Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Determination of the Present-Day Atmospheric Noble Gas Compositions

1.1 Absolute Determinations of the Atmospheric 3He/4He

The terrestrial air contains He with a concentration of 5.24 ppm by volume (Gluckauf 1946; Gluckauf and Paneth 1946), considerably lower than would be predicted from solid Earth degassing of radiogenic 4He, because He escapes from the high temperature exosphere to inter-planetary space as a result of its light mass (Kockarts and Nicolet 1962) Helium is nevertheless supplied to the atmosphere by the solid Earth through volcanic activity (Craig et al. 1975), fault movements, erosion and groundwater circulation (Sano et al. 1998). The helium content and isotopic composition of the atmosphere are therefore the result of a balance between degassing from the mantle and crust, and escape of He isotopes to space. Since the residence time of He in the atmosphere is approximately 106 years (Torgersen 1989) and is significantly longer than the mixing time in the atmosphere of a few years (Tans et al. 1989), the air 3He/4He ratio is believed to be constant on a global scale (Lupton 1983). On this basis most laboratories engaged in helium isotope measurements use atmospheric helium as a natural isotopic standard (Lupton and Graham 1991). The 3He/4He ratios of terrestrial samples are usually expressed as a unit of Ra where Ra is the atmospheric 3He/4He ratio.

Aldrich and Nier (1948) were the first to measure the atmospheric 3He/4He ratio although their measurements were not calibrated against an absolute standard (i.e. pure 3He and 4He were not available); Coon (1949) also reported 3He/4He ratios of atmospheric and natural gases using neutron activation analysis where 3He was detected by counting ionization pulses arising from the disintegration products of the reaction 3He(n,p)3H induced by thermal neutrons. Their measured atmospheric 3He/4He ratio was (1.3 ± 0.2) × 10−6, which agreed well with the value of (1.2 ± 0.2) × 10−6, reported by (Aldrich and Nier 1948).

There are only three absolute 3He/4He measurements of atmospheric helium in the literature. In 1970, Mamyrin et al. (1970) reported that the atmospheric 3He/4He ratio in St. Petersburg, Russia, was (1.399 ± 0.013) × 10−6 (1σ) and that air helium ratio did not vary with latitude, longitude, and altitude (up to 10 km) within the error margins of the individual measurements of about 2 ~ 3 %. Clarke et al. (1976) re-determined the atmospheric 3He/4He ratio in Ontario, Canada. The value of (1.384 ± 0.006) × 10−6 (1σ) was somewhat lower but in agreement, within error, of the St Petersburg value. The average of the two measurements, 1.39 × 10−6 was accepted as the air 3He/4He ratio until the end of 1980s when Sano et al. (1988) determined an atmospheric 3He/4He ratio of (1.343 ± 0.013) × 10−6 (1σ) at Ueno Pak, central Tokyo, Japan. This value was apparently lower than the previous data. Either experimental artifacts and/or natural phenomenon accounted for the difference. The International Union of Pure and Applied Chemists (IUPAC) committee currently uses Sano et al’s value (1.343 ± 0.013) × 10−6). In 1990, Davidson and Emerson (1990) measured the absolute 3He content of air by mass spectrometry. When combined with an earlier measurement of the atmospheric 4He abundance (Holland and Emerson 1987), this provides an additional estimate of the atmospheric 3He/4He ratio of (1.393 ± 0.014) × 10−6 (1σ). The average of the above determinations taking into their weighted errors is (1.382 ± 0.005) × 10−6 (1σ) and this is likely the best estimate of atmospheric helium isotopic composition (Ra) for the period 1970–1990.

It should also be noted that Matsuda et al. (2002) have prepared an artificial mixture of 3He and 4He to be used as an internal standard for precise measurements of mantle-derived samples and for cosmogenic dating. Since He is depleted in air and the isotopic composition is low compared with mantle-derived or cosmogenic samples, a large volume of air is needed to obtain an adequate 3He signal in the mass spectrometer, creating purification difficulties. It is in any case preferable to have the composition of the standard as close as possible as the samples. As a result, an artificial standard with a high 3He/4He ratio was required. The standard prepared for the purpose is called the “He Standard of Japan” (HESJ) and the recommended ratio is 20.63 ± 0.10 Ra (1σ) calculated by the weighted mean of six independent determinations relative to air in four noble gas laboratories.

1.2 Absolute Measurements of Atmospheric Ar Isotope Ratios

Of all the noble gases, accurate determination of the atmospheric Ar isotopic composition is probably the most important; not only is Ar the most abundant atmospheric noble gas but is certainly the most measured noble gas. For example, the K–Ar and Ar–Ar methods for dating rocks and minerals depend on knowledge of the atmospheric 40Ar/36Ar and 40Ar/38Ar ratios for mass discrimination corrections (Turrin et al. 2010) and for correcting for atmospheric contamination.

