Helium, Neon & Argon

Introduction

The group of elements known as the rare, inert or noble gases possess unique properties that make them important as geodynamic tracers. The daughter isotopes fractionate readily from their parents, they are inert, they give information about the degassing history of the mantle, the formation of the atmosphere, and about mixing relationships between different mantle components, and they differ from other geochemical tracers in being gases and diffusing relatively rapidly.

The study of noble gases in basalts has led to the concept of a largely undegassed, ancient, “primordial reservoir” in the lower mantle. However, the best places on Earth to find the most “primordial” material (i.e., high concentrations of noble gases, including ³He and ²⁰Ne and high ³He/⁴He and ²⁰Ne/²²Ne ratios) are on mountain tops, in the stratosphere, and in deep-ocean sediments. Evidence for recent additions of noble gases to the Earth comes from deep sea sediments where high concentrations of ³He (written [³He]), and high ³He/⁴He are found (Merrihue,1964; Nier & Schlutter, 1990). However, mantle-derived basalts and xenoliths having similar “primordial” or “solar” characteristics are generally attributed to plumes from an undegassed primitive lower mantle reservoir. This is based on the assumption that high ³He/⁴He and ³⁶Ar/⁴⁰Ar ratios in basalts require high ³He and ³⁶Ar contents in their sources.

The critical issues are when and how the “primordial” noble gases entered the Earth and whether transport is always outwards. It is therefore important to understand the noble gas budget of the Earth, both when it formed and as it aged. The major part of terrestrial formation probably occurred at high temperature from dry, volatile-poor, degassed particles. Elements that formed the Earth during this time are known as “primordial”. During the early stages of planetary formation, core differentiation and massive bolide impacts, one of which was large enough to form the moon, raised the Earth to temperatures that may have been high enough for a substantial part of it to have melted, and some to be vaporized. This is suggested by the fact that a large fraction (30 to 58%) of the incompatible elements (e.g., U, Th, Ba, K, Rb) are in the Earth’s crust (Anderson, 1983; 1989), which is consistent with an extensively differentiated mantle. Strong degassing of the primordial volatiles occurred during this period (Figure 1). The siderophiles have very low abundances in crustal and mantle rocks (the bulk silicate Earth) because they are in the core.

Figure 1: Abundances of elements in the crust+mantle relative to those in chondritic meteorites (from Anderson, 1989; Figure 8.2)

The subsequent 4.55 billion years of planetary evolution was probably more conducive to retention and accumulation of volatiles (except for H and He, and possibly some Ne, which can escape from the atmosphere to space), including the noble gases. Some of the components now in the mantle fell to Earth in a late veneer, mostly around 3.8 Ga, the period of late bombardment, that added material to the Earth after completion of core formation (Kimura et al., 1974). This probably introduced most of the volatiles and trace siderophiles (iron-lovers) that are in the upper mantle today. However, what fraction of the noble gases in mantle basalts is truly primordial vs. a late veneer – or even later additions of cosmic dust – or recent contamination is unknown.

The hot-accretion model of Earth evolution contrasts with the “standard model” of noble gas geochemistry that assumes that the bulk of the mantle accreted in a primordial undegassed state with chondritic, or cosmic, abundances of the noble gas elements. Such material is assumed to make up the deep mantle today. This theoretical deep, isolated, undegassed or little-degassed lower-mantle reservoir is considered to have high [³He], [²²Ne], [³⁶Ar], ³He/⁴He and ²⁰Ne/²²Ne, low ²¹Ne/²²Ne and to be tapped by plumes. The upper mantle, the presumed source of mid-ocean ridge basalts (MORB), is considered to convect vigorously, to be well-stirred, homogeneous and to have noble gas characteristics the inverse of those of the lower mantle, listed above. Any high-³He/⁴He or so-called “solar” neon (see below) found along mid-ocean ridges is presumed to have been transported into the upper mantle by plumes.

Isotopic Ratios

The most useful light noble gas isotopic ratios are shown in Table 1

Table 1: The most useful light noble gas isotopic ratios. signifies radiogenic or nucleogenic.*
³He/⁴He*	³He/²²Ne	³He/²⁰Ne	³He/³⁶Ar
⁴He/²¹Ne	⁴He/⁴⁰Ar
²⁰Ne/²²Ne
²¹Ne*/²²Ne
⁴⁰Ar*/³⁶Ar

Basic data concerning the noble gases He, Ne, & Ar are given in Table 2.

