V13A-2076 INVITED
Fractionation Of Mercury Isotopes In The Environment
Mercury is a toxic global pollutant. Although, from flux and concentration based models, it is well understood that atmospheric deposition plays a cardinal role in the global mercury cycle and that anthropogenic contributions to the annual atmospheric budget rival those of natural processes; crucial uncertainties in the mercury cycle still exist. Recently it has been demonstrated that mercury undergoes both mass dependent (MDF) and mass independent fractionation (MIF) in the environment. A peat core representing about 2200 years of accumulation from an ombrotrophic peat bog of Penido Vello, Spain was serially sampled for Hg isotope analysis. There is an approximately 10-fold increase of Hg concentration in the top 5 cm of the peat core (representing mostly anthropogenic mercury released from coal plants and other modern industries) compared to the bottom 10 cm of the core (representing mercury from pre-industrial times). The peat core shown an average enrichment of approximately 4.00‰ for 198Hg/202Hg (MDF) with respect to NIST SRM 3133 Hg standard and Almaden Cinnabar. The MIF exhibits that there is depletion in nearly equal amounts of the odd neutron number isotopes 199Hg and 201Hg (up to -0.5‰). In spite of the large temporal increase in concentration and temporal change in the nature of sources, the MDF and MIF composition of the top 5 cm of the core is very similar to the MDF and MIF composition of the bottom 10 cm of the peat core. The peat core has mostly derived its mercury from atmospheric deposition and it seems that both MDF and MIF are tracing processes that are transforming mercury in the cycle more than provenances. The potential for isotopic tracing of Hg speciation in the environment should have important implications for a better understanding of the behavior of mercury in the environment and eventually provide important constraints on models of the mercury cycle, including both natural and anthropogenic influences.
V13A-2077 INVITED
Certification of new Pb iCRM (Candidate ERM-38xx series) via Gravimetric Isotope Mixtures and MC-ICP-MS Measurements
Lead is known to be a particularly toxic chemical element. Mining and smelting of Pb and its domestic use over the centuries have contaminated the surface of Earth and jeopardized the health of humans, domestic animals and wildlife. Omnipresence of Pb has however an advantage. It can be used as an isotopic tracer of pollution sources and pathways in the environment. Pb isotopic Certified Reference Materials (iCRM) with undisputed characteristics are then required to validate Pb isotope ratio measurements. The materials currently available worldwide were produced in the 1960's by the National Bureau of Standards (now NIST, USA) and are now the object of polemics regarding the accuracy of some of the accompanying certified values. Moreover, new materials with lower relative uncertainty statements are demanded by users. This presentation is centred on the production and the certification at the EC-JRC-Institute for Reference Materials and Measurements of a new series of Pb iCRM (candidate ERM-38xx series). The production included six Pb gravimetric isotope mixtures, a common Pb material dedicated to routine calibration work and a series of four 207Pb slightly enriched natural-like Pb materials for the validation of the δ-scale method below 0.01%. The gravimetric isotope mixtures method, originally supported by mass spectrometry measurements performed on GS-MS or TI-MS instruments, was applied successfully to MC-ICP-MS. The measurements and certification methods we developed were validated in several ways, including a systematic investigation on possible significant sources of uncertainty and comparisons of results obtained by different laboratories on identical samples. Relative uncertainties on isotope ratios obtained for the newly produced Pb iCRM are as low as 0.017%, which is between 2 and 4 times smaller than the uncertainties carried by the NIST-981 material. Going below 0.01% was not possible mainly because of uncompressible uncertainties coming from the weighing stage. We suggest that the newly produced 207Pb series, for which the δ values span from 0.001% to 0.013%, could be used as validation tool for isotopic variations below 0.01%. Results obtained from an inter-laboratory comparison involving 5 laboratories on δ-scale measurements using our candidate iCRMs show that relative isotopic variations below 0.005% are hardly measurable accurately.
V13A-2078 INVITED
Multiple ion Counting for Noble gas Mass Spectrometry
In geo- and cosmochemistry noble gases are analyzed for their abundance and isotopic composition by mass spectrometry in the static mode (no pumping). This results in that they are among the elements with the highest detection efficiency. This is especially so for xenon, where during the course of a normal measurement almost all atoms will have been ionized. Not all of them will have been detected, however, even for 100 percent transmission of the mass spectrometer, if individual isotopes are measured by stepped change of the magnetic field setting. On the other hand, except for special cases, noble gas isotopes other than radiogenic 4He and 40Ar usually have abundances too low for useful detection by multicollection using Faraday cups. Multiple ion counting, i.e. parallel detection of all isotopes of a given noble gas element by ion counting, is the obvious choice to improve this situation and gain (in case of Xe with its 9 isotopes) up to an order of magnitude in sensitivity. We have exchanged the standard detection system of our Nu instruments Noblesse noble gas mass spectrometer with a multi-ion counting unit from Nu instruments, the second they have assembled (the other is at Washington University, St. Louis). The detector contains seven channeltrons and (at the high mass side) a Faraday collector. Xenon can be measured by multiple ion counting in two magnetic field steps, one for the seven even-numbered isotopes (masses 124-136) and a second for the high abundant odd- numbered isotopes 129,131Xe, for which we use a shorter counting time. Static measurement of xenon easily leads to memory effects from previously analyzed samples, depending on the history. Hence, often the real improvement is in the gain in speed by which useful data can be obtained, rather than the actual yield, since extrapolating to "time zero" (gas-inlet into the mass spectrometer) can be performed more reliably. "Real" samples will therefore often be measured for a shorter time than required to detect the maximum number of ions. I will report on first experiences with standard samples. Actual samples planned to be analyzed include presolar grains from meteorites, micrometeorites, and interplanetary dust particles.
V13A-2079 INVITED
Considerations in the Application of Multiple Ion Counting for the Trace Analysis of Plutonium and Uranium Isotope Ratios Using Thermal Ionization and Inductively-Coupled Plasma Mass Spectrometry
The use of simultaneous multiple-ion counting for the analysis of small samples of plutonium and uranium has been investigated using three different instruments, the ThermoElectron Neptune inductively-coupled plasma mass spectrometer, the ThermoElectron Triton thermal ionization mass spectrometer, and the Isotopex Iso-T thermal ionization mass spectrometer. The Neptune and Triton instruments utilize identical multiple ion counter arrays, with ions impinging directly on the channeltron surface. The Isotopex instruments utilize a different style of channeltron detector. The most significant difference in the Isotopex detectors is the presence of a conversion dynode at the entrance to the channeltron. Results suggest that the performance of the ThermoElectron MIC system varies between the Neptune and Triton instruments, which probably reflects both differences in the inherent characteristics of plasma and thermal sources and the performance of the MICS themselves. Differences in performance and stability between the '"naked"' and conversion dynode equipped channeltrons on the two thermal ionization instruments support these observations. These differences suggest that different analytical approaches to calibration of the multiple-ion counters may be required. Differences in potential analytical strategies employing multiple ion counters on the different instruments, including calibration schemes, precision and accuracy limits, and analytical strategies that can be employed, will be discussed. Results from both thermal ionization and inductively-coupled plasma sources suggest that the dominant source of uncertainty in isotope ratio measurement using multiple ion counting shifts from counting limitations for very small signals to uncertainties in gain calibration and detector drift among the ion counters at higher count rates. These characteristics place limits on the applicability of multiple ion counters; results from mixed Faraday/multiple ion counting analysis will illustrate the potential to overcome some of these limitations.
