Experimental determination of the xenon isotopic fractionation during adsorption

Authors

  • Yves Marrocchi,

    Corresponding author
    1. Université de Lorraine, CRPG, UMR 7358, Vandoeuvre les Nancy, France
    2. CNRS, CRPG UMR 7358, Vandoeuvre les Nancy, France
    • Corresponding author: Y. Marrocchi, CNRS, CRPG UMR 7358, 15 rue Notre Dame des Pauvres, BP 20, Vandoeuvre les Nancy, FR-54501, France. (yvesm@crpg.cnrs-nancy.fr)

    Search for more papers by this author
  • Bernard Marty

    1. Université de Lorraine, CRPG, UMR 7358, Vandoeuvre les Nancy, France
    2. CNRS, CRPG UMR 7358, Vandoeuvre les Nancy, France
    Search for more papers by this author

Abstract

[1] Adsorption of noble gases into solids is often posited to account for their abundance patterns in meteorites, terrestrial rocks, and planetary atmospheres. Since these elements present isotope variations among geochemical reservoirs, we have experimentally tested the possibility that adsorption of neutral noble gases could result in isotopic fractionation. Our experiment consists of a cycle of adsorption/desorption processes in which xenon is progressively lost from a reservoir by equilibrium adsorption on kerogen and on montmorillonite. Any isotopic fractionation would then be amplified by the Rayleigh-like distillation experiment. The fractionation factors α are extrapolated to be −0.18‰ ± 0.08‰ per u and −0.22‰ ± 0.07‰ per u (both at the 2σ level) for kerogen and montmorillonite, respectively. Thus, adsorption of neutral noble gases alone cannot account for the specific isotopic composition of noble gases trapped in meteorites, nor for the isotopic composition of xenon in the terrestrial and martian atmospheres.

1 Introduction

[2] Among the physical processes that drive the noble gas behavior, adsorption is a retention mechanism by which the noble gases interact with solid surfaces via van der Walls attractive force arising from mutual electric dipole induction [Ozima and Podosek, 2002]. Such a process is dominant below a threshold energy corresponding to the energy at which a noble gas atom or ion starts to penetrate the first atomic layer (≈ 50 eV) [Smentkowski, 2000]. Adsorption tends to be more efficient for the heavy noble gases than for the light ones, due to the higher polarizabilities of the former. As a result, adsorption is frequently posited as a cause of noble gas captures in solids, in environments as different as the Earth, the Moon, and the accretion disk that surrounded the early Sun 4.56 Ga ago [Ozima and Podosek, 2002]. It has been proposed that adsorption could have led to the establishment of: (i) the noble gas abundance observed in lunar breccias (especially the enrichment of heavy noble gases relative to the light ones normalized to the solar composition) [Bernatowicz and Podosek, 1991; Leich and Niemeyer, 1975; Lightner and Marti, 1974; Niederman and Eugster, 1992; Niemeyer and Leich, 1976], (ii) the Earth's atmosphere composition (i.e., the depletion of atmospheric xenon as the result of Xe preferential adsorption on sediments) [Bernatowicz et al., 1984; Podosek et al., 1981], and (iii) the noble gas signature measured in insoluble organic meteoritic matter (i.e., the so-called phase Q, also enriched in the heavy noble gases relative to light ones when compared to the solar composition) [Fanale and Cannon, 1972; Marrocchi et al., 2005a, 2005b, 2011; Wacker et al., 1985; Wacker, 1989; Yang and Anders, 1982a, 1982b; Yang et al., 1982; Zadnik et al., 1985]. However, variations of the noble gas elemental patterns among solar system reservoirs and objects are often accompanied by significant isotope differences relative to the inferred original compositions (Figure 1). For instance, meteoritic noble gases differ from the solar ones by enrichments in the heavy isotopes of 1% or more per u (here u: atomic mass unit; Figure 1). Likewise, terrestrial xenon differs from both chondritic and solar compositions by 3%–4% per u enrichments in the heavy Xe isotopes relative to light ones (Figure 1). Trapping of ionized noble gases in solids is known to result in significant isotopic fractionation [Bernatowicz and Hagee, 1987; Frick et al., 1979; Hohenberg et al., 2002; Marrocchi et al., 2011; Ponganis et al., 1997], and ionized xenon is more efficiently trapped than neutral xenon [Marrocchi et al., 2011]. Such a mechanism induces significant noble gas isotopic fractionation but may be limited to specific ionizing environments (e.g., accretion disk and planetary atmospheres). In contrast, Xe isotopic fractionation during adsorption of neutral Xe has not been detected so far [Bernatowicz et al., 1984; Bernatowicz and Podosek, 1986; Marrocchi et al., 2011; Podosek et al., 1981]. However, any limited isotopic fractionation could become a significant effect in the residual gas reservoir if amplified by distillation. Ozima et al. [1998] proposed that some form of Rayleigh distillation could have generated the mass-dependent elemental and isotopic fractionation observed in phase Q relative to Solar. To investigate this possibility, we have developed a distillation experiment aimed at measuring at the subpermil level Xe isotopic fractionation during adsorption of neutral xenon.

