garbe-Schönberg, C.D., Lechtenberg, F., Manceau, A., Tamura, N., Kulik, D.A., Kersten, M., 2002. Trace metal fluxes to ferromanganese nodules from the western Baltic Sea as a record for long-term environmental changes. Chem. Geol. 182, 697–709. Hoffmann, E., Lüdke, C., Scholze, H., 1997. Is laser-ablation-ICP-MS an alternative to solution analysis of solid samples? Fresenius J. Anal. Chem. 359, 394–398. Hunter, J.D., 2007. Matplotlib: a 2D graphics environment. Comput. Sci. Eng. 9, 90–95. Kashiwabara, T., Takahashi, Y., Tanimizu, M., 2009. A XAFS study on the mechanism of isotopic fractionation of molybdenum during its adsorption on ferromanganese oxides. Geochem. J. 43, e31–e36. Koch, J., Günther, D., 2011. Review of the state-of-the-art of laser ablation inductively coupled plasma mass spectrometry. Appl. Spectrosc. 65, 155A–162A. Konz, I., Fernández, B., Fernández, M.L., Pereiro, R., Sanz-Medel, A., 2012. Laser ablation ICPMS for quantitative biomedical applications. Anal. Bioanal. Chem. 403, 2113–2125. Konz, I., Fernández, B., Fernández, M.L., Pereiro, R., Gonzá-lez, H., Álvarez, L., Coca-Prados, M., Sanz-Medel, A., 2013. Gold internal standard correction for elemental imaging of soft tissue sections by LA-ICP-MS: element distribution in eye microstructures. Anal. Bioanal. Chem. 405, 3091–3096. Koschinsky, A., Halbach, P., 1995. Sequential leaching of marine ferromanganese precipitates: genetic implications. Geochim. Cosmochim. Acta 59, 5113–5132. Koschinsky, A., Hein, J.R., 2003. Uptake of elements from seawater by ferromanganese crusts: solid-phase associations and seawater speciation. Mar. Geol. 198, 331–351. Kunzendorf, H., Glasby, G.P., Stoffers, P., Plüger, W.L., 1993. The distribution of rare earth and minor elements in manganese nodules, micronodules and sediments along an east–west transect in the southern Pacific. Lithos 30, 45–56. Lear, J., Hare, D.J., Fryer, F., Adlard, P.A., Finkelstein, D.I., Doble, P.A., 2012. High-resolution elemental bioimaging of Ca, Mn, Fe, Co, Cu, and Zn employing LA-ICP-MS and hydrogen reaction gas. Anal. Chem. 84, 6707–6714. Lyle, M., 1982. Estimating growth rates of ferromanganese nodules from chemical compositions: implications for nodule formation proceses. Geochim. Cosmochim. Acta 46, 2301–2306. Manceau, A., Tamura, N., Celestre, R.S., Macdowell, A.A., Geoffroy, N., Sposito, G., Padmore, H.A., 2003. Molecular-scale speciation of Zn and Ni in soil ferromanganese nodules from loess soils of the Mississippi basin. Environ. Sci. Technol. 37, 75–80. Mohwinkel, D., Kleint, C., Koschinsky, A., 2014. Phase associations and potential selective extraction methods for selected high-tech metals from ferromanganese nodules and crusts with siderophores. Appl. Geochem. 43, 13–21. Moreno-Gordaliza, E., Giesen, C., Lázaro, A., Esteban-Fernández, D., Humanes, B., Cañas, B., Panne, U., Tejedor, A., Jakubowski, N., Gómez-Gómez, M.M., 2011. Elemental bioimaging in kidney by LA-ICP-MS as a tool to study nephrotoxicity and renal protective strategies in cisplatin therapies. Anal. Chem. 83, 7933–7940. Palumbo, B., Bellanca, A., Neri, R., Roe, M.J., 2001. Trace metal partitioning in Fe–Mn nodules from Sicilian soils, Italy. Chem. Geol. 173, 257–269. Rizescu, C.T., Georgescu, G.N., Duliu, O.G., Szobotka, S.A., 2001. 3-D dual gamma-ray computer axial tomography investigation of polymetallic nodules. Deep Sea Res., Part I 48, 2529–2540. Russo, R.E., Mao, X., Liu, H., Gonzalez, J., Mao, S.S., 2002. Laser ablation in analytical chemistry—a review. Talanta 57, 425–451. Shaheen, M., Gagnon, J.E., Yang, Z., Fryer, J.B., 2008. Evaluation of the analytical performance of femtosecond laser ablation inductively coupled plasma mass spectrometry at 785 nm with glass reference materials. J. Anal. At. Spectrom. 23, 1610–1621. Shaheen, M.E., Gagnon, J.E., Fryer, B.J., 2012. Femtosecond (fs) lasers coupled with modern ICP-MS instruments provide new and improved potential for in situ elemental and isotopic analyses in the geosciences. Chem. Geol. 330-331, 260–273. Terashima, S., Usui, A., Imai, N., 1995. Two new GSJ geochemical reference samples: syenite JSy-1 and manganese nodule JMn-1. Geostand. Newslett. 19, 221–229. Urgast, D.S., Ellingsen, D.G., Berlinger, B., Eilertsen, E., Friisk, G., Skaug, V., Thomassen, Y., Beattie, J.H., Kwun, I., Feldmann, J., 2012. Multi-elemental bio-imaging of rat tissue from a study investigating the bioavailability of bismuth from shotgun pellets. Anal. Bioanal. Chem. 404, 89–99. von Blanckenburg, F., O'Nions, R.K., Belshaw, N.S., Gibb, A., Hein, J.R., 1996. Global distribution of beryllium isotopes in deep ocean water as derived from Fe-Mn crusts. Earth Planet. Sci. Lett. 141, 213–226. Wang, X., Gan, L., Wiens, M., Schloßmacher, U., Schröder, H.C., Müller, W.E.G., 2012. Distribution of microfossils within polymetallic nodules: biogenic clusters within manganese layers. Mar. Biotechnol. 14, 96–105. Wang, H.A.O., Grolimund, D., Giesen, C., Borca, C.N., Shaw-Stewart, J.R.H., Bodenmiller, B., Günther, D., 2013. Fast chemical imaging at high spatial resolution by laser ablation inductively coupled plasma mass spectrometry. Anal. Chem. 85, 10107–10116. Zhi, X.H., Hong, H.S., Chu, H.Z., Feng, J.J., Hong, W.X., 2007. Element distribution in cobaltrich crust profile by laser ablation-inductively coupled plasma-mass spectrometry. Chin. J. Anal. Chem. 35, 1099–1104. Zhu, Y., Hiroki, A., Itoh, A., Umemura, T., Haraguchi, H., Chiba, K., 2012. Relative enrichment of Mo in the radicle of peanut seed (Arachis hypogaea), observed by multielemental imagining with LA-ICP-MS. Anal. Sci. 28, 1121–1124.