V13A-2080
Advances in low level uranium and plutonium isotope mass spectrometry using multiple ion counting and filament carburization
After upgrading IRMM's mass spectrometric capabilities for certification measurements for uranium and plutonium using large sample sizes during the previous years, in 2006-2007 we focused on necessary improvements in the area of low-level isotopic analyses for uranium and plutonium. This project was driven firstly by the need for reliable verification measurements for the Nuclear Signatures Measurement Evaluation Programme (NUSIMEP) samples at IRMM, secondly by the need for verification measurements on single uranium oxide reference particles and thirdly by the request from the IAEA's Safeguards Analytical Laboratory (SAL) to provide assistance for this type of analyses through the EC support programme. Improving low-level isotope mass spectrometry for uranium and plutonium at IRMM consisted of three steps. First a new thermal ionization mass spectrometer was acquired in order to have an instrument which can be used for peak-jumping measurements in ion counting mode, and which can be subsequently upgraded with a "Multiple Ion Counting" (MIC) system. This detector system allows the simultaneous detection of up to seven small ion beams with currents of 10-19 – 10-14 Ampere in ion counting mode, corresponding to count rates of 1–60.000 counts per second. As a result of test measurements with the MIC system it turned out that static measurements using the MIC system with a sample-versus-standard type external calibration can be associated with uncertainties even higher than in peak-jumping mode. The second step of improvement to tackle this situation was to implement the principle of "multi-dynamic" measurements for both uranium and plutonium measurements. This "multi- dynamic" measurement procedure provides an internal calibration of the MIC system and therefore circumvents the need for complicated inter-calibration routines. As a third step, a filament carburization procedure was implemented by which the ionization efficiencies for uranium and plutonium were improved by a factor of 3 and 10, respectively. Results for measurements performed on samples of previous NUSIMEP campaigns will be shown in comparison to results from various techniques employed by participating laboratories.
V13A-2081
Overview of Uranium Isotopic Reference Materials at IRMM
For many applications in the geological sciences, in particular in geochemistry research, isotope ratio measurements play a significant role. For instance, in geochronology isotope abundances of uranium and its daughter products thorium and lead have been used since more than five decades to determine the age of various samples of geological interest. However, in order to validate mass spectrometric measurement procedures and to calibrate detector systems, suitable isotope reference materials are needed. IRMM is a well recognized provider for nuclear isotope reference materials to the nuclear industry and nuclear safeguards authorities, which can also be used for geological applications. This paper gives an overview of isotope reference materials for uranium prepared and certified at IRMM. These materials are synthetic isotope reference materials prepared based on proven methods of purifying and mixing highly enriched oxides. Firstly, the double spike IRMM-3636 with a 233U/236U ratio of 1:1 was prepared which allows internal mass fractionation correction for high precision 235U/238U ratio measurements. The 234U abundance of this double spike material is low enough to allow an accurate and precise correction of 234U/238U ratios, even for measurements of close to equilibrium uranium samples. The double spike IRMM-3636 is offered in 3 concentrations: 1mg/g, 0.1mg/g and 0.005mg/g. Secondly, the 236U single spike IRMM-3660 was prepared and is offered in 3 concentrations: 1mg/g, 0.1mg/g and 0.01mg/g. Thirdly, a quad-isotope reference material, IRMM-3101, has been prepared which is characterized by 233U/235U/236U/238U=1/1/1/1. This material is useful for checking Faraday cup efficiencies and inter- calibration of MIC (multiple ion counting) detectors. The quad-IRM is offered in 3 concentrations: 1mg/g, 0.1mg/g and 0.01mg/g.
V13A-2082
Linearity Testing of Secondary Electron Multipliers using the IRMM-073/-074 Series of Isotope Reference Materials
For linearity testing as well as dead time correction of secondary electron multipliers (SEM), special sets of gravimetrically prepared synthetic isotope reference materials have been prepared by IRMM. Firstly, a set of 10 mixtures of 233U, 235U and 238U was made in which the 235U:238U ratios were kept at 1:1 and the 233U/235U ratios varied from 1.0 to 10-6 (IRMM-072/1-15). After the IRMM-072 series was exhausted, it has been replaced by the IRMM-073/1-15 and IRMM-074/1-10 series. In this paper two procedures for the application of IRMM-073/-074 for linearity testing are presented. First, the static procedure is ideal for ICPMS instruments to circumvent the influence of any plasma instabilities. The 233U within IRMM-073/1-15 (or IRMM-074/1-10) is detected at various intensity levels using the secondary electron multiplier, whereas the 235U and 238U isotopes are detected using a Faraday multi- collector at a uniform intensity level for each of the samples of the IRMM-073/1-15 series. The inter- calibration of the secondary electron multiplier versus the Faraday multi-collector is in each case performed using a jump of 235U into one Faraday cup. This static procedure is also applicable for TIMS instruments. However, because it is often more difficult in TIMS to realize a uniform intensity level for different standard solutions loaded on different filaments, the dynamic procedure appears to be the better choice for TIMS instruments. For the dynamic procedure one suitable solution of the series, e.g. IRMM-074/3 with a 233U/235U ratio of 0.01, is measured in peak-jumping mode only using the secondary electron multiplier. Results for the static and the dynamic procedures for both ICPMS and TIMS instruments will be presented, and the uncertainty of a dead time determination be evaluated.
V13A-2083
Increasing the Accuracy in the Measurement of the Minor Isotopes of Uranium: Care in Selection of Reference Materials, Baselines and Detector Calibration
The minor isotopes of uranium (U-233, U-234, U-236) are increasingly useful for tracing a variety of
processes: movement of anthropogenic nuclides in the environment (ref 1), sources of uranium ores (ref 2),
and nuclear material attribution (ref 3). We report on improved accuracy for U-234/238 and U-236/238 by
supplementing total evaporation protocol TIMS measurement on Faraday detectors (ref 4)with multiplier
measurement for the minor isotopes. Measurement of small signals on Faraday detectors alone is limited by
noise floors of the amplifiers and accurate measurement of the baseline offsets. The combined detector
approach improves the reproducibility to better than ±1% (relative) for the U-234/238 at natural
abundance, and yields a detection limit for U-236/U-238 of <0.2 ppm. We have quantified contribution of
different factors to the uncertainties associated with these peak jumping measurement on a single detector,
with an aim of further improvement. The uncertainties in the certified values for U-234 and U-236 in the
uranium standard NBS U005, if used for mass bias correction, dominates the uncertainty in their isotopic ratio
measurements. Software limitations in baseline measurement drives the detection limit for the U-236/U-238
ratio. This is a topic for discussion with the instrument manufacturers. Finally, deviation from linearity of the
response of the electron multiplier with count rate limits the accuracy and reproducibility of these minor
isotope measurements.
References:
(1) P. Steier et al(2008) Nuc Inst Meth(B), 266, 2246-2250.
(2) E. Keegan et al (2008) Appl Geochem 23, 765-777.
(3) K. Mayer et al (1998) IAEA-CN-98/11, in Advances in Destructive and Non-destructive Analysis for
Environmental Monitoring and Nuclear Forensics.
(4) S. Richter and S. Goldberg(2003) Int J Mass Spectrom, 229, 181-197.
V13A-2084
Limitation of secondary electron multiplier non-linearity on accurate U-Th isotopic determination by MC-ICP-MS
Contemporary multicollector-inductively coupled plasma mass spectrometry (MC-ICP-MS) with discrete dynode secondary electron multipliers (SEMs) can offer U-Th isotopic determinations with subpermil-permil- level precision in femtogram quantities. However, accurate isotopic measurement requires fully understanding SEM mass and intensity biases. In additional to dead-time effect, Richter et al (2001, Int. J. Mass Spectrom., 206, 105-127) reported a nonlinearity on SEMs produced by ETP and MasCom for count rates > 20 thousand counts per second (cps). We evaluated the possible biases for ion beams of 500- 1,600,000 cps on a latest MasCom SEM, SEV TE-Z/17, with more effective ion optical acceptance area (>50%) and better peak shape than previous models, used in a MC-ICP-MS, Thermo Fisher NEPTUNE. With the retarding potential quadruple lens (RPQ) turned off, ion beam intensity can be biased by only dead- time effect, which can be precisely corrected online or offline. With the RPQ on, two additional biases, an exponential-like increase of ion beam intensity from 100-100,000 s cps and an apparent dead-time effect (-2 to 2 ns) at high count rates, are observed. They are likely caused by the slightly defocused ions with a wide kinetic energy spread of ~5 eV, 10 times worse than that with thermal source, passing through the RPQ lens to the SEM, which is installed behind the focal plane. Fortunately, the two biases, which are stable during the daily measurements with the same settings of inlet system, source lenses, zoom optics, and RPQ, can be corrected effectively offline to earn accurate U-Th isotopic measurement.