Figure 1.

Xenon isotopic composition of phase Q [Busemann et al., 2000] and the Earth's atmosphere [Ozima and Podosek, 2002] normalized to Solar Xe [Wieler and Baur, 1994]. Xe-Q is enriched in favor of the heavy isotopes by a factor of +1.3% per u relative to Solar while atmospheric xenon is even more fractionated relative to Solar (i.e., 3.5% per u).

2 Experimental Setup

[3] Adsorption experiments were performed on terrestrial type III kerogen because it is considered to be the best chemical analog of phase Q studied so far [Gardinier et al., 2000] and on montmorillonite, an analog of oceanic sediments. Because these two samples are characterized by contrasting surface chemistry and area (4 m2/g for kerogen and 80 m2/g for montmorillonite, respectively) [Marrocchi et al., 2005b], the experiments were performed on different amounts of material (3.6 g for kerogen and 1.3 g for montmorillonite, respectively).

[4] The experimental device developed at Centre de Recherches Pétrographiques et Géochimiques (CRPG) consists of an electropolished stainless steel volume of 10 L (volume A; Figure 2) that is connected to a quartz finger of 3 cm3 (volume B; Figure 2). Both volumes are isolated from each other by automatic valves and are connected to a turbo-molecular pump in order to reach the high vacuum required for carrying out the experiments (Figure 2). The sample to be investigated was weighted and placed within the quartz finger (Figure 2). The whole system was baked at 373 K for 48 h in order to reach a residual pressure of 10−8 mbar. The system was then left to cool down to room temperature, volumes A and B were isolated from each other by closing valve 2 (Figure 2), and xenon was introduced at a pressure of ≈ 0.1 mbar in volume A from a manufactured xenon gas bottle (purity of 99.96%). Volume A was then pumped out down to a xenon pressure of around 5 × 10−5 mbar, to provide the starting gas of the experiment. Given that this procedure might induce significant isotopic fractionation, we measured the xenon isotopic composition of this starting gas by taking an aliquot (0.1 cm3; Figure 2) of xenon at a pressure of 5 × 10−5 mbar from volume A. The aliquot was first expanded in a dilution line, purified using Ti-Zr getters to remove active gases (10 min at 800°C, 10 min at room temperature), and then analyzed by noble gas static mass spectrometry [Marrocchi et al., 2005a]. The xenon isotopic composition of three aliquots separated from volume A shows a mass-dependent fractionated isotopic composition of 1.43% and 1.68% per u relative to air Xe for kerogen and montmorillonite experiments, respectively. The three aliquots present a good reproducibility that allows the composition of the starting gas to be well known (Tables 1 and 2).

Figure 2.