Table 2: Noble gas isotopes, their sources and half-lives.
Isotope	Source	Half life
³He	stable
⁴He*	decay of U+Th. ²³⁸U=8(⁴He)+²⁰⁶Pb, ²³⁵U=7(⁴He)+²⁰⁷Pb, ²³²Th=6(⁴He)+²⁰⁸Pb	10³ – 10⁴ Ma
²⁰Ne	stable
²¹Ne*	nucleogenic decay chains of ²⁴Mg, ¹⁸O when irradiated by U+Th decay products. Thus produced at the same rate as ⁴He
²²Ne	some is primordial and some produced in a similar way to ²¹Ne but only in very small amounts
³⁶Ar	stable
³⁷Ar*	from fission
³⁸Ar	stable
³⁹Ar	cosmic rays
⁴⁰Ar*	decay of ⁴⁰K & ⁴⁰Ca. Comprises 99.6% of the Ar in the atmosphere.	10³ Ma

Helium

There are few constraints on how much helium may be in the Earth or when it arrived. It is readily lost from the atmosphere via escape into space. It is therefore not very useful for studying the outgassing history of the Earth. Neon and argon are potentially more useful in this regard since they do not escape. We know, for example, that most of the argon in the Earth is now in the atmosphere, even though much of it was produced over time and was not in the mantle during the earliest high temperature phases of accretion and evolution. This argues against there being a large undegassed reservoir in the Earth.

Both melting and degassing fractionate the noble gases, and separate the parent from the daughter isotopes. Helium and the light noble gases are more readily retained in magma than the heavy noble gases during degassing. Thus, the He/Ne and He/Ar ratios of different basalt types contain information about their degree of degassing. The He/Ne and He/Ar ratios of MORB are generally higher than OIB (Figure 2) (Ozima & Podosek, 2002) suggesting that MORB is more degassed than OIB. Nevertheless, [He] is higher in MORB than in OIB, which suggests that OIB initially contained less [He] and possibly less [Ne] than MORB, prior to MORB degassing. Some of the more noble-gas-rich igneous rocks are the “popping rocks” found along the mid-Atlantic ridge (e.g., Sarda & Graham, 1990; Javoy & Pineau, 1991; Staudacher et al., 1989). Ordinary MORB appear to be degassed versions of these popping rocks, and even some of these gas-rich rocks appear to have lost some gas.

Figure 2: Histogram of ³He/²²Ne in MORB and OIB. “Solar” indicates the cosmic abundance ratio. MORB and OIB are fractionated in opposite directions relative to the solar ratio [from Ozima & Igarashi, 2000].

A plot of ⁴He/⁴⁰Ar vs [⁴He] and related inter-element plots (e.g., ³He/²¹Ne vs. [He]) shows a continuous transition from MORB to OIB to air (e.g., Figure 3). The MORB to OIB trends could be interpreted as air-contamination of noble gases degassed from MORB and trapped in vesicles or cumulates.

Figure 3: Plot showing relation between helium and argon isotopes in MORB and OIB and the effects of degassing and “contamination” with air and older CO₂-rich vesicles (Anderson, 2000).

So-called primitive ratios of noble gas isotopes may be preserved by ancient separation of He and Ne from U+Th and storage of the gas in low-U+Th environments such as depleted lithosphere or olivine cumulates (Anderson, 1998a, 1998b; Natland, 2003). In this way, the radiogenic isotopes ⁴He and ²¹Ne are not added to the gas and old isotopic ratios can be “frozen in”. This contrasts with the situation in more fertile materials such as MORBs or undepleted peridotite, where U+Th contents are relatively high and continued radiogenic and nucleargenic ingrowth occurs (see also Helium Fundamentals page).

Neon

Neon isotope ratios for mantle rocks are typically displayed on a “3-isotope plot” of ²⁰Ne/²²Ne vs. ²¹Ne/²²Ne (Figure 4; see Graham, 2002). Neon isotope ratios are more susceptible to atmospheric contamination than helium isotope ratios because Ne is more abundant in the atmosphere. Most basalts plot on what appear to be mixing lines between atmospheric values (~10), and solar wind – SW or interplanetary dust particle – IDP values (13.7) of ²⁰Ne/²²Ne. The corresponding values for the ²¹Ne/²²Ne ratios are 0.029 (air), > 0.04 (OIB) and > 0.06 (MORB).