V13A-2085
Application of Uncertainty in Measurement (GUM) to Isotope Mass Spectrometry: Introduction, Implemention, and Examples
As the measured value and its unit are integral parts of a measurement, so is a statement of the associated measurement uncertainty. The importance of providing an uncertainty that can reasonably be attributed to the measured value is often underrated. An assessment of uncertainty provides confidence in the value of the measurement, judgement on significance of differences between measurement results, information regarding the capability of the measurement procedure, and quality assurance. The limitations of the classical error analysis were seen as a hindrance to communication of scientific and technical measurement results, initiating the development of the Guide to the Expression of Uncertainty in Measurement (GUM) in the late 1970s. Just as the use of the International System of Units brings coherence to measurements, the International Organization for Standardization Guide to the Expression of Uncertainty in Measurement recommends a standardized way of expressing uncertainty in all kinds of measurements. Consequently, GUM has been adopted by most of the national metrology institutes in the world. A short introduction to GUM and the logical steps leading to its development will be presented, as well as a comparison between classical error analysis and GUM. Examples related to mass spectrometry for isotopic and elemental analysis will be discussed. The merits of GUM - transparency of the uncertainty evaluation, the treatment of uncertainties in a consistent logical way, and the presentation of an uncertainty budget resulting in a feedback to the analyst (i.e. identifies the dominant components of uncertainty and allows better understanding and improvement of the measurement process) - will be emphasised.
V13A-2086
A Smorgasbord of Double Spikes for the Discerning Gourmet
The natural mass-dependent isotope fractionation (MDF) of many elements can be quantified by addition of a tracer consisting of two artificially enriched isotope spikes prior to analysis. The double spike (DS) technique is the method of choice whenever possible, in terms of simplicity and accuracy, since it is free of nearly all instrumental artifacts, including the instrumental mass bias. Moreover, the elemental concentration is obtained by isotope dilution as a by-product. A more widespread use of double spikes has been hampered by several issues. The first of these is the erroneous idea that post-analysis data reduction for the MDF originally in the sample is difficult, which it is not. More serious is the choice of DS and mixing proportions to use for a given element, whereby the aim is to minimize uncertainties on the calculated MDF. This DS choice is not straightforward, not least because it is unclear what constitutes a "good" versus a "bad" DS. A more troublesome practical issue is the sheer number of candidate DS to look into for some elements. Here, I remedy the situation by providing a comprehensive catalogue of optimized, "good" DS compositions to use, and how much to mix in, based upon some simple objective criteria. The natural fractionation is solved for via mass balance of 3 independent isotope ratios (effectively a 3D isotope space). Thus, the DS technique can be used for any element with four or more isotopes - or 33 elements in all. In the case of Fe, having 4 isotopes. there are only 4 isotope spaces to consider, while for Sn (10 isotopes) there are 840 possible isotope spaces. In all, there are 33 elements that can be double spiked, and a total of 3868 potential isotope spaces to examine. The "optimization" was performed by considering a penalty function which depends on the normal N (unspiked), mixture M and spike S compositions in a given 3D isotope space. Normal compositions (N) were taken from IUPAC. Four criteria are built into the penalty: (1) M should lie far from S and close to N, (2) the angle between the fractionation vector through N and the mixing line should be large (optimally 90°), (3) the so-called "intersection angle" should be large (optimally 90°), and (4) the absolute magnitudes of the 3 isotope ratios of M should lie close to unity. The composition of M was solved for numerically (Nelder-Mead simplex method) that minimized the penalty function. This approach yields, first, an estimate of the "best" mixture (M) to use within a given 3D isotope space (and its corresponding spike, S), and second, a "figure of merit" (the penalty) enabling comparison between different isotope spaces of each element. The methodology contrasts with that of previous studies in two ways: first, the "optimal" mixture is explicitly solved for rather than the spike composition, and second, far more rigorous and realistic criteria are used. Since all the potential DS for a given element have been explored and evaluated here, this allows a greater objective choice of DS in practical applications. Examples of where this is useful include situations where isobaric interferences preclude certain isotopes from consideration, or where the natural or instrumental fractionation exhibits mass-independent effects.
V13A-2087
Fractionation behavior of neodymium isotopes during thermal ionization mass spectrometry
With the introduction of new generation multi-collector thermal ionization mass spectrometer (TRITON) a number of studies have investigated the 142Nd isotope excesses/deficits in terrestrial and extra- terrestrial materials. A key issue is the extent to which the fractionation behavior of Nd in the ion-source affects the accuracy and the ultimate experimental precision to which the Nd isotopes can be measured using thermal ionization mass spectrometry. We investigated if there is a systematic fine structure to the observed mass dependent isotope fractionation of Nd isotopes during thermal ionization, which violates the exponential law and requires its further refinement for high precision Nd isotope data. We find that: 1. The exponential fractionation law is fully adequate in correcting for mass dependent fractionation introduced in the TIMS source at the level of precision presently obtainable. This is not the case for other fractionation laws, including the Rayleigh law. 2. Seemingly mass dependent correlations between fractionation corrected isotope ratios are correlated errors caused by counting statistics uncertainty. These correlations are not related to the fractionation behavior of the sample. 3. In the case of Nd where variations in the light isotopes 142Nd and 143Nd are of primary interest, more precise data are obtained when the isotope pairs with greater mass range (i.e., 150Nd/144Nd or 150Nd/142Nd) are used to assess fractionation factor. However, this only increases the precision with ~20%. The downside of using this normalization is an increase in the demand of the chemical purification of Nd to reduce/eliminate Sm (and Ce) interference. 4. Rreservoir mixing can introduce variations in isotope ratios as large as 50% of the external reproducibility. However, the fractionation behavior of a sample is not necessarily a good measure of the amount of mixing between reservoirs on the filament. For samples showing reverse fractionation, there appear to be mixing at the same extent before, during, and after the period of reverse fractionation. There is likely significant mixing taking place in samples showing only normal fractionation. 5. It is imperative to measure all the isotopes of Nd with the best possible precision to assess whether measured anomalies (positive or negative) in 142Nd are real or could be the product of reservoir mixing on the filament.
V13A-2088
Isotopic analysis of small Pb samples using MC-ICPMS: The limits of precision and comparison to TIMS
Multicollector ICP-MS is a mainstream method for precise isotopic analyses of large (over 10-8 g) quantities of Pb, and is becoming increasingly popular for very rapid, even if relatively imprecise, U-Pb dating of U-bearing minerals using laser ablation. At the same time, high precision U-Pb geo- and cosmo- chronology mainly utilizes isotope dilution thermal ionization mass spectrometry, recently enhanced by application of double spikes for both Pb and U. Here we explore the suitability of MC-ICPMS for analysis of 10-11-10-9 g quantities of radiogenic Pb, contained in small single grains of zircon and other U- bearing minerals, and in chondrules, refractory inclusions and mineral fractions from meteorites. Analyses were performed at the Geological Survey of Canada using a Nu Plasma with DSN-100 desolvating nebulizer, at Arizona State University using a Neptune with Apex nebulizer, and at Harvard University using an Isoprobe P with Apex nebulizer. A total ion yield of 0.4-0.5% was achieved in all three instruments in 2.5-4 minute analyses. The fractions of SRM-981 and SRM-983 standards, spiked with 202Pb-205Pb- 233U-235U [1], containing between 3*10-11 and 10-9 Pb, were analyzed in all three labs. Precision of 207Pb/206Pb ratios in SRM-981 was 0.1-0.3% for 3*10-11 g fractions, 0.03-0.1% for 10-10 g fractions, and 0.006-0.013% for 10-9 g fractions. Precision of the best MC-ICPMS analyses was similar to precision of average TIMS analyses from the same quantities of Pb. Reproducibility of analyses depends on accurate blank and background subtraction as much as on the counting statistics. A series of analyses of the same solution run within a short period of time (i.e. with constant background) yielded a reproducibility similar to that of TIMS, whereas the analyses of a series of separately prepared aliquots were less reproducible. Our data demonstrate that the quality of analyses of 10 11 – 10 9 g Pb fractions by modern MC ICPMS approaches the quality of TIMS analyses. Further enhancement in precision can be expected when the total ion yield is increased to 1.5-1.7%, the highest values reported for MC-ICPMS [2, 3], and with using ion beam detection systems optimized for measuring medium to small signals. [1] Amelin Y. and Davis W.J. (2006) JAAS 21, 1053-1061. [2] Thirlwall M.F. (2002) Chem. Geol. 184, 255–279. [3] Makishima A. et al. (2008) JAAS 23, 1102–1107
V13A-2089
Laser Ablation U(-Th)-Pb Geochronology at the Sub-Micron Scale
Accessory phase minerals with inherited cores surrounded by very thin (often <10μm) and irregular rims are common throughout the geologic record. Current in-situ U-Th-Pb analytical techniques (e.g. SIMS, LA-ICPMS) generally employ a static 'spot' analysis method which makes it difficult to confidently sample only the outer rim of such material. In addition, this approach has the potential to introduce a significant component of external common-lead, either from the grain surface itself and/or from the mounting medium. These limitations make accurate and precise age determination of thin (<20μm) rim domains particularly challenging. In order to overcome this issue, we have developed a novel analytical strategy that involves mounting whole crystals (with thin rims intact) on double-sided sticky tape and analysing them by LA- MC-ICP-MS. This presents the laser with the flat outer surface of the grain which can then be ablated using single pulses of a laser beam. Data acquisition relies on a low-volume ablation cell that facilitates 1) rapid transfer (~1.2 s) of ablated material to the mass spectrometer and; 2) a short (~0.5 s) sample washout time. Data from reference zircons obtained by this method indicate that it is possible to consistently measure 206Pb/238U and 207Pb/206Pb ratios with external reproducibilities of 3% and 2.5% (2SD) respectively. The level of uncertainty is achieved on up to 90% less material than 'multi- second' static ablation protocols making it an ideal way to rapidly generate accurate and precise age information from complex accessory phase minerals at the sub-micron scale. Further, it is comparable in terms of both volume of material consumed (0.6ng zircon / laser pulse) and precision achieved to other in- situ (e.g. SIMS) methods.
