Quantitative Chemical Imaging of Element Diffusion into

May 31, 2011 - sample to gain improved statistics. Due to the 1А2 s washout time of the aerosols, LA was stopped for 3 s before measurement of the ne...
1 downloads 0 Views 4MB Size
Download PDF
ARTICLE pubs.acs.org/ac

Quantitative Chemical Imaging of Element Diffusion into Heterogeneous Media Using Laser Ablation Inductively Coupled Plasma Mass Spectrometry, Synchrotron Micro-X-ray Fluorescence, and Extended X-ray Absorption Fine Structure Spectroscopy H. A. O. Wang,†,§ D. Grolimund,*,† L. R. Van Loon,‡ K. Barmettler,|| C. N. Borca,† B. Aeschlimann,§ and D. G€unther§ microXAS Beamline Project, Swiss Light Source, and ‡Laboratory for Waste Management, Paul Scherrer Institut, 5232 Villigen PSI, Switzerland § Trace Element and Micro Analysis Group, ETH Z€urich, Wolfgang-Pauli-Strasse 10, 8093 Z€urich, Switzerland Soil Chemistry, Institute of Biogeochemistry and Pollutant Dynamics, ETH Z€urich, Universit€atstrasse 16, 8092 Z€urich, Switzerland

)

†

bS Supporting Information ABSTRACT: Quantitative chemical imaging of trace elements in heterogeneous media is important for the fundamental understanding of a broad range of chemical and physical processes. The primary aim of this study was to develop an analytical methodology for quantitative high spatial resolution chemical imaging based on the complementary use of independent microanalytical techniques. The selected scientific case study is focused on high spatially resolved quantitative imaging of major elements, minor elements, and a trace element (Cs) in Opalinus clay, which has been proposed as the host rock for high-level radioactive waste repositories. Laser ablation inductively coupled plasma mass spectrometry (LA-ICPMS), providing quantitative chemical information, and synchrotron radiation based micro-X-ray fluorescence (SR-microXRF), providing high spatial resolution images, were applied to study Cs migration into Opalinus clay rock. The results indicate that combining the outputs achievable by the two independent techniques enhances the imaging capabilities significantly. The qualitative high resolution image of SR-microXRF is in good agreement with the quantitative image recorded with lower spatial resolution by LA-ICPMS. Combining both techniques, it was possible to determine that the Opalinus clay sample contains two distinct domains: (i) a clay mineral rich domain and (ii) a calcium carbonate dominated domain. The two domains are separated by sharp boundaries. The spatial Cs distribution is highly correlated to the distribution of the clay. Furthermore, extended X-ray absorption fine structure analysis indicates that the trace element Cs preferentially migrates into clay interlayers rather than into the calcite domain, which complements the results acquired by LA-ICPMS and SR-microXRF. By using complementary techniques, the quantification robustness was improved to quantitative micrometer spatial resolution. Such quantitative, microscale chemical images allow a more detailed understanding of the chemical reactive transport process into and within heterogeneous media to be gained.

S

mall-scale chemical and physical structures are gaining more and more attention, promoted by the boost of nanotechnology and related unexpectedly intriguing phenomena. A wide variety of techniques aiming at micrometer and nanometer scale imaging (this refers to chemical and physical as well as topographical images) have been developed.1 7 For daily or routine applications, high throughput even under reduced spatial resolution and with lower sensitivity is required for fast screening to process larger areas within samples or a large number of samples. Many techniques have been developed to aid in achieving this goal. Different excitation beams and detection schemes have been employed, resulting in various probe depths, spatial resolutions, and sensitivities. For example, surface-sensitive particleinduced X-ray emission (PIXE), which is based on charged r 2011 American Chemical Society

particle beam excitation, has the advantage of a resolution in the submicrometer range and allows detection limits within the range of 0.1 10 μg/g for a large number of elements.8 10 Secondary ion mass spectrometry (SIMS) coupled to time of flight (TOF) mass spectrometry has become a routine technique for solid-state sample surface imaging, with high spatial resolution in the range from a few micrometers11 down to tens of nanometers12 as well as high sensitivity.12 14 Inductively coupled plasma mass spectrometry (ICPMS) is currently one of the most sensitive techniques for elemental analyses and Received: April 8, 2011 Accepted: May 31, 2011 Published: May 31, 2011 6259

