Roadmap | The review process | Home
   Helium Fundamentals
Helium: Fundamental models

Don L. Anderson1, G. R. Foulger2 & Anders Meibom3

1Seismological Laboratory, California Institute of Technology, Pasadena, CA 91125

2Dept. Geological Sciences, University of Durham, Durham DH1 3LE, U.K.

3Laboratoire d'Etude de la Matiere Extraterrestre, USM 205 (LEME), Case Postale 52, Museum National d'Histoire Naturelle, 61 rue Buffon, 75005 Paris


Helium is rare in the Earth but has become an important geochemical tracer since is it widely considered to be the only unambiguous geochemical indicator of a lower mantle and plume component in surface rocks. In practice, observations of 3He/4He higher than ~ 10 times the atmospheric value are generally interpreted as evidence for a plume from the lower mantle, even in the absence of supporting data. This, however, is strictly an assumption.

There is evidence that high-3He/4He may originate in the upper mantle. This model is more consistent with other observations at “hotspots” where a shallow-mantle origin is strongly supported and suggests that high-3He/4He cannot be used to prospect for plumes.

3He is a so-called primordial isotope. It was made in the Big Bang and incorporated into Earth during its initial accretion and in the subsequent long-term acquisition of “late veneer” material. 3He is not produced in any large quantities by radiogenic decay, and is thus not being added to Earth’s inventory at a significant rate. Nevertheless, a small amount is constantly being added to the surface of the Earth by interplanetary dust particles [Anderson, 1993] and by cosmic rays. This so-called “cosmogenic” 3He may be important in rocks that have lain at the surface of the Earth for long periods.

Figure 1: Industrial helium is mostly 4He and is mined from helium mines

4He is a product of alpha decay of U and Th, and accumulates over time. This accumulation is most rapid in rocks that are rich in U+Th, but may be very slow in rocks that contain little U+Th. The U+Th content of mantle rocks and recycled material varies by 3 or 4 orders of magnitude and the opportunity therefore arises to develop large variations in 3He/4He ratios.

The Earth is constantly degassing, which transports helium from the crust and mantle into the oceans and atmosphere. Because it is such a light atom, helium escapes from Earth and is thereby continually lost from the atmosphere. The approximate lifetime of helium in the atmosphere is ~ 1 to 2 Myr.

The absolute abundance of helium in rocks is difficult to interpret since helium is so mobile. Thus, the 3He/4He ratio (R) is usually used as a proxy for 3He content. R is generally expressed as a multiple of the present-day atmospheric 3He/4He ratio, Ra, which is 1.38 x 10-6.

The average value of 3He/4He in Earth when it originally formed may have been ~ 200 Ra, but it has decreased subsequently as a result of the addition of 4He by U+Th decay over the 4.5-billion-year lifetime of the planet. The amount by which it has decreased in a given rock sample is related to the integrated U+Th content of the sample and, importantly, the absolute abundance of helium. For example, if 4He is added at a given rate to a sample rich in helium, the value of 3He/4He will change only slowly. On the other hand, if the sample is poor in helium, its 3He/4He ratio will decrease rapidly. Because rocks in Earth are continually and variably experiencing degassing and metasomatism (which may involve the addition of gas), even if U+Th is known, the rate of change of 3He/4He with time cannot be calculated accurately. The ratio in the atmosphere is also constantly changing because it represents a balance between extra terrestrial input, degassing and atmospheric escape.

Figure 2: Trade-off curves between U/4He and time, which give high and low 3He/4He reservoirs [from Anderson, 1998b].

The observed value of 3He/4He varies in terrestrial rocks. Some typical values for R/Ra are:

  • Continental rocks (high-U+Th) << 1
  • Commonly assumed for mid-ocean ridge basalt (MORB) 8 ± 2
  • Average spreading ridge basalt 9.1 ± 3.6
  • Ocean island basalt (OIB) (“hotspot” rocks) ~ 5 - 42

The highest, non-cosmogenic value for R/Ra reported for “hotspot” rocks anywhere on Earth (42) is from Iceland [Breddam & Kurz, 2001].

The standard model

Values for 3He/4He greatly exceeding those typical of MORB were first observed in some basalts from Hawaii. It was suggested that they were indicative of a lower mantle component, since Hawaii was postulated to be underlain by a plume, and plumes were assumed to originate in the lower mantle to explain the apparent fixity of hotspots [Morgan, 1971]. It was thus concluded that the lower mantle is characterised by high 3He/4He – perhaps 25 or 30 Ra. This number has been continually upward-adjusted as higher and higher 3He/4He values have been found, e.g., from Iceland, and in ancient rocks such as flood basalts and komatiites.