The atmospheric Ar isotope ratio was originally reported in a seminal paper by Nier (1950a) and has more recently been revisited by two independent studies. Absolute isotope determinations consist of creating reference gases with known isotopic ratios by mixing known quantities of the pure (or as near-pure as possible) isotopes. The reference gases are then used in order to characterize instrumental mass fractionation (IMF) in the mass spectrometer. Once the IMF is well known then the unknown (atmospheric Ar) can be measured and corrected for the mass spectrometer fractionation in order to estimate the true isotope composition.

Nier (1950a) mixed isotopically pure 40Ar and 36Ar produced by selective thermal diffusion; the respective quantities were determined volumetrically. The discrimination of two different mass spectrometers (MS1 and MS2) were constrained using four mixtures with different 40Ar/36Ar ratios; this mass discrimination was measured at 1.6 ± 0.3 per mil per mass unit for the MS1 mass spectrometer (the more accurate of the two mass spectrometers). Nier measured values of 295.5 ± 0.5 (1σ) and 0.1,880 ± 0.0003 for 40Ar/36Ar and 38Ar/36Ar respectively on laboratory air (dried over hot lithium metal); commercially available gas (Air Reduction Sales Corporation) was also analyzed but found to be fractionated (depleted in 36Ar and 38Ar relative to air).

The isotopic composition of air was re-investigated in 2006 by Lee et al. (2006) who used considerably more precise gravimetric methods to constrain the masses of the isotopic mixtures, and used multicollection mass spectrometry (a MAT271 gas source mass spectrometer) to better constrain the discrimination and isotopic ratios. Lee et al. proposed (40Ar/36Ar)air = 298.6 ± 0.3 (1σ) and (38Ar/36Ar)air = 0.1885. Most recently, Valkiers et al. (2008) using a similar methodology to Lee et al. (gravimetric mixtures but starting with different gases) confirmed the redetermined 40Ar/36Ar value (obtaining a value of 298.7 ± 0.1) but disagreed for the (38Ar/36Ar)air measurement, finding a value of 0.1898 ± 0.0001.

Without constraining the IMF, Mark et al. (2011) recently measured 40Ar/36Ar and 38Ar/36Ar on different air aliquots reasoning that since IMF results in mass-dependent fractionation, then the true atmospheric isotopic composition should lie on the mass fractionation line (the slope defined by √(40 − 36)/(38 − 36)) passing through the raw (non-IMF corrected) mass spectrometer data. Assuming that this hypothesis is correct, it would appear to confirm that the Lee et al. (2006) measurements are the best estimate of modern air (Fig. 1).

Fig. 1
figure 1figure 1

Estimates of atmospheric Ar isotope ratios (modified after Mark et al. 2011). Lee et al. (2006) and Valkiers et al. (2008) refined the original Nier 40Ar/36Ar ratio but were unable to agree on an atmospheric 38Ar/36Ar value. Mark et al. measured atmospheric Ar without constraining fractionation, therefore Mark et al’s value (the value shown is the average of 180 measurements with the reported standard error) should fall all on the mass fractionation line passing through the true value. The slanted grey bars indicate the mass fractionation line (and associated uncertainty) passing through the three (Nier, Lee and Valkiers) absolute measurements. Mark et al’s data are consistent with the Lee et al. estimate of the atmospheric Ar

Full size image

It is important to note that commercial processing of air can fractionate Ar isotopes: Valkiers et al. (2008) observed a spread in 40Ar/36Ar of ~1 % in commercial sources of Ar produced by distilling air. However, Mark et al. (2011) found no difference between commercial Ar (produced by distilling air) and locally sampled air. Similar variations have been observed for Ne isotopes (Pavese et al. 2005). Mass fractionation for distilled gases depends in fact on the distillation process itself and may vary from tank to tank.

1.3 Absolute Measurements of Ne, Kr and Xe Isotopic Compositions

Nier (1950b) determined the isotopic compositions of Ne, Kr and Xe in the atmosphere using the same mass spectrometer (MS1) as that used for the air Ar isotope measurements see above). Mass discrimination was estimated either by using the 36Ar++/40Ar++ ratio (for Ne) or assuming that the mass discrimination in the spectrometer for Ar (which was accurately determined using synthetic mixtures of 36Ar, 38Ar and 40Ar) applied to all the heavy noble gases: individual isotopes were not available for the other noble gases, preventing direct determination of mass discrimination for each element. The Ne isotopic composition of air was re-determined by (Eberhard et al. 1965) and was found to be essentially identical to Nier’s measurement.

The Ne, Kr and Xe isotopic measurements by Nier in the late 1940s (Nier 1950b) have not changed since, although the errors have been reduced by modern redeterminations (Table 1): the fact that these values are still valid over 60 years after their determination are testament to a highly skilled analyst.