Figure 4: “Three-isotope plot” of ²⁰Ne/²²Ne vs. ²¹Ne/²²Ne.

The mixing line for MORB extends to greater values of ²¹Ne/²²Ne than the mixing line for OIB, indicating more ²¹Ne ingrowth in MORB than in OIB. This is analogous to the higher ⁴He/³He ratios found in some MORB samples compared to the extremes found in some OIB samples. The ²⁰Ne/²²Ne ratios in mantle derived rocks, both MORB and OIB, extend from atmospheric to the solar or IDP ratio (Anderson, 1993; Allegre et al., 1993). The identification of solar-like (“primordial” or IDP) Ne isotopic ratios in some OIB and MORB samples implies that solar neon trapped within the Earth has remained virtually unchanged over the past 4.5 Gyr (the standard “primordial mantle” model) or, alternatively, that noble gases have been added to the mantle or the samples more “recently” (i.e., since 2 Ga), perhaps from noble-gas-rich sediments. Some basalts and some diamonds essentially have pure “solar” (SW or IDP) ²⁰Ne/²²Ne ratios. The fact that the ²⁰Ne/²²Ne ratio of rocks or magmas thought to come from the mantle is greater than the atmospheric ratio is a paradox for standard models of mantle degassing and atmospheric evolution because ²⁰Ne and ²²Ne are essentially primordial in the sense that ²⁰Ne is stable and only very small amounts of ²²Ne are produced. Their ratio is thus expected to be uniform throughout the Earth (Craig & Lupton, 1976), if the solar system reservoir is uniform and invariant with time.

Neon and Helium

The ³He/²²Ne and ⁴He/²¹Ne ratios observed in mantle xenoliths and basaltic glasses vary by orders of magnitude and define a linear correlation with a slope of unity which passes through the point defined by the mean primordial ³He/²²Ne ratio (= 7.7) and the radiogenic ⁴He to nucleogenic ²¹Ne* production ratio (= 2.2 x 10⁷) (Figure 5, see Dixon et al., 2001). The linear correlation implies that the elemental fractionation event (perhaps degassing of magma), which enriched MORB glasses in [He] with respect to [Ne], is recent, otherwise in-growth of radiogenic ⁴He and nucleogenic ²¹Ne would have systematically shifted the data points from the correlation line. Basalts exhibit positive correlations between ³He/²²Ne and [³He], and between ⁴He/²¹Ne* and [⁴He]. The ³He/²²Ne, ⁴He/²¹Ne* and ⁴He/⁴⁰Ar* ratios in MORB glasses are systematically higher than the primordial ³He/²²Ne ratio or the radiogenic ⁴He*/nucleogenic ²¹Ne* and radiogenic ⁴He*/radiogenic ⁴⁰Ar* production ratios.

Figure 5: Plot of ³He/²²Ne vs. ⁴He/²¹Ne (from, Dixon et al., 2001).

The production rate of ²¹Ne parallels that of ⁴He since they both result from the decay of U+Th (Table 2). The 3-isotope plot showing OIB (Loihi) as forming one trend from atmosphere to solar and MORB as forming another (Figure 4) can be interpreted as a rotation of the OIB trend by nucleogenic ingrowth, i.e., MORB source is aged OIB source. In this simple model MORB = OIB + time. OIB and MORB could have evolved from a common parent at some time in the past. Old MORB gases (from ancient degassed MORB) stored in a depleted refractory host (e.g., olivine crystals, U+Th-poor lithosphere) will have high ³He/⁴He and low He/Ne compared to current MORB magmas. One of the mantle “reservoirs” for noble gases may be isolated gas-filled inclusions or vugs. This “reservoir” would contribute nothing besides He and CO₂ and perhaps some Ne, thereby decoupling He form other isotopic tracers.

Argon

The main isotopes of argon found on Earth are ⁴⁰Ar, ³⁶Ar, and ³⁸Ar. Naturally occurring ⁴⁰K with a half-life of 1.250 x 10⁹ years, decays to stable ⁴⁰Ar.

Atmospheric He has a very low abundance due to its complete escape from the atmosphere, and even atmospheric Ne is believed to have been significantly depleted in the atmosphere by intense solar irradiation. The Ne three-isotope plot allows atmospheric contamination to be monitored for this isotope. On the other hand, the heavy rare gases (argon, krypton and xenon) have accumulated in the atmosphere over Earth history, and contamination is a serious problem in the study of mantle values.

