Determination of the H Isotopic Composition of Individual Components

Nov 4, 2012 - Earth and Planetary Science Letters 2016 435, 36-44 ... D/H heterogeneity in chondrite matrices: A signature of the pristine solar syste...
0 downloads 0 Views 3MB Size
Article pubs.acs.org/ac

Determination of the H Isotopic Composition of Individual Components in Fine-Scale Mixtures of Organic Matter and Phyllosilicates with the Nanoscale Secondary Ion Mass Spectrometry Laurette Piani,*,† Laurent Remusat,† and François Robert† †

Laboratoire de Minéralogie et Cosmochimie du Muséum (LMCM), Muséum National d’Histoire Naturelle, 61 rue Buffon, 75005 Paris, France S Supporting Information *

ABSTRACT: When organic matter is mixed on a nanometer scale with clay minerals, the individual D/H ratios of the two H-bearing phases cannot be directly measured even with the nominal spatial resolution of nanoscale secondary ion mass spectrometry (NanoSIMS, 50−100 nm). To overcome this limitation, a new analytical protocol is proposed based on the deconvolution of the D−/H− and 16OD−/16OH− ionic ratios measured by NanoSIMS. Indeed, since the yields of H− and 16 OH− are different for organics and clays, it should be theoretically possible to determine the mixing ratio of these two components in the area analyzed by the ion probe. Using organics with different D/H ratios, the interdependence of the D−/H− and 16OD−/16OH− ionic ratios was determined in pure samples. Then using the H− and 16OH− yields and the isotopic ratios measured on pure organic matter and clays, the expected D−/H− and 16OD−/16OH− variations as a function of the mixing proportions were determined. These numerical predictions are consistent with measurements on laboratory prepared mixtures of D-rich organic matter and D-poor phyllosilicates, validating both the proposed experimental protocol and its use for meteorites. With an improvement of the precision of the ionic ratios by a factor of 10, it should possible to expend this protocol to samples having natural terrestrial D/H variations. Such an improvement could be attainable with the development of synthetic deuterated reference samples. The fine-grained matrix of primitive meteorites (chondrites) is an emblematic example of a fine mixture of organic matter and phyllosilicates − the two main H-bearing components in chondrites. Occurrences down to some tens of nanometres wide9 of organic matter are surrounded by phyllosilicates. De facto the primary ion beam of the NanoSIMS cannot spatially resolve both phases. Nevertheless, large differences are expected between the D/H ratio of phyllosilicates and organic matter from measurements on the isolated organic matter and mass balance calculation in some chondrites: the D/H ratio is estimated to be more than 2 times higher in organic matter than in phyllosilicates of carbonaceous chondrite.7 Moreover, huge isotopic heterogeneities are observed in the isolated organic matter at the nanometre scale.10,11 Evidence of isotopic heterogeneities in SIMS studies - i.e. a lack of equilibrium isotope distribution - has been reported for phyllosilicates of the matrix of two chondrites12 thanks to the higher H+ yield of phyllosilicates compared to that of the organic matter with a negative oxygen ion primary beam.

T

he nanoscale secondary ion mass spectrometer (NanoSIMS) is the latest generation of SIMS instruments for imaging and quantification of elemental and isotopic distribution at high sensitivity and with an optimal spatial resolution of ∼50−100 nm. However in natural samples some components are so intimately mixed and some concentrations are so low that the spatial resolution and the sensitivity of the NanoSIMS are not sufficient to distinguish individual components. This is illustrated, for instance, by the nanoscale size of clay minerals intimately mixed with organic matter in terrestrial1 or extraterrestrial material.2−5 To understand the link between these phases (genetic link, postformation equilibration etc.) a powerful tool is the comparison of their isotopic ratios. However, these isotopic ratios cannot be measured at the mixing scale (some tens of nanometres) with the current analytical techniques. Even at a larger scale (scale of the rock), the hydrogen isotopic composition of these phases is difficult to determine. In contrast to organic matter, which can be isolated by acid treatment, no reliable method exists to separate phyllosilicates from the organic contribution.6 Consequently, in the literature,7,8 the hydrogen isotopic composition of phyllosilicates is calculated by mass balance between the whole rock sample and isolated organic compounds. However, local links between organic matter and clays are lost during organic matter isolation. © 2012 American Chemical Society