dx.doi.org/10.1021/ac200899x | Anal. Chem. 2011, 83, 6259–6266

Analytical Chemistry isotope ratio determinations. It is well suited for quantitative trace elements with limits of detection in the range from the subpicograms per gram level to the femtograms per gram level when bulk samples are dissolved prior to analyses. Advanced mass analyzers for ICPMS include double-focusing multicollector sector-field and TOF mass analyzers and provide various advantages, including high mass resolution and (quasi)simultaneous ion detection, respectively. Routinely used quadrupole mass analyzers have a limited mass resolution, but benefit from lower operational cost. In addition to solution analyses, laser ablation (LA) has become one of the most applied direct solid sampling techniques providing medium (micrometer range) spatial resolution.9,15 17 Since the laser technology allows stoichiometric and matrix-independent sampling, the combination of LA with ICPMS has been used to generate spatially resolved quantitative analysis. However, in comparison to SIMS or PIXE, its detection capability with respect to the sampled mass would require 100 1000-fold improved sensitivity to generate images of trace elements at a spatial resolution in the low 1 μm or nm range. Many reports on LA-ICPMS imaging are focused on bio/biomedical applications.15,18 Therefore, a selection of papers, closely related to heterogeneous solid samples, will be discussed here. The utilization of LA-ICPMS as a twodimensional (2D) imaging technique was reported by Treble et al.19 They studied the elemental distribution in stalagmite samples by using a rectangular slit aperture of 5  50 μm2 and a circular spot size of 32 μm, respectively, and imaged trace element distributions of in total seven elements. The raw intensities were first moving averaged with three points, then internally calibrated by Ca, and finally calculated into concentration using NIST612 as an external calibration standard. Furthermore, Woodhead et al.20 demonstrated the potential of LA-ICPMS by generating a qualitative surface image of trace elements on speleothem samples using a 157 μm spot size. In addition, isotope ratio imaging (87Sr/86Sr) has been shown with LA coupled to multicollector ICPMS on a barramundi otolith sample by employing a spot size of 71 μm, which is mainly required for precise isotope ratio determination. More recently, the introduction of near-field laser ablationICPMS, a technique based on a focused laser beam to a diameter of a few hundreds of nanometers, demonstrated that various arrangements7,21 can be applied for solid sampling in the nanometer domain. However, the sensitivity of ICPMS allows only the detection of major elements (upper weight percentage) when using a 100 nm spatial resolution (stated by Becker et al.22 and in the erratum23). However, despite the advanced focusing techniques, spatial resolution, sensitivity, and data acquisition speed remain limitations of LA-ICPMS to achieve high resolution images. Synchrotron radiation based micro-X-ray fluorescence (SRmicroXRF) analysis is more and more accessible for routine analysis. With a high flux of photons, tunable wavelengths, and advanced X-ray focusing techniques, SR-microXRF provides semiquantitative images of trace elements in various media, which can be acquired even under atmospheric conditions. Furthermore, it can be adapted to various sample shapes and environments and is a well-suited technique for acquiring high spatial resolution data at fast speed. For example, using a 1 μm step, one can image an area of 500  500 μm2 in less than 10 h. However, the sensitivity is still much lower than commonly obtained in mass spectrometry, and quantitative analyses are limited to homogeneous samples where matrix-matched calibration

ARTICLE

materials are available. Nevertheless, SR-microXRF is a widely used technique for chemical imaging, and intensive studies have been carried out on various types of samples, including archeological samples,24 biological tissues,25,26 and environmental/geological samples.27 29 It is often combined with other X-ray microprobe techniques, such as synchrotron radiation based micro-X-ray diffraction (SR-microXRD), or synchrotron radiation based micro-X-ray absorption spectroscopy (SR-microXAS), which complement the chemical images with crystallographic and molecular information.28,30 A recently developed confocal microscopic SR-microXRF technique allows imaging of samples in three dimensions (3D).31 33 However, X-ray microprobe techniques are often prone to suffer from matrix effects and interferences, which limit the quantification capabilities, precision, and accuracy of the images, especially when applied to heterogeneous media. This study was focused on exploring the capabilities of SRmicroXRF and LA-ICPMS for acquiring qualitative and quantitative distribution information (1D and 2D) of the trace element Cs in a heterogeneous material (Opalinus clay rock). Both complementary techniques were optimized and applied to identical sample areas of synthetically treated Opalinus clay rock to measure the distribution and concentration of elements. These measurements were used to determine the capabilities of each individual technique and their combination with respect to quantitative high spatial resolution imaging. The heterogeneous sample of interest in this study was a natural Opalinus clay rock treated with natural abundance 133Cs as a radioactive tracer simulant. Several countries consider clay formations as host rocks for deep geological nuclear waste repositories.34 The safe disposal of hazardous nuclear waste represents a very important challenge for modern societies. Therefore, it is of major importance to study the migration of radionuclides, such as Cs (and other hazardous chemicals), through this natural barrier material. The aim of this study was to develop an analytical methodology for quantitative high spatial resolution images of the Cs distribution after an extended migration time into a heterogeneous medium with respect to the precision and accuracy of quantification. To complement the results achieved with the microbeam techniques, extended X-ray absorption fine structure (EXAFS) analysis aiming for Cs molecular speciation information in Opalinus clay rock was included in the measurement protocol.

’ MATERIAL AND METHODS Samples. The samples investigated in this study originated from a unique field-scale migration experiment conducted at the Mont Terri Underground Rock Laboratory, located in northwestern Switzerland.35,36 Stable and natural abundance 133Cs (as a chemical analogue to radioactive Cs isotopes) from a multitracer solution was injected into a borehole, which was drilled in the Opalinus clay rock formation.35 After approximately 9 months of diffusion, an overcore including the injection borehole was obtained from the rock formation. Several secondary (experimental) drill cores were taken from this overcore. Specially for microprobe analyses, an experimental core was embedded in epoxy resin and further sectioned using a diamond wheel.36 A typical cross-section is shown in Figure 1 a, which shows the interface between the quartz-filled injection borehole and the rock sample. The Opalinus clay rock generally shows 6260

dx.doi.org/10.1021/ac200899x |Anal. Chem. 2011, 83, 6259–6266

Analytical Chemistry

ARTICLE

Figure 1. Optical image of the sample section (a), where the quartz-filled injection borehole is on the left of the interface and Opalinus clay rock is on the right of the interface. The white line indicates the line scan position, and the white box indicates the 2D matrix scan position. The scanning electron microscopy (SEM) images show the trace of a line scan (b) and matrix scan (c), as well as a magnified SEM image (d) of the matrix scan.