V13A-2090
Evaluation of Solid Geologic Reference Materials for Uranium-Series Measurements via LA-ICPMS
Uranium-series geochemistry and geochronology have a wide range of applications in paleoclimatology, volcanology and other disciplines. To further explore these fields, the geoanalytical community has now begun to exploit recent advances in in situ, micron-scale sampling via laser ablation-ICPMS. Unfortunately, improvements in instrumentation have generally outpaced development of the appropriate geologic reference materials required for in situ U-series work. We will report results for uranium and thorium isotopic ratios and elemental concentrations measured in a suite of solid standards from the USGS (e.g., BCR-2G, BHVO-2G, GSD-1G, MACS-1, NKT-2G), as well as those from the MPI-DING series (e.g., ATHO-G, T1-G, StHs6/80-G). Specifically created for microanalysis, two of these standards are synthetic (GSD-1G, MACS-1) and the remainder are naturally-sourced glasses. They cover a range of compositions, ages (± secular equilibrium), elemental concentrations and expected isotopic ratios. The U-series isotopics of some powdered source materials have been characterized (e.g., BCR-2, BHVO-2), although there is no confirmation of the same ratios in the glass. Bulk measurement of these solid standards via TIMS and solution multicollector-ICPMS can then be used to assess the performance of LA-ICPMS techniques which require matrix-matched solid standards for correction of U-series elemental and isotopic ratios. These results from existing, widely-available reference materials will also facilitate quantification and comparison of U-series data among laboratories in the broader geoscience community.
V13A-2091
U-Pb Dating and Ion Imaging of Zircon by SHRIMP: The Serendipity of Finding the Earliest Hadean Zircons by Single-Spot Analysis
Traditionally, SHRIMP analysts have targeted the most pristine regions of a zircon to give the best estimate of crystallisation age. All zircons older than 4.2 Ga appear to have multiple age domains. These ages have been interpreted in terms of multiple magmatic overgrowths and/or episodic lead loss. This results in a problem for single spot analysis in that it introduces an element of chance in finding the oldest components, and in the comparison of individual SHRIMP ages with bulk zircon analyses as produced by TIMS and ICPMS. In order to unify the Hadean dataset and to resolve issues of technique bias, we have developed a multi-spot U-Pb and Pb-Pb SHRIMP dating protocol that in effect takes a 2-D slice through the whole zircon. In doing so, a picture emerges of the event information stored by that zircon. The ages in a number of slices can also be compared to assess commonality and the role of potential major geological inputs. This technique has application in the analysis of younger complex geological terranes as well. SHRIMP ages in individual zircon slices show as much age diversity as the grain population as a whole. Grains identified from single spots as being 4.35 Ga commonly show that the concordant area is very localised and ancient and recent U-Pb and Pb-Pb disturbance is extensive and not symmetric about the grain core. Therefore targeting grain cores alone will only provide a stochastic chance of finding the oldest component. The age distribution in a slice can be integrated to allow comparison with bulk single-grain analyses. For zircons with single-spot ages as old as 4.35Ga, the integrated slice age for that zircon is typically 4.1Ga (simple mean of 207Pb/206Pb ages), which is identical to that measured in whole grains.
V13A-2092
Uranium Isotopic Fractionation in Soil Extractions
Acid leachable and residual fractions in soils were separated and used for precise U isotopic measurements, as well as other major and trace elements. Two soil profiles (HS and GS) located in a river terrace in central Taiwan were selected to examine the effect of α-recoil, as well as U mobilization during weathering and soil formation. The major and trace elements in leachable and residual fraction were determined by HR- ICP-MS and the U isotopic compositions of 234U/238U and 238U/235U were measured by MC-ICPMS. Prior to isotopic analyses, dissolved U was purified using TRU-spec resin. The external reproducibility of 234U/238U for IAPSO and in-house U standard (sample size of 2-5 ng U) is 1.1 ‰ (2σm, n=10) and 0.6 ‰ (2σm, n=36), respectively compared with that of 0.2 ‰ for 238U/235U (2σm, n=36). The U concentration shows no distinguishable difference at soil horizon in both profiles. The average (234U/238U)AR ratio in the leachable fraction, 1.40, deviates largely from the equilibrium value and appears to be affected by soil maturation and degree of weathering. On the other hand, the average (234U/238U)AR in the residual fraction (0.93) is also much lower than the equilibrium ratio. Further study will focus to examine if these ratios provide information on soil formation rate and/or chemical weathering rate.
V13A-2093
Using Natural and Synthetic Zircons to Test the Reliability of Hf Isotopic Measurements During Laser Ablation MC-ICP-MS Analyses
Hf isotopic analyses of zircons in combination with U-Pb age information constitute a powerful tool for constraining the formation and evolution of the continental lithosphere. A rapidly growing data-set of in-situ Hf isotopic measurements by LA-MC-ICP-MS has accumulated since the early 2000s. Such in-situ analyses require Lu and Yb isobaric interference corrections. In order to assess the viability of the isobaric corrections on the Hf isotopic measurements, we prepared a set of synthetic zircons with variable amounts of REE and Y. Analyses of these zircons and natural zircons show that accuracy and precision of Hf isotopic measurements in zircons decrease systematically with increasing REE/Hf ratios and Y contents. The synthetic zircon experiments indicate that at Lu-Yb/Hf ratios resulting in around 20 percent or less difference between uncorrected and Lu-Yb corrected 176Hf/177Hf ratios, we can routinely obtain accurate and precise Hf isotopic data (+/- 1 epsilon Hf at 2 sd). Grain to grain reproducibility increases to +/-4 epsilon units at corrections of 60 percent and accuracy suffers slightly as well. At corrections of around 80 percent, the grain to grain reproducibility is +/- 9 epsilon units and the 176Hf/177Hf ratios are highly inaccurate. This is probably related to formation of REE oxides and possibly Y dimers, which directly interfere with all masses of interest during the Hf isotopic analyses. As a result, above certain REE/Hf threshold the Lu-Yb peak stripping alone becomes an inadequate proxy for all isobaric interferences. In order to better assess these effects in natural zircons, we analyzed a number of zircon standards, including 91500, BR266, MAD, SAME, FC-1, and R33. FC-1 has the highest REE/Hf ratios that yielded accurate and precise Hf isotope ratios (+/- 1 epsilon Hf, 2 sd). R33 has higher than FC1 REE/Hf ratios and as a result both accuracy and precision suffer (+/- 2 epsilon units Hf, 2 sd). We attribute this to spatial heterogeneity in REE contents and possibly of 176Hf/177Hf. Overall, our experiments indicate that we can routinely obtain accurate and precise Hf isotopic data for natural and synthetic zircons with REE/Hf ratios that require 20 percent or less correction between uncorrected and Lu-Yb corrected 176Hf/177Hf ratios. Obtaining accurate in situ Hf isotopic compositions from zircons, therefore, hinges on having appropriate standards and a careful evaluation of trace element abundances in the unknowns.
