An alternative hypothesis for the origin of the Moon

R.J. de Meijer Stichting EARTH, de Weehorst, 9321 XS Peize 2, the Netherlands Department of Physics, University of the Western Cape, Private Bag X17, Bellville 7535, South Africa (E-mail: demeijer@geoneutrino.nl)

W. van Westrenen Faculty of Earth and Life Sciences, VU University Amsterdam, De Boelelaan 1085, 1081 HV Amsterdam, the Netherlands (E-mail: wim.van.westrenen@falw.vu.nl)

(Submitted to Earth, Moon and Planets, 23 October 2008; revised 21 July 2009)

Abstract: Recent high-precision measurements of lunar samples show a very high degree of similarity between the elemental and isotopic compositions of Earths mantle and the Moon. This similarity, which is exhibited by both light and heavy elements and their isotopes, is difficult to reconcile with the currently favoured giant impact hypothesis for lunar formation. We propose an alternative explanation for the compositional correspondence, namely that the Moon was formed from the ejection of terrestrial mantle material in a heat-propelled jet, triggered by a run-away natural georeactor at Earths core-mantle boundary. The energy produced by the run-away reactor supplies the missing energy term in the fission hypothesis for lunar formation first proposed by Darwin (1879). Our hypothesis straightforwardly explains the identical isotopic composition of Earth and Moon for both lighter (oxygen, silicon, potassium) and heavier (chromium, neodymium and tungsten) elements. Keywords: Moon formation, georeactor, Core-Mantle Boundary, isotope chronology, isotopic composition, angular momentum Earth-Moon system, missing energy, two-body system.
Abbreviations: BSE; Bulk Silicate Earth: CMB; CaPv; Calcium Perovskite: Core-Mantle Boundary: DRW; Darwin-Rise-Ringwood: Ma; Mega year: Ga; Giga year: TW; Terawatt (1012 W).

1. Introduction
Measurements of the oxygen (Clayton and Mayeda 1996; Wiechert et al. 2001), chromium (e.g. Shukolyukov and Lugmair 2000), potassium (Humayun and Clayton 1995), and silicon (Georg et al. 2007) isotopic composition of lunar rocks show that the bulk silicate Earth (BSE, i.e. mantle + crust) and the inferred bulk lunar composition show a very high degree of similarity. Recent compositional analyses of new lunar meteorites (Warren 2005) and high-precision isotopic ratio measurements on extinct radionuclide systems such as Hf-W (Touboul et al. 2007; 2009) in various lunar samples and Sm-Nd in terrestrial and lunar samples (Boyet and Carlson 2005) are reinforcing the notion (e.g. ONeill 1991 and references therein) of a very high correspondence between BSE and lunar rock compositions. Even the water content of the interior of the Moon could be as high as that of the Earths mantle (Saal et al. 2008). These results are very difficult to reconcile with the currently favoured giant-impactor hypothesis for the formation of the Moon (Hartmann and Davis 1975; Cameron and Ward 1976). This hypothesis is supported by high-resolution smooth-particle hydrodynamic simulations (e.g. Canup and Esposito 1996; Canup 2004; Canup 2008), although in detail the great majority of these simulations overestimate the resulting angular momentum of the Earth-Moon system by 1020%. At present, the best angular momentum match is obtained for collisions involving a retrograde rotating Earth (Canup 2008). Regardless of the collision parameters, all successful simulations indicate that approximately 80% of the Moon would originate from the impactor, with only 20% originating in the Earth.