a complex mineralogical composition dominated by clay minerals (illite, smectite, kaolinite, and chlorite, about 70 wt %). The nonclay mineral fraction is dominated by calcite and quartz, while trace amounts, among others, of dolomite, siderite, K-feldspars, albite, and pyrite are observed.37 The sample material employed in this study showed a distinct separation between sediment layers rich in clay minerals and domains dominated by calcium carbonate precipitates. The two distinct domains can be localized readily by chemical imaging (see below) on the basis of characteristic elemental distributions. The calcium carbonate zones yield a high Ca content and typically lowest trace element concentrations. On the other hand, the clay mineral rich regions yield an elevated concentration of Al and Si as structural elements of the clay minerals, of Na as the dominant natural counterion at the geological setting of Mont Terri, and of trace elements such as Fe or Ti (typical structural substituents in clays or amorphous surface coatings frequently observed on clay particles). LA-ICPMS. LA-ICPMS experiments were conducted using an ArF excimer laser system (Lambda Physik, Goettingen, Germany) coupled to an Agilent 7500cs ICP-quadrupole-MS instrument (Agilent Technologies, Waldbronn, Germany). The laser system is based on a 193 nm ArF excimer laser with a pulse duration of 15 ns. The laser fluence was adjusted to 23.6 J/cm2. The ablation spot size for both line scan ablations and 2D imaging was 10 μm. All measurements were carried out using a laser repetition rate of 10 Hz. The line scans were recorded at a continuous sample translation of 15 μm/s. The square 2D images were obtained from 64 subimages, each of them consisting of a 5  5 LA matrix scan. Each of the pixels in the matrix scans consists of five laser pulses into the same position on the sample to gain improved statistics. Due to the 1 2 s washout time of the aerosols, LA was stopped for 3 s before measurement

of the next pixel. The time-dependent signals consist of 5  5 = 25 pulses, whose signals were integrated separately. In total, 8000 laser shots were integrated into 1600 data points, from which a 2D image of 40  40 pixels was calculated and reconstructed. For all measurements the ICP was operated at a power of 1500 W and the quadrupole MS instrument was operated in a peak hopping mode. For the line scans, a 50 ms integration time for each isotope was used. For the matrix scans, a reduced integration time of 5 ms was employed. The ablation cell was an in-house, fast washout cell, which has an effective cell volume of 10 cm3. To improve signal stability, a poly(tetrafluoroethylene) (PTFE) tube with an inside diameter of 4 mm was chosen for the 1.5 m connection of the LA cell to the ICPMS instrument.38 Helium was employed as the sample carrier gas at a flow rate of 1 L/min. In a mixing bulb located directly in front of the ICP, the sample carrier gas was mixed with argon at a flow rate of 0.85 L/min. SR-microXRF. Qualitative chemical imaging of the Opalinus clay sample was conducted at two synchrotron based X-ray microprobe facilities: (i) Sector 20 (Pacific Northwest Consortium Collaborative Access Team, PNC-CAT) at the Advanced Photon Source (APS), Argonne, IL (results are given in the Supporting Information), and (ii) microXAS beamline at the Swiss Light Source (SLS), Paul Scherrer Institute, Villigen, Switzerland. Both beamlines are insertion device lines dedicated to microprobe analysis. They use double-crystal monochromators to convert the X-ray beam to monochromatic radiation and reflective optics in the Kirkpatrick Baez arrangement to focus the incident X-ray beam onto the sample. The beam size utilized at the APS was 4  3 μm2, while at the SLS the beam size was 3  3 μm2. At both microprobe facilities, chemical images were obtained by recording two-dimensional X-ray fluorescence maps using 6261

dx.doi.org/10.1021/ac200899x |Anal. Chem. 2011, 83, 6259–6266

Analytical Chemistry fully motorized x y sample scanning stages. At both facilities fluorescence emission was monitored simultaneously by two detector systems: (i) a Microspec wavelength dispersive X-ray (WDX) spectrometer and (ii) an energy dispersive X-ray (EDX) spectrometer. At the APS, a 13-element Ge EDX detector (Canberra) was used. At the SLS, a single-element Si drift diode EDX detector (Ketek) was used. While the typical energy resolution of the EDX detector is around 130 150 eV (at 8 keV), the WDX spectrometer has an energy resolution as narrow as a few electronvolts. For the analysis of the reactive transport pattern of Cs in the Opalinus clay sample, this high resolving power is required to resolve the fluorescence emission of trace concentrations of Cs (LR-line, 4.29 keV) from the interferences of the major elements Ca and Ti in the sample matrix (Ca Kβ-line, 4.01 keV, and Ti KR, 4.51 keV). The need for high energy resolution is illustrated in Figure S-1 (Supporting Information), which shows the fluorescence emission spectra measured with the Ge EDX detector and with the WDX spectrometer. The measured fluorescence radiation originates from an Opalinus clay sample excited with 7.1 keV of radiation. Figure S-1 clearly demonstrates that high spectral resolution obtained using the WDX spectrometer is required to resolve the Cs emission from the major element emission of Ca and Ti. No Cs L-line fluorescence emission signal is resolvable when using the EDX solid-state detector. At both microprobe facilities, ion chamber systems were used to monitor the incident exciting intensity. Synchrotron Radiation Based X-ray Absorption Spectroscopy. EXAFS spectra were collected to investigate the chemical speciation of Cs in the Opalinus clay material. Elucidating the chemical speciation is an important step in determining the geochemical process controlling the retardation of Cs within the Opalinus clay rock. EXAFS measurements were performed at PNC-CAT of the APS. To avoid interferences with absorption resonances of other matrix elements of the clay (especially Ti), Cs EXAFS spectra were recorded at the Cs K-edge (35.985 keV) using a pair of Si(311) monochromator crystals. The beam size was limited by two pairs of slits of approximately 150  150 μm2. The Cs K-edge EXAFS spectra were recorded in fluorescence mode using a 13-element Canberra Ge detector. For the EXAFS measurement, the X-ray beam was localized ∼5 mm from the borehole rock interface, yielding the highest Cs fluorescence signal. Calibration Materials and Quantitative Procedure. A standard addition method based on mill-mixing of homogenized Opalinus clay rock material and additions of known Cs-bearing minerals (CsAlSiO4 or Cs2CO3) was of limited use due to a considerable spatial heterogeneity (in the range of a few micrometers) of the mixed material. The dilution (distribution) of CsAlSiO4 or Cs2CO3 mineral grains in the pristine Opalinus clay rock materials results in the observation of pronounced Cs hot spots. Therefore, a wet chemical preparation was used which was based on the adsorption of aqueous Cs cations onto the surface of Opalinus clay rock material powder obtained by ball milling. A significant improvement in standard homogeneity was obtained. However, for most of the material prepared from pressing powder particles, grain-to-grain heterogeneity and the corresponding grain size distribution remained a limit to the applicability of such materials for use as calibration standards for microbeam techniques. This is problematic for “homogenized” heterogeneous materials. Nevertheless, three powder standards were prepared by the wet chemical approach (∼400, ∼1600, and ∼6000 μg/g Cs). The Cs concentration in the solid was