The values of ⁴⁰Ar/³⁶Ar (or ⁴⁰Ar*/³⁶Ar) measured, or estimated, in various materials are:

air ~ 300
OIB atmospheric to ~ 13,000
MORB atmospheric to ~ 44,000

In the standard model, which assumes a lower-mantle source for OIB, these high ratios for OIB and MORB, compared with air, have been taken to indicate that both the upper (MORB) and the lower mantles (OIB) have been degassed such that little ³⁶Ar remains, but large quantities of ⁴⁰Ar have been produced by radiogenic decay and are retained. The higher ⁴⁰Ar/³⁶Ar in mantle rocks than in the atmosphere contrasts with the situation for He, where ⁴He/³He is usually much higher in the atmosphere than in either OIB or MORB as a result of the preferential escape of the lighter ³He atom. Much of the He in the atmosphere entered it in the last million years and is therefore dominated by ⁴He.

Low ⁴⁰Ar/³⁶Ar plays the same role in the standard noble gas model as high ³He/⁴He in that low radiogenic ratios are taken as proxies for less degassing and more primitive reservoirs. Others have argued, however, that this does not place strong constraints on the argon isotope ratio of the “lower mantle undegassed reservoir” because plumes may have been contaminated with radiogenic argon during ascent through the MORB source.

The concept of a relatively less degassed mantle reservoir with respect to heavy rare gases was contested by Fisher (1983; 1985) who argued that low ⁴⁰Ar/³⁶Ar is a result of atmospheric or seawater contamination. Seawater has 2 – 4 orders of magnitude more ³⁶Ar than mantle basalts and two orders of magnitude less ³He. Hence, Ar isotope ratios in magmas can be significantly changed by seawater without affecting their He signature. Measured [³⁶Ar] covers the same range for MORB and OIB (Ozima & Igarashi, 2000), and [³He] is higher for MORB than for OIB. This also argues against OIB arising from a reservoir much less degassed than the MORB reservoir. This model for Ar is at odds with the standard model for He that considers the lower mantle to be little degassed. The lower ratios for OIB can thus be explained by more extreme atmospheric contamination, or a potassium-poor- or younger source. Usually, however, unradiogenic ratios are attributed to high abundances of the primordial isotope, i.e., to undegassed or primordial reservoirs. The ²⁰Ne/²²Ne ratios for mantle rocks that are significantly higher than atmospheric indicate that the source of noble gas in both OIB and MORB cannot be mainly present-day atmospheric.

⁴⁰Ar is derived from ⁴⁰K and is retained by the atmosphere. The initial amount of ⁴⁰Ar contained in the Earth is negligible compared with the amount that was produced subsequently radiogenically. This makes it a useful tool with which to study mantle degassing. The amount of K in the crust and mantle may be estimated by calculating the mix of mantle and crustal components that satisfies the cosmic ratios of the refractory elements (Anderson, 1989). The amount of K in the silicate Earth (the mantle and crust) so determined is 151 ppm of which 46% is in the crust. The amount of ⁴⁰Ar in the atmosphere represents 77% of that produced by the decay of this amount of potassium over the age of the Earth. This implies that either 23% of the Earth is still undegassed or that the degassing process is not 100% efficient i.e., that there is a delay between the production of ⁴⁰Ar and its release into the atmosphere. Ozima (1998) has suggested that Ar may be compatible at depths > ~ 150 km and thus not be easily extracted from the Earth by volcanism. A similar phenomenon appears to occur with ⁴He, which does not degas at a high enough rate to correspond to the outflow of heat generated by U+Th decay. This is known as the helium heat-flow paradox. Since it is more difficult to degas the mantle now than during the early, high-temperature period of Earth history, it is likely that the stable isotopes – those that have not accumulated over time – are more degassed than the radiogenic and nucleogenic ones e.g., ⁴He and ⁴⁰Ar.

The “two reservoir” model

The partially degassed or “two reservoir” model for the mantle was originally proposed to explain argon isotope systematics (Hart et al., 1979), and was applied to helium by Kaneoka & Takaoka (1980). Kaneoka & Takaoka (1980) based their case on rare gas compositions in phenocrysts from Haleakala volcano, Hawaii; these were subsequently shown to be contaminated with atmospheric argon and cosmogenic helium (Fisher, 1983; Kurz, 1986a). Some ocean island basalts, most notably from Hawaii, Iceland and the Galapagos have elevated ³He/⁴He values compared to most MORB samples. Hence the two-reservoir model for mantle helium was resurrected and is now widely accepted.

