Extending 10Be applications to carbonate-rich and mafic environments
Published:January 01, 2006
- PDF LinkChapter PDF
Régis Braucher, Pierre-Henri Blard, Lucilla Benedetti, Didier L. Bourlès, 2006. "Extending 10Be applications to carbonate-rich and mafic environments", In Situ-Produced Cosmogenic Nuclides and Quantification of Geological Processes, Ana María Alonso-Zarza, Lawrence H. Tanner
Download citation file:
Quantitative geomorphologic studies using cosmogenic nuclides in carbonate-rich and mafic environments have up to now been restricted to the cosmogenic radio-nuclide 36Cl (T1/2 = 0.31 m.y.), and to the stable 3He and 21Ne cosmogenic nuclides, respectively. To extend the time span and erosion rate range quantifiable in carbonate-rich environments, and to provide the possibility to decipher complex exposure histories by differential radioactive decay over several Ma in mafic environments, the in situ production rate of 10Be (T1/2 = 1.5 m.y.), the nuclide with the longest half-life of the well-established terrestrial cosmogenic radionuclides, has been determined in calcite and clinopyroxenes.
The development of new chemical decontamination procedures efficiently removing meteoric 10Be from carbonates and altered clinopyroxenes allows determining 10Be production rates. A 10Be production rate in clinopyroxenes of 3.1 ± 0.8 atoms/g/yr at sea level and high latitude is proposed from measurements of 10Be and 3He concentrations in K-Ar-dated Quaternary basaltic flows of Etna volcano.
Through measurements of 10Be and 36Cl concentrations in the same calcite samples and of 10Be concentrations in depth profiles of flint from the same erosional surface, a value of 37.9 ± 6.0 atoms/g/yr has been determined for the 10Be production rate in calcite at sea level and high latitude. Approximately sixfold higher than production in the coexisting flint, this higher rate of production may be due to high production cross sections for C spallation by cosmic rays with energies below 50 MeV. These results also open the possibility of dating burial events in carbonate-rich environments by differential radioactive decay of 10Be and 36Cl.
Cosmogenic nuclides produced in terrestrial rocks provide an efficient way to quantify Earth surface processes governing landscape evolution and to date geological events that exhume material from depth, such as glaciation, mass wasting, volcanic and tectonic activity, … (see review in Gosse and Phillips, 2001). The peculiar physical responses to impinging particles of each of the various targets involved in the production mechanisms, as well as the differing chemical properties, imply that measurement of in situ–produced nuclides is easier and more reliable for specific pure mineral phases. The selected phase must indeed be able to be efficiently decontaminated from the atmospheric variety (if it exists) (Brown et al., 1991), it must fully retain the studied cosmogenic nuclide (Brook et al., 1995), and the production rate in that phase must be particularly well constrained (Masarik and Reedy, 1995). Over the past 15 yr, these considerations led to the use of the following stable and radioactive in situ cosmogenic nuclide systems: (1) 10Be (T1/2= 1.5 Ma), 26Al (T1/2 = 0.73 Ma), and 21Ne (stable) in quartz (e.g., Brown et al., 1991; Staudacher and Allègre, 1991), (2) 36Cl (T1/2 = 0.31 Ma) in carbonates (Stone et al., 1996), and (3) 3He (stable) and 21Ne in olivines and clino-pyroxenes (e.g., Kurz, 1986; Marti and Craig, 1987).
However, the time span and erosion rate range that can be evaluated using a given radioactive cosmogenic nuclide increase with its half-life. Constraining the in situ production rate of 10Be within carbonates can thus significantly extend the range of cosmic-ray exposure ages and erosion rates that can be determined in carbonate-rich environments, given the half-life of this radionu-clide (T1/2 = 1.5 Ma). Furthermore, the ability to measure couples of cosmogenic nuclides with different half-lives in the same mineral phase allows complex exposure histories to be deciphered by differential radioactive decay. This approach, successfully applied using 26Al and 10Be in quartz-rich environments, will have thus the potential to be extended to carbonate-rich environments using 10Be and 36Cl. Similarly, studies using cosmogenic nuclides in mafic environments are limited to the stable isotopes 3He and 21Ne. Developing the ability to measure in situ–produced 10Be in olivines or clinopyroxenes might therefore have important implications on the characterization and dating of burial events over several m.y. in these environments.