A model for helium has been developed with a depleted, degassed and homogenized upper mantle, “the convecting mantle”, and a lower mantle that is little- or undegassed, and contains much more 3He than the upper mantle. By this reasoning, the high 3He/4He ratios were attributed to an abundance of 3He in the “primordial undegassed” mantle.

This model is flawed for a number of reasons:

1. Mass balance calculations show that it predicts an unreasonably high concentration of 3He, relative to other volatiles and incompatible elements, in the lower mantle [Anderson, 1989]. This follows because over the lifetime of Earth, radioactive decay of U+Th in the lower mantle has generated a large amount of 4He. In order for the value of 3He/4He in the lower mantle to have remained high, the concentration of 3He there must be high. Calculations show that it is predicted to be about an order of magnitude less than that in chondritic meteorites, relative to refractory elements [Kellogg & Wasserburg, 1990]. This is contrary to expectations that it should be several orders of magnitude less (see Figure at right). The problem of the high predicted relative concentration of 3He in the lower mantle becomes greater as the maximum 3He/4He observed at the surface increases. Most recently, a value of 3He/4He > 50 Ra was reported from Baffin Island [Stuart, 2003], which is far higher than the value of ~ 30 Ra assumed by Kellogg & Wasserburg [1990].

Figure 3: Abundances of elements in the crust+mantle relative to those in chondritic meteorites (from Anderson, 1989; Figure 8.2)

A high abundance of [3He] in the lower mantle is also at odds with high-temperature models of planetary accretion. The Earth is thought to have accreted from meteoric material, but subsequently huge amounts of heat were generated by gravitational compaction, segregation of the core, and impact with a Mars-sized body that caused the Moon to form. The Earth thus went through one or more phases where it was largely or completely molten. Extensive degassing would have occurred at these times, which is supported by the observation that Earth is strongly depleted in volatiles such as Na, K, Cl, CO2, H2O, and Rb (see Figure). By this reasoning, a model whereby a large part of Earth is undegassed is unlikely to be right. Furthermore, mass balance calculations show that about 70% of the mantle must be depleted in the volatile and crust-forming elements, and this is consistent with the amount of 40Ar in the atmosphere [Anderson, 1989]. The volume above the 650 km discontinuity, the usual choice for the depth to the top of the depleted, undegassed reservoir, is only 30% of the mantle. This choice therefore cannot be correct, since most of the mantle must have contributed to the incompatible elements in the crust and the 40Ar in the atmosphere.

2. If the concentration of helium in the lower mantle is very large – much larger than in the upper mantle, then helium abundances in OIB are predicted to be much higher than in MORB. This is not observed. The helium abundances in OIB are 2 – 3 orders of magnitude less than in MORB. It has been suggested that this is due to preferential degassing of OIB as a result of their shallow depth of eruption. This can be checked by looking at the relative concentrations of the heavier noble gases.

The heavy noble gasses are expected to have a greater tendency to degas upon eruption than helium, so He/Ne, He/Ar etc. should be higher, the more degassed a rock is. It is found that He/Ne and He/Ar are lower in OIB than in MORB, however, suggesting OIB is less degassed than MORB, not more [Moreira & Sarda, 2000; Ozima & Igarashi, 2000]. This is supported by CO2:He systematics at Hawaii. CO2 is a helium carrier, and the quantities of CO2 being degassed at Hawaii are not consistent with the source having a high abundance of helium. Most of the 3He being degassed from the mantle is from mid-ocean ridges and island arcs. The model whereby OIB were originally high in helium, which has been lost by degassing accompanying eruption, is therefore not supported – mid-ocean ridges exhale more 3He than hotspots [Anderson, 1998a; Anderson, 1998b].

Figure 4: 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].

Figure 5: 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, 2000a).

An alternative model

In response to a growing body of observations that support shallow sources for some high-3He/4He hotspots, models that attribute an upper-mantle origin to high 3He/4He have been developed. It has been suggested that high 3He/4He results from a deficit of 4He consequential to storage of old helium in a low time-integrated U+Th host rock [Anderson, 1998a; Anderson, 1998b]. The controlling parameters are then U/3He, U/Th and time. There are several lines of evidence that support this model:

a) high 3He/4He is observed in Samoan xenoliths that are known to be of upper mantle origin,

b) high 3He/4He has been measured in diamonds known to have been mined from pipes. (High 3He/4He has been reported in diamonds of unknown origin, but in these cases it is has been suggested that they may be “detrital” diamonds, i.e., they may have lain on the surface for a long time and acquired “cosmogenic” 3He. Although this is conjecture, it is safer to use diamonds known to be from pipes.)

c) high 3He/4He is observed at Yellowstone, where extensive work has provided a strong case that the magmatic system there is lithospheric only [see Yellowstone page & Christiansen et al., 2002].