Models of solar system evolution show that it is highly unlikely for the chemical composition of the Earth and impactor to be identical (e.g. Clayton 1993; Pahlevan and Stevenson 2007). The Moon therefore either consists almost entirely of material from Earths mantle, or complete isotopic homogenisation of terrestrial and lunar material occurred after the impact. Even if homogenisation in a turbulent exchange between partially molten and vapourised Earth and Moon shortly after the impact is able to explain the similarity in oxygen isotopes (as claimed by Pahlevan and Stevenson 2007, but contested by e.g. Zindler and Jacobsen 2009), it is improbable that such a mechanism would work for much heavier, refractory elements including Cr, Sm, Nd, Hf and W. A more likely possibility for the large degree of compositional similarity between the BSE and Moon is that the Moon derives directly from terrestrial material. In one of the first lunar formation hypotheses (Darwin 1879), the Moon indeed originated from a hot, fast-spinning Earth. In Darwins model, the centrifugal forces marginally exceeded the equatorial attraction, and the Moon was formed from resonant effects of solar tides. In the beginning of the 20th century Moulton (1909) and Jeffreys (1930) showed that the present angular momentum of the Earth-Moon system was too low for separation to have occurred, and that tidal frictions limited the height of any terrestrial tidal bulge. Subsequently, Ringwood (1960) and Wise (1963) updated Darwins explanation by including models for the thermal evolution of the Earth. They suggested that core-mantle differentiation led to a reduced moment of inertia of the Earth and hence to a larger angular velocity. In these modified models, the proto-Earth is a rapidly rotating body (rotation period of 2.65 h) with gravitational forces at the Earths surface only barely exceeding centrifugal forces. In this situation, a slight 4

increase in angular velocity would allow part of Earths equatorial mass to be ejected into space. The Darwin-Ringwood-Wise (DRW) model however still lacks sufficient energy to eject our Moon. As a result, this so-called fission hypothesis was abandoned. In this paper we re-examine the energetics of initial Earth-Moon separation and quantify the missing energy term. We show that this missing energy could be supplied by a supercritical georeactor in Earths core-mantle boundary (CMB) (de Meijer and van Westrenen 2008), producing sufficient heat to vaporize and eject part of the bulk silicate earth. This hypothesis straightforwardly explains the chemical similarities between Earth and Moon, and explicitly connects Moon formation with processes that took place following Earths early internal differentiation.

2. Earth-Moon dynamics
Consider the Earth and Moon as a gravitationally bound two-body system, where in the groundstate, just prior to separation, the Moon is a part of the Earth. In the excited state, just after separation, the Moon circles the Earth at a distance, rEM, which we consider as a variable. From radar-reflection data it has been established that rEM currently increases by about 4 cm per year (Williams and Dickey 2002). Assuming a linear extrapolation backwards in time to 4.5 Ga ago, yields rEM 2*108 m. The actual value of rEM is expected to be considerably smaller because the separation is an effect of tidal interaction, and tidal forces (proportional to the third power of the distance) were considerably stronger shortly after separation. Hence we will consider the upper value of rEM to be 2*108 5

m. Contrary to the present situation, at these distances the gravitational force on the Moon exerted by the Earth dominates over the attraction by the Sun by a factor of two or more. The gravitational force, Fg, between Moon (mass mM) and Earth (mE) can be expressed by:

Fg =

mM mE , 2 rEM

(1)

where is the gravitational constant. The corresponding gravitational potential energy, Eg,, for this two-body system becomes:

E g =

m M mE . rEM

(2)

with the assumption that Eg = 0 at infinite distance. The first Kepler law states that the orbit of an orbiting planet/moon is an ellipse. In our first-order approach we take a particular case, namely the circle, and according to the second Kepler law F will point to the centre of the circle. We attach the reference frame to the centre of the body with mE, where mE >> mM such that we may replace the reduced mass of the system by mM. In this case the centripetal force may be written as:

mM v 2 Fc = , rEM

(3)

where v is the velocity of mM relative to the centre of mass. Defining the binding energy of the two bodies as the sum of their internal energies, Eb, and denoting the rotational motion of the two bodies by their moments of inertia, I, and the rotational frequency, , we may write for the total energy, Etot, of the two body system: 6

1 m m 2 2 Etot = Eb + ( I M M + I E E + mM v 2 ) M E . 2 rEM
Since in a stationary orbit Fg + Fc = 0,

(4)

mM v 2 =
equation (4) reduces to:

mM m E , rEM

(5)

1 m m 2 2 Etot = Eb + ( I M M + I E E M E ). 2 rEM

(6)

The total angular momentum of the system, L, is according to Steiners theorem given by:

L = I M M + I E E + mM rEM v.