However, cosmogenic 10Be is also produced in the atmosphere through cosmic-ray particle reactions with atmospheric 14N and 16O with an average flux of atmospheric 10Be orders of magnitude higher than the integrated rate of in situ 10Be produced in 1 cm2 of surficial rock (Monaghan et al., 1986). Atmospheric 10Be may then be adsorbed on minerals and/or incorporated into weathered mineral zones via superficial circulation of meteoric waters. Reliable measurements of the in situ–produced 10Be concentration within carbonates, weathered olivines, or clinopyroxenes thus require an efficient decontamination procedure of this meteoric 10Be variety. For quartz, chemical cleaning based on hydrofluoric acid (HF) step dissolutions proved to remove efficiently this undesirable meteoric component (Brown et al., 1991). A similar step-dissolution method showed efficiency to decontaminate olivines (Nishiizumi et al., 1990; Shimaoka et al., 2002) but appeared to be unsuitable for clinopyroxenes (Ivy-Ochs et al., 1998). New specific decontamination procedures have been thus developed for removing meteoric 10Be contamination from calcite and altered clinopyroxenes. They were tested on Tortonian calcite from southeastern France (43°N) and on variously altered clinopyroxenes from exposed Quaternary lava flows of Etna volcano (Sicily, 38°N).
Cosmogenic 36Cl measured in the same calcite samples and in situ–produced 10Be measured within associated flints allow validating the cleaning procedure developed on carbonates and proposing an empirical 10Be production rate in CaCO3. Similarly, the clinopyroxenes' 10Be decontamination procedure was assessed both by measuring the cosmogenic 3He concentrations in the same samples and by K-Ar dating of the sampled lava flows. Once the procedure had been validated, empirical 10Be production rates were computed for clinopyroxenes and compared to the modeled production rate (Masarik, 2002).
METEORIC 10Be DECONTAMINATION PROCEDURES
Because the average atmospheric 10Be flux to Earth's surface is orders of magnitude higher than the integrated total rate of in situ 10Be produced in a 1 cm2 column of surficial rock, measurement of the accumulated in situ–produced 10Be concentration requires elimination of the meteoric 10Be. Specific cleaning procedures to remove meteoric 10Be were thus developed on carbonates and variously altered clinopyroxenes.
A cleaning procedure to remove meteoric 10Be was developed on Oligocene carbonates sampled at Limans, France (43°58′N, 5°43′E; altitude 671 m) (LIM02, LIM04, LIM08, and LIM09 in Table 1). X-ray diffraction indicates that all these carbonate samples are pure calcite. Their measured bulk density is 2.61 g cm−3.
To test different cleaning procedures, each sample was crushed, sieved (250–500 µm), and divided into five subsamples. The first group was stirred for 24 h in ultrapure water (18 MΩ-cm), dried, and weighed. This cleaned material was prepared for 10Be measurement by complete dissolution in 15 M HNO3 followed by addition of 300 µg 9Be carrier (Merck 1000 mg/L Be standard) and subsequent purification by solvent extractions and alkaline precipitations (Bourlès, 1988). The second group, after being subjected to the same deionized water treatment, was partially dissolved by gradual addition of sufficient 1.5 M HNO3 to dissolve ∼10% of the calcite. The pH was monitored and maintained below 5 throughout this step. The cleaned calcite was then dried, weighed, and prepared for 10Be measurement as described above. The other three groups were treated similarly, but were subjected to two, three, and four repetitions of the partial dissolution treatment in 1.5 M HNO3.10Be/9Be ratios were measured by accelerator mass spectrometry (AMS) at the Tandétron facility of Gif-sur-Yvette, France. Measured ratios were calibrated directly against the National Institute of Standards and Technology (NIST) standard reference material SRM 4325 using its certified 10Be/9Be ratio of (26.8 ± 1.4) × 10−12. It has been noted that the ratio reported by NIST is incompatible with the ICN standards used at the University of Pennsylvania and at the University of California (Middleton et al., 1993). We have thus normalized our 10Be concentrations by a factor of 1.143 ± 0.039 to make them directly comparable to those based on ICN standards. 10Be uncertainties (1σ) include a 3% contribution conservatively estimated from observed standard variations during the runs, a 1σ statistical uncertainty in the number of 10Be events counted, and the uncertainty of the blank correction (associated 10Be/9Be blank ratio was (2.4 ± 1.2) × 10−15).