Schematic diagram of the Rayleigh-type distillation apparatus used in this work. It consists of an electropolished stainless steel volume of 10 L (volume A) connected to a quartz finger of 3 cm3 (volume B). The experiment consists in performing adsorption cycles by allowing xenon equilibration between volume A and the sample located within the quartz finger. The isotopic composition of volume A was measured sequentially after several adsorption steps. If any isotopic fractionation is induced by adsorption, it should be visible in the evolution of the isotopic composition of xenon in the volume A as a function of the number of cycles.

Table 1. Evolution of the Amount of 130Xe Absorbed and the Xenon Isotopic Ratio as a Function of the Number of Adsorption Cycle Performed on Kerogena
# Cycles130Xe (moles)% Adsorbed124Xe/130Xe126Xe/130Xe128Xe/130Xe129Xe/130Xe131Xe/130Xe132Xe/130Xe134Xe/130Xe136Xe/130Xe
  1. a

    Air composition data are from Ozima and Podosek [2002].

  2. b

    The value for the starting gas (corresponding to the 0 cycle line) represents the average of three measurements. Isotopic ratio × 10.

Air  0.233 ± 0.00080.218 ± 0.00114.715 ± 0.00764.96 ± 0.0652.13 ± 0.0666.07 ± 0.0525.63 ± 0.0421.76 ± 0.02
0b1.12 × 10−900.215 ± 0.0020.208 ± 0.0024.543 ± 0.00663.54 ± 0.0252.62 ± 0.0367.79 ± 0.0227.13 ± 0.0423.84 ± 0.05
35.49 × 10−1050.920.217 ± 0.0040.209 ± 0.0044.561 ± 0.00863.52 ± 0.0352.81 ± 0.0467.56 ± 0.0327.26 ± 0.0523.86 ± 0.06
53.22 × 10−1071.260.209 ± 0.0040.211 ± 0.0044.582 ± 0.00863.61 ± 0.0352.77 ± 0.0367.81 ± 0.0327.19 ± 0.0623.82 ± 0.06
101.32 × 10−1088.220.213 ± 0.0040.208 ± 0.0044.551 ± 0.00863.73 ± 0.0352.91 ± 0.0367.78 ± 0.0327.09 ± 0.0623.88 ± 0.06
127.71 × 10−1193.110.217 ± 0.0040.213 ± 0.0054.536 ± 0.00963.42 ± 0.0352.73 ± 0.0367.70 ± 0.0327.17 ± 0.0623.81 ± 0.06
141.73 × 10−1198.450.214 ± 0.0050.209 ± 0.0054.516 ± 0.00963.59 ± 0.0352.84 ± 0.0467.65 ± 0.0327.22 ± 0.0623.86 ± 0.07
161.04 × 10−1299.90.211 ± 0.0050.212 ± 0.0054.584 ± 0.00963.74 ± 0.0452.66 ± 0.0467.59 ± 0.0427.06 ± 0.0623.86 ± 0.07
Table 2. Evolution of the Amount of 130Xe Absorbed and the Xenon Isotopic Ratio as a Function of the Number of Adsorption Cycle Performed on Montmorillonitea
# Cycles130Xe (moles)% Adsorbed124Xe/130Xe126Xe/130Xe128Xe/130Xe129Xe/130Xe131Xe/130Xe132Xe/130Xe134Xe/130Xe136Xe/130Xe
  1. a

    See the Table 1 footnotes for further explanation.