Table 1 IUPAC approved noble gas isotopic compositions of the atmosphere (Laeter et al. 2003)
Full size table

Subsequent absolute determinations of the Ne, Kr and Xe isotopic compositions have been made by the Institute for Reference Materials and Measurements (IRMM) group, Geel, Belgium (Aregbe et al. 1997, 1998, 1996b; Valkiers et al. 1998); these measurements are currently used by IUPAC (Laeter et al. 2003) as the best estimates for atmospheric Kr and Xe isotopic composition. The IRMM approach is to prepare synthetic volumetric or gravimetric mixtures of gases from single isotope gases in order to calibrate instrumental mass fractionation in the MAT271 mass spectrometer used for the determinations (e.g. near-pure 86Kr was added to atmospheric Kr to create known variable isotope ratios). Potential diffusive fractionation of Kr and Xe isotopes during sample inlet were monitored and corrected for (and found to deviate from ideal gas behavior by only a few ppm). Bottles of Kr and Xe were prepared from purified air (by the Messer Greisheim company) which were then calibrated and certified as primary isotope gas standards by IRMM (IRMM-2030 for Kr and IRMM-2000 for Xe), with traceable and certified isotopic compositions (Aregbe et al. 1996b).

Concerning the atmospheric Ne isotope compositions IUPAC refers to a 1984 paper by (Bottomley et al. 1984) and not the IRMM study (Valkiers et al. 2008). Table 1 provides both the IUPAC preferred values and the more recent IRMM determinations, which differ from IUPAC only in the measurement uncertainty.

Thus it appears that the atmospheric compositions of Ne, Ar, Kr and Xe have been well calibrated and for Ar, Kr and Xe there are reference materials available for laboratories interested in making SI-traceable determinations. Nevertheless, all the measurements were made on commercially produced gases (made by distilling air) supplied by Messer which could have systematic offsets relative to air; this has been established for the case of neon (Pavese et al. 2005). It should be noted that most laboratories normalize their analyses to the quoted atmospheric values and essentially report only relative isotope ratio variations, therefore the potential errors in reported data (which are only relative isotope shifts and not absolute measurements) are relatively minor.

2 Heterogeneity of the Modern Atmosphere

2.1 Possible Temporal Variation of Atmospheric Helium Isotopes

Oliver et al. (1984) carefully measured and reported the He concentration in the Earth’s lower atmosphere. The value of 5.222 ± 0.017 ppm by volume (1σ) agrees well with the certificated number of 5.239 ± 0.004 ppm (1σ) determined by (Gluckauf 1946). Although the He contents are identical, Oliver et al. (1984) suggested that (3–12) × 1016 ccSTP of He had been released into the atmosphere from natural gas production between 1939 and 1981. This anthropogenic He should have increased the 1981 inventory of atmospheric He by 0.1–0.6 % compared to (Gluckauf’s 1946) determination, even though the excess was not detectable in their measurements. Stimulated by this work, Sano et al. (1989) measured the 3He/4He ratios of 20 air samples at several different sampling sites and dates. There was a possible change of the atmospheric helium isotope ratio from 1.362 × 10−6 in December 1977 to 1.339 × 10−6 in September 1988, or a decrease in 3He/4He of about 0.74 ‰/year. Although this change could be attributed to local/non-global effects or experimental artifacts, the observations are also consistent with a significant anthropogenic release of crustal He with low 3He/4He ratio (~3 × 10−8). Lupton and Graham (1991) countered that this estimated rate of decrease in 3He/4He was not statistically different from a zero time rate of change (TROC). They also reported California marine air data which showed a TROC of (+0.09 ± 0.43) ‰/year (2σ), i.e. that there was no evidence that the atmospheric He isotope ratio is changing with time. Sano et al. (1991) re-evaluated the original Sano et al. data and reported a non-zero TROC of (−0.81 ± 0.62) ‰/year (2σ). Taking into account the TROC of three independent determinations (Davidson and Emerson 1990; Mamyrin et al.1970; Sano et al. 1989) as well as that of Lupton and Graham, the conclusion of Sano et al. (1991) was that the average TROC based on laboratory measurements of contemporary air was (−0.245 ± 0.055) ‰/year (2σ).

Hoffman and Nier (1993) reported an absolute helium isotopic ratio which was measured by a double focusing mass spectrometer in November 1956. Their value of (1.371 ± 0.019) × 10−6 (2σ), when combined with the previously reported results above, effectively extending the integration time to a 32 year period, gave a regression line corresponding to a TROC of (−0.3 ± 8.7) ‰/year (2σ). Sano (1998) reported a TROC of (−0.9 ± 1.6) ‰/year (2σ) based on the comparison between GEOSECS Pacific Expedition 3He/4He data (Site 263 and 251) which was calibrated against California marine air collected in 1973 (Lupton and Graham 1991) and South Pacific deep seawater values (SA-7 and SA-6 of Sano et al. (1995)) against air from Chiba, Japan in August 1992.