V13A-2094
Evaluation of d-DIHEN for MC-ICP-MS Isotope Ratio Measurements
Multiple Collector Inductively Coupled Plasma Mass Spectrometry (MC-ICP-MS) is today's method of choice for measurements of small relative isotopic variations (below 10-4). For elements such as Nd, isotope ratio relative repeatability can be as low as few parts per million when comfortable isotope signals (>1V) are acquired. For low concentrations and/or limited volume samples, it might be necessary to enhance the instrument sensitivity to keep such analytical performances. While desolvation systems are commonly interfaced with MC-ICP-MS instruments, the use of Direct Injection Nebulizer (DIN) is rarely reported for isotope ratio measurements. As described earlier, for standard sample introduction systems - including a spray chamber and a pneumatic nebulizer - only 15% of the analyte is transmitted to the ICP, the remaining being collected to the drain. Using a DIN could thus potentially increase the instrument sensitivity by a factor ~6 as compared to conventional systems. In this case, the primary aerosol is injected into the plasma, therefore the optimisation of the spray conditions is essential to get a stabilised signal and a competitive isotope ratio repeatability. In this study, we evaluate the performance of the demountable version of the Direct Injection High Efficiency Nebulizer (d-DIHEN) for the measurements of isotope ratios by MC-ICP-MS. A comparison of figures of merit is described with experiments conducted with a classical mini-cyclonic spray chamber (Cinnabar) and a desolvation system (APEX). The d-DIHEN shows similar or better sensitivity than the APEX. Also, isotope ratio repeatability and accuracy for both systems are equivalent. As expected, rinse out profiles during d-DIHEN experiments do not reveal memory effects and lower backgrounds are observed. Repeatabilities for internally corrected isotope ratios are equivalent for all systems, while long term mass discrimination stability appears to be twice better for the d-DIHEN. Additionally to easy maintenance and relatively low cost, these experiments showed that the d-DIHEN can be an interesting alternative to desolvation devices for MC-ICP-MS isotope ratio measurements in micro-volume samples.
V13A-2095
USGS GSD-1G: A Geological Reference Glass for in situ Elemental and Isotopic Analysis
To overcome the drawback of synthetic calibration materials for microanalysis of geological samples, the
USGS prepared basaltic GS reference glasses, which are doped with many trace elements in similar
abundances. The GSD-1G glass is of special interest, because the trace elements have appropriate
concentrations of about 50 μg/g. Meanwhile first analytical results have been published which all agree
very well. The GeoReM preferred values (http://georem.mpch-mainz.gwdg.de) have therefore a high level of
confidence.
For in situ isotope analysis an increasing need exists for well characterized reference glasses to satisfy the
requirements for new LA-ICP-MS and SIMS applications. We have therefore determined the isotopic
composition of Li, B, Si, Ca, Sr, Nd and Pb in GSD-1G by high-precision bulk and microanalytical techniques
(Table 1). To check isotopic homogeneity, three different glass splits were analyzed.
The highly precise δ11B, δ44/40Ca, 87Sr/86Sr, 143Nd/144Nd, 206Pb/204Pb values are
indistinguishable within uncertainty limits indicating isotopic homogeneity. Small Pb isotopic heterogeneities
between splits are detected in the TIMS triple spike data but not resolved in the LA-ICP-MS Pb data due to
the larger uncertainty, a factor of 10 greater than that of TIMS. Nevertheless, the mean LA-ICP-MS
207Pb/206Pb and 208Pb/206Pb agree well with the TIMS data. δ29Si and δ30Si were
determined by LA-MC-ICP-MS using a femtosecond laser. The δ7Li and Li concentration (43.8
μg/g) values obtained by MC-ICP-MS also agree well with previously published MC-ICP-MS, TIMS and
SIMS Li isotopic data analyzed on different splits (30.3‰ Jochum et al. (2006); 31.14, 31.7, 31.3‰
Kasemann et al. (2005)) and the GeoReM preferred Li concentration (43 μg/g) indicating isotopic and
elemental homogeneity of Li in GSD-1G.
Table 1: Compilation of isotope data of USGS GSD-1G reference glass.
δ7Li(‰ LSVEC): 30.4 using MC-ICP-MS
δ11B(‰ NIST951: 10.07; 9.80; 10.74 using TIMS
δ29Si(‰ NIST8546): -0.080 using LA-MC-ICP-MS
δ30Si(‰ NIST8546): -0.224 using LA-MC-ICP-MS
δ44/40Ca(‰ NIST915a): 0.79; 0.90; 0.89 using TIMS
87Sr/86Sr: 0.709396; 0.709403; 0.709403 using TIMS
143Nd/144Nd: 0.511513; 0.511514; 0.511515 using TIMS
206Pb/204Pb: 19.578; 19.579; 19.578 using TIMS-triple spike
207Pb/204Pb: 15.7436; 15.7446; 15.7459 using TIMS-triple spike
208Pb/204Pb: 38.906; 38.910; 38.916 using TIMS-triple spike
207Pb/206Pb: 0.80416; 0.80515; 0.80424 using TIMS-triple spike
208Pb/206Pb: 1.9872; 1.9873; 1.9876 using TIMS-triple spike
207Pb/206Pb: 0.8042 using LA-ICP-MS
208Pb/206Pb: 1.987 using LA-ICP-MS
V13A-2096
The Absolute Isotopic Composition of Zn in Terrestrial Materials Determined Using Double Spike Thermal Ionization Mass Spectrometry
Although long suspected to be widespread in nature, until recently, little was known about the extent of the variation of the isotopic composition, or isotopic fractionation, of Zn in natural materials. During the last decade an increasing number of high precision Zn isotopic fractionation data have been reported using MC- ICP-MS (MARECHAL et al., 1999; PETIT et al., 2008; PICHAT et al., 2003), but none have been reported on an absolute scale which is essential for interlaboratory comparison of results. In this work we report sub- permil Zn fractionation in a range of natural materials relative to the internationally proposed absolute Zn isotopic reference material (δ zero) (PONZEVERA et al., 2006)using the Thermal Ionization Mass Spectrometry double spike technique. Repeated double spike analysis of the laboratory standard relative to itself demonstrated a long term reproducibility of +0.006 ± 0.039 permil amu-1. The measured isotopic composition of Zn in minerals and igneous rocks SRMs was found to be the same as the proposed absolute (δ zero) which makes it possible to consider the proposed absolute Zn isotopic standard as being representative of "bulk earth" Zn. A significant and consistent fractionation of ~+0.3 permil amu-1 was found in 5 sediments from a range of localities. The results obtained for metamorphic SRMs indicate that the fractionation of Zn in these rocks is the same as found in igneous rocks but are different from the Zn found in sedimentary rocks. A clay SRM sample TILL-3 appears to exhibit a consistently Zn fractionation of +0.12 ± 0.10 permil amu-1. The isotopic composition of Zn was also measured in two plant SRMs and one animal SRM sample. The fractionation of (-0.088 ± 0.070 permil amu-1) of Zn in the Rice (a C3 type plant material) sample suggested that Zn may be used to study Zn systematics in plants. The result obtained for MURST-ISS-A2 (Antarctic Krill) was +0.21 ± 0.11 permil amu-1 relative to the laboratory standard which is similar to the average Zn fractionation results of +0.281 ± 0.083 permil amu-1 obtained for marine sediments. The fractionation of Zn in seven ultra pure Zn standard materials was also measured relative to the laboratory standard and found to range from -5.11 ± 0.36 permil amu-1 for AE 10760 to +0.12 ± 0.16 permil amu-1 for Zn IRMM 10440 confirming that that significant care must be exercised in the selection of Zn isotope laboratory standards (TANIMIZU et al., 2002). A pilot study to determine the concentration and the isotopic composition of Zn in river and tap water, and a number of processed materials was also performed. The implications and applications of these results, such as on the atomic weight of Zn will be presented.