Air  0.233 ± 0.00080.218 ± 0.00114.715 ± 0.00764.96 ± 0.0652.13 ± 0.0666.07 ± 0.0525.63 ± 0.0421.76 ± 0.02
01.23 × 10−900.210 ± 0.0020.201 ± 0.0034.574 ± 0.00663.54 ± 0.0252.74 ± 0.0367.69 ± 0.0227.31 ± 0.0423.90 ± 0.05
109.91 × 10−1019.390.214 ± 0.0040.209 ± 0.0044.570 ± 0.00863.52 ± 0.0352.61 ± 0.0467.81 ± 0.0327.19 ± 0.0523.86 ± 0.06
407.49 × 10−1039.040.217 ± 0.0040.205 ± 0.0044.559 ± 0.00863.61 ± 0.0352.75 ± 0.0367.59 ± 0.0327.30 ± 0.0523.90 ± 0.06
753.56 × 10−1071.010.211 ± 0.0040.223 ± 0.0044.574 ± 0.00863.73 ± 0.0352.61 ± 0.0367.66 ± 0.0327.36 ± 0.0523.90 ± 0.06
1001.15 × 10−1090.630.2215 ± 0.0040.215 ± 0.0044.556 ± 0.00863.42 ± 0.0352.90 ± 0.0367.51 ± 0.0327.37 ± 0.0523.89 ± 0.06
1206.58 × 10−1194.650.209 ± 0.0040.207 ± 0.0044.579 ± 0.00863.59 ± 0.0352.59 ± 0.0367.64 ± 0.0327.32 ± 0.0523.89 ± 0.06
1404.36 × 10−1196.450.211 ± 0.0040.211 ± 0.0044.553 ± 0.00863.74 ± 0.0352.76 ± 0.0467.56 ± 0.0327.36 ± 0.0523.90 ± 0.06
1602.89 × 10−1197.650.207 ± 0.0050.226 ± 0.0054.597 ± 0.00963.74 ± 0.0352.64 ± 0.0467.71 ± 0.0327.31 ± 0.0623.84 ± 0.06
1801.28 × 10−1198.960.202 ± 0.0050.214 ± 0.0054.579 ± 0.00963.74 ± 0.0452.81 ± 0.0467.82 ± 0.0427.40 ± 0.0623.90 ± 0.07
2004.42 × 10−1299.640.215 ± 0.0050.208 ± 0.0054.582 ± 0.00963.74 ± 0.0452.70 ± 0.0467.74 ± 0.0427.30 ± 0.0623.93 ± 0.07

[5] The experiment consisted in performing adsorption cycles in order to mimic a distillation. Xenon from volume A was expanded in static mode in volume B and therefore put in contact with the sample located in the quartz finger. Two temperature conditions were used. The kerogen sample was cooled down to 196 K using dry ice, whereas the montmorillonite sample was left at room temperature. The sample was equilibrated with the gas for 15 min, then the larger and smaller volumes were isolated by means of valve 2  (Figure 2). Then, an aliquot was taken off from volume A, and the xenon isotopic composition was measured following the same procedure as the starting gas. Before each new cycle, the sample was heated to 200°C under pumping from volume B for 5 min (with valve 2 still closed and valve 1 open) in order to degas residual xenon from the absorbant.

[6] The adsorption cycle was repeated several times in order to adsorb and then pump off most of the initial xenon. This process of repeated removal of xenon by adsorption is inferred to represent a Rayleigh distillation. Depending on the specific surface of the samples and the experimental conditions, between 16 and 200 cycles of adsorption were required to adsorb more than 99% of xenon. The isotopic composition of volume A was measured regularly after several adsorption steps. If any isotopic fractionation was induced by adsorption, it should be seen in the evolution of the isotopic composition of xenon in volume A as a function of the number of cycles. It is important to note that the dead volume (i.e., the residual gas that is not adsorbed) located in the quartz tube corresponds to less than 0.03% of the amount adsorbed and therefore has a negligible influence on the experiments.