The constant 10Be concentration of the stepwise dissolutions, even after the first water washing step, indicates that meteoric 10Be has been removed by the cleaning procedure (Table 1; Fig. 1), even by the mildest treatment (deionized water). These results are in agreement with previous observations in marine environments (Bourlès et al., 1989; Southon et al., 1987) showing that atmospherically produced 10Be is not adsorbed on carbonates.
A cleaning procedure to remove the meteoric 10Be contamination was developed on variously altered clinopyroxenes from surficial samples of exposed Quaternary basaltic flows of Etna volcano (Sicily, 38°N) (see Fig. 2). Two surficial samples were collected on the pahoehoe Nave flow (SI41, 830 m, and SI27a, 1190 m) at the top of plurimetric tumuli (Table 2). SI43, from the Piano Della Lepre site (2070 m) was collected at the top of a pahoehoe flow, today shielded by the emplacement of an overlying flow (Table 2).
In order to decontaminate from meteoric 10Be, which can be hosted both into mineral overgrowth (Friedmann and Weed, 1987) and into secondary clay minerals (Kabata-Pendias and Pendias, 2000; Wedepohl, 1974), we had to deal with a specific drawback linked to the clinopyroxenes crystallographic structure. Ivy-Ochs et al. (1998) indeed highlighted that sequential dissolution with HF of monomineralic grains (0.2–0.5 mm) was unable to efficiently remove this contamination. This result was interpreted as a proof of the meteoric 10Be accumulation within the grains, in micrometric locked pits where weathering generates clay formation. As a consequence, the originality of the proposed cleaning method is to overcome this difficulty through a preliminary crushing step releasing most of the contaminated zones.
Basalt samples from Etna were crushed and sieved, the 0.6–1 mm fraction being processed. Extraction of mafic phenocrysts from the basaltic groundmass was performed by magnetic separation and density partition in heavy liquids. Since olivines and clinopyroxenes have close densities (∼3.3 g cm−3) and magnetic susceptibilities, handpicking under binocular microscope improved the separation of both species. Approximately 25 g of >95%-pure clinopyroxenes (<5% olivines) were isolated for each sample, except for the SI27a sample (65% clinopyroxenes, 35% olivines). These phenocrysts were then crushed in an agate mortar and sieved to keep the fraction smaller than 80 µm.
To optimize the meteoric 10Be removal without dissolving unaltered minerals, lattice powder samples were leached at 95 ± 5 °C during 10 h in a 0.04 M solution of hydroxylammonium chloride (NH2OH.HCl) in 25% acetic acid. This solution dissolves metallic oxides (Bourlès et al., 1989) and allows the grain-adsorbed 10Be to be released in the aqueous phase. After centrifugation, the leachate solution is poured into a Teflon bottle as well as the three rinses performed using the same leaching solution. The remaining solid phase is ovendried (6 h at 100 °C) and precisely weighed (±1 mg) to determine the mass of material eliminated by the leaching step.
Powder is next laden within a 60 mL Teflon jar containing 25 mL of 1 M HCl and shaken for 24 h. After separation of the supernatant by centrifugation, the remaining solid phase is rinsed three times with ultrapure deionized water that is added to the supernatant.
Oven-dried and weighed, the remaining powder sample undergoes five sequential dissolution steps in 4 M HF in a 60 mL Teflon jar shaken during 24 h. After each sequential HF dissolution, ∼10 mL of concentrated 15 M nitric acid is added to dissolve coprecipitated fluorides (MgF2, CaF2). Twenty minutes later the acid floating is extracted, the solid phase is rinsed three times using ultrapure water that is added to the supernatant, and the remaining powder sample is dried and weighed.
The observed mean value of 0.48 ± 0.14 g of clinopyroxenes dissolved per mL of concentrated (48%) HF added is consistent with the stoichiometric value of ∼0.5 g mL−1, suggesting that the dissolving reactions are close to completion (Table 3).