Figure 6: Maximum value of Rvs. the spread in values showing that high-R islands also have the largest range in values (figure provided by A. Meibom).

An alternative way of looking at the helium data is via histograms. Both OIB and MORB samples define smooth distributions, with the MORB distribution being more Gaussian-like and with smaller variance. The so-called “high-He” hotspots exhibit very large variance (Figure 6). This is suggestive of the predictions of the central limit theorem and implies that mid-ocean ridges sample larger volumes of a heterogeneous mantle than do oceanic islands (Anderson, 2000).

Another test of the two-reservoir model was made by comparing helium and heat fluxes (O’Nions & Oxburgh, 1983). These fluxes must be related because the decay of uranium and thorium produces both radiogenic helium (alpha particles) and heat. O’Nions & Oxburgh (1983) calculated the concentration of U necessary to generate the observed helium and heat fluxes, assuming an instantaneous steady-state. The results were surprising. The amount of uranium required to generate 88% of the Earth’s oceanic helium flux only produces 3% of the oceanic heat flow. O’Nions & Oxburgh (1983) proposed that a boundary layer inhibits upward transport of helium from the “primordial” reservoir (assumed to be very deep) much more effectively than the transport of heat. They envisaged this boundary layer near 650 km depth, at the phase-change boundary separating the upper and lower mantle. This implies that the whole lower mantle is a kind of primordial helium reservoir.

The other alternative is that helium, along with CO₂, from degassing magmas, is trapped in the shallow mantle which, along with U+Th, contributes to the low ³He/⁴He of some upper mantle components, and the ubiquitous carbonatitic metasomatism. O’Nions & Oxburgh (1983) ignored secular cooling, the diversion of heat from continental regions toward the thin-lithosphere oceanic regions, and the delay of heat reaching the surface. Nevertheless, their results suggest that helium is not as mobile an element as generally thought, opening up the possibility that helium in various upper mantle components with differing He/U ratios, can diverge in situ in their isotopic ratios, removing the requirement for ancient or large gas-rich sources (see also Helium Fundamentals page)

A challenge for the two-reservoir model is to explain the respective concentrations of helium and other rare gases in the reservoirs. If OIBs come from an undegassed source, they would be expected to contain more helium than MORB glasses from the degassed upper mantle. However, OIB glasses typically have ten times less ³He than MORB (Fisher, 1985). This has been explained as degassing, but that cannot explain the higher He/Ne in MORB than in OIB. This observation has sometimes been called one of the Helium Paradoxes, and there are several (Anderson, 1998a, 1998b). The dynamics of mantle convection and melt segregation under ridges must be different from under oceanic islands and seamounts. Ridge magmas probably collect helium from a greater volume of mantle and blend magmas from various depths, during the melting and eruption process. This would even out the extremes in MORB basalts. Most workers, however, favor the two-reservoir model for the noble gases, although the evidence is far from definitive, and the assumption of an undegassed lower mantle causes more problems that it solves.

Reservoirs, components and recycling

The concepts of primordial and homogeneous degassed reservoirs in the Earth were challenged by Anderson (1993) and Meibom et al. (2003) who argue that there are mantle components with a large range of helium isotopic ratios – because of the distribution in ages and ³He/U ratios – rather than two reservoirs with distinctive compositions. The high “solar” Neon component has several possible sources including IDP and “old” gas trapped in depleted U+Th rocks or minerals.

“Primitive” signatures have been attributed to the subduction of IDP (Allegre et al., 1993; Anderson, 1993). These particles accumulate in ocean-floor sediments (Merrihue, 1964). Cosmic dust has very high ³He/⁴He ratios and [³He] and can fall to Earth without burning up in the atmosphere (Nier & Schlutter, 1990). In this way ocean-floor sediments develop a “primordial” helium isotope signature. This would have been even more true in the past if ET flux was higher then, and oceanic organic productivity lower.