A possible low-U+Th host material is the residuum left after basalt melt is extracted from mantle peridotite [e.g., Brooker et al., 2003]. Residuum contains only traces of U+Th, and thus helium stored for hundreds of millions of years would preserve its older, higher 3He/4He ratio little changed [Anderson, 1998a; Anderson, 1998b]. One model for high 3He/4He involves trapping of gases from ascending MORB magmas in recycled oceanic lithosphere or olivine-rich cumulates.

Another possible host is individual olivine crystals [Natland, 2003; Brooker et al., 2003]. These grow so as to encapsulate gas bubbles containing He, but olivine itself contains essentially no U+Th. Alpha particles generated by U+Th decay in surrounding minerals have insufficient energy to penetrate olivine crystals, and thus the encapsulated helium is effectively shielded from the addition of 4He at the time of decay. Diffusion of 4He probably has little effect also. Solid state diffusion coefficients for most lithophile elements (including U and Th) are of the order of 1013 cm2s-1 and even on the time scale of the age of the Earth, this only allows a characteristic length scale of diffusion of the order of a few kilometers. On time scales less than a few 100 My diffusion should thus have little effect in domains of moderate size. Furthermore, diffusion is driven by differences in chemical potential, not simply by concentration gradients, and thus high partition coefficients can block diffusion even in the presence of large concentration gradients. He is volatile and highly soluble in trapped CO2-rich bubbles in olivine, but isessentially insoluble in olivine itself. Thus, even though there might be a high concentration of He in a trapped bubble inside an olivine crystal, He will not tend to diffuse out. 4He would have a greater tendency to diffuse into a trapped bubble, but there would be little 4He present in thick, olivine-rich cumulate lithologies from which melt, and thus U+Th, had been squeezed out. Unradiogenic He trapped in olivine crystals in this way will be well protected from the addition of 4He, even over long time scales.

This suggests a “He in olivine” model, whereby noble gases are trapped in olivine and pyroxene in a cumulate olivine-gabbroic layer that forms the lowermost oceanic crust. This layer comprises densely packed cumulates that have compacted and squeezed out the interstitial melt, thereby expelling essentially all the U+Th. It may be subducted, and if young and warm will be buoyant and remain in the upper mantle. When it is once more transported into a melting region, it will produce voluminous, high-3He/4He melts.

The resulting He isotope composition of the melt will depend on the age and abundance of the olivine-gabbroic material, how much helium is trapped, and how much 4He is generated in the surrounding rock. Stochastic sampling of such a source would predict Gaussian distributions of 3He/4He isotope ratios similar to those expected for MORB but with more variability and no global correlation with large-ion lithophile elements and isotope systems, though local correlations may occur. This kind of distribution is precisely what is observed [Meibom et al., 2003].

A schematic diagram showing a model for the evolution of helium isotope ratios in the Earth is shown in Figure 6.

Figure 6: The Primordial mantle curve shows the growth of 3He/4He in a hypothetical primordial reservoir from Big Bang values (right) to present (left). This is how the lower mantle is predicted to have evolved in the Standard Model. At various times magma is extracted (MAGMAS). In this magma, or the lavas formed when it freezes, 3He/4He evolves faster because U and Th were preferentially partitioned into the magma and gas was lost to the atmosphere or retained in the residual mantle. Simply put, differentiation and degassing changes the 3He/(U,Th) ratio of the products and resulting evolution of 3He/4He. The magmas lose more gas as they rise to shallow depths, decreasing the He/U ratio still further and increasing the He/Ar and He/Ne ratios. Some of the gas lost ends up in olivine-rich cumulates and gas inclusions in peridotites, imparting to them relatively high He/U and 3He/4He. Ancient 3He/4He ratios may thus be frozen into whatever material the gas ends up in, while the residual degassed magma goes on to evolve to low values. The trajectories of He (and CO2) and U (and Th) following melting and degassing events are shown schematically. Each melting/degassing event gives this kind of fractionation. The dashed lines indicate schematically how ancient 3He/4He ratios are frozen in as He is separated from U and Th and stored elsewhere.