The solutions obtained were spiked with 300 µL of a 1 × 10−3 g g−1 9Be Merck solution for 10Be measurement. Iron is extracted from the aqueous phase by organic separation in diisopropylether. Beryllium is then separated following the chemical extraction procedure described in (Bourlès et al., 1989). 10Be/9Be ratios were also measured at the AMS facility of Gif-sur-Yvette, France. During this run, the associated 10Be/9Be blank value was (4.9 ± 2.8) × 10−15.
The K-Ar dating and cosmogenic 3He (3Hec) data are presented in Table 2. The K-Ar ages are 32 ± 4 ka and 33 ± 2 ka for the Nave flow samples SI27a and SI41, respectively, and the past exposure interval is 10 ± 3 ka for the Piano Della Lepre site (SI43). Scaled for latitude and altitude with the Stone (2000) factors, the 3Hec production rate of 108 ± 10 atoms/g/yr (Masarik and Reedy, 1996) yields 3Hec cosmic-ray exposure ages in agreement with the K-Ar ages for SI41 (Nave bottom) and SI43 (Piano Della Lepre) (Table 2).
However, the 24 ± 3 ka exposure age calculated from 3Hec measured on SI27a (Nave top) is significantly lower than the 32 ± 4 ka K-Ar age of this flow, suggesting that a mean erosion rate of 11 ± 3 m/m.y. has affected this surface since the flow emplacement.
Table 3 presents the 10Be concentrations measured in aqueous phases obtained during the cleaning procedure for SI27a and in the aqueous phases resulting from the last HF dissolution for all the samples. Uncertainties, reported as 1σ on the 10Be concentrations take into account counting statistics, observed external reproducibility of the AMS (3%), and blank correction.
Assuming a 3Hec/10 Bec production ratio of ∼24 (Table 4), the SI27a 3Hec concentration (Table 2) thus implies a theoretical in situ–produced 10Be concentration of (2.54 ± 0.08) × 105 atoms/g. This theoretical value can be used to check the cleaning procedure efficiency.
The leaching step performed on SI27a yields a 10Be concentration of (6.34 ± 0.43) × 107 atoms/g. This concentration is two orders of magnitude higher than the theoretical in situ–produced concentration, evidencing a contamination by atmospherically produced 10Be. The hydroxylamine leaching thus removes a significant part of the meteoric contamination from the crushed clinopyroxenes. The solution resulting from the second cleaning step in HCl has a concentration of (4.3 ± 0.6) × 106 atoms/g, lower than that of the leaching solution but nevertheless higher than the theoretical in situ–produced concentration. Such a 10Be release may result from the dissolution of remaining secondary minerals (oxides, clays) by HCl.
The three analyzed HF dissolution steps agree within 1σ and have 10Be concentrations averaging (2.55 ± 0.29) × 105 atoms/g (weighted mean using relative uncertainties) (Fig. 5). These concentrations agree within uncertainties with the (2.54 ± 0.08) × 105 atoms/g theoretical in situ–produced 10Be concentration. Such an agreement strongly suggests that the atmospherically produced 10Be was efficiently removed and that the concentrations measured from HF dissolutions represent the in situ– produced component. The tested chemical cleaning procedure therefore appears suitable to allow measurement of in situ–produced 10Be in clinopyroxenes.
10Be IN SITU PRODUCTION RATES
To estimate the 10Be in situ production rate in carbonates, coexisting depth profiles of calcite and flint from the vertical southern limb of the Trevaresse anticline were sampled (43°37′N, 5°25′E; altitude 375 m; Tortonian lacustrine series; Chardon and Bellier, 2003) (Fig. 3). The calcite and flint samples were collected along separate vertical profiles less than 1 m apart within a single trench.
In situ–produced 10Be concentrations were measured in both flint and calcite profile samples. The flint samples were decontaminated using sequential HF dissolutions (Brown et al., 1991), and the calcite samples as described above. All solutions obtained after total dissolution of the material remaining after the specific decontamination procedures were spiked with 300 µL of a 1 × 10−3 g g−1 9Be Merck solution for 10Be measurement. Beryllium was then extracted and purified by successive solvent separations and alkaline precipitations (Bourlès et al., 1989).