(7)

In our model we make a transition from a groundstate in which the system is a single body:
' Etot = Eb +

1 ' ' 2 ' ' I E ( E ) , and L = I E E , 2

(8)

to a two-body state with energy and angular momentum given by equations 6 and 7. In the transition the total angular momentum is conserved. The binding energy Eb in equations (4), (6), and (8) is energy released when assembling the respective bodies from their constituents (see below). In equation (8) we implicitly assume that the binding energy of the one- and two-body states are approximately equal. Since the surface energies are small compared to the volume energies, this approximation is justified. The energy difference between the one- and two-body states follows from the expressions for the energy in equation 8 and equation 6,

respectively. The radius of the orbit and the energy required for the transition are coupled since angular momentum conservation is imposed. As no external torques are present in the system, the angular momentum of the total system remains unchanged during the transition. Angular momentum of the proto-Earth in the one-body state is redistributed between the Earth and Moon in the two-body state. The binding energy Eb. which contains a volume and surface term, sets an upper limit to the rotational energy. For our first-order estimate we consider the Earth and Moon to consist of quartz and derive the volume and surface terms from lattice-energy considerations (Barron et al. 1982) and from the measured surface energy of silicon spheres (Hawa and Zachariah 2004). This corresponds to the mechanical energy required to break up all minerals present in the system into their constituting elements. As expected the volume term dominates and leads to Eb = -2.3*1031 J and -2.9*1029 J, for the Earth and the Moon, respectively. The angular momentum of the present Earth-Moon system is a vectorial sum of the angular momentum due to the Moons rotation around Earth (2.9*1034 kg m2 s-1) in a plane with a 5.1 angle with respect to the plane of ecliptics, and the Earth rotation around its axis (5.9*1033 kg m2 s-1) tilted 23.5. The resultant angular momentum of 3.5*1034 kg m2 s-1 has a tilt of 9.7. If angular momentum is conserved and the shape of the proto-Earth is assumed to be an oblate ellipsoidal with a longer axis twice as long as the shorter axes, Earths rotation period would be 2.3 h. For the Earth-Moon groundstate with the Moon inside the proto-Earth at r ~ 5*106 m, the sum of the rotational and gravitational energy totals 2.87*1030 J. This value is an order of magnitude smaller than the binding energy, confirming that the DRW-model is energetically unfeasible. 8

For the excited state, shortly after separation, the gravitational energy depends on the distance between the Moon and Earth. Equation (1) shows that this energy is directly related to the orbital velocity of the Moon, and hence to its rotational frequency. Since in the excited state the moment of inertia is well approximated by mM*(rEM)2, the angular momentum, L, rotational frequency and rotational energy of the rotating Earth are directly related to the Earth-Moon distance. The result also depends on the degree to which angular momentum is conserved. Table 1 shows the sum of the rotational and gravitational energy for a combination of rEM and L values ranging from 0.9 to 2 times that of the present. Because the value of L also directly affects the energetics of the groundstate, the bottom line of Table 1 indicates the corresponding binding energy of the groundstate. The values in Table 1 printed in italics show parameter combinations for which the Earth-Moon system is less bound than the groundstate, a prerequisite for an excited state. Therefore only those values represent physically relevant cases. Table 1 shows that the physically relevant cases, for the Earth-Moon system 4.5 Ga ago, are not sensitive to the assumption on the extent of the conservation of angular momentum. A value of rEM around 1*108 m seems to be reasonable for L ~ Lp. It is estimated that the EarthMoon system angular momentum decreased by a few to ten per cent since the Moon was formed (1.0 Lp < L < 1.1 Lp, Canup 2008) resulting in the energy required to launch the Moon of the order of 0.4-0.5*1030 J. For L = 2Lp the total energy approaches the estimated value of the binding energy of the Earth. At rEM = 1*108 m and L = Lp the gravitational potential energy Eg has a value of -1.2*1029 J and the corresponding velocity of the proto-Moon becomes, 9