V13A-2097
High-precision calculation of the branching ratio of the 40K decay constant.
40K is of great importance in Earth science, particularly for K/Ar, 40Ar/39Ar and K/Ca
geochronology. The decay scheme of the 40K includes two different modes of decay, beta and electron
capture followed by gamma-ray emission, which yield two different products, 40Ca* and 40Ar*. The
relative probability that 40K decay following one of the two schemes is known as the branching ratio. An
original method of calculation to obtain the value of the 40K branching ratio (λβ-
/λtot) based on the K/Ar technique, is proposed. λβ- /λtot is obtained
by combining the 40Ar*/40K value of Fish Canyon sanidine (FCs) secondary standard derived
from four primary 40Ar/39Ar standards, with the current best estimates of the age of FCs and the
value of the 40K total decay constant. The latest estimation of the 40K total decay constant and the
age of FCs by Mundil et al. (2006), through comparison with U/Pb ages, yields a λβ-
/λtot value of 89.59 ± 0.03% (1σ; relative error = ± 0.035%). Indirect
measurement of the age of FCs by orbital tuning (Kuiper et al., 2008) combined with the value of 40K
total decay constant measured by liquid scintillation counting by Kossert and Gunther (2004) yields a
statistically indistinguishable value for the branching ratio of 89.61 ± 0.03%, with an average between
the two values of 89.60 ± 0.04%. The method proposed here allows can easily be applied to further
constrain the value of the 40K branching ratio as future refinements of the 40K decay constant and
FCs age are produced, although it is expected that the adopted value will be close to λβ-
/λtot = 89.60 ± 0.04%.
Kossert and Gunther, 2004. Appl. Radiat. Isot. 60, 459–464.
Kuiper et al., 2008. Science 320, 500-504.
Mundil et al. 2006, Eos Trans. AGU, 87(52)
V13A-2098
Effectiveness of Combined K-Ar and 40Ar/39Ar Dating Methods in the 14C age Range
Several studies achieved during the last decade evidence the strong endeavor to push toward younger and younger ages the 40Ar/39Ar chronology. In parallel, all these progress have extended the scope of application of the 40Ar/39Ar chronology especially in paleoclimatic studies to provide Astronomical independent ages for oxygen isotope stages. Most of these promising results are based on K-rich feldspars, from tephra layers, which allow a direct comparison between oceanic and continental records. However, K-rich feldspars-bearing tephra are scarce and improvements to date effusive volcanic rocks, covering much wider area, have to be done to enlarge the field of chronological studies in Quaternary times. To participate to this challenge a new 40Ar/39Ar laboratory has been set up at the LSCE to complete the already existing K-Ar facility. The unspiked K-Ar and the 40Ar-39Ar methods have each their own drawbacks and advantages and the combination of these two methods is very efficient to obtain accurate and reliable age determinations as demonstrated by the recent dating of the Laschamp excursion (Guillou et al., 2004, EPSL). To extend this approach toward younger ages, we studied two lava flows which emplaced around 30 ka, based on preliminary datings within the Teide - Pico Viejo volcanic complex in the Canary Islands. The first sample, which has a conventional radiocarbon age of 32,360 +/- 800 yrs B.P. is K-Ar dated at 33 +/- 2 ka and 40Ar/39Ar dated at 33 +/-2 ka. The second sample is K-Ar dated at 32 +/- 2 ka and 39Ar-40Ar dated at 35 +/-3ka. These similar ages and therefore successful comparison between the two methods validate the performances of the new LSCE 39Ar-40Ar facility and is a very promising approach to calibrate the Quaternary timescale.
V13A-2099
The use of Fluorinated Graphite Polymer in the Oxygen-Isotope Analysis of Silicates by Continuous Flow Isotope Ratio Mass Spectrometry (CF-IRMS)
To date the majority of analytical procedures for δ18O analysis of silicates require relatively expensive apparatus involving the use of hazardous fluorine or halogen fluoride gases. Carbon reduction offers the potential to avoid the use of hazardous gases; however, it has been plagued by low yields and variable isotopic fractionation leading to significant differences in isotopic composition compared to conventional fluorination techniques (CFT). We adopt a carbon reduction method, which includes a fluorinated graphite polymer (F>60%) to help oxidize the silicate, thus increasing the percent yield as well as the precision of the analyses. Three silicate standards (NBS-28; UWG-2; GBW044190) were analyzed to compare this method against CFT. Silicate samples of 0.2mg are combined with 2mg of fluorinated graphite and sealed in a silver cup. The cups are then heated instantaneously to 1470°C in a high temperature conversion elemental analyzer (TC/EA). The evolved oxygen reacts with carbon in the furnace resulting in carbon monoxide which is transported via a helium carrier gas to a series of chemical and cold traps prior to introduction into the gas chromatograph. An Ascarite II and Mg(ClO4)2 chemical trap removes acids (e.g. HF) and water, followed by an LN2 cold trap to remove SiF4. The He carrier gas flow is kept low (0.7bar) to maximize reaction time, whereas He flow to the open split is increased (0.7bar) in order to reduce peak tailing in the IRMS. Using this method individual analysis can be completed in ~7 minutes. Oxygen yields range from 54-92%; however reproducibility better than 0.5‰ is normally achievable. Measured values are consistently enriched in 18O compared with accepted values; however, this offset is systematic and can be corrected for using a linear correction factor. Corrected values are: 9.11±0.3‰, 5.95±0.5‰, and -1.78±0.4‰ for NBS-28, UWG-2, and GBW044190 respectively. As the number of laboratories with CF-IRMS and TC/EA apparatus increases the development of this method has the potential to make δ18O analysis of silicates more widely available.
V13A-2100
Multi-Step CA-TIMS U-Pb Zircon Analysis of the Relative Decay Constants of 235U and 238U
The U decay constants of Jaffey et al. (1971) are widely regarded as the "gold standard" for geochronology, given their exceptional precision and accuracy compared with earlier U decay constant measurements and with the decay constants for other isotopic systems used in geochronology. Several years ago, it became clear that analytical methods had improved to the point where measurements of carefully selected natural zircon samples could provide a useful independent check of the accuracy of the 235U and 238U decay constants relative to each other. Using this approach, Mattinson (2000) and Schoene et al. (2006) both determined that the accepted decay constant for 235U is too low by ca. 0.09%, resulting in 207Pb/206Pb ages that are systematically too high, relative to 206Pb/238U ages. The relative accuracy of the two decay constants is crucial for the interpretation of "concordance", i.e., agreement of ages calculated from the two independent decay systems. The present study applies detailed multi-step CA-TIMS measurements to a suite of zircon samples selected for apparent perfect concordance; no Pb loss after removal of radiation- damaged zircon by early CA-TIMS steps, and no evidence of any older inherited component. Most samples were selected from a 400Ma to 600Ma age range, a range where errors from tracer calibration uncertainties and corrections for intermediate daughter product disequilibrium are both low. The new measurements strongly confirm the ca. 0.09% error in the 235U decay constant, relative to the Jaffey et al. (1971) 238U decay constant. Use of this recommended correction of the 235U decay constant will allow more accurate assessment of concordant to near-concordant zircon data, and also more accurate concordia intercept ages for more discordant data.