ARTICLE

calculated from the concentration difference between the initial and equilibrium states of the solution. The aqueous Cs concentrations were determined using two independent methods, inductively coupled plasma optical emission spectroscopy (ICP-OES) and ion chromatography (IC). The solid in-house calibration materials were also analyzed using two additional independent methods: (i) LA-ICPMS, using an external calibration with non-matrix-matched NIST SRM 610 glass and a total mass normalization procedure,39 and (ii) laboratory source based macro-X-ray fluorescence (macroXRF, Spectro X-LAB 2000 Kleve, Germany), with quantification using a fundamental parameter method. The results of these measurements agreed within 10%, which validated the use of NIST610 as an external standard and the total mass normalization procedure as an appropriate calibration approach for these analyses. Accordingly, LA-ICPMS measurements of the Opalinus clay sample were based on the same method as for the calibration materials. SR-microXRF analyses were calibrated using wet chemistry prepared matrixmatched calibration materials. Calibration curves were fitted to a linear equation and used to calculate the concentrations for all determinations.

’ RESULTS AND DISCUSSION Line Scans. The first comparison of the imaging capabilities of LA-ICPMS and SR-microXRF were carried out using 30 mm line scans across the Opalinus clay sample section in a continuous scanning mode. Line scan analyses were started at the borehole rock interface and followed the direction of the Cs diffusion into the rock sample, as shown in Figure 1 a. Figure 2 a,e represent the typical behavior of Cs diffusion into the rock. A general agreement between the two independent techniques was observed. The quantitative LA-ICPMS measurements start at Cs concentrations of approximately 200 μg/g followed by the typical diffusion decay toward the geological Cs background level (∼6 μg/g)37 within a distance of 30 mm. The major difference between the two analytical techniques is in their spatial resolutions (Figure 2, red). The high spatial resolution analyses obtained using SR-microXRF consist of approximately 12 000 individual data points per 30 mm and of approximately 3800 data points acquired using LA-ICPMS over the same distance. Employing a continuous (on-the-fly) scanning mode, the spatial resolution in SR-microXRF analysis is related to the spot size of the X-ray beam (3 μm), the translation rate of the stage (1.25 μm/s), and the integration time of the detector (2 s). Therefore, a spatial resolution of approximately 3 μm was achieved using SR-microXRF. Although LA-ICPMS analyses were carried out using a 10 μm spot size, the spatial resolution also depends on the integration time of the MS instrument and the washout time of the lasergenerated aerosol through the transport system, including the ablation cell and transport tubing. During a period of 0.53 s used for one data reading of 10 elements (including a 3 ms quadrupole settling time for each mass/charge ratio change), the stage was moved 7.95 μm. The aerosol washout procedure is approximately described by an exponential decay,40 which also influences the spatial resolution. On the basis of the LA-ICPMS line scan analysis, one data point will theoretically tail onto the next four data points (typical signals decrease from 12 000 counts per second (cps) to a background of 2.6 cps in 2 s). Accordingly, an influence length (the distance on the sample where significant memory of previous sampling contributes to the total intensity) 6262

dx.doi.org/10.1021/ac200899x |Anal. Chem. 2011, 83, 6259–6266

Analytical Chemistry

ARTICLE

Figure 2. Comparison of Cs concentrations in line scans obtained by (a) LA-ICPMS and (e) SR-microXRF, where the transient signal represents the raw data calculated in concentration and the black line is the moving average. (b, f) Cs, (c, g) Ca, and (d, h) Ti transient concentrations are compared in detail for both techniques indicated by black arrows in (a) and (e).