The rare gases in cosmic dust particles are encapsulated in silicate and magnetite grains – the magnetite grains are easier to collect from the seafloor and have been most studied – which are relatively resistant to thermal degassing (Matsuda et al., 1990). Therefore, the cosmic helium in ocean-floor sediments might survive the subduction process and be transported into the shallow mantle, along with other “fragile” and “mobile” materials such as water and ¹⁰Be. In contrast, atmospheric rare gases trapped in ocean-floor sediments are very susceptible to thermal degassing.

Arguments against subduction of “primordial” noble gas components of cosmic and ET origin into the mantle include:

1. rapid diffusive loss from individual grains,
2. the low average ³He/⁴He ratio in current MORB and island arc basalts,
3. the low current ET influx of ³He, and
4. the large current rate of mantle ³He loss.

Diffusive loss of a component from a rock or sediment does not mean it is completely lost from the system. It could also become trapped in other host components in the upper mantle at subduction zones. The low average ³He/⁴He ratio of MORB can be explained by large volume averaging of a heterogeneous mantle with a wide range of ³He/⁴He values (Meibom & Anderson, 2003). 3. and 4. are uniformitarianism arguments, i.e., the current ET and mantle fluxes are assumed to have remained constant throughout Earth history and ancient sediments are assumed to have had the same composition as modern sediments. Many basalts with so-called primitive or primordial noble gas signatures show, additionally, evidence for atmospheric contamination, e.g., ⁴⁰Ar/³⁶Ar in some OIB is similar to the atmospheric value. This combination of surface characteristics with ones conventionally attributed to the deep mantle is a paradox in the standard model for noble gases.

If cosmic dust and oceanic sediments survive the “subduction barrier” against atmospheric rare gases, they have the potential to deliver He with a primordial signature into the mantle, probably its shallow levels (Allegre et al., 1993, Anderson, 1993). The ³He in OIB is a small fraction of the total ³He degassed from the mantle and in volume is comparable to the small amount of ³He brought onto the seafloor by IDP. However, much of the ³He in the mantle may have come in as part of the late veneer; it is subduction, or the Archean equivalent, at some 1 – 2 Ga that is relevant, not today’s flux (Anderson, 1993). Allegre et al. (1993) noted that the ³He/²⁰Ne ratio in cosmic dust is one to two orders of magnitude lower than ³He/²⁰Ne in the upper mantle. Helium has a much greater diffusivity than Ne, which would promote its preferential degassing from cosmic dust grains during subduction (Hiyagon, 1994) although this does not mean that it immediately escapes the mantle. It appears that subduction of cosmic dust having the properties of today’s dust cannot contribute more than a small fraction of the mantle ³He budget without causing excessive enrichment of ²⁰Ne in submarine glasses. The preferred explanation for the high ³He/⁴He components in both OIB and MORB mantle is, therefore, evolution in low-U+Th environments such as olivine cumulates (Natland, 2003) or depleted lithosphere (Anderson, 1998a, 1998b).

While Loihi, Iceland, Yellowstone and the Galapagos have the highest ³He/⁴He components – usually called the most primordial He compositions – some ocean islands such as Tristan, Gough, islands in the SW Pacific and the Azores, and some components in Yellowstone, Iceland and the Galapagos, have ³He/⁴He ratios lower (more radiogenic) than MORB. These low ³He/⁴He ratios require a component of radiogenic He from a long-lived U+Th-rich source such as recycled oceanic crust or sediments in the mantle, as inferred from lithophile isotope data. Parallel recycling of refractory peridotite or olivine-rich cumulates could then be responsible for high ³He/⁴He ratios and these components will not have enriched signatures for other isotopes. None of these components need to be recycled very deep into the mantle.

Some of the arguments against recycling involve steady-state or uniformitarian principles such as constancy of ET input, volcanic output and sedimentary and IDP compositions. Kellogg & Wasserburg (1990) assumed a steady state between supply and degassing in order to determine the residence time of He in the upper mantle. They argued that ridges are the principal sites of helium escape from the upper mantle (whereas hot spots would dominate in outgassing a deep, large, gas-rich reservoir in “the lower mantle”).

Another approach to this problem is to dismantle the two reservoir model entirely. Anderson (1989), Coltice & Ricard (1999), Anderson (2001), Meibom & Anderson (2003) and Seta et al. (2001) suggested that a “primordial” or “relatively undegassed” reservoir does not exist. Instead, they argued that all the mantle is degassed, but some parts (e.g., the MORB “reservoir”) are more enriched in radiogenic helium due to high-U+Th concentrations combined with great age. Anderson (2000) and Meibom & Anderson (2003) suggested that OIB are derived from a heterogeneous shallow mantle through sampling that is different from that which produces MORB. In contrast, other workers assume a lower-mantle plume source with very high ³He/⁴He and [³He], which must be isolated and preserved against convective homogenisation with the more radiogenic upper mantle.