Received: April 26, 2012 Accepted: November 4, 2012 Published: November 4, 2012 10199

dx.doi.org/10.1021/ac301099u | Anal. Chem. 2012, 84, 10199−10206

Analytical Chemistry

Article

respectively, the yield of M− in the mixture M−Mix should follow the relation

However no isotopic information has been obtained for the organic matter spatially related to these phyllosilicates. With a cesium ion primary beam, organic material has a higher H− yield than that of phyllosilicates, while the opposite is observed for 16OH−.13−15 In this respect, several attempts are in progress in laboratories to use the 16OD−/16OH− ratio as a proxy for the D/H composition of silicates.16 In this paper, we investigate the use of H−, D−,16OH−, and 16 OD− measurements in unresolved mixtures of organic matter and phyllosilicates. As schematically shown in Figure 1, the

M − Mix = x1 × ⟨M1−⟩ + (1 − x1) × ⟨M 2−⟩ −

(1)



thus we can write the D /H ratio of a mixture of organic matter (IOM) and phyllosilicate (Phy) as ⎛ D− ⎞ ⎜ ⎟ = ⎝ H − ⎠ Mix









D − ⟩( H ) ( HD )IOM + (1 − xIOM)⟨HPhy Phy

− xIOM⟨HIOM ⟩

− − xIOM⟨HIOM ⟩ + (1 − xIOM )⟨HPhy ⟩

(2a)

In this equation, ⟨D−IOM,Phy⟩ is replaced by (D−/H−)IOM,Phy × ⟨H − IOM,Phy ⟩ assuming (D − /H − ) IOM,Phy ≅ (⟨D − IOM,Phy ⟩/ ⟨H−IOM,Phy⟩) the isotopic ratio measured for the component IOM or Phy. A similar equation is obtained for the 16OD−/16OH− ratio: ⎛ OD− ⎞ ⎜ ⎟ ⎝ OH − ⎠ Mix =





OD OD − ⟩( OH ) ( OH )IOM + (1 − xIOM)⟨OHPhy Phy

− ⟩ xIOM⟨OHIOM



− ⟩ xIOM⟨OHIOM



+ (1 −

− ⟩ xIOM )⟨OHPhy

(2b)

Therefore, the comparison between isotopic measurements performed on a mixture and the corresponding calculated hyperbola will provide information on the isotopic composition of the two components of the mixture that could not be distinguished by the spatial resolution of the ion probe. To calculate the mixing hyperbolae, it is necessary to measure on pure samples of organic matter and phyllosilicates (i.e., samples mainly composed of organic matter or of phyllosilicates) the mean yields of H− and 16OH− and the mean D−/H− and 16 OD−/16OH− ratios. In order to compare the measured isotopic ratios in mixtures with the calculated values obtained by the mixing equations, eqs 2a and 2b are formulated with ionic ratios and not with ratios corrected for the mass instrumental fractionation ratio. This assumes that the instrumental mass fractionations (ratio of the measured isotope ratio over the reference isotope ratio) of IOM and phyllosilicates are not affected by the occurrence of multiple phases in the mixture. We have checked (1) that the 16OD−/16OH− ratio is a linear function of the D−/H− ratio in pure organic compounds having a large range of D/H ratios and (2) that a laboratory mixture of organic matter and phyllosilicate of known isotopic compositions follows the trend predicted by eqs 2a and 2b. However, a better fit is obtained by introducing an empirical correction on the 16OH− yield coming from the IOM as a function of the mixing proportions.

Figure 1. Schematic images of a mixture of organic matter (IOM) and phyllosilicates (Phy). On the left, the images represent the mixture (upper part) and its isotopic composition (lower part). On the right, four images are schematic NanoSIMS images of this mixture. The square (noted Unr.) at the bottom right corner stands for a mixture of the 2 components unresolved by the primary beam. On the top: the H− yield is higher for the IOM than for phyllosilicates and vice versa for the 16OH− yield. On the bottom: the consequence for the D/H ratios is shown, namely: the IOM-Phyllosilicate mixture exhibits a difference in its D−/H− and 16OD−/16OH− ratios.

D−/H− and 16OD−/16OH− ratios depend on the phase proportions of the mixtures and on the ion yields of H− and 16 OH− in both components. In principle, the deconvolution of the D−/H− and 16OD−/16OH− ratios should permit us to estimate the isotopic compositions of both phases in a spatially unresolved mixture.