of approximately 30 μm needs to be considered. Taking this into account, the first subsequent point had a 12% increase in intensity, while the second subsequent data point gained only 1.5% from the original data point. The tailing of intensity memory on the further acquired data can be neglected. Therefore, using the parameters previously stated, the LA-ICPMS operating conditions used provide a spatial resolution of approximately 15 μm. The mentioned overlap of the signals in time in LA-ICPMS were corrected by subtracting the contribution of the previous two data values, on the basis of an exponential decay of the washout. This correction allowed generation of the deconvoluted data set, which provided spatially resolved LA-ICPMS data points as independent from washout as those of SRmicroXRF. Due to the slow washout of the LA aerosol, the shot to shot overlap was first deconvoluted (see the previous discussion). Considering the different spatial resolutions of the two techniques, adjacent mean calculations were applied to LA-ICPMS data (20 points in a window) and to SR-microXRF data (60 points in a window). The 150 μm averaging distance is approximately the length of the geological heterogeneity of the sample. As shown in Figure 2, in both cases, a nearly ideal diffusion behavior of Cs was observed. The obtained microscopic results are consistent with the general (“coarse”) Cs diffusion profile observed by several antecedent macroscopic XRF analyses (Figure S-2, Supporting Information). However, at distinct locations (e.g., at 13, 17, and 25 mm) the Cs distribution deviates from an ideal diffusion profile. These deviations are triggered by the chemical heterogeneity of the

sample matrix. To stress this observation, as illustrated in Figure 2b d,f h, a selected range of the sample (1.35 1.50 mm) corresponding to the black arrow in Figure 2a,e was magnified to show the pronounced change in the elemental concentration by 1 2 orders of magnitude within a few micrometers. From the comparison, a correlation between Cs and Ti (indicating the clay domain) was observed, and an anticorrelation between Cs and Ca (indicating the calcite domain) is apparent as well. In terms of quantification, the sensitivity, limit of detection (LOD), and availability of matrix-matched reference materials are most crucial. The sensitivity of LA-ICPMS refers to 60 cps/ μg/g for 133Cs, when using a 10 μm diameter laser spot, which resulted in an LOD of ∼0.7 μg/g. More importantly, the NIST610 as a non-matrix-matched reference material was applicable for the calibration, which allowed accurate quantification of Cs when applying a total mass normalization approach.39 The SR-microXRF sensitivity was on the order of 0.1 cps/μg/g, nearly 2 orders of magnitude less than that obtained using LA-ICPMS. The LOD in SR-microXRF is below 10 μg/g, which is higher than that obtained in LA-ICPMS, even when considering the volume analyzed. Consequently, this low Cs count rate for SR-microXRF limited its capability for fully quantitative trace element analysis. This is mainly due to the limited acceptance (small solid angle) of the WDX detector. On the other hand, the EDX detector provided a higher acceptance solid angle, but the energy resolution was not sufficient to resolve the large interferences from the major elements Ca and Ti. In contrast to LA-ICPMS, quantification on heterogeneous samples is difficult using SR-microXRF, since no homogeneous matrix-matched 6263

dx.doi.org/10.1021/ac200899x |Anal. Chem. 2011, 83, 6259–6266

Analytical Chemistry

ARTICLE

Figure 3. Comparison of 2D images obtained using SR-microXRF (a d) and LA-ICPMS (e h). Elemental correlations were plotted between (i) Cs and Al, (j) Fe and Ca, and (k) Cs and Ca, in which three colors indicate three distinct populations in the Fe Ca correlation map, which also correspond to the colors in the other two correlation maps. Ovals are eye guides.

reference materials were available and a total mass normalization approach was not feasible on this sample. However, a calibration curve was established using averaged signals of line scans on the matrix-matched reference materials, and semiquantitative results were calculated. Imaging. Figure 3a h shows 2D images of the elemental distributions for the matrix elements Ca, Fe, and Ti, as well as for the trace element, Cs. Both SR-microXRF and LA-ICPMS images depict the same 800  800 μm2 area on the Opalinus clay sample section. Two distinct types of chemical domains can easily be distinguished. In the first domain Ca is the only major element, while in the second domain a chemically more complex matrix composition occurred. Major elements include Na, Mg, Al, Si, K, Ca, Ti, Mn, and Fe. On the basis of the complementary X-ray diffraction studies, the first domain is calcite and the second domain is the clay matrix. The two domains are separated by sharp boundaries (Figure 3; Figures S-3 and Figure S-4, Supporting Information). The element Cs is found mainly in the clay domain, in agreement with well-known experimental geochemical findings.37 The element Fe shares a pattern similar to that of the clay domain, but in the top-middle part of the image, it also shows a characteristic region which is assigned to a third Fe-rich domain. The 2D SR-microXRF measurements were carried out on 3 μm spaced parallel line scans with a 3 μm/point advance in each line, giving a spatial resolution of 3  3 μm2. The spatial resolution in the 2D LA-ICPMS image was restricted by the 20 μm spacing between adjacent matrix scan pixels. A moderate spacing for LA-ICPMS was chosen for two important reasons. First, the impact of uptaking the redeposited material around the craters may be significant, if the spacing of adjacent craters is made too small. Second, to map a relatively large area within a reasonable time, the resolution increments needed to be enlarged. However, each point represents information on an area with a diameter of approximately 10 μm rather than the full