The latter model has several problems. Mantle convection could allow large lumps of the lower mantle to be preserved intact, without being homogenised (van Keken & Ballentine, 1998; 1999) but realistic models cannot preserve lower-mantle domains large enough to preserve a significant primordial helium reservoir. On the other hand it is difficult to convectively homogenize the upper mantle (Meibom & Anderson, 2003). Extremely high ³He/⁴He data from Baffin Bay (³He/⁴He = 49.5 R/R_A) and Iceland (³He/⁴He = 42.2 R/R_A) require that the hypothetical ancient undegassed reservoir must be even more gas-rich than previously thought (Stuart et al., 2003; Foulger & Pearson, 2001). Higher ³He/⁴He and solar Ne ratios, in the standard model, require even less degassing. In alternate models, these components are viewed as having been isolated in [He]-poor, U+Th-poor environments rather than in [He]-rich ones.

Finally, low ³He/⁴He values observed at some hotspots may result from shallow mixing with radiogenic sources. The average ³He/⁴He value of the undegassed source might need to be even higher than proposed. In the recycling, heterogeneous model both the high- and low-³He/⁴He components arise from recycling, but the extreme values – high and low – are averaged out at ridges by large-volume mantle sampling. Primordial He may thus be present in the mantle but a large, coherent, ancient, primordial, undegassed region is unlikely.

Solar neon in the Earth

Evidence for non-atmospheric neon in the mantle was presented by Craig & Lupton (1976) on samples of MORB and volcanic gases. These are enriched in ²⁰Ne, as well as nucleogenic ²¹Ne, relative to ²²Ne. These ²⁰Ne-enriched components were attributed to primordial components in the Earth, possibly solar neon. Elevated ²⁰Ne abundances were also found in diamonds (Honda et al., 1987; Ozima & Zashu, 1988; 1991). Diamonds represent in situ solid samples of the mantle. Therefore, Ozima & Zashu (1988; 1991) suggested that diamonds sample a solar neon reservoir in the Earth, whereas the present-day atmosphere has been depleted in ²⁰Ne by mass fractionation. They argued that bombardment of the early Earth by radiation caused massive blow-off of a primitive solar-type atmosphere, leaving a residue enriched in heavy neon. Most MORB samples can be explained by three-component mixing of solar-type, atmosphere-type and nucleogenic neon (Figure 4).

Basalt glasses from Loihi and Kilauea reveal a range of Ne isotope ratios, stretching from atmospheric to ²⁰Ne-enriched compositions (Honda et al., 1991; Hiyagon et al., 1992). The enriched end of this array approaches the solar wind composition. Honda et al. (1991) and Hiyagon et al. (1992) attributed all Ne in the Earth’s interior to mixing between solar and nucleogenic isotopes. The arrays of MORB and OIB were attributed to variable atmospheric contamination of this solar + radiogenic mantle Ne. Loihi Ne samples were attributed almost entirely to atmospheric contamination.

Allegre et al. (1993) attributed these signatures to the subduction of cosmic dust particles accumulated in deep-sea sediments. These dust particles become impregnated with Ne from the solar wind during their exposure in space. Analysis of this material in the atmosphere and in deep sea sediment reveals ²⁰Ne/²²Ne ratios which span the range between atmospheric and solar compositions (e.g., Nier & Schlutter, 1990). Matsuda et al. (1990) suggested that these particles could survive the “noble gas subduction barrier” and deliver cosmic (solar) Ne to the deep mantle. This could explain the high ²⁰Ne/²²Ne ratios of submarine glasses without having to invoke primordial Ne in the Earth. Experimental studies by Hiyagon (1994) suggested that Ne would be completely extracted from cosmic dust (the magnetic portion) in a partial vacuum within 3 years at 500°C. But this assumes that no other host in the sediments or mantle can take up the gases, and that the non-magnetic fractions of cosmic dust are not more retentive. Evidence from Iceland (Dixon et al., 2000; Moreira et al., 2001) implies that some basalts, as well as oceanic sediments, contain essentially pure solar or IDP neon components, termed “primordial” by the authors.