APPROACH The starting point of our work is the speciation of the ions according to the phase they come from: the yield of an ion M∓ for a given primary ion beam and a given content of M is not identical in 2 different matrices. Hence for different ion species, one can apply the classical mixing models to measured ion ratios (details on the mixing hyperbolae can be found in Supporting Information, Figure S1). For a mechanical mixture, the position of the measured ratios on the mixing hyperbola will depend to a first approximation, on x1 and x2, the fractions of the surface covered by the components 1 and 2, respectively (with x1 + x2 = 1 in a two component mixture). The curvature of the mixing hyperbola depends on the difference of the ion yields of the species between both components. Defining ⟨M−1 ⟩ and ⟨M−2 ⟩ as the average yield of the ion M− in the components 1 and 2,



EXPERIMENTAL SECTION Samples and Mixtures. The two components of the mixture used to test the method are counterparts of organic matter and phyllosilicates found in the matrix of hydrated ordinary chondrites.17,18 The insoluble organic matter (IOM) isolated from the L3.2 Antarctic ordinary chondrite Grosvenor Mountains 9550219 (hereafter referred to as GRO95502) was used for the organic matter component. By HF/HCl treatment,20 we isolated around 50 mg of IOM from 20 g of chondrite provided by the Antarctic meteorite collection (NASA). This meteorite was chosen for its large mass and its D-rich IOM (δD = 3290 ‰21). A set of organic materials was 10200

dx.doi.org/10.1021/ac301099u | Anal. Chem. 2012, 84, 10199−10206

Analytical Chemistry

Article

Table 1. List of Samples Used in This Study, Their Abbreviations (cf. Samples and Mixtures), Their D/H Ratios (Reference Values Obtained by Isotope-Ratio Mass Spectrometry (IRMS) after Pyrolysis of the Solid Sample: Internal Standards and Literature21), and the Mean D−/H− and 16OD−/16OH− Ratios Measured by NanoSIMS in This Studye A. organic matter

abbreviation

D/Href (× 10‑6)

D−/H− (× 10‑6)c

16

OD−/16OH− (× 10‑6)c

Nd

IOM GRO95502 IOM Bishunpur IOM Orgueil IOM Murchison charcoal Type III Kerogen B. pure components

IOM-Gro IOM-Bish IOM-Org IOM-Murch charcoal Ker abbreviation

669.0b 583.9b 307.0 ± 276.8 ± 143.0 ± 141.5 ± D/Href (×

376 ± 12 241 ± 66 182 ± 10.5 136 ± 7 108 ± 5 115 ± 3 D−/H− (× 10‑6)c

466 ± 18 406 ± 53 289 ± 10 235 ± 3 200 ± 9 203 ± 10 16 OD−/16OH− (× 10‑6)c

8 12 8 8 6 8 N

montmorillonite IOM GRO95502

Mtm IOM-Gro

163 ± 16 351 ± 32

220 ± 13 477 ± 27

28 24

0.3b 4.2b 0.2a 0.3a 10‑6)

142.9 ± 0.3a 669.0b

a

Internal standards. bData from the literature: the H of the samples was converted to H2 by pyrolysis before measurement of D/H ratios by gas phase mass spectrometry,21 errors stand for reproducibility on duplicated measurements (accuracy of the D/H ratios estimated at 1.6 × 10−6 at 1σ for terrestrial D/H and larger for D-rich compositions21). cMean value, 1σ standard deviation. dNumber of 3.9 × 3.9 μm2 regions of interest. eThe organic matter (part A) and the mixed components (part B) were measured in two different analytical session; the IOM of GRO95502 was measured in both sessions.