20  20 μm2 pixel area. Such a matrix scan with incomplete analytical coverage of the entire sampling area results in the inconsistency in fine scale features when compared to the SRmicroXRF image. However, the overall distribution pattern was maintained. Clearly, SR-microXRF analysis captured signals from the entire surface, which provides a resolution on the order of 3 μm. Another reason for the inconsistency at a fine scale is the different sampling depths of the two different incident energies. X-ray can penetrate deeper into the sample (from several to tens of micrometers, depending on the X-ray excitation and emission energies), yielding a different shape and volume of the sampled material compared to those for LA, which ablates to a depth of ∼3 μm in this study. The sample volume is much more clearly defined in LA (Figure 1c,d) compared to SR-microXRF emission methods, due to the simple volume ablated in LA and a very complex excitation volume and emission volume in X-ray methods, which is a function of the photon energy of the excitation X-ray beam and the fluorescence emission energies. On the other hand, the volume of LA sampling is also complicated due to the mentioned washout time and material deposition on the surface from pervious ablations. Although the images obtained using SR-microXRF show a higher spatial resolution, the quantification accuracy is limited due to the lack of appropriate matrix-matched calibration standards, which are required for this technique, and the low sensitivity and high backgrounds, which result in higher limits of detection. The Cs sensitivity of SR-microXRF was calculated from the calibration curve, and the result was on the order of 0.1 cps/μg/g with a background of 1.3 cps. When using LA-ICPMS, the application of the total mass normalization approach39 provides an enhanced robustness of quantification. Since no matrix-matched calibration material was used, Cs sensitivity was calculated from the calibration curve plotted by using the corresponding count rates and concentrations on each one of the 1600 matrix pixels. The result was on the order of 6264

dx.doi.org/10.1021/ac200899x |Anal. Chem. 2011, 83, 6259–6266

Analytical Chemistry

Figure 4. Chemical speciation of Cs present in the Opalinus clay rock by means of X-ray absorption spectroscopy: (top) Cs K-edge EXAFS spectrum, (bottom right) Fourier transform, and (bottem left) corresponding wavelet modulus.

10 cps/μg/g. Consequently, the obtained images using these two complementary techniques improved quantitative chemical information and allowed generation of a spatially highly resolved elemental distribution pattern. Elemental Correlation. For geochemistry (and many other scientific fields), it is important to analyze elemental correlations in 2D images. Figure 3i k depicts the elemental correlation using data from the LA-ICPMS image. Each scatter point corresponds to the concentrations of elements in one pixel. Due to the quantitative nature of the LA-ICPMS data, the observed correlations are not affected by artifacts. The linear correlation between Cs and Al shown in Figure 3i agrees with the previous assumption that Cs has a preferred affinity to the clay phase. The Fe vs Ca correlation depicted in Figure 3j shows a general negative correlation, but can be grouped into three distinct populations: (i) the upper (red) population indicates the calcite domain; (ii) the bottom left (green) one indicates the clay domain, where low concentrations of Ca and Fe can be found; (iii) the bottom right one (blue) indicates the Fe-rich domain (corresponding to the high Fe concentration in Figure 3g). Applying this color coding to the Cs Ca correlation map, where an anticorrelation can be found between Cs and Ca, hardly any scatter points from Ca-rich and Fe-rich domains fall in high Cs concentration. This gives evidence that Cs is prone to be preferentially located in the clay domain rather than calcite or Fe-rich domains while migrating through the Opalinus clay barrier material. X-ray Absorption Spectroscopy. Molecular-level information about the chemical speciation of Cs present in the rock material can be obtained by X-ray absorption spectroscopy techniques, in particular EXAFS. Knowing the molecular coordination environment (speciation) of Cs is very important and crucial information needed to understand the geochemical reactivity of the Cs Opalinus clay system. Complementary to the quantitative chemical images presented in this study, Cs K-edge EXAFS spectra were recorded to gain molecular-level information. Figure 4 shows the k3-weighted EXAFS spectrum and the corresponding Fourier transform (FT) of Cs present within the Opalinus clay material. Additionally, the corresponding wavelet modulus41 is also shown. Three spectroscopic features are apparent in the FT and, even more distinctly, in the wavelet modulus image. These observed spectroscopic signatures are consistent with a splitting of the nearest coordination sphere. Two distinct oxygen distances were observed, with Cs O bond lengths of 3.2 and 4.2 Å. A third feature was also observed, which is a weak second-shell

ARTICLE

contribution determined at a distance of approximately 4.5 Å. Unfortunately, on the basis of the available data, an unambiguous identification of the responsible scattering pair is not possible. However, wavelet analysis42,43 reveals additional relevant details. A most important observation is that the center of the second-shell contribution is found at rather low k, near ∼4.5 Å 1. None of the scattering pairs Cs Al, Cs Si, and Cs Ti produce such a wavelet modulus. Accordingly, a direct covalent chemical bonding of Cs to a solid surface can be ruled out. Similarly, the proposition of a two Cs+ two H2O complex44 can be rejected as there is no evidence of a Cs Cs scattering pair in the wavelet modulus. However, the observed bonding environment can be rationalized on the basis of migration of Cs into clay interlayers. For Cs+ present in 2:1 clay interlayers, Monte Carlo (MC) and molecular dynamics (MD) investigations propose that indeed Cs coordination involves interlayer water and hydroxyl groups of the clay layers.44 47 Consistent with the presented EXAFS data, MC and MD simulations reveal a splitting of the Cs O coordination. Furthermore, MD simulations provide evidence for Cs being “trapped” by basal oxygen hexagonal cavities.44,45 Such well-defined, rigid molecular geometries would favor multiple scattering. Indeed, the best match to the observed “second”-shell contribution is obtained by considering a multiple scattering process. The present spectroscopic data provide strong evidence that the retardation of Cs within the Opalinus clay rock is due to an intercalation into clay interlayers. This finding is consistent with the characteristic domain association of Cs revealed by the chemical images and fully in line with the elemental correlation results.