In addition, in situ–produced 36Cl concentrations were measured in the calcite profile samples. After grinding, leaching, and chemical separation of chlorine by precipitation of silver chloride, the 36Cl and chloride concentrations in the carbonate were determined for all samples by isotope dilution accelerator mass spectrometry at the Lawrence Livermore National Laboratory CAMS (Center for Accelerator Mass Spectrometry) facility. Blanks were two orders of magnitude lower than the samples. Uncertainties of 36Cl concentration include the statistical uncertainties of measurements and the blank correction. 36Cl production rates from calcium of Stone et al. (1998) were used. Those production rates were calculated for our site latitude and altitude using Stone (2000) coefficients.
The 10Be concentrations in the flint (Table 5) and calcite (Table 6) samples from the Trevaresse profiles decrease exponentially with depth, consistent with in situ production by cosmic rays (Fig. 4).
Furthermore, ratios of 10Be to in situ 36Cl concentrations (Table 6) measured within the decontaminated calcite samples are constantStone et al., 1996), this independently suggests that the 10Be concentrations measured in the decontaminated calcite correspond to the in situ–produced 10Be.
The 10Be profile in calcite indicates an apparent attenuation length of ∼325 g/cm2. The profile is not deep enough for quantitative differentiation between fast and slow muons. Nevertheless, at steady state under the erosional conditions of the site, this apparent attenuation length may correspond to various combinations of both types of muon contributions within ranges of 0%–6% for the slow muons and of 0%–2% for the fast muons.
Using the well-constrained local erosion rate, and assuming contributions for fast and slow muons that yield an apparent attenuation length of ∼325 g/cm2, equation 1 can be solved for the production rate. This yields a 10Be production within calcite of 50.5 ± 8.0 atoms/g/yr, 6.3 ± 1.3 times higher than that of 7.9 ± 1.2 atoms/g/yr within flint based on the polynomials of Stone (2000). This corresponds to a value of 37.9 ± 6.0 atoms/g/yr (sea level and high latitude). The calculated uncertainty includes propagation of a 6% uncertainty on the 10Be production rate used to estimate the erosion rate within flint and the analytical uncertainties (∼5%) of 10Be in calcite.
Elemental compositions of minerals play a major role in inducing higher in situ 10Be production in calcite relative to quartz. In quartz, the targets responsible for 10Be production are O, through the reactions 16O(n, 4p3n)10Be and 16O(µ−, αpn)10Be, and Si, through the reactions 28Si(n, x)10Be and 28Si(µ−, x)10Be. Relative elemental production rates can be estimated by dividing the product of the production cross section and the mass fraction of the relevant element by its atomic weight. For example, in SiO2, O represents ∼53.3% of the mass, so ∼92% of 10Be in SiO2 is produced from O (Leya et al., 2000b). In calcite, O represents ∼47.5% of the mass, so production by O spallation should produce almost the same amount of 10Be as in SiO2. Another target must be considered. As spallogenic reactions favor product masses that are either slightly less than the target or much less, such as protons and neutrons (Gosse and Phillips, 2001), the most favorable target in carbonate is C.
The occurrence of 10Be production from carbon has been demonstrated in meteorites (Nagai et al., 1993) and studied in diamonds (Lal et al., 1987). 10Be is produced from C by protons mainly via the reaction 12C(p, 3p)10Be. Secondary neutrons produce 10Be from carbon via the reaction 12C(n, 2pn)10Be. To determine production of nuclides by galactic cosmic ray (GCR) or solar cosmic ray (SCR) particles, models based on physical approaches have been developed (Leya et al., 2000a; Masarik and Reedy, 1995; Michel et al., 1995). Amounts of produced nuclides or production cross sections (Leya et al., 2000b; Raisbeck and Yiou, 1977) have also been estimated after irradiation of pure or composite targets. These studies indicate that, for particles having energy higher than 500 MeV, production of 10Be from C is 2–4 times higher than that from O (Leya et al., 2000a; Nagai et al., 1990).
More recently, from experimental cross sections in natural carbon targets, Kim et al. (2002) reported that the production rate of 10Be in carbon at 40 MeV is 22 times higher than that in oxygen, while, in agreement with studies in meteorites, it is only ∼2 times higher at 400 MeV. As the energy of the particles (mainly secondary neutrons) that impinge on Earth's surface is lower than 500 MeV (Lal, 1958, 1988), this altogether may well explain the ratio found in the Trevaresse samples. By compiling production cross sections from Michel et al. (1995), Sisterson et al. (1997), Leya et al. (2000b), and Kim et al. (2002), the production ratio of carbon over oxygen can be estimated for spallation by protons with energies ranging from 30 to 600 MeV. This indicates systematically higher production of 10Be from C than O, the highest production ratio values being associated with low energies.