Pierson-Wickman et al. (2001) measured the isotopic composition of He trapped in historical metallurgical slag from France and the United Arab Emirates in an attempt to investigate variations in the atmospheric 3He/4He ratio with time. Iron is manufactured through the oxydo-reduction of Fe-ore reduced over charcoal with air being pulsed and circulated through the furnaces. The best results in terms of sample reproducibility was obtained for blast furnaces that appeared during the Middle Age in which the temperatures reached 1,600 °C and the mass of circulated air was comparable to the mass of processed Fe-ore, assuring efficient ventilation and therefore noble gas exchange with air. Air was trapped in 0.1–1 mm size vesicles contained in the slag matrix having a composition of forsterite therefore with negligible helium diffusivity at environmental temperatures. Pierson-Wickman et al. found that the He isotopic composition of air trapped in vesicles from industrial slags from the Middle Age and from the beginning of the 20th century was 30 ± 12 ‰ (at 95 % conf. level) higher than Today. They considered various natural sources and sinks of He isotopes in the period concerned and concluded that these could account for this significant change in atmospheric 3He/4He. Thus they concluded that crustal helium rich in radiogenic 4He has been released to the atmosphere due to industrial activities (mostly mining and hydrocarbon exploitation), and that the inferred fluxes from anthropogenic activity were consistent with the observed variation. They suggested a TROC of (−0.14 ± 0.07) ‰/year (2σ).

Lupton and Evans (2004) compared the 3He/4He ratio of samples of Pacific marine air collected over the 30 year period from 1973 to 2003 against a secondary geothermal gas standard and also against a running air standard. They concluded that the rate of change of air 3He/4He over the past 30 years fell between –0.10 and +0.02 ‰/year, which is a TROC of (−0.042 ± 0.061) ‰/year. Figure 2 shows the compilation of the time rate of atmospheric 3He/4He change in the literature. When we take into account all TROCs, the weighted mean of the various estimates is (−0.086 ± 0.045) ‰/year (2σ). This value is apparently distinct from the zero rate of change, even though it corresponds to only 2.6 ‰ change in the atmospheric 3He/4He ratio over the past 3 decades, which is below the detection limit of most laboratories in the world.

Fig. 2
figure 2figure 2

A compilation of the time rate of atmospheric 3He/4He changes in the literature. Top and bottom show the change (‰/year) on large and small scales, respectively

Full size image

Very recently Matsuda et al. (2010) reported that the 3He/4He ratios in some old Chinese porcelain were significantly higher than the present-day air value. They claimed that porcelain can be used as a paleo-atmosphere time capsule over historical times. Taking into account their data, the average 3He/4He ratio is 1.516 ± 0.043 × 10−6 (1σ, assuming that present Ra = 1.382 × 10−6 from Table 1) at 1400AD and 1.447 ± 0.030 × 10−6 (1σ) at 1700AD. They took an average of both data, 1.476 ± 0.025 × 10−6 (1σ) as a pre-1700AD atmospheric value (i.e. before the Industrial Revolution) and suggested a TROC of (−0.34 ± 0.18) ‰/year (2σ) under the assumption that the air 3He/4He ratio decreased linearly during the 200 years since the middle of the 18th century to 1970AD. This TROC is significantly larger than the average TROC of (−0.086 ± 0.045) ‰/year (2σ) as calculated in the above discussion.

Sano et al. (2010) have reexamined the 3He/4He ratios of air presumed to be trapped in historical metallurgical slag derived from Japan. They observed that it is difficult to use the Ra unit when discussing a possible temporal variation of the air 3He/4He ratio (because this is Ra itself!) but instead expressed the data relative to the He Standard of Japan (HESJ; see above). The mean 3He/4He ratio of gases extracted from metallurgical slag found at refineries in operation between AD1603 and 1907 in Japan was (5,106 ± 108) × 10−5 RHESJ (2σ) and was consistent with the reported value of (5,077 ± 59) × 10−5 RHESJ in historical slags from France and United Arab Emirates (Pierson-Wickmann et al. 2001). These values are about 40 ‰ higher than that of average modern air (4,901 ± 4) × 10−5 RHESJ from AD 2006 to 2008 (this value is obtained by the average of air samples with various locations as discussed below).