’ CONCLUSIONS A complementary method for quantitative high spatially resolved chemical imaging of trace elements in heterogeneous media is described in this work. A case study of determining the trace element Cs distribution in an Opalinus clay rock sample was performed using LA-ICPMS and synchrotron-based microprobe analysis. Characteristic spatial distribution patterns were readily apparent in both of the images which were obtained. The Cs distribution and concentration were validated by combining the complementary information contained in the two images, which were both of high spatial resolution. The elemental distribution pattern increases the reliability level of reactive transport modeling and improves our knowledge of chemically reactive transport or physical processes in heterogeneous systems. The efficiency of LA-ICPMS measurements was low, due to the slow washout of the ablation cell. Therefore, a faster aerosol transportation system could improve the spatial resolution. This could also help the separation of the peaks in concentration versus time plots used for 2D imaging, so that signal deconvolution may be less necessary in the future, resulting in higher quality imaging. With a higher data acquisition efficiency using LA-ICPMS, the current experiment could be extended to record a higher resolution 2D image, or possibly 3D image. SR-microXRF is fast in terms of data acquisition; nevertheless, the quantification procedure needs improvement. The need for calibration using matrix-matched standards is an impediment to quantitative results. Also being able to calculate the chemical concentration at each point using a fundamental parameter approach and perform a full spectra analysis would further increase the robustness of the quantification in SR-microXRF. Finally, the use of a polycapillary half-lens attached to the detector head would provide 3D imaging capabilities. 6265

dx.doi.org/10.1021/ac200899x |Anal. Chem. 2011, 83, 6259–6266

Analytical Chemistry

’ ASSOCIATED CONTENT

bS

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: +41 (0)56 310 4551.

’ ACKNOWLEDGMENT Financial support by the Swiss National Foundation is acknowledged (Project 200021-119779). Sample material was provided by the Mont Terri Underground Rock laboratory, and related efforts by Paul Wersin and Olivier Leupin are appreciated. Henry Longerich is acknowledged for his valuable comments on an earlier version of the manuscript. Julien Zimmermann is acknowledged for expert support during SEM imaging. We thank Werner M€uller for the preparation of the in-house Cs standards. A part of the experiments were performed on the microXAS beamline at the Swiss Light Source, Paul Scherrer Institut, Villigen, Switzerland. Initial experiments were performed at the Pacific Northwest Consortium Collaborative Access Team’s (Pacific Northwest Consortium/ X-ray Operations and Research, PNC/XOR) beamline at the Advanced Photon Source at the Argonne facility. Special thanks are due to Steven Heald (PNC/ XOR) for excellent local support. ’ REFERENCES (1) Krivanek, O. L.; Chisholm, M. F.; Nicolosi, V.; Pennycook, T. J.; Corbin, G. J.; Dellby, N.; Murfitt, M. F.; Own, C. S.; Szilagyi, Z. S.; Oxley, M. P.; Pantelides, S. T.; Pennycook, S. J. Nature 2010, 464, 571–574. (2) Jermy, A. Nat. Rev. Microbiol. 2010, 8, 312–312. (3) Hieftje, G. M. Nat. Chem. 2009, 1, 10–11. (4) Mimura, H.; Handa, S.; Kimura, T.; Yumoto, H.; Yamakawa, D.; Yokoyama, H.; Matsuyama, S.; Inagaki, K.; Yamamura, K.; Sano, Y.; Tamasaku, K.; Nishino, Y.; Yabashi, M.; Ishikawa, T.; Yamauchi, K. Nat. Phys. 2010, 6, 122–125. (5) Shvyd’ko, Y. V.; Stoupin, S.; Cunsolo, A.; Said, A. H.; Huang, X. Nat. Phys. 2010, 6, 196–199. (6) Haessler, S.; Caillat, J.; Boutu, W.; Giovanetti-Teixeira, C.; Ruchon, T.; Auguste, T.; Diveki, Z.; Breger, P.; Maquet, A.; Carre, B.; Taieb, R.; Salieres, P. Nat. Phys. 2010, 6, 200–206. (7) Zoriy, M. V.; Becker, J. S. Rapid Commun. Mass Spectrom. 2009, 23, 23–30. (8) Ryan, C. G. Int. J. Imaging Syst. Technol. 2000, 11, 219–230. (9) Lobinski, R.; Moulin, C.; Ortega, R. Biochimie 2006, 88, 1591–1604. (10) Kuisma-Kursula, P. X-Ray Spectrom. 2000, 29, 111–118. (11) Lee, T. G.; Park, J.-W.; Shon, H. K.; Moon, D. W.; Choi, W. W.; Li, K.; Chung, J. H. Appl. Surf. Sci. 2008, 255, 1241–1248. (12) Chabala, J. M.; Soni, K. K.; Li, J.; Gavrilov, K. L.; Levi-Setti, R. Int. J. Mass Spectrom. 1995, 143, 191–212. (13) Chandra, S. Appl. Surf. Sci. 2008, 255, 1273–1284. (14) Carado, A.; Kozole, J.; Passarelli, M.; Winograd, N.; Loboda, A.; Bunch, J.; Wingate, J.; Hankin, J.; Murphy, R. Appl. Surf. Sci. 2008, 255, 1572–1575. (15) Becker, J. S.; Zoriy, M.; Matusch, A.; Wu, B.; Salber, D.; Palm, C.; Becker, J. S. Mass Spectrom. Rev. 2010, 29, 156–175. (16) Petrelli, M.; Perugini, D.; Alagna, K. E.; Poli, G.; Peccerillo, A. Period. Mineral. 2008, 77, 3–21.