used at the entrance of the mass analyzer, and an aperture slit of 150 μm (AS3) reduced the beam divergence. The energy slit was not used. Under these conditions, 50 000 counts per second were collected on 16OH− for IOM-Gro with a mass resolution MRPcam = 12 000, where MRPcam is the Cameca definition of the MRP (definition given in Fletcher et al.;24 see details in the next section). Automatic secondary beam alignment (EOS, Cy, and P2/P3) was performed on the ions 16 OH−. During a second step, NanoSIMS images of H− and D− were acquired using an aperture diaphragm of 150 μm (D1−4) and L0 and L1 settings to obtain a ∼20 pA primary current and a beam diameter of ∼400 nm. An entrance slit of 30 μm (ES3) was used at the entrance of the mass analyzer, and an aperture slit of 200 μm (AS2) reduced the beam divergence. Under these conditions, 300 000 counts per second were collected on H− for IOM-Gro with a mass resolution MRPcam = 5200. Secondary beam alignment (EOS, Cy, and P2/P3) was performed on the ion H−. Before the second step a brief acquisition using two magnetic fields and collecting H−, 13C−, 16 OH−, and 28Si− was performed to precisely locate the areas analyzed in the first step. Comparable analytical conditions were used on organic matter having a range of D/H ratios in a different analytical session (an analytical session corresponds to a set of analyses done with given analytical conditions and during a short period of time - at a maximum of a week) with similar high mass resolution conditions as described above. 16 OD− Measurement with the NanoSIMS. The theoretical mass resolving powers (MRP = M/ΔM) to separate 18O− and 17OH− from 16OD− are 1830 and 8740, respectively. Although 18O− and 16OD− are clearly resolved with our analytical conditions on the NanoSIMS, specific attention must be given to 17OH−. NanoSIMS instruments are mainly used in multicollection mode; each detector has a three discrete position exit slit corresponding to three widths (for our instrument: 25, 50, and 80 μm) with a constant height (2400 μm). Therefore the standard definition of instrumental mass resolution power, depending on the width of the exit slit, does not provide information on the mass dispersion of the mass spectrometer of the NanoSIMS. Cameca’s definition of the MRP (MRPcam) was proposed to overcome this difficulty. Nevertheless it does not

also used to quantify the effect of increasing the D/H ratio on the measured D−/H− and 16OD−/16OH− ratios: the IOM of the carbonaceous chondrites Orgueil22 (CI1) and Murchison (CM2), the IOM of the ordinary chondrite Bishunpur23 (LL3.15), and two terrestrial organic samples: a type III kerogen from Virginia and a charcoal. In Table 1, the different samples used in this study with their D/H values determined by gas phase mass spectrometry (hereafter referred to as D/Href) are listed. A smectite (montmorillonite) of Camps Bertaux (Morocco) was chosen as a mixture component similar to phyllosilicates in chondrite. Its H content and D/H ratio were determined by gas phase mass spectrometry (H2O = 4.61 ± 0.25 wt.%, D/H = [142.9 ± 0.3] × 10−6). The IOM of GRO95502 (hereafter referred to as IOM-Gro) and the montmorillonite (Mtm) were crushed and mixed in an agate mortar. Two mixtures were prepared: the first contained 50 wt.% of Mtm and 50 wt.% of IOM (IOM50), the second contained 90 wt.% of Mtm and 10 wt.% of IOM (IOM10). All the powders were pressed on cleaned indium foils and gold coated. The mixtures were imaged with a Scanning Electron Microscope (SEM): the sizes of organic matter and phyllosilicate grains range from less than 1 to 5 μm. NanoSIMS Analytical Procedure. NanoSIMS images were acquired on the different samples with a 16 keV Cs+ ion primary beam on 10 × 10 μm2 areas divided in 64 × 64 pixels and with a counting time per pixel and per cycle from 2 to 5 ms according to the ion yields of the samples. A cycle corresponds to the total raster of the analyzed area; each image is a stack of 20 to 40 cycles. Before the analyses, the samples were rastered with a 400 pA primary beam during 5 min on a 15 × 15 μm2 area to clean the surface and reach the sputtering equilibrium. The electron gun was necessary to improve charge compensation for the insulating samples (samples containing phyllosilicates). The difference of isotopic ratios measured with and without the electron gun for the conductive sample is insignificant in our analytical uncertainty. During a first step, NanoSIMS images of 13C−, 16OH−, 16 OD−, and 28Si− were acquired using an aperture diaphragm of 300 μm (D1−2) and primary lens tunings (L0 and L1) to obtain a current of ∼70 pA resulting in a beam diameter of ∼500 nm. The ions were collected on the detectors 2 to 5 with a single magnetic field. An entrance slit of 10 μm (ES5) was 10201