Production ratios can be evaluated by integrating the excitation functions for oxygen and carbon (Kim et al., 2002) over the ground-level energy spectra for neutrons and protons (Ashton et al., 1971; Lal, 1958, 1988). This can be expressed by the following equation:
We propose two potential sources for this inconsistency. A significantly lower threshold for 10Be production by C spallation than that for O spallation, coupled with the high ground-level particle fluxes at low energies (i.e., <25 MeV where few production cross sections data are available), could lead to significantly higher production in calcite than we calculate above. Secondly, experimental cross sections have generally been determined for protons, whereas neutrons dominate ground-level production of cosmogenic nuclides. The fact that we consider cross sections for neutrons equal to those for protons may underestimate production of 10Be from both C and O. Basic theoretical and statistical principles indicate that the reaction 16O(n, 3n4p)10Be should have higher reaction cross sections than 16O(p, 2n5p)10Be and that this effect is more substantial for 12C(n, n2p)10Be compared to 12C(p, 3p)10Be.
Similarly, measuring cosmogenic nuclide concentrations in independently dated lava flows for which prior exposure to cosmic rays can be excluded (N10Be (t0) = 0 atoms/g) allows calculating empirical production rates. Equation 3 (Lal, 1991) indeed describes the evolution of the surficial in situ–produced 10Be concentration, N10Be (atoms/g), as a function of the exposure duration, t (yr); of the surficial in situ 10Be production rate at the sampling site, P 10 (atoms/g/yr); and of the erosion rate, ϵ (cm yr−1):
If erosion is negligible (ρϵ/Λ << λ), P10 can be calculated knowing only the exposure age, t, and the 10Be concentration, N10Be (t), at the surface. For a lava flow, the exposure age, t, can be assumed to be the same as the K-Ar age of the eruption.
If the flow surface denudation has been significant (ϵ > 1 m/ m.y. for flows younger than 50 ka), erosion has to be determined independently to avoid underestimating the production rates. This can be done by measuring in the same sample the concentration of the stable cosmogenic nuclide 3Hec, which is described byMasarik and Reedy, 1996), equations 3 and 4 can be combined to eliminate the erosion term ϵ and, therefore, to calculate P10.
Both approaches were used to calculate empirical in situ 10Be production rates (P10). For this, the measured in situ 10Be concentrations (Table 3) were first adjusted for their sampling depth assuming an attenuation length of 150 g cm−2 in basalt (Gosse and Phillips, 2001). This correction is less than 15% for the thickest sample (SI41, Nave bottom, 15 cm thick). Topographic mask corrections are negligible. The computed rates were finally adjusted from sampling latitudes and elevations to SLHL applying the Stone (2000) scaling factors (Table 3). Despite various clinopyroxenes' weathering degrees, all the so-calculated empirical values agree within the 1σ level (Table 3), which strongly supports the efficiency of the decontamination procedure. Using the calculation based only on the 10Be and K- Ar data, calculated SLHL P10 values range from 2.2 ± 0.6 to 3.5 ± 0.6 atoms/g/yr. When the erosion term is eliminated using the 3Hec data, calculated SLHL P10 values then range from 2.3 ± 0.6 to 4.6 ± 0.9 atoms/g/yr. Means weighed to relative uncertainties are 2.7 ± 0.5 and 3.1 ± 0.8 atoms/g/yr, respectively. The effect of erosion on this production rate calibration is thus limited, because underestimates of erosion remain inside the 1σ uncertainty for all the samples.
Table 4 presents the computed SLHL P10 using the physical model developed by Masarik (2002). The values of the SLHL P10 modeled from the SI41, SI27a, and SI43 clinopyroxene (diopside) compositions, 4.58, 4.58, and 4.60 atoms/g/yr, respectively, are similar. The uncertainties attached to these simulated rates are assumed to be ∼10%, leading to a mean (1σ) of 4.6 ± 0.5 atoms/g/yr. Therefore, these simulated values are slightly higher than the weighed mean of the empirical values, 3.1 ± 0.8 atoms/ g/yr, but both values nevertheless remain statistically compatible within uncertainties.