according to equation (5), 1.9*103 m s-1. This corresponds to a rotation period of 92 h or ~3.8 d around the proto-Earth and an angular momentum carried by the proto-Moon of 1.3 kg m2 s-1, which corresponds to about 40% of the total angular momentum Lp of the Earth-Moon system. Immediately after the separation of the proto-Earth and proto-Moon, the rotation period of the proto-Earth becomes 5.7 h. Due to tidal forces, energy and angular momentum are transferred from the Earth to the Moon until they have, with their increased mass, the present properties. In our first-order approach we considered circular orbit. Elliptical orbits would constitute a refinement and impose some additional limitations set by the Roche limit. Tidal forces form an additional complication. The faster rotation of Earth compared to Moon induces friction, causing a substantial part of the tidal bulges to be offset from the line passing through the centres of Earth and Moon. As a consequence a torque exists between Moon and Earth, which slows down Earth rotation and enlarges the radius of the Moons orbit. In a two-body system, this effect would continue until the rotation of Earth equals the orbital period of the Moon, i.e. the Moon would be fixed over a certain position of Earth. Friction is an essential part of the tidal interaction. Because of frictional heat loss, energy is no longer conserved. The torque between Moon and Earth is internal and therefore angular momentum remains conserved. The fact that the tidal bulge promotes Moon in its orbit means that energy and angular momentum are transferred from Earth to the orbital motion of the Moon. A further complication comes from three-body dynamics. Incorporating the Sun into our analysis would have comparable implications for energy and angular momentum exchange, but would violate the assumption of the 10

conservation of energy and angular momentum within the Earth-Moon system. The three-body problem is a longstanding and complex problem in physics and astronomy. Wang (1991) presented a solution for the n 3 problem in terms of mathematical chaos theory. This theory describes dynamic systems that are extremely sensitive to minute changes in the initial starting conditions, which may lead to an exponential growth of disturbances leading to unpredictable system behaviour. In this paper we restrict ourselves to a first-order approach and left our these second order effects.

3. Nuclear excitation
We recently assessed the feasibility of natural georeactors (Herndon 1992; Hollenbach and Herdon 2004) in Earths CMB (de Meijer and van Westrenen 2008; see also Ball 2008). These georeactors are a natural consequence of concentrating significant proportions of our planets U and Th budgets in the CMB region, as suggested by current models of Early Earth evolution (see Boyet and Carlson 2005; Tolstikhin and Hofmann 2005; Tolstikhin et al. 2006). We showed that selective incorporation of this fissionable material by the mineral calcium silicate perovskite (CaPv) leads to a concentration of fissionable material which is only a factor of twenty lower than required for igniting and maintaining a nuclear breeder reactor, even if U, Th and CaPv are assumed to be distributed homogeneously throughout the CMB. Given the identification of many small-scale heterogeneities in the coremantle boundary region (e.g., the recent study of van der Hilst et al. 2007), it is clear that the CMB is far from homogeneous, making a concentration factor of 11