3 Results

[7] Kerogen—As the specific surface area of this sample is relatively low (i.e., 4 m2 g−1) [Marrocchi et al., 2005b], adsorption experiments were performed at 196 K using dry ice. At this temperature, the amount of xenon adsorbed is rather important for the first cycles (e.g., 50% of xenon adsorbed after three cycles; Table 1). The evolution of the amount of adsorbed xenon as a function of the number of cycles follows a logarithmic relationship until > 99% of xenon is adsorbed after 16 cycles (Figure 3a). The isotopic composition of xenon measured after 3, 5, 10, 12, 14, and 16 cycles does not show any detectable isotopic variation (Figure 3a; Table 1). The maximum isotopic fractionation factor can be calculated from this data. In the case of a Rayleigh distillation, the evolution of the isotopic ratios is linked to the remaining fraction of gas in the system as described by the following equation:

display math(1)

where Ri stands for the initial isotopic ratio, Rf is the final isotopic ratio (i.e., after 3, 5, 10, 12, 14, and 16 cycles), f is the fraction of xenon remaining in the system, and α is the fractionation factor. In order to calculate the maximum fractionation factor, we considered the 136Xe/130Xe isotopic ratio because it presents the larger mass difference with respect to the normalizing isotope (conventionally, 130Xe at CRPG Nancy) among the other xenon isotopic ratios. Equation (1) can be rewritten as

display math(2)
Figure 3.

(a) Evolution of the amount of 130Xe adsorbed on kerogen versus the 136Xe/130Xe isotopic ratio in volume A as a function of the number of cycles (isotopic ratio × 10). (b) Evolution of the 136Xe/130Xe isotopic ratio as a function of the remaining gas fraction (f) in the system during the experiment of xenon adsorption on kerogen. The slope permits the determination of the fractionation factor α (see text).

[8] Thus, in a diagram of ln(Rf/Ri) versus ln(f), the slope of the correlation (if any) permits the computation of (α − 1) directly and therefore the fractionation factor α. For kerogen, the maximum fractionation factor for the 136Xe/130Xe ratio was calculated to be −1.08‰ ± 0.56‰ (2σ) (Figure 3b). Hence, considering a difference of 1 u, the maximum fractionation factor is determined to be: αkerogen= −0.18‰ ± 0.08‰ per u (2σ).

[9] Montmorillonite—In order to test the influence of the structural parameter of the sample on the isotopic fractionation, we performed an experiment on synthetic montmorillonite [Marrocchi et al., 2005b]. This sample presents a higher specific surface area (80 m2/g−1) than kerogen and, more importantly, has an important concentration of micropores that may favor an isotopic fractionation of xenon. As the adsorption cycles were performed at room temperature, 200 cycles were required to adsorb more than 99% of the starting xenon. However, the results were the same as those observed with kerogen: (i) the evolution of the amount of xenon adsorbed follows a logarithmic relationship (Figure 4a and Table 2) and (ii) no isotopic fractionation was measured (Figure 4a and Table 2). Following the same approach than for kerogen, the maximum isotopic fractionation factor determined for montmorillonite is −1.32‰ ± 0.48‰ (2σ) (Figure 4b), which yields a αmontmorillonite value of −0.22‰ ± 0.07‰ per u (2σ).

Figure 4.

(a) Evolution of the amount of 130Xe adsorbed on montmorillonite versus the 136Xe/130Xe isotopic ratio in volume A as a function of the number of cycles (isotopic ratio × 10). (b) Evolution of the 136Xe/130Xe isotopic ratio as a function of the remaining gas fraction (f) in the system during the experiment of xenon adsorption on montmorillonite. The slope permits the determination the fractionation factor α (see text).