Simulated SLHL production rates are, however, valid for the present cosmogenic neutron flux. Discrepancies between simulated and calibrated rates may thus result from the past geomagnetic fluctuations as the natural calibration sites are affected by the so-induced neutron flux variations over the exposure duration. Virtual dipole moment (VDM) variations from Carcaillet et al. (2004) database and dipole position fluctuations from Lanza and Zanella (2003) were used to calculate the past neutron flux according to the equations given in (Dunai, 2001). This signal was thus integrated over the appropriate exposure interval to compute the geomagnetic corrections for each studied flow. This correction has to be considered only for the production rate estimate based on equation 3. Indeed, the calculation based on both equations 3 and 4 necessarily integrates magnetic variations through the 3Hec data input. Nevertheless, given the considered exposure durations (last 33 ± 2 k.y. for the Nave flow and from 20 ± 1 to 10 ± 3 ka for the Piano Della Lepre site) at sampling latitudes and altitudes, geomagnetic variations do not induce shifts larger than 3%. Such a limited effect can therefore be considered as insignificant given the uncertainty attached to the empirical SLHL P10.
Finally, the data published in (Nishiizumi et al., 1990) may also be used to compute empirical SLHL P10 in olivines (Forsterite80). The authors indeed measured both 3Hec and 10Be in olivines from eroded Maui lavas erupted ca. 500 ka. Combining equations 3 and 4, and using a SLHL P3 of 110 ± 10 atoms/g/yr from Masarik and Reedy (1996) scaled with the Stone (2000) factors, these data yield a SLHL P10 of 3.2 ± 1.0 atoms/g/yr for olivines. This SLHL P10 computed for olivines is similar to that determined for clinopyroxenes in this study, suggesting that the compositional differences between olivines Fo80 and diopside clinopyroxenes do not induce significant in situ 10Be production rate discrepancies.
However, further studies could be done in order to strengthen the evidence that in situ 10Be can be measured accurately in mafic phenocrysts. These investigations should particularly focus on (1) the potential loss of in situ 10Be enhanced by the preliminary crushing step and (2) the in situ 10Be overestimation due to the trapping of atmospheric 10Be from the magmatic chamber.
The new chemical procedures proposed in this study to decontaminate carbonates and clinopyroxenes from atmospheric 10Be have been tested on calcite from a depth profile sampled in the Tortonian lacustrine series from the Trevaresse anticline (France, 43°N), and on K-Ar-dated basaltic flows of Etna volcano (Sicily, 38°N), respectively.
Measurement of 10Be concentrations within both flint and calcite from two depth profiles sampled below the same erosional surface and measurement of 36Cl concentrations within the same calcite samples demonstrate that the cleaning procedure developed to remove meteoric 10Be contamination from calcite is efficient. The effective attenuation length along the calcite profile implies that production induced by muons plays a significant role in the in situ 10Be production in calcite. The in situ 10Be production rate in calcite thus computed at the sampling site is 50.5 ± 7.5 atoms/g/yr, 6.3 ± 0.3 times higher than that in flint. This corresponds to an in situ 10Be production rate in calcite at sea level and high latitude of 37.9 ± 6.0 atoms/g/yr. Leading to a production ratio significantly higher than can be explained by existing production cross sections, this work emphasizes the need for additional measurements at low energies.
The efficiency of the chemical procedure proposed to decontaminate weathered clinopyroxenes from atmospheric 10Be is attested to not only by the 10Be concentrations reaching a plateau during final HF dissolutions, but also by a strong agreement between the measured 10Be concentrations and those inferred from theoretical production rates (Masarik, 2002). The key step of the proposed chemical decontamination procedure relies on a preliminary crushing step of the pyroxenes (<80 µm) performed to maximize the release of the contaminated zones. To limit the undesirable dissolution of clinopyroxenes during the decontamination steps, the crushing step is followed by leachings of the sample in hydroxylamine and HCl solutions.
The combined 10Be and 3Hec data obtained from K-Ar-dated basaltic flows of Etna volcano allowed in situ 10Be production rates to be calculated for clinopyroxenes. The yielded sea-level high-latitude mean in situ 10Be production rate of 3.1 ± 0.8 atoms/ g/yr is similar, within uncertainties, not only to simulated values (Masarik, 2002), but also to the other empirical values.