twenty not unreasonable. Our work concluded that natural georeactors in the CMB are feasible. Here we discuss the consequences of one of these reactors becoming supercritical, and providing the missing energy to eject the Moon. The size of the run-away reactor can be estimated if the assumption is made that the required excitation energy of ~0.5*1030 J is supplied by nuclear fission. In Table 2 the amounts of 232Th, 235U and 238U according to a commonly used Bulk Silicate Earth (BSE) compositional model is tabulated (McDonough 2003) for both the present and 4.5 Ga ago. From Table 2 one may calculate that fission of 1 kg of a natural mixture at t = -4.5 Ga of 232Th, 235U and 238U yielded 8.21*1013 J. Consequently, it requires fission of 6.1*1015 kg of the natural (U+Th) mixture to separate the Moon from the Earth at t = -4.5 Ga. The concentration of (U+Th) to allow a georeactor to become critical is estimated to be of the order of 150 ppm (U+Th) (de Meijer and van Westrenen 2008). Hence the corresponding mass of CMB material involved is of the order of 4*1019 kg. At a density close to 5*103 kg m-3, this mass corresponds to a sphere with a radius of approximately 120 km. Of course the shape of the real reactor is not necessarily spherical, but this calculation demonstrates that such a volume is fully compatible with our present understanding of the dimensions of the CMB. A mass of 6.1*1015 kg of the natural (U+Th) mixture corresponds to 1% of the fissionable (U+Th) in the BSE at t = -4.5 Ga. If approximately half of the BSE (U+Th) budget was concentrated in the CMB, as proposed by Tolstikhin and Hofmann (2005) and Tolstikhin et al. (2006), this corresponds to ~ 5% of the CMB (U+Th) content. This percentage does not seem unrealistic, and at the same time limits the number of possible georeactors of this size in the CMB.

12

In the following we sketch, in a qualitative way, a scenario for the ejection of the Moon. Quantitative models would require detailed dynamical calculations, but these will be hard to constrain due to the general lack of of data on the precise conditions in Earths deep interior 4.5 Ga ago. The heat produced by the run-away reactor is assumed to be produced faster than its dissipation by conduction or convection. With a heat capacity of 835 J kg-1 K-1, 0.5*1030 J produced by fissioning 6.1*1015 kg (U+Th) would raise the temperature of the corresponding reactor mass of 4*1019 kg from about 5000 K (a conservative estimate of the regular temperature of the CMB 4.5 Ga ago) to 1*1011 K. Even at a CMB pressure of 125 GPa, all materials will vaporise at this temperature. The vapour will want to expand and move upwards (in the direction of the largest density gradient). On its way to the surface, heat will be transferred to the surrounding materials of the overlying mantle and crust. If it is eventually transferred to a mass equal to that of the Moon (~7*1022 kg), the temperature of this lunar mass would be raised from an estimated mantle temperature of ~2,000 K to about 10,000 K. This temperature rise would be sufficient to vaporise most non-refractory materials. The upwards movement of the vaporized material towards the surface will be along the maximum density and minimal friction gradient and therefore more or less radially. Since it is replaced by denser material, the vapourised material gains gravitational potential energy, which is partially converted to kinetic energy. The material accelerates until it reaches a equilibrium velocity at which the buoyancy force equals the friction force. Friction will add to transfer of heat from the vaporised material to the surrounding mantle, causing an increase in the volume of the vaporised material. Near the surface overlying mantle and crust 13

material is launched into orbit. (This process is comparable to water-vapour bubbles ascending in a pan of boiling water.) In this separation process the Moon takes with it about 40% of the systems angular momentum (see section 2), causing a decrease in Earth rotation rate, and less elliptical shape. After this relaxation the binding energy of the Earth has increased and the system becomes too tightly bound to produce another moon. Strong tidal forces would lock the Moons rotation rapidly. A logical consequence of our hypothesis is that conservation of the total angular momentum with an orientation of 9.7 to the plane of eclipse (see section 2), leads to a present tilting angle of the Earths rotation axis of 23.5, if the Moon was emitted into its present orbiting plane with a tilt angle of 5.1. This requires that at the time of formation the Moon was ejected at -4.6 (slightly into the southern hemisphere). The precise angle depends among others on the path of least resistance through Earths mantle and crust of the material that forms the Moon. This result suggests the run-away georeactor was located near the equatorial plane, which is consistent with our hypothesis that centrifugal forces played a role in concentrating U and Th to levels sufficient to ignite georeactors.