2.2 Anthropogenic Sources of Atmospheric Helium

All slag 3He/4He data as well as porcelain are plotted against the year together with the modern air value in Fig. 3 (top panel). When we assume all data are correct, the simple linear regression taking into account the weight of each experimental error yields a slope of –(0.41 ± 0.11) × 10−5 RHESJ/year (2σ), which is equivalent to a TROC of (−0.084 ± 0.022)  ‰/year (2σ). Since the uncertainty in the composition of the modern surface air, (4,901 ± 4) × 10−5 RHESJ, is significantly smaller than that of slag and porcelain samples, the regression line should be forced to pass through the point of the air value in Fig. 4a. The resulting TROC is identical to the weighted mean of data, (−0.086 ± 0.045)  ‰/year in Fig. 2.

Fig. 3
figure 3figure 3

Secular variation of atmospheric 3He/4He ratio estimated from historical metallurgical slag and old Chinese porcelain. A dotted line in the top shows a simple linear regression taking into account of weighted errors. The curve in the bottom shows a simulation of the 3He/4He change based on fossil fuel consumption (see text). The gray area in the lower panel is obtained by the uncertainty of ±65 % based on the He/C values of natural gas and petroleum

Full size image
Fig. 4
figure 4figure 4

The changes of atmospheric 3He/4He ratio with latitude (a), and longitude (b). The 3He/4He ratios appear to increase from north to south but they are constant from east to west

Full size image

It is possible that the decrease of the air 3He/4He ratio is due to a natural perturbation of the steady-state atmosphere resulting from the geomagnetic reversal at 0.79–0.76 Ma (Izett and Obradovich 1994) which probably severely depleted air helium at that time and the atmosphere is now slowly returning to equilibrium (Kockarts 1973; Sano et al. 1989). Pierson-Wickmann et al. (2001) suggested that degassing of the solid Earth could have been increased by a factor of 10 since the last deglaciation (15,000 years ago) due to recirculation of groundwater in previously frozen areas which accumulated radiogenic He while frozen, then release this during melting. Similarly, melting of glaciers that might have accumulated crustal 4He could increase the rate of 4He release to the atmosphere during post-glacial times. Therefore the TROC of −0.086 ‰/year may be attributable to natural processes.

In addition to natural processes, anthropogenic He sources to a non-steady-state atmosphere should be taken into account (Oliver et al. 1984). Possible sources and sinks of atmospheric He (from (Sano et al. 2010)) were estimated and are listed in Table 2. Commercial production of He gas for cryogenic purposes, welding and space missions was 7.6 × 109 mol/year in 2006 (US Geological Survey 2009, http://minerals.usgs.gov/minerals/pubs/commodity/helium). This value is ten times greater than the natural crustal degassing rate of 6.1 × 108 mol/year (Ozima and Podosek 1983), clearly indicating that there are likely anthropogenic sources of He in the atmosphere.

Table 2 Possible sources and sinks of atmospheric helium
Full size table

A further major source of crustal He can be attributed to fossil fuel exploitation (Pierson-Wickmann et al. 2001; Sano et al. 1989). Consumption of natural gas, petroleum and coal have been well documented by the Carbon Dioxide Information Center (http://cdiac.ornl.gov). It is reasonable to assume that the rate of consumption of fossil fuel is similar to that of production on a global scale. However, the He content of fossil fuel is known to be highly variable and difficult to constrain. Based on ~3,000 analyses of natural gas samples containing less than 8 % He in the U.S.A., (Zartman et al. 1961) reported that the He abundance showed a log-normal distribution with a mean concentration of 0.217 volume  %. The mean He content in natural gas from outside the U.S.A. is not well documented, but might be 0.1 % (Oliver et al. 1984). The He/total C molar ratio of crude oil in the Paris basin, France was 800 ppm (Pinti and Marty 1995). On the other hand, there is no available datum in coal samples. We assume the He/C ratio of 7 × 10−11 for coal (an average U content = 2 ppm for coal, Th/U = 4 (Swaine and Goodarzi 1997) and an average age of coal of 360 Ma). Thus 4He from coal is probably negligible.

It is then possible to calculate the anthropogenic release of crustal He due to fossil fuel production (see Table 2), resulting in a total fossil fuel flux of 3.4 × 1011 mol He/year in 2006, which is 3 orders of magnitude larger than the natural flux. On the other hand, the monotonous decrease of atmospheric 3He/4He ratio (the trend in Fig. 3a), i.e. a TROC of −0.084  ‰/year implies a He input to the atmosphere at 8 × 1010 mol/year based on the simple mass balance calculation where crustal He has a 3He/4He ratio of 3 × 10−8 (Mamyrin and Tolstikhin 1984). Thus the He flux estimated from the global fossil fuel production statistics is about 4 times higher than the TROC calculation.