The MORB and OIB histograms are plotted at arbitrary positions on the time axis but represent current values. The samples that make up the OIB field represent a mix of magmas and gases of various ages stored in the mantle. It is closer to the actual mantle distribution than the suite of values measured in MORB because OIB melt batches are smaller than MORB batches. Large samples of material with OIB-like 3He/4He distributions, homogenised, will have a MORB-like 3He/4He distribution (Anderson, 2002b) because large-batch homogenising supresses outliers. For example, the mantle today might contain regions where R/Ra varies from, say, 1 to 100, but the variation in MORB might be no more than 5 to 10 simply as a result of sampling and homogenising. OIB extracted from the same mantle might have R/Ra values that vary from, say, 2 to 30, as a result of the extraction of smaller-volume samples.


The model whereby high 3He/4He is attributed to a lower-mantle source, and is thus effectively an indicator of plumes from the lower mantle, is becoming increasingly untenable as evidence for a shallow origin for many high-3He/4He hotspots accumulates. Shallow, low-4He models for high-3He/4He are logically reasonable, cannot be ruled out, and need to be rigorously tested if we are to understand the full implications of this important geochemical tracer.


and for those who are really into all this.....

  • Anderson, D.L., The statistics of helium isotopes along the global spreading ridge system and the central limit theorem, Geophys. Res. Lett., 27, 2401-2404, 2000.
  • Anderson, D.L., The statistics and distribution of helium in the mantle, Int. Geology Rev., 42, 289-311, 2000.
  • Brooker, R.A., V. Heber, S.P. Kelley and B.J. Wood, Noble gas partitioning behaviour during mantle melting: A possible explanation for “The He Paradox”?, EOS Trans. AGU Fall Meet. Suppl., Abstract, V31F-03, 2003.
  • Dodson A., Kennedy B.M., and DePaolo D.J., Helium and neon isotopes in the Imnaha Basalt, Columbia River Basalt Group: Evidence for a Yellowstone Plume Source, Earth Planet. Sci. Lett.,150, pp. 443-451, 1997.
  • Ellam, R.M. & F.M. Stuart, Isotope geochemistry of the North Atlantic Igneous Province: a Tertiary to Recent record of a mantle plume, J. Petrol., 141, 919-932, 2000.
  • Foulger, G.R. and D.G. Pearson, Is Iceland underlain by a plume in the lower mantle? Seismology and helium isotopes, Geophys. J. Int. (Fast Track), 145, F1-F5, 2001.
  • Graham, D. W., J. E. Lupton, F. J. Spera & D. M. Christie, Upper mantle dynamics revealed by helium isotope variations along the Southeast Indian Ridge, Nature, 409, 701-703, 2001.
  • Graham, D. W., Noble gas isotope geochemistry of mid-ocean ridge and ocean island basalts; characterization of mantle source  reservoirs. In: Noble Gases in Geochemistry and Cosmochemistry, eds. D. Porcelli, R. Wieler & C. Ballentine, Reviews in Mineralogy and Geochemistry, Mineral. Soc. Amer., Washington, D.C., pp. 247-318, 2002.
  • Hanan, B.B. & Graham, D.W., Lead and Helium Isotope Evidence from Oceanic Basalts for a Common Deep Source of Mantle Plumes, Science 272, 991-995, 1996.
  • Ozima, M. & F.Podosek, Noble Gas Geochemistry, Cambridge University Press, New York, 2002.
  • Stuart, F.M., R.M. Ellam, P.J. Harrop, J.G. Fitton & B.R. Bell, Constraints on mantle plumes from the helium isotopic composition of basalts from the British Tertiary igneous province, Earth Planet. Sci. Lett., 177, 273-285, 2000.

Volcano noble gas database

A new USGS online database: USGS-NoGaDat – A Global Dataset of Noble Gas Concentrations and Their Isotopic Ratios in Volcanic Systems, by A.A. Abedini, S. Hurwitz, and W.C. Evans, U.S. Geological Survey DigitalData Series, 202.

This database contains almost 5,000 entries of published information on noble gas concentrations and isotopic ratios from volcanic systems in mid-ocean ridges, ocean islands, seamounts, and oceanic and continental arcs. Where available the isotopic ratios of strontium, neodymium, and carbon were also included. The database is sub-divided both into material sampled (e.g., volcanic glass, different minerals, fumarole, spring), and into different tectonic settings (MOR, ocean islands, volcanic arcs). Included is also a large reference list.

last updated 2nd September, 2006