dx.doi.org/10.1021/ac301099u | Anal. Chem. 2012, 84, 10199−10206

Analytical Chemistry

Article

Figure 2. A. Mass spectrum recorded at 18 a.m.u. with the thinnest exit slit (the exit slit width is 25 μm) on the detector 4: 18O− is totally separated from 17OH− and 16OD− that overlap. λ5% corresponds to the width of the peak at 5% of the maximum intensity of 18O− and hL5% corresponds to 90% of the beam width entering the detector. B. Enlargement of the mass spectrum around the 17OH− and 16OD− group for IOM50: λ5% and hL5% measured on 18O− peak were carried forward to the 17OH− and 16OD− peaks and d5% represents the widths where 16OD− could be collected with an interference contribution of less than 5% of the maximum intensity of 17OH−. Vertical gray lines indicate the centers of the mass peaks of 17OH− and 16 OD−; hM corresponds to the distance between the centers.

allow us to determine if two masses are fully separated or not.24 Figure 2 shows two mass spectra recorded around 18 a.m.u. using the thinnest exit slit on the detector 4. On the first mass spectrum (Figure 2 A), it can be observed that 18O− is totally separated (with a ∼0% valley) from 17OH− and 16OD− that are grouped together. The mass spectrum of the ions 17OH− and 16 OD− is then shown with a linear scale (Figure 2B). To estimate the maximum contribution of 17OH− on the 16OD− measurement, we carried the measured width λ5% that corresponds to the width of the 18O− peak at 5% of the maximum intensity forward to the peak of 17OH−. In our analytical conditions, there exists a distance d5% = 9 μm, where the 16OD− can be measured with a contribution of less than 5% of the maximum of the intensity of 17OH−. It corresponds to a maximum relative uncertainty (=[uncertainty of the measurement]/measurement) of around 0.082 for 16OD−/16OH−. This uncertainty was taken into account in the error calculation for the 16OD−/16OH− ratio. Data Processing. The distributions of D−/H− and 16 OD−/16OH− ratios and of H−, 13C−, 16OH−, and 28Si− are deduced from the NanoSIMS images by data processing performed with the L’IMAGE software package developed by Larry Nittler, Carnegie Institution Washington.

A NanoSIMS image is made up of a large number of pixels, and each pixel corresponds to an individual analysis of each collected ion. The pixels are usually grouped into regions of interest (ROIs) according to the image properties and the required counting rates. In our images, D− and 16OD− may have very low counting rates: for instance, 0.04 counts per pixel per cycle of 16OD− were collected in Mtm. Therefore, we choose to define 4 ROIs of 25 × 25 pixels (3.9 × 3.9 μm2) in the center of the images by summing the counting rate of all the cycles: the statistical error given as a relative precision (=[uncertainty of the measurement]/measurement) is 0.04 for the 16OD−/16OH− ratio and 0.09 for the D−/H− ratio for each ROI in the least favorable case (Mtm).



RESULTS AND DISCUSSION Isotopic Ratios in Pure Organic Samples. Table 1 gives the mean D−/H− and 16OD−/16OH− ratios measured on pure organic matter; results will henceforth be given as measured D−/H− and 16OD−/16OH− ratios. In our samples, the reference D/H ratio spans a range from terrestrial values (∼141 × 10−6) up to 669 × 10−6 (IOM-Gro). A higher D-enrichment is observed as a D-rich hotspot occurring in both the D−/H− image and the corresponding 16OD−/16OH− image of the same 10202