V13A-2101
New Approach to Estimate 40Ar/36Ar Ratio in Shocked Meteorites
Measurement of a 40Ar/39Ar age in a meteorite requires proper estimation of a 40Ar/36Ar "initial" ratio. For a primordial one, it is reported to be an order of 10-4, suggesting that one can ignore it in age calculation. However, there is no basis for applying this value as individual meteorites have different histories. In terms of 40Ar/39Ar studies on various meteorites, very few studies discuss how the ratios were estimated. Using terrestrial value of 295.5 is invalid. The only case in which the ratio does not affect on the results is when an age of a meteorite is in an order of billion years, resulting in a large raw 40Ar/36Ar ratio. A proper approach to estimate the ratio is application of isochron analyses. When a small amount of sample (a few micrograms) is studied in laserprobe analysis, this often becomes difficult. A shocked meteorite Y-75097 has developed maskelynite veins, which were identified in optical microscope, electron microprobe and thermoluminescence studies. We have applied 40Ar/39Ar dating method using both stepwise heating by a continuous laser and pinpoint dating by a pulsed laser. Stepwise heating results of several fragments (about 50 micrograms) of the meteorite yielded various old and young ages. Most pinpoint analyses gave the raw 40Ar/36Ar ratio less than the terrestrial value with sporadically and anomalously old ages. Among the data, three points isochron from a maskelynite vein yielded about 300 Ma with "initial" ratio of 180. To confirm this result, plagioclase and olivine data near the vein were plotted in a correlation diagram, but an isochron is not well defined due to relatively large analytical errors. A new method to analyze these data is proposed. Once 40Ar/39Ar ratio is determined from an experiment, the only factor which determines the age is the "initial" ratio in the 40Ar/39Ar age equation. Instead of using a known fixed ratio, we set the ratio (y-axis) and the age (x-axis) as variables. In this diagram, a single data point defines a curve. If the data belong to the same group with the same ratio and age, they share a single common crossing point. This method helps identifying different groups visually. In fact, crossing points of three curves from the maskelynite vein created a relatively tight group. Furthermore, plagioclase and olivine data formed a loose group of 300 Ma age and ratio value of zero. These results suggest that the different minerals in a few millimeters recorded the same shocked age with different "initial" ratio. Examples of several other studies on shocked meteorites are compared and a possible origin of anomalous old ages will be discussed.
V13A-2102
Evaluation of double and triple spike Fe isotope measurements and results for basaltic lavas
The precision of Fe isotopic analyses by MC-ICP-MS largely depends on the efficacy of the correction for instrumental mass fractionation, and the ability to correct for polyatomic (Ar-N-O) interferences. Polyatomic interferences can be addressed with high-resolution mass spectrometry (e.g. Nu Plasma 1700), assisted by a desolvating nebulizer. In order to correct for instrumental fractionation, spiking can be used to directly monitor fractionation during analyses. We evaluated the potential precision of Fe double and triple spike analyses, and two data reduction routines from the literature. Using a Monte Carlo type error analysis, we evaluated the potential precision of any spike composition from the full range of 58Fe-54Fe, 58Fe-57Fe, and 57Fe-54Fe double spikes and the full range in a 58Fe-57Fe- 54Fe triple spike. The optimal spike compositions that follow from this error calculation are double spike compositions, where particularly the 58Fe-57Fe double spike has attractive properties. For example, the standard deviation of the difference between true composition and Monte Carlo results does not increase above 0.025 ‰ within a range of spike addition of ~0.05-0.95 mol fraction. Therefore, we prepared a 58Fe-57Fe double spike, and calibrated its composition with a series of measurements of a standard with an increasing amount of double spike added. The true composition can then be found through reverse isotope dilution calculations and minimizing the differences in double spike-reduced compositions and instrumental fractionation factors of the mixtures. Using this double spike we measured the Fe isotope composition of a number of rock standards as well as lavas from a range of tectonic settings. Our sample measurements show most samples have Fe isotope compositions near average igneous rocks, although some Samoan lavas show relatively large fractionation. Based on replicate analyses of igneous rock standards our reproducibility is ~0.05 ‰ (2σ).
V13A-2103
Matrix Independent Standardization by LA-ICP-MS
LA-ICP-MS analysis combines laser ablation with inductively coupled plasma mass spectrometry for the rapid chemical analyses of solid samples at spatial resolution better than 100 μm. Standardization in LA-ICP- MS involves both internal and external calibration, and is limited by the small number of reliable standards available and by the diversity of solid matrices. In laser ablation, the laser produces an aerosol that is ionized by the argon plasma source, so that ionization efficiency depends primarily on the plasma and not on the ablation process. Thus, internally standardized measurements by laser ablation should, in principle, be insensitive to matrix composition. The observations that large ablated particles are not fully vaporized by the plasma and that particle size is correlated with the sample transparency, however, introduce matrix effects. Using a UP213 (213 nm) laser ablation system coupled to high resolution ICP-MS (Element1 and Element XR) we examined the effects of sample transparency, mode of ablation, and laser energy on the relative sensitivities of a suite of elements across a wide range of a silicate standards, from the nearly transparent SRM 614, to the almost opaque BCR-2G. We compare these sensitivities to those obtained by solution ICP- MS to document that laser ablation of low Fe glasses (< 500ppm) results in the decreased measured sensitivities of refractory elements. The matrix effects observed for transparent glasses diminish with increasing energy output of the laser, and are not apparent in glasses with Fe concentrations above ~500 ppm. Additionally we find that measured elemental relative sensitivities factors (RSFs) in metals, sulfides, and iron-bearing silicates agreed well, indicating that RSFs are largely independent of matrix in any standard which absorbs laser energy readily. This demonstration of matrix-independent calibration for non-transparent samples greatly expands the potential of LA-ICP-MS. For example, RSFs for PGEs derived from iron meteorite standards can be applied to PGE analysis in silicates and sulfides; RSFs from silicate MPI-DING glasses can be applied to the analysis of trace elements in carbonates, and analysis of a wide range of cosmochemically interesting elements from the NASA Stardust particles embedded in an aerogel matrix is now possible using RSFs from a diverse array of metal, sulfide and silicate standards.
V13A-2104
Argon diffusion in plagioclase and implications for thermochronology
Despite being the most abundant mineral in the Earth's crust and one commonly used for 40Ar/39Ar geochronology, plagioclase has been little studied with respect to Ar diffusion kinetics and remains unexploited in most thermochronologic studies. We present results of 39Ar and 37Ar diffusion experiments and 40Ar/39Ar geochronology on both plutonic and volcanic plagioclase encompassing a variety of compositions and microstructural states. Single 300-1200 μm neutron- irradiated plagioclase crystals were wrapped in platinum or niobium packets and incrementally heated (in most cases cyclically) with a diode laser between 400 and 1600 °C. Temperature was controlled and monitored with an optical pyrometer, calibrated to ± 10 °C (1σ). We calculated diffusion coefficients (D/a2) from fractional release data assuming a spherical geometry. From least-squares regression of the low temperature values (<1000 °C; the first ~10-30% of the total 39Ar and 37Ar released), we quantified the diffusion parameters Ea and D0/a2. Like K- feldspar, microstructure, grain size, and compositional zoning heavily influence argon diffusion in plagioclase, highlighting the importance of single-crystal analyses. Over 15 experiments with different heating schedules yield Ea of 30-70 kcal/mol, positively correlated with D0/a2, which varies between –1.0 and 20.0. Arrhenius plots reveal a range of behaviors interpretable in terms of either single diffusion domains, multiple diffusion domains, fast track diffusion pathways, and/or microstructures evolving during the heating experiment. These diffusion parameters correspond to closure temperatures, Tc, of 225-375 °C for cooling rates of 5-100 °C/Ma. Our results indicate that plagioclase can be used to constrain thermal histories in almost all geologic settings, including those lacking alkali feldspars (e.g. meteorites and mafic/ultramafic intrusions). When multiple single-crystal analyses are considered in conjunction, detailed low-temperature thermal histories can be quantified. We illustrate this with analyses of plagioclase grains from the Bushveld Complex, South Africa.