4. Timing of Moon formation

If the Moon originates directly from the Earths mantle, the temporal evolution of the two bodies must be intimately connected. Our hypothesis is consistent with the timescales of Earth differentiation and lunar formation. Coremantle differentiation must have preceded Moon formation because of the relatively low iron content of the Moon. The timing of terrestrial core formation is derived from the interpretation of tungsten isotopic data for terrestrial and 14

meteorite samples. Most recent models date core segregation in the Earth at t = 30-50 Ma (Kleine et al. 2002; Yin et al. 2002; Kleine et al. 2004). Based on Hf-W analyses of lunar rocks Touboul et al. (2007, 2009) conclude that the Moon was formed at t = 50-150 Ma, after completion of most of the core-mantle differentiation of Earth. A second prerequisite for the presence of georeactors at the CMB is the availability of sufficient concentrations of uranium and thorium. As outlined in our previous paper (de Meijer and van Westrenen 2008), such elevated concentrations of U and Th accompany the formation of a hidden reservoir in the CMB. Evidence for the timing of the formation of this reservoir is provided by the 146Sm/142Nd chronometer. As both Sm and Nd are lithophile elements one expects no difference between the Sm/Nd ratio of mantle samples and undifferentiated meteorites. However Boyet and Carlson (2005) showed that terrestrial mantle samples have an enhanced Sm/Nd ratio. The mantle should therefore contain a complementary reservoir, with a low Sm/Nd ratio, which has remained isolated from the rest of the mantle since its formation. The only viable option for the location of this reservoir is the core-mantle boundary region. Using the 146Sm/142Nd chronometer, one arrives at a date for the formation of the hidden reservoir of around t = 30 Ma (Boyet and Carlson 2005). Again, this is consistent with the timing of Moon formation. The completion of core-mantle differentiation, the formation of a hidden reservoir, and the formation of the Moon all took place in a narrow time-interval. Hence one may question whether these events occurred independently or if some of the events are correlated. In our hypothesis this sequence of events is

15

necessarily correlated. In the giant impact hypothesis, the narrow time-interval is a coincidence.

5. Supporting evidence
The dominant supporting evidence for our hypothesis is the correspondence in isotopic and elemental composition between the BSE and (surface) lunar rocks. Positive identification of nuclear fission products from the georeactor in lunar material would strongly support our hypothesis. In our preceding paper (de Meijer and van Westrenen 2008) we quantitatively assessed a wide range of potential changes in isotopic abundances due to georector activity. We showed that helium and xenon isotope ratios are the primary messengers of georeactor activity. The magnitude of changes in these ratios depends critically on the proportion of supercritical georeactor material that is eventually ejected, a parameter that is poorly constrained at present. In principle, the strongest indicator would be the presence of high concentrations of 3He in the Moon. Unfortunately, the continuous influx of 3He from the solar wind means surface lunar rocks cannot be used to measure the indigenous lunar 3He levels. Any 3He found at greater depths in the Moon will be a strong indication of the involvement of a georeactor in the origin of the Moon. Elevated levels of 136Xe are also expected if georeactor fission products were incorporated in the ejected material that formed the Moon. Marti et al. (1970) do report a small excess of 136Xe, which Boulos and Manuel (1971) subsequently coupled to the activity of extinct 244Pu. We have previously shown (de Meijer and van Westrenen, 2008) that the amount of 136Xe is too large for
244

Pu to be the source of the surplus 136Xe in terrestrial gas wells, and that instead 16

136

Xe is an indicator of georeactor activity. Due to the similarity in composition

between Earth and Moon this argument holds even more strongly for the Moon, as, with a half-life time of 80 Ma, the 244Pu content at the time of Moon formation would already have been reduced by at least a factor of two compared to the initial terrestrial concentration. We interpret the measured xenon isotopic composition in the lunar surface sample analysed by Marti et al. (1970) in terms of a mixture between solar wind xenon and internal xenon. Our estimate is based on a very crude model in which the internal Xe abundance and isotopic composition is the sum of xenon produced in the run-away georeactor and the xenon present in terrestrial mantle material (Lodders 2003). The isotopic composition in the mantle material is taken from Busemann et al. (2000). Our estimate indicates that about 70% of the xenon in this lunar soil sample originates from solar wind. Analysis of soil and rock samples from greater depths will be more conclusive. As approximately 5% of the U and Th in the CMB is assumed to be involved in the run-away georeactor, one would at first glance expect a difference in the 235U/238U and Th/U ratios between Moon and Earth, which is not consistent with observations (e.g. Tatsumoto and Rosholt 1970). There are however a number of reasons why such differences are not too likely. The main reason is that in a breeder type georeactor, both 235U and 238U (as well as 232Th) disappear by conversion to fissile materials. The probability for interaction (mainly fission) by fast neutrons for these three isotopes is virtually identical (6.84, 7.10 and 7.00 barn at En = 1 MeV for 235U, 238U and 232Th, respectively, e,g, Rinard 1991). The actual change in the 235/238U or Th/U ratio will depend on the precise shape of the energy distribution of the neutrons, on the temperature and pressure at the 17