From the He content in natural gas and petroleum as stated above, it is possible to simulate the secular variation of the atmospheric 3He/4He ratio (based on the fossil fuel consumption per year from AD 1751 to 2006 supplied by the Carbon Dioxide Information Analysis Center (http://cdiac.ornl.gov)). Little change is expected between AD 1400 and 1751 because of minimal consumption and/or production of fossil fuel compared to after the Industrial Revolution. The solid black curve in Fig. 3b shows the calculated change in atmospheric 3He/4He ratio together with observed slag and porcelain data. All data except for the Chinese porcelain in 1400AD (2nd crushing of POR2 in Matsuda et al. (2010)) are consistent with the calculated 3He/4He change within the error margins resulting from uncertainties in He/C ratios, suggesting that the simulation is valid. However, the simulation predicts an average decrease at 0.32 ‰/year for the last 10 years, significantly higher than the TROC value of −0.086 ± 0.045 ‰/year obtained by measuring contemporary air (Fig. 2). In reality, the true TROC is probably located between the monotonic decrease in the atmospheric 3He/4He ratio (dotted line in Fig. 3a) and the recent catastrophic decrease (a solid curve in Fig. 3b). Even so, taking the largest variation of 0.32 ‰/year, this is still below the detection capabilities of most laboratories. Thus atmospheric He is still suitable as a laboratory standard.

2.3 Possible Spatial Variation of Atmospheric Helium Isotopes

Changes of the atmospheric 3He/4He ratio with latitude and longitude were assessed by the pioneering work of (Mamyrin et al. 1970). They reported that the spatial variation with both location and altitude (up to 10 km) was within the error margin of individual measurements of about 3 %. More recently Sano et al. (2010) measured the 3He/4He ratios of 21 surface air samples collected around the world with a precision of 2–4 ‰ (2σ) using a newly developed analytical system (Sano et al. 2008). Even though the data available are not sufficient to cover the entire planet, the sites cover high, low and mid-latitudes (ranging from Svalbard, Norway, to far Showa base, Antarctica). The atmospheric 3He/4He ratios generally increase from north to south (Fig. 4a), but they do not change significantly from east to west (Fig. 4b). The result is attributable to the fact that mixing of long-lived tracers between northern and southern hemispheres is a much slower process than east to west transport (Jacob 1999; Muller and Brasseur 1995).

By fitting the atmospheric 3He/4He ratios variation with latitude by eye (dotted curve in Fig. 4a), one obtains an apparently constant 3He/4He ratio from 38°S to 38°N, with high values in Antarctica and low values at high latitudes in the northern hemisphere (from 38°N to 46°N). When we take the uniform 3He/4He ratio in mid-latitudes, the average becomes (4,901 ± 4) × 10−5 RHESJ, which could be adopted as a representative value of the present atmosphere (see Sect. 2.2.4). This would result in a 3He/4He ratio in HESJ of 20.404 ± 0.017 (2σ) in Ra units. There is a relatively high 3He/4He ratio in Antarctica but not in the Arctic. The major source of atmospheric 3He is auroral precipitation into the upper atmosphere in polar regions (Johnson and Axford 1969). Therefore, these higher 3He/4He ratios may be attributable to the auroral effect, but the apparently lower ratio in the Arctic needs to be investigated further.

The average 3He/4He ratio of air samples at high latitudes from 38°N to 46°N is (4,886 ± 7) × 10−5 RHESJ (2σ), 0.31 ± 0.14 % lower than the average value of (4,901 ± 4) × 10−5 RHESJ, although there is a possibility of experimental artifacts in Sano et al. (2010). The atmospheric volume at high latitudes (38°N ~ 46°N) is calculated to be about 5.2 % of the global atmosphere, assuming that the volume is proportional to surface area. The He inventory of the region is estimated to be 4.8 × 1013 mol, (5.2 % of the total inventory of 9.3 × 1014 mol (Ozima and Podosek 1983)). In order to make the regional atmospheric 3He/4He ratio 0.31 ± 0.14 % lower than the mid-latitude helium, a local input of crustal He of (1.5 ± 0.7) × 1011 mol is required. This amount is significantly smaller than a single year of anthropogenic He (5.9 × 1011 mol/year). Because atmospheric circulation and mixing homogenize the atmosphere (Jacob 1999), the released anthropogenic He could rapidly dissipate, causing the air 3He/4He to be higher than expected. Note that the 0.31 ± 0.14 % anomaly is below the detection limit of most noble gas laboratories in the world.