4 Discussion

[10] The main carrier of meteoritic heavy noble gases, known as phase Q, is formed of carbonaceous material not fully characterized, obtained by acid attack of chondrites by HF/HCl [Busemann et al., 2000; Lewis et al., 1975]. Q-gases are characterized by several specific properties [Busemann et al., 2000], especially: (i) an elemental fractionation of noble gas relative to solar wind; light noble gases being depleted relative to heavy ones and (ii) a mass-dependent isotopic fractionation relative to solar wind of 1.3% per u in favor of the heavy isotopes relative to the light ones. Over three decades, several adsorption experiments succeeded in reproducing the Q-like elemental abundance pattern [Fanale and Cannon, 1972; Marrocchi et al., 2005b; Wacker et al., 1985; Wacker, 1989; Yang and Anders, 1982a, 1982b; Yang et al., 1982]. These results lead adsorption to be frequently invoked as a plausible process at the origin of meteoritic noble gases [Marrocchi et al., 2005b; Wacker, 1989]. However, any model that derives Q-gases from solar nebula gases should reproduce, in addition to the elementary pattern, the isotopic noble gas fractionation observed in phase Q relative to Solar. Our results clearly do not support equilibrium adsorption as an isotopically fractionating process for noble gases, even when amplified by a Rayleigh distillation. Thus, the isotopic difference of +1.30% ± 0.06% per u observed between Xe-Q and Solar Xe [Busemann et al., 2000; Wieler and Baur, 1994] is unlikely to have been generated by adsorption of neutral xenon onto carbonaceous surfaces in the early solar system. To the contrary, our data confirm previous experiments that pointed out the dominant role of noble gas ions during trapping processes in generating Q-gas isotopic characteristics [Marrocchi et al., 2011]. Thus, the Q-gas properties imply a physical environment allowing efficient ionizing of gaseous species, probably at the periphery of the solar system that underwent UV light ionizing conditions from massive O/B stars or from the Sun itself at the surface of the disk [Bally et al., 2005].

[11] Xenon in the terrestrial and martian atmospheres is under-abundant compared to other noble gases and normalized to Chondritic, by a factor of about 20 relative to krypton. Because xenon has the largest adsorption capability among noble gases [Ozima and Podosek, 2002], its atmospheric depletion could be related to adsorption onto surface material with high specific surface areas such as ice or sediments. This possibility, however, faces a mass balance problem because there are not enough ices and sedimentary rocks at the Earth's surface to adsorb the xenon presently missing in the atmosphere [Bernatowicz et al., 1982, 1984; Podosek et al., 1981]. The present data also dismiss this possibility: atmospheric Xe is isotopically fractionated by 3%–4% per u relative to Chondritic, or Solar, which is 2 orders of magnitude larger than the maximum isotopic fractionation of −0.02% per u determined in this study. Assuming a Rayleigh distillation, for instance if surface sediments having adsorbed xenon would be continuously recycled into the mantle, does not solve the problem. In equation (1), the fraction f corresponding to a 20-fold depletion is 0.05 and the ratio Rf/Ri corresponding to a 3.5% per u is 1.035. Therefore, the fractionation factor α becomes 0.988, corresponding to a Xe isotope fractionation of −1.2% per u, still 2 orders of magnitude higher than the limiting value determined in this work. We conclude that isotopic fractionation during adsorption, if any, cannot account for the isotopic composition of atmospheric xenon. Other processes, like Xe preferential ionization in the upper atmosphere [Pujol et al., 2011] are required to account for its specific isotopic composition.

5 Concluding Remarks

[12] We have estimated the maximum fractionation factor of xenon isotopes during adsorption onto kerogen and montmorillonite to be −0.2‰ ± 0.1‰ per u. Hence, adsorption of neutral noble gases cannot account for the specific isotopic composition of noble gases trapped in meteorites, nor for the unique isotopic composition of xenon in the terrestrial and martian atmospheres, which both differ from their presumed source compositions by one or several percents. As a consequence, any adsorption model that requires a noble gas isotopic fractionation relative to a starting gas should appeal to another mechanism, probably ionization, in order to generate a mass-dependent isotopic fractionation.

Acknowledgments

[13] Laurent Zimmerman and Laurent Michot are thanked for assistance during sample preparation and analysis. This study was supported by grants from the Programme National de Planétologie (INSU-CNRS) and from the Agence Nationale de la Recherche (ANR grant T-Tauri). This is CRPG-CNRS contribution 2250.

[14] The Editor thanks two anonymous reviewers for their assistance evaluating this paper.

Ancillary