dx.doi.org/10.1021/ac301099u | Anal. Chem. 2012, 84, 10199−10206

Analytical Chemistry

Article

were used to determine the coefficient of the regression line that links the D−/H− and 16OD−/16OH− ratios: the type III kerogen (Ker) and the IOM of GRO95502. It must be noted that, when plotted versus the reference D/ H ratio measured by classical gas phase mass spectrometry, the D−/H− and 16OD−/16OH− ratios measured by NanoSIMS both define a nonzero intercept line (Table 1). This nonzero intercept line could have different sources: nonorganic contributors in the acid residue, terrestrial H adsorption on the sample surface, difficulty to determine accurate D/H ratios with gas phase mass spectrometry25 due to the lack of standards with high D/H ratios, etc. If we consider that the gas phase mass spectrometry D/H ratios represent the absolute D/H ratios of the samples, we can interpret the nonzero intercept line as the result of a systematic contribution of a terrestrial contaminant on the total H−, D− and 16OH−, 16OD− signals collected on the organic matter. This should then allow us to intercalibrate the NanoSIMS D/H ratios with those obtained by classical gas phase mass spectrometry techniques. The issue of intercalibration for the isotopic measurements obtained by different analytical techniques has to be taken into account for all the studies dealing with the hydrogen isotopic composition of samples, whose D/H ratios fall outside the range of values commonly found on Earth. Since the proposed mixing model is only tested using the measured isotopic ratios (ionic ratios) and considering that the contribution would be equivalent on each component, the intercalibration protocol of these ratios is not addressed in the present paper. Images of Laboratory Prepared Mixtures. An example of NanoSIMS images of H−, 13C−, 16OH−, and 28Si− as well as D−/H− and 16OD−/16OH− images is reported for the mixture IOM50 in Figure 4A. The NanoSIMS images illustrate the similarities between the distributions of H− and 13C−: H− is predominantly present in C-rich zones that correspond to IOM-Gro. In these zones, high D-enrichments are observed in contrast to montmorillonite rich areas (image Δ of the D−/H− ratio). The ion 16OH− is present on the whole image, but it correlates with 28Si− intense regions. The clear distinction between the H− and 16OH− distributions

area of IOM-Gro: we estimated the D/H ratio of this hotspot as 1160 ± 117 × 10−6. Measured D−/H− and 16OD−/16OH− ratios define a linear correlation (Figure 3) with a correlation coefficient of 0.98 over

Figure 3. Correlation between the D − /H − ratio and the 16 OD−/16OH− ratio in pure organic matter with D/H ratio going from 141 × 10−6 (Ker) to 1160 × 10−6 (a hotspot in IOM-Gro). The solid line is the regression line over all the data shown on the graph, and the dashed lines are the boundaries of the prediction interval at the 95% confidence level (the prediction interval corresponds to the domain where a new measurement will fall with identical conditions of measurement).

the large range of isotopic values. IOM-Org plots slightly above the correlation line compared to the other samples; the reason for this discrepancy is unclear. Interestingly, the measured ratios of Mtm also fall on the regression line close to the kerogen, indicating that matrix effects are not detectable within our analytical precision. To reduce the measurement duration during the second analytical session, only two organic samples

Figure 4. A. NanoSIMS images of the 50/50 mixture (IOM50): H−, 13C−, 16OH−, and 28Si− distributions and isotopic distributions. The D−/H− and 16 OD−/16OH− images are associated with “Δ images”. In these images, at each pixel i having a ratio Ri and an error Ei we can calculate Δi = |Ri − RMtm|/Ei (with RMtm the ratio measured in the pure montmorillonite). The white squares correspond to the same area on the different images. B. Same area image with a scanning electron microscope (SEM) with backscattered electrons: the light gray grains are large phyllosilicate grains. 10203

dx.doi.org/10.1021/ac301099u | Anal. Chem. 2012, 84, 10199−10206

Analytical Chemistry

Article

Table 2. Intensity of Ions H−, 13C−, 16OH−, and 28Si− (Mean ± Standard Deviation over n ROIs) Expressed in Counts Per Second per pA for IOM-Gro, Mtm, and the Mixtures IOM50 and IOM10 H− Mtm IOM-Gro IOM50 IOM10