ARTICLE

(17) Hetzinger, S.; Halfar, J.; Zack, T.; Gamboa, G.; Jacob, D. E.; Kunz, B. E.; Kronz, A.; Adey, W.; Lebednik, P. A.; Steneck, R. S. Palaeogeogr. Palaeoclimatol. 2011, 302, 81–94. (18) Neilsen, J. L.; Abildtrup, A.; Christensen, J.; Watson, P.; Cox, A.; McLeod, C. W. Spectrochim. Acta, B 1998, 53, 339–345. (19) Treble, P. C.; Chappell, J.; Shelley, J. M. G. Geochim. Cosmochim. Acta 2005, 69, 4855–4863. (20) Woodhead, J. D.; Hellstrom, J.; Hergt, J. M.; Greig, A.; Maas, R. Geostand. Geoanal. Res. 2007, 31, 331–343. (21) St€ockle, R.; Setz, P.; Deckert, V.; Lippert, T.; Wokaun, A.; Zenobi, R. Anal. Chem. 2001, 73, 1399–1402. (22) Becker, J. S.; Gorbunoff, A.; Zoriy, M.; Izmer, A.; Kayser, M. J. Anal. At. Spectrom. 2006, 21, 19–25. (23) Becker, J. S.; Gorbunoff, A.; Zoriy, M.; Izmer, A.; Kayser, M. J. Anal. At. Spectrom. 2007, 22, 222–222. (24) Grolimund, D.; Senn, M.; Trottmann, M.; Janousch, M.; Bonhoure, I.; Scheidegger, A. M.; Marcus, M. Spectrochim. Acta B, 2004, 59, 1627–1635. (25) Geraki, K.; Farquharson, M. J.; Bradley, D. A. Phys. Med. Biol. 2002, 47, 2327. (26) Zoeger, N.; Wobrauschek, P.; Streli, C.; Pepponi, G.; Roschger, P.; Falkenberg, G.; Osterode, W. X-Ray Spectrom. 2005, 34, 140–143. (27) Vespa, M.; Wieland, E.; D€ahn, R.; Grolimund, D.; Scheidegger, A. M. Cem. Concr. Res. 2007, 37, 1473–1482. (28) Singer, D. M.; Zachara, J. M.; Brown, G. E., Jr. Environ. Sci. Technol. 2009, 43, 630–636. (29) Frisia, S.; Borsato, A.; Fairchild, I. J.; Susini, J. Earth Planet. Sci. Lett. 2005, 235, 729–740. (30) Denecke, M. A.; Somogyi, A.; Janssens, K.; Simon, R.; Dardenne, K.; Noseck, U. Microsc. Microanal. 2007, 13, 165–172. (31) Perez, R. D.; Sanchez, H. J.; Perez, C. A.; Rubio, M. Radiat. Phys. Chem. 2010, 79, 195–200. (32) Janssens, K.; Proost, K.; Falkenberg, G. Spectrochim. Acta, B 2004, 59, 1637–1645. (33) Malzer, W.; Kanngieer, B. Spectrochim. Acta, B 2005, 60, 1334–1341. (34) Verstricht, J.; Bl€umling, P.; Merceron, T. Repository concepts for nuclear waste disposal in clay formations. Proceedings of the Sixth International Symposium on Field Measurements in Geomechanics, A.A. Balkema Publishers: Oslo, Norway, 2003. (35) Wersin, P.; Van Loon, L. R.; Soler, J. M.; Yllera, A.; Eikenberg, J.; Gimmi, T.; Hernan, P.; Boisson, J. Y. Appl. Clay Sci. 2004, 26, 123–135. (36) M€ori, A.; Inderbitzin, L.; Nussbaum, C.; Eikenberg, J.; R€uthi, M.; Aubry, C.; Brisset, N.; Steiger, H.; Bossart, P. Mont Terri Project; Technical Note 20, 2003. http://www.mont-terri.ch. (37) Van Loon, L. R.; Baeyens, B.; Bradbury, M. H. Appl. Geochem. 2009, 24, 999–1004. (38) Kovacs, R.; G€unther, D. J. Anal. At. Spectrom. 2008, 23, 1247–1252. (39) Guillong, M.; Hametner, K.; Reusser, E.; Wilson, S. A.; G€unther, D. Geostand. Geoanal. Res. 2005, 29, 315–331. (40) Bleiner, D.; G€unther, D. J. Anal. At. Spectrom. 2001, 16, 449–456. (41) Addison, P. S. The Illustrated Wavelet Transform Handbook, 1st ed.; Institute of Physics Publishing: Bristol, U.K., 2002. (42) Farges, F.; Munoz, M.; Argoul, P. Wavelet transform of XAFS spectra. http://perso.univ-mlv.fr/farges/wav/. (43) Funke, H.; Chukalina, M. Wavelet transform for EXAFS. http://www.esrf.fr/UsersAndScience/Experiments/CRG/BM20/ Software/Wavelets/Wavelets. (44) Sutton, R.; Sposito, G. Geochem. Trans. 2002, 3, 73–80. (45) Nakano, M.; Kawamura, K.; Ichikawa, Y. Appl. Clay Sci. 2003, 23, 15–23. (46) Smith, D. E. Langmuir 1998, 14, 5959–5967. (47) Sutton, R.; Sposito, G. J. Colloid Interface Sci. 2001, 237, 174–184. 6266

dx.doi.org/10.1021/ac200899x |Anal. Chem. 2011, 83, 6259–6266