georeactor site, and on the presence of 244Pu. In first order there will hardly be any difference of the ratios between lunar and terrestrial samples. A second factor that will reduce differences in isotopic ratios is the dilution of the georeactor material by mixing with terrestrial mantle material. As pointed out in section 3 the mass of the initial georeactor is three orders of magnitude smaller than the final mass of Moon. A third aspect is the fact that U is a highly refractory element, meaning that it is one of the first elements to condense during the cooling of the vaporised georeactor materials on their way to the surface. In summary, no significant difference in 235/238U ratio is expected between lunar and Earth materials.

6. Conclusions
A run-away georeactor in the CMB can provide the missing energy source for the Darwin-Ringwood-Wise fission model for Moon formation. Our hypothesis provides an alternative explanation for the striking similarity in elemental and isotopic composition of the Earths mantle and lunar rocks, and is consistent with the sequence of differentiation events during our planets earliest history. Future Moon missions returning lunar samples from greater depths may contain conclusive evidence for the validity of our hypothesis. The 3He contents and xenon isotopic compositions in particular, will be a crucial test. Acknowledgements: Profs. Klaus Jungmann, Rob Timmermans and dr. Ricky Smit are gratefully acknowledged for the many discussions on the angular momentum of the Earth-Moon system. Prof. Dave Stevenson is thanked for his critical comments which led to improvements in the presentation of our model.

18

We are indebted to Prof. Walter Seifritz for his critical evaluation of georeactor criticality.

References P. Ball, Nature News, (2008) doi:10.1038/news.2008.822 T.H.K. Barron, J.F. Collins, T.W. Smith, G.K. White, Journal of Physics C 15, 4311 (1982) J.L. Birck, M. Rotaru, C.J. Allgre, Geochimica et Cosmochimica Acta 63, 4111 (1999) M.S. Boulos, O.K. Manuel, Science 174, 1334 (1971) M. Boyet, R.W. Carlson, Science 309, 576 (2005) H. Busemann, H. Baur, R. Wieler, Meteoritics and Planetary Science 35, 949 (2000) A.G.W. Cameron, W.R. Ward, Proc. Lunar Planet. Sci. Conf 7th, 120 (1976) R.M. Canup, Icarus 196, 518 (2008) R.M. Canup, Icarus 168, 433 (2004) R.M. Canup, L.W. Esposito, Icarus 119, 427 (1996) R.N. Clayton, Annual Reviews of Earth and Planetary Science 21, 115 (1993) R.N. Clayton, T.K. Mayeda, Geochimica et Cosmochimica Acta 60, 1999 (1996) G.H. Darwin, Phil. Trans. Roy. Soc. (London) 170, 1 (1879) R.B. Georg, A.N. Halliday, E.A. Schauble, B.C. Reynolds, Nature 447, 1102 (2007) W.K. Hartmann, D.R. Davis, Icarus 24, 504 (1975) T. Hawa, M.R. Zachariah, Journal of Chemical Physics 121, 9043 (2004) J.M. Herndon, Naturwissenschaften 79, 7 (1992) 19