1100 15000 14800 2500

± ± ± ±

13 −

OH−

16

C

400 2500 1800 1400

0.5 1000 870 100

± ± ± ±

0.3 120 30 100

700 1200 1640 1130

in both phases of the mixture is fully consistent with the prediction illustrated by Figure 1; the 16OD−/16OH− ratio appears to be lower in the C-poor zones but with a lower contrast than for the D−/H− ratio. We however observed that silicon distributions do not exactly match between images by NanoSIMS and by scanning electron microscopy (SEM) in backscattered electron mode or X-ray elemental maps on the same areas of the mixtures (Figure 4B). In contrast to 13C−, the 28Si− intensity is not homogeneous but is higher at the boundary of the silicates (Figure 4A). Hence, in the mixtures, an additional contribution of 28Si− and 16 OH− coming from the edges of the silicate boundary is added to the total 28Si− and 16OH− yields. This observation can explain the higher mean intensity of 28Si− and 16OH− measured on the mixtures compared with pure montmorillonite and IOM-Gro (Table 2). This effect is nevertheless not observed for H− and 13C−. Unexpected high 28Si− (and probably the 16OH−) yield at the silicate grain boundary is not observed in pure montmorillonite and seems to be an artifact due to the proximity of carbonaceous matter. A comparable observation was reported by Ghosal et al.26 They have performed NanoSIMS depth profile analyses on a C−Si multilayer sample (sample with alternating layers of C and Si of known thicknesses) and observed that silicon intensity increased at the interface of C− Si layers when a cesium ion primary beam is used. This effect remains poorly constrained. It could result from the difference in charge evacuation between conductive organic matter and insulating silicates in spite of the use of electron gun and gold coating. The presence of organic matter in close association with silicate grains could facilitate charge evacuation and modify the secondary ion yields coming from the silicates. In samples like meteorites, where rare occurrences of organic grains (15%) fall significantly out of the prediction interval. The points of IOM50 have similar or lower isotopic ratios than IOM-Gro. The points of IOM10 have values similar or lower than the lowest values of IOM50, and 2 points are close to Mtm. The isotopic ratio measured for IOM-Gro exhibits a large dispersion (relative standard deviation at 1σ ∼ 0.09 for the D−/ H− ratio) limiting the precision of the method. This dispersion is certainly linked to isotopic heterogeneities of the IOM in meteorites where some terrestrial contaminants and D-rich hotspots have been observed.10,11 Measured values of IOM50 and IOM10 can be compared with mixing hyperbolae calculated with eqs 2a, 2b, and 3 using the mean yields (Table 2) and isotopic ratios (Table 1B) measured in the pure components Mtm and IOM-Gro for different values of xIOM. To take into account the isotopic variations of the IOM, we calculated two hyperbolae using the isotopic ratio plus or minus two standard deviations for IOM-Gro. The two hyperbolae define a 95% confidence domain taking into account 2 standard deviations of the isotope ratios of IOM-Gro. The two data



CONCLUSION We have tested the combined use of two couples of isotopologues in a fine mixture of two phases. The comparison between measured and calculated values shows that qualitative information can be obtained. The precise determination of 10205

dx.doi.org/10.1021/ac301099u | Anal. Chem. 2012, 84, 10199−10206

Analytical Chemistry

Article

(8) Alexander, C. M. O'D.; Newsome, S. D.; Fogel, M. L.; Nittler, L. R.; Busemann, H.; Cody, G. D. Geochim. Cosmochim. Acta 2010, 74, 4417−4437. (9) Le Guillou, C.; Brearley, A. J.; Floss, C.; Stadermann, F. Meteorit. Planet. Sci. Suppl. 2010, 73, 5342. (10) Busemann, H.; Young, A. F.; Alexander, C. M. O’D.; Hoppe, P.; Mukhopadhyay, S.; Nittler, L. R. Science 2006, 312, 727−730. (11) Remusat, L.; Robert, F.; Meibom, A.; Mostefaoui, S.; Delpoux, O.; Binet, L.; Gourier, D.; Derenne, S. Astrophys. J. 2009, 698, 2087− 2092. (12) Deloule, E.; Robert, F. Geochim. Cosmochim. Acta 1995, 59, 4695−4706. (13) Deloule, E.; France-Lanord, C.; Albarède, F. In Stable isotope Geochemistry: a tribute to Samuel Epstein; Taylor, H. P., O’Neil, J. R., Kaplan, I. R., Eds.; Geochemical Society: 1991; pp 53−62. (14) Tenner, T. J.; Hirschmann, M. M.; Withers, A. C.; Hervig, R. L. Chem. Geol. 2009, 262, 42−56. (15) Mosenfelder, J. L.; Le Voyer, M.; Rossman, G. R.; Guan, D. R.; Bell, Y.; Asimow, P. R.; Eiler, J. M. Am. Mineral. 2011, 96, 1725−1741. (16) Deloule, E. CNRS research director at CRPG, France, personal communication, 2012. (17) Hutchison, R.; Alexander, C. M. O’D.; Barber, D. J. Geochim. Cosmochim. Acta 1987, 51, 1875−1882. (18) Alexander, C. M. O’D; Barber, D.; Hutchison, R. Geochim. Cosmochim. Acta 1989, 53, 3045−3057. (19) Grossman, J. N. Meteorit. Planet. Sci. Suppl. 1998, 33, 221−239. (20) Gardinier, A.; Derenne, S.; Robert, F.; Behar, F.; Largeau, C.; Maquet, J. Earth Planet. Sci. Lett. 2000, 184, 9−21. (21) Alexander, C. M. O’D.; Fogel, M.; Yabuta, H.; Cody, G. D. Geochim. Cosmochim. Acta 2007, 71, 4380−4403. (22) Remusat, L.; Derenne, S.; Robert, F.; Knicker, H. Geochim. Cosmochim. Acta 2005, 69, 3919−3932. (23) Yang, J.; Epstein, S. Geochim. Cosmochim. Acta 1983, 47, 2199− 2216. (24) Fletcher, I. R.; Kilburn, M. K.; Rasmussen, B. Int. J. Mass Spectrom. 2008, 278, 59−68. (25) Alexander, C. M. O’D.; Newsome, S. D.; Fogel, M. L.; Nittler, L. R.; Busemann, H.; Cody, G. D. Geochim. Cosmochim. Acta 2010, 74, 4417−4437. (26) Ghosal, S.; Fallon, S. J.; Leighton, T. J.; Wheeler, K. E.; Kristo, M. J.; Hutcheon, I. D.; Weber, P. K. Anal. Chem. 2008, 80, 5986− 5992. (27) McMahon, G.; Saint-Cyr, H. F.; Lechene, C. J. Am. Soc. Mass. Spectrom 2006, 17, 1181−1187.