2.4 Anthropogenic Kr and Xe

Fission of nuclear fuel produces Xe and Kr that are isotopically distinct from that of the atmosphere, the precise isotopic pattern depending on the fuel characteristics and subsequent operation of the nuclear pile. Weiss et al. (1986) demonstrated that the atmospheric 85Kr (t1/2 = 10.73 years) budget is dominated by 85Kr produced during fission. Aregbe et al. (1996a) modeled the stable Kr and Xe isotopes expected from mixtures between air and Kr and Xe produced by two different nuclear pathways: from civil nuclear energy production in a Pressurized Water Reactor (PWR) and from plutonium enrichment. However, the lower fission yields and higher atmospheric abundance of Kr means that stable Kr isotopes are not particularly suited to tracing nuclear atmospheric inputs. Normalising to 132Xe, Agrebe et al considered that the most sensitive Xe isotopes for tracing fission produced Xe are 136Xe/132Xe and 129Xe/132Xe with production ratios greater and lower than atmosphere respectively (however, it would seem more logical to the authors of this chapter to normalize to 130Xe as this is nearly completely shielded); Pu-enrichement and UOX (PWR) produce characteristically different Xe isotope patterns. Aregbe et al. (1996a) compared the calculated isotope anomalies generated by 1 g of fission Xe variably diluted by air Xe with the (impressive) measuring capacity of the MAT 271 mass spectrometer at IRMM, Belgium to detect Xe isotope anomalies (about 0.02 per mil for both 129Xe/132Xe and 136Xe/132Xe). They concluded that 1 g of fission Xe diluted in 108 m3 of air (~108 kg air) would be detectable, and at 1 g fission Xe in 106 m3 of air it would be possible to distinguish between uranium processed for civil or weapons purposes. Thus Xe isotopes have the possibility for identifying nuclear activity if air samples are available close to the reactor. Nevertheless, anthropogenic Xe is highly unlikely to have affected the isotopic composition of atmospheric Xe, given an atmospheric mass of the order 5 × 1018 kg and an annual global U production of the order 50 × 106 kg (2009 data: source: “World Uranium Mining”. World Nuclear Association.), approximately 1,000 years of U extraction would be required in order to create a measurable excess in the global 136Xe/132Xe.

2.5 Temporal Variation of Atmospheric Argon Isotopes

Apart from helium, argon is the only other noble gas for which temporal variations have been reported. Argon has three natural stable isotopes, 36Ar, 38Ar and 40Ar with well measured relative abundances (see Table 1). While the first two are primordial, that is, have been trapped from one or several cosmochemical reservoirs when the Earth formed. There was virtually no 40Ar at that time and this isotope was produced in the silicate Earth (mantle and crust) by the decay of radioactive 40K. Thus 40Ar accumulated in the atmosphere due to magmatic and tectonic degassing of the mantle and crust. Determining the degassing rate of radiogenic 40Ar through time is important for a number of geoscientific problems such as the growth regime of the continental crust, the potassium content of the silicate reservoirs, variations in mantle convection rate and global tectonics. Consequently, there have been several attempts to determine the rate of 40Ar degassing from measurements of the Ar isotope composition of ancient atmosphere presumably preserved in sedimentary rocks and ice of known ages. The first unambiguous result was obtained for the Devonian Rhynie chert by (Cadogan 1977) and later confirmed by further analysis of the same unit by other groups. Cadogan (1977) measured a 40Ar/36Ar ratio of 293.0 ± 1.6 (renormalized to the new value of the modern 40Ar/36Ar ratio of 298.6, Lee et al.) for argon trapped in this 396 Ma-old chert. Assuming that this value represents the Ar isotope composition of air at that time and from its difference relative to the present-day 40Ar/36Ar value, it is straightforward to compute a 40Ar flux from the solid Earth of 6 × 107 mol/year for the last 400 Ma. It is interesting to compare this flux with that from the mantle. The flux of 3He from the mantle was initially estimated at 1,000 mol/year Craig et al. (1975) but has been scaled down to about 500 mol/year by (Bianchi et al. 2010). With a mantle 4He/40Ar ratio of 2 ± 1 (Moreira et al. 1998) and a mid-ocean ridge 3He/4He value of 8 ± 1 Ra (Graham 2002), one finds that the contemporaneous flux of 40Ar from the mantle is within the range (1.2–8.6) × 107 mol/year. The flux value obtained from the Rhynie chert is within this range, but the large uncertainty on these numbers prevents one to derive conclusions on the importance of other 40Ar sources such as the continental crust. A more positive approach to these estimates is to consider that there is a surprisingly good match despite the fact that these flux values address drastically different periods of time.