R.D. van der Hilst, M.V. de Hoop, P. Wang, S.H. Shim, P. Ma, L. Tenorio, Science 315, 1813 (2007) D.F. Hollenbach, J.M. Herdon, Proc.Natl. Acad. Sci. USA 98, 11085 (2001) M. Humayun, R.N. Clayton, Geochimica et Cosmochimica Acta 59, 2131 (1995) H. Jeffreys, Monthly Notices Roy. Astron. Soc. 91, 169 (1930) T. Kleine, C. Mnker, K. Mezger, H. Palme, Nature 418, 952 (2002). T. Kleine, H. Palme, K. Mezger, A.N. Halliday, Science 310, 1671 (2005) K. Lodders, Astrophys. J. 159,1220 (2003) K. Marti, G.W. Lugmair, H.C. Urey, Science 167, 548 (1970). W.F. McDonough, in Treatise of Geochemistry, (Elsevier, Amsterdam, 2003), volume 2, chapter 18, R.J. de Meijer, W. van Westrenen, South African Journal of Science, 104, 111 (2008) F.R. Moulton, in Tidal and Other Problems, Carnegie Inst. Wash. Publ., 107, 77 (1909) H.St.C. ONeill, Geochimica et Cosmochimica Acta 55, 1135 (1991) K. Pahlevan, D.J. Stevenson, Earth and Planetary Science Letters 262, 438 (2007) P.M. Rinard, Los Alamos Technical Report NUREG/CR-5550, LA-UR-90-732, March 1991 A.E. Ringwood, Geochimica et Cosmochimica Acta 20, 241 (1960) A.E. Saal, E.H. Hauri, M. Lo Cascio, J.A. Van Orman, M.C. Rutherford, R.F. Cooper, Nature 454, 192 (2008) A. Shukolyukov, G.W. Lugmair, Space Science Review 92, 225 (2000) M. Tatsumoto, J.N. Rosholt. Science 167, 461 (1970).

20

I. Tolstikhin, A.W. Hofmann, Physics of the Earth and Planetary Interiors 148, 109 (2005) I. Tolstikhin, J.D. Kramers, Chemical Geology 226, 79 (2006) M. Touboul, T. Kleine, B. Bourdon, H. Palme, R. Wieler, Nature 450, 1206 (2007) P.H. Warren, Meteoritics and Planetary Science 40, 477 (2005) Q.D. Wang, Celestial Mechanics and Dynamical Astronomy 50, 73 (1991) U. Wiechert et al., Science 294, 345 (2001) J.G. Williams, J.O. Dickey, 13th International Workshop on Laser Ranging, October 7-11, 2002, Washington, D. C., 17 pages. D.U. Wise, Journal of Geophysical Research 68, 1547 (1963) Q. Yin, et al., Nature 418, 949 (2002) A. Zindler, S.B. Jacobsen, Lunar and Planetary Science Conference XV, 2542 (2009)

21

Table 1: The sum of the rotational and gravitational energy, in units of 1030 J, of the EarthMoon system 4.5 Ga ago as function of the Earth-Moon distance and the angular momentum in units of its present value (Lp).

rEM (108 m) 0.8 0.9 1.0 1.25 1.5 2.0 Groundstate

0.9 Lp 2.63 2.46 2.30 1.96 1.67 1.21 1.87

1.0 Lp 3.67 3.47 3.29 2.87 2.52 1.94 2.87

1.1 Lp 4.89 4.65 4.43 3.95 3.53 2.84 3.97

1.2 Lp 6.26 6.13 5.75 5.19 4.71 3.91 5.17

2.0 Lp 23.3 22.8 22.3 21.1 20.1 18.4 18.6

Table 2. Masses and isotopic abundances of Th and U isotopes in the Bulk Silicate Earth (BSE) (McDonough 2003)
232

Th

235

238

Total mass

t1/2 (Ga) m (1017 kg) (t = 0) Isotopic abundance (t = 0) m (1017 kg).(t = -4.5 Ga) Isotopic abundance (t = -4.5 Ga)

14.05 3.15 100% 3.94 100%

0.70 5.87*10-3 0.73% 0.52 24.3%

4.47 0.80 99.27% 1.62 75.7% 6.06 3.95

22