isotopic ratios of the pure components is hampered by (1) the modification of the ions yields (28Si−, 16OH−) at the interface of the components likely due to their difference of surface state and (2) the modification of the IOM yields of 16OH− bearing the hydrogen isotopic signature of the IOM. Based on the uncertainties on the determination of the 8 parameters dictating the mixing, the present protocol is suitable for the large isotopic differences existing in meteorites, between the phyllosilicates and the organic matter. It is nevertheless conceivable to improve the precision using synthetic deuterated samples so that this protocol can be used with terrestrial samples.



ASSOCIATED CONTENT

S Supporting Information *

The first section gives details concerning the calculation of mixing hyperbolae. In the second section, the data are compared to the model without empirical correction. NanoSIMS data measured on pure components (Mtm, IOM-Gro, Ker) and mixtures (IOM50, IOM10) are given in the third section. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We gratefully thank the Antarctic meteorite collection curators who have provided us with the GRO95502 ordinary chondrite and the members of the Bioemco laboratory (UPMC Paris) for their help during the IOM isolation. Aurélien Thomen, Anders Meibom, and Roger Hewins are thanked for helpful discussions and manuscript corrections. We are really grateful to the anonymous reviewers for their comments and reviews that helped us to clarify and improve the manuscript and to Editorin-Chief Jonathan V. Sweedler for careful editing. This work has been supported by the Programme National de Planétologie (PNP INSU) and the ANR T-Tauri Chem. The National NanoSIMS facility at the Muséum National d’Histoire Naturelle was established by funds from the CNRS, Région Il̂ e de France, Ministère délégué à l’Enseignement supérieur et à la Recherche, and the Muséum itself.



REFERENCES

(1) Lagaly, G.; Ogawa, M.; Dékány, I. A. In Handbook of Clay Science; Bergaya, F., Theng Benny, K. G., Lagaly, G., Eds.; Elsevier: 2006; Vol. 1, pp 309−377. (2) Aléon, J.; Engrand, C.; Robert, F.; Chaussidon, M. Geochim. Cosmochim. Acta 2001, 65, 4399−4412. (3) Pearson, V. K.; Sephton, M. A.; Kearsley, A. T.; Bland, P. A.; Franchi, I. A.; Gilmour, I. Meteorit. Planet. Sci. 2002, 37, 1829−1833. (4) Eiler, J. M.; Kitchen, N. Geochim. Cosmochim. Acta 2004, 68, 1395−1411. (5) Garvie, L. A. J.; Buseck, P. R. Meteorit. Planet. Sci. 2007, 42, 2111−2117. (6) Hyeong, K.; Capuano, R. M. Geochim. Cosmochim. Acta 2000, 64, 3829−3837. (7) Robert, F.; Epstein, S. Geochim. Cosmochim. Acta 1982, 46, 81− 95. 10206

dx.doi.org/10.1021/ac301099u | Anal. Chem. 2012, 84, 10199−10206