Recently, a novel approach to quantify the flux of 40Ar from the solid Earth to the atmosphere has been developed by (Bender et al. 2008) using the record of air trapped in deep ice cores from Antarctica. Air trapped in ice is, however, gravitationally fractionated and direct measurements of ice-trapped noble gas isotope ratios are not representative of ancient atmospheric compositions. Additionally, the ice record extends to less than one Ma and the expected Ar isotope variations are much lower than the precision attained by most current noble gas mass spectrometers. In order to circumvent these problems, Bender et al. first developed high precision analysis of argon isotopes by adapting stable isotope mass spectrometry (dynamic mass spectrometry, i.e. with pumps open and a variable leak inlet) since the amount of sample was large by noble gas standards. They achieved an amazing precision of 0.024  ‰ (2σ) on Ar isotope ratio measurements. Second, they made use of the fact that Ar has 3 isotopes and the ratio between two of them, 36Ar and 38Ar, can be safely assumed to be constant through time, at least during the last Ma, since they are not significantly produced or destroyed by nuclear reactions. Diffusion is a mass-dependent process and, in the case of Ar, can be corrected for by measuring the relative proportions of the 3 isotopes of argon trapped in ice and assuming a mass-dependent law for isotopic fractionation. They found a rate of change of the Ar isotope ratio of air (40Ar/38Ar in their case) of 0.066 ± 0.014  ‰/Ma, which yields a 40Ar outgassing flux of 11 ± 1 × 107 mol/year, in fact not very different from Cadogan (1977)’s value of 6 × 107 mol/year averaged over the last 400 Ma. Bender et al. concluded that their flux could be well accounted for by the sum of degassing fluxes from the mantle and the continental crust. Additionally, they proposed that their method provides a new way to date old ice, especially when other age records have been erased, although uncertainties are large (±180 Kyr).

3 Future Directions

The future of studying variations in atmospheric noble gas compositions relies on developing instrumentation in order to better measure the small differences involved and (particularly for addressing the temporal evolution of atmospheric noble gases) identifying suitable samples.

Considerable advances have been made in the last ten years and multicollector noble gas mass spectrometers are now commercially available. Although commercially available noble gas mass spectrometers are typically used in ‘static’ (pumps closed) mode, data from Mark et al. (2011) show that high precision noble gas isotope ratio measurements are possible on small gas quantities (of the order ±0.5 per mil on the 40Ar/36Ar ratio with 2 × 10−12 mol Ar). These mass spectrometers can be relatively easily adapted for ‘dynamic’ (pumps open) measurements, resulting in the precisions available using adapted stable isotope machines (of the order 0.05 per mil for Ar (Bender et al. 2008; Lee et al. 2006). Dynamic mass spectrometry has the ability to improve precision relative to static mass spectrometers, at the cost of requiring much (103) larger samples and of inducing mass fractionation during inlet of the gas into the mass spectrometer. Large sample sizes are difficult to handle, particularly for He and Ne which are difficult to separate by distillation, instead requiring chemical purification of large volumes of air. Instrumental mass fractionation resulting from the inlet leak is furthermore a major problem for measurement of He isotopes in air, because, with only 2 isotopes, it is difficult to quantify the magnitude of this fractionation during analysis.

Identification of suitable archives for noble gas (particularly He) isotope analysis is critical in order to determine temporal changes in noble gas compositions. For the present-day atmosphere, air sampled since 1976 at Cape Grim, Tasmania, under the auspices of the CSIRO, constitutes an archive of choice.

Suitable natural archives need to be able to preserve atmospheric noble gases without significant in situ production or destruction of isotopes, and without isotopic fractionation or, at least with the possibility of correcting for such fractionation with accuracy. Thus the nature of suitable archives depends on the targeted time interval. For the last couple of 1,000 years, only changes in the He isotopic composition are expected to be detectable with modern mass spectrometry systems. Helium has, however a high diffusivity in many of the materials of interest, including glass and organics. Greater predicted variations in 3He/4He are offset by higher precision of Ar isotope measurements and ability to correct for mass discrimination.

For the last millenium, some success has been achieved using industrial slags and porcelain, and there are certainly other possibilities such as for example manufactured metal tools, munitions, scientific instruments etc.

For variations over the last Myr the ice record has proven to be exploitable for Ar isotopes and although this may be possible also for He isotopes, this poses a considerably greater challenge as (a) helium may diffuse through ice over these timescales; (b) helium has only 2 isotopes and correction for significant isotope fractionation will be model-dependent and (c) the low abundance of He in the atmosphere and atmosphere-equilibrated liquids require large sample sizes.

For the 108–109 years time intervals, the sedimentary record seems to be the only way to get ancient air but in many cases the noble gas record is likely to have been altered by isotope exchange and in situ production. So far only one sedimentary unit (the Devonian Rhynie chert) seems to have preserved an ancient atmospheric noble gas record. There is little doubt that other time capsules exist and should be investigated such as amber, evaporites, chert, and barite. Recently, Pujol et al. (2009) claimed that noble gases trapped in Archean barite were representative of the Archean atmosphere. In particular, they found that the stable isotope composition of xenon was intermediate between that of the modern atmosphere and that of primitive meteorites, suggesting time-dependent evolution for such composition.

With the advent of modern noble gas mass spectrometry and of novel analytical techniques, there is no doubt that we are opening a new chapter in the field of noble gas geochemistry with precise studies of the atmospheric evolution in space and time.