V13A-2105
The Isotopic Composition of Atmospheric Argon and K-Ar Based Geochronology
K-Ar and 40Ar/39Ar geochronologists use the isotopic composition of atmospheric argon to correct for non-radiogenic 40Ar and to determine instrumental mass discrimination, among other purposes. The currently accepted atmospheric composition is based on measurements of Nier (1950), as filtered by Steiger and Jager (1977). A redetermination of the 40Ar: 38Ar: 36Ar of atmospheric argon by Lee et al. (2006), with 40Ar/36Ar = 298.56 ±0.31 (1 σ here and throughout) compared with Nier's (1950) value of 296.0 ±0.7, offers a more precise and likely more accurate reference standard that should be adopted by the K-Ar and 40Ar/39Ar communities. Adopting the revised atmospheric composition has little impact on most age calculations as changes resulting from the magnified atmospheric correction are exactly balanced by those from the adjusted mass discrimination correction, in cases where air aliquots are used to determine mass discrimination. More significant are the effects of modified discrimination on reactor-produced isotopes, though in most realistic cases the resulting age difference is a few thousand years or less. The Lee et al. (2006) data provide a more sensitive benchmark for evaluating both natural and experimental fractionation of argon isotopes. Non-atmospheric, isotopically light initial argon turns out to be widely manifest in volcanic ejecta in which 40Ar/36Ar vs. 38Ar/36Ar correlations appear to follow a kinetic mass fractionation trend. This phenomenon has been recognized for decades (e.g., Krummenacher, 1970; Matsumoto et al., 1989; Ozawa et al, 2006) but has somehow escaped notice of the broader community. New data presented here suggest that the most extreme fractionation occurs in silicic glasses (obsidians), in which we have measured initial 40Ar/36Ar (isochron intercept) and 38Ar/36Ar (unirradiated aliquots) as low as 271.38 ±3.00 and 0.18117 ±0.00055, respectively. We suggest that the data of Lee et al. (2006) be adopted as a new standard for geochronology, and that these data be implemented in efforts to better characterize instrumental mass discrimination and to understand argon isotope fractionation in nature. Krummenacher, D., 1970, Earth Planet. Sci. Lett. 8: 109– 117; Lee, J.-Y., et al., 2006, Geochim. Cosmochim. Acta 70: 4507-4512; Matsumoto, A., et al., 1989, Mass Spectroscopy 37: 353-363; Nier, A.O., 1950, Phys. Rev. 77: 789-793; Ozawa, A., et al., , 2006, Chem. Geol. 226: 66-72; Steiger, R.H., and Jäger, E., 1977, Earth Planet. Sci. Lett. 36: 359-362.
V13A-2106
Mg Isotopes of USGS Igneous Rock Standards
Magnesium has three stable isotopes, 24Mg, 25Mg, and 26Mg with abundances of 78.99%, 10.00%, and 11.01%, respectively. It is one of the most abundant elements in the crust and mantle. As advancements of analytical techniques using MC-ICP-MS have dramatically advanced our ability to measure isotope ratios of Mg with greater precision, Mg isotopes can now be applied to study a variety of fundamental geological processes, such as continental crust weathering, chemical diffusion, and chondrule formation. Therefore the need for well characterized Mg isotope ratios for geological materials is increasingly important. Routine measurement of readily-available USGS rock standards is a viable way for inter-lab comparison to show the quality of data. However, the Mg isotope data for USGS standards reported in the literature are limited and inconsistent. USGS standards reported by different MC-ICP-MS labs have a range of Mg isotopic data outside of the normal external error of 0.1‰ (2σ). Mg isotopes of USGS igneous rock standards (dunite, DTS-1; basalts, BCR-1, BCR-2, BHVO-1; and andesite, AGV-1) were measured by a sample-standard bracketing method using a low resolution MC-ICP- MS (Nu-Plasma HR). The method has a large tolerance of matrix bias with Na/Mg and Al/Mg > 100% only changing the δ26Mg by less than 0.1‰. Dilution effects do not cause significant error (< 0.1‰) until the concentration difference between standard and sample is greater than 25%. The isobaric interference of CN+ on 26Mg was avoided by measuring Mg signal on the low mass shoulder. Only purified samples with excellent yields (>99.5%) and acceptable concentrations of matrix (mainly Na, Al, Ca, and Fe) are included in these results. Duplicate analyses of independently processed standards yielded the following results (δ26MgDSM-3 (‰)): BCR-2 (-0.306±0.144, - 0.290±0.116, -0.283±0.048, -0.288±0.057), BCR-1 (-0.399±0.079, -0.346±0.046), AGV-1 (-0.295±0.110, -0.307±0.086, -0.339±0.068), BHVO-1 (-0.308±0.076, - 0.299±0.103), and DTS-1 (-0.299±0.163, -0.368±0.059). δ26MgDSM-3 of measured USGS standards are consistent within error (2σ).
V13A-2107
Strontium ion exchange mechanism in synthetic sitinakite from time resolved studies
The mechanics of ion exchange within zeolitic materials are not well understood due to the speed at which ion diffusion takes place and the difficulty in probing the sample in real-time. We used time resolved X-ray diffraction to study ion exchange in synthetic sitinakite, a porous structure known for its very high ion selectivity. Experiments were carried out at the X7B beamline of the National Synchrotron Light Source to examine strontium ion diffusion in the hydrogen and sodium forms of the titanosilicate. Our initial diffraction results illustrate that ion exchange proceeded in two steps in both forms of sitinakite. First there was a unit cell contraction as strontium migrated into the crystal structure, and then the unit cell parameters expand to accommodate the new interstitial ion. We will present a model to explain the structural positions, time dependence of multiple site ion exchange, and the mechanism that is responsible for extraordinary ion selectivity in sitinakite.
V13A-2108
Zircon U-Pb analyses by TIMS and LA-ICPMS on the same material
A number of U-Pb geochronological techniques are now available, each of which has pros and cons associated with spatial resolution, accuracy, precision and also operating cost and throughput. Zircon U-Pb age analyses using TIMS (Thermal Ionization Mass Spectrometry) can yield geological dates with uncertainties at or below the 1 permil level (precision and accuracy), but consume relatively large amounts of material (entire crystals of fragments) and are cost and time intensive. Conversely, LA-ICPMS (Laser Ablation Inductively Coupled Plasma Mass Spectrometry) geochronology is a relatively inexpensive and fast micro-sampling technique, but individual and combined uncertainties are significantly greater. The choice of a specific approach depends primarily on the nature of the geological problem; a combination of several techniques is often advisable. We present new TIMS and LA-ICPMS results on zircons from an intraformational conglomerate within the Chinle Fm (Late Triassic), and from a volcanic ash within the Late Jurassic Morrison Fm (with additional 40Ar/39Ar analyses on phenocrystic sanidine). Preliminary TIMS analyses on annealed and chemically abraded zircons from the Chinle Fm conglomerate indicate an age of ca 215.3 ± 0.9 Ma (further analyses are expected to reduce the uncertainty significantly). Previous analyses using LA-ICPMS techniques on non-polished zircons yield an age which is significantly younger (212 ± 1.5 Ma). TIMS analyses were done on the same crystals as the LA-ICPMS analyses, only one of four crystals yields an overlapping age with the ICPMS date (most of the TIMS ages are older than the corresponding LA-ICPMS ages). Part of this discrepancy may be due the fact that the LA-ICP ages preferentially suffer from surface correlated Pb loss. This effect may be mitigated by LA-ICPMS analyses on polished crystals; however, it has been shown that Pb loss also occurs within the crystal interior. Zircons from the Morrison show a similar bias in sense and magnitude when TIMS and LA-ICPMS ages are compared, and fail to agree within uncertainty (151.3 ± 0.7 Ma versus 145.6 +2.3/-3.1 Ma, (1)). A 40Ar/39Ar age is in agreement with the TIMS age (after correction for a systematic bias from a miscalibrated decay constant of 40K), again indicating that the LA-ICPMS age may be affected by open system behavior. The origin of these discrepancies remain as yet speculative. We are confident that further comparative studies will help us understand the nature of this discrepancy. We suggest that existing datasets are critically evaluated; in some cases, quoted uncertainties may fail to take complications such as Pb loss into account. Further analytical progress, for example applying the chemical abrasion technique to zircons prior to LA- ICPMS analyses, may yield more accurate ages, although potential complications arising from Pb/U fractionation due to increased surface area have to be considered. The study on the zircons from the Chinle conglomerate, which contains zircon populations from several different sources, demonstrates that a useful approach may be to apply LA-ICPMS to a large number of zircons, followed by TIMS analyses on selected crystals, to attain the highest possible precision on the depositional. 40Ar/39Ar on K-bearing minerals yields additional constraints but systematic biases have to be taken into account. 1. B. J. Kowallis, B. Britt, E. H. Christiansen, A. L. Deino, paper presented at the GSA Joint Meeting, Las Vegas 2008.