A critical multidisciplinary view on the speculative hypothesis of deep-earth georeactor and its impact on the global climate

Scientific literature is overwhelmed by hypotheses claiming to disprove generally accepted ideas about the causes of climate change. Until recently, a vast majority of climate revisionists just forced themselves not to consider human activity as a driving force of global warming, while the role of natural climate cycles has been overemphasized.

The concept of naturally emerged georeactor operating in a pulse mode brings us to the next level of climate revisionism. Authors of this hypothesis dare to associate cyclic climate changes with the repetitive occurrence of nuclear fission reactions in the deep Earth’s interior rather than with long-term variations in Earth's orbit and axial tilt, commonly referred to as the Milankovitch cycles [Milankovitch, 1941]. A proposed operating cycle of hypothetical deep-Earth georeactor includes: 1) exsolution of actinide (uranium (U) and thorium (Th)) particles from iron melt in the Earth's outer liquid core; 2) deposition of these particles onto the inner solid core; 3) initiation of nuclear fission as the actinide layer reaches a critical thickness; 4) termination of fission due to the dispersion of fissionable material within liquid core; 5) re-deposition of actinide particles onto the inner-outer core boundary and so on.

It was speculated that the alternation of glacial and interglacial periods in the Earth's geological history is synchronized with georeactor operating cycles. According to this hypothesis’s proponents, radiogenic heat from nuclear fission is transferred through the silicate rock mantle and warms the Earth’s crust, thereby triggering a massive release of carbon-containing greenhouse gases (GHG) from their natural reservoirs and subsequent global temperature increase. Although other georeactor concepts have been proposed in the literature before [Herndon, 1993; Rusov et al., 2007; Anisichkin et al., 2008; Ludhova et al., 2015; Meijer & van Westrenen, 2008], none of them suggested a pulsed mode of operation or considered the possible influence of radiogenic heat production on the Earth’s climate.

As demonstrated in this discussion paper, the feasibility of nuclear fission reactions in the deep Earth’s interior should be rejected for geochemical reasons, and the core georeactor hypothesis described above is the easiest to disprove. Being refractory litophile elements, U and Th should have been partitioned into the primitive mantle rather than into the core during metal-silicate segregation in the magma ocean (MO) stage of Earth's evolution [McDonough & Sun, 1995]. They are considered to form oxides at the mantle conditions, whereas their residual amounts appeared as pure metals within the core. Partitioning experiments in the laser-heated diamond anvil cell performed at high pressure-temperature conditions relevant to primordial MO have shown that no more than 2.5-3.5 ppb U could be dissolved in iron melt during liquid core formation [Chidester et al., 2017; Blanchard et al., 2017]. Such trace amounts of U tended to be further segregated into solid inner core as it crystallized.

Once exsoluted from the liquid portion of the core, U has been irreversibly incorporated into the solid one as pure grains and/or point defects in crystalline Fe. This scenario, confirmed by density functional theory calculations, rules out any chance for U particles to form a layer of critical thickness on the inner-outer core boundary [Botana et al., 2025]. Although such experimental and theoretical evidences are lacking for thorium, Th and U concentrations in the newly formed liquid core should have been of the same order due to the comparable values of metal-silicate partition coefficients [Faure et al., 2020]. Since both these elements are characterized by very low solubility in molten iron, Th was expected to exsolute from the remaining liquid core portion in the same manner as U. Thus, almost negligible content of actinides within the Earth’s core, along with their irreversible incorporation into its solid portion, leaves no room for the concept of core georeactor operating in a pulse mode, but what about the bulk of them segregated into the molten silicate phase during MO differentiation?

The current U and Th content in bulk silicate Earth (mantle plus crust, BSE) was estimated to be ~20 and ~80 ppb, respectively. Extrapolation of these values back to 4.5 Ga (the age of core-mantle differentiation) would result in ~54 ppb U and ~99 ppb Th in initial BSE (primitive mantle before crust formation). Such decrease in the concentration of these elements is due to their radioactive decay with time. Low relative abundances of actinides in initial BSE, constrained by the composition of chondritic meteorites regarded as the main building blocks of the Earth, did not allow for U and Th oxides to crystallize from silicate melt as pure phases. Instead of such crystallization, these oxides were prone to be incorporated into the lattice of mantle silicate minerals such as CaSiO3 perovskite (Ca-perovskite) [Gautron et al., 2006; Gréaux et al., 2009]. Apart from other conditions required for the initiation and sustained operation of hypothetical mantle georeactor, there should exist concentration factors of several orders of magnitude to reach criticality without the formation of distinct mineral phases by actinide compounds.

First of all, U and Th oxides are thought to reside in the lowermost part of the mantle (so-called D" layer) which comprises 5 wt% (percentage by weight) of initial BSE and is almost unaffected by convective processes [Tolstikhin et al., 2006]. Assuming that D" layer stores one-fifth of the total BSE inventory of actinides, we easily concluded that this geochemical reservoir must be enriched with U and Th by a factor of four. Then, these elements have been found to be incorporated much more readily (by a factor of 104-105) into the crystal lattice of Ca-perovskite than of the other lowermost mantle minerals (ferropericlase and post-perovskite) [Walter et al., 2004; Corgne et al., 2005]. Since Ca-perovskite phase constitutes only ~5 wt% of D" layer, a further 20-fold enrichment in actinides was to be achieved.

The resulting concentration factor of ~80 corresponds to ~4.3 ppm U and ~7.9 ppm Th upon the formation of Ca-perovskite reservoir within D" layer. These values are still several orders of magnitude less than those required for georeactor initiation. If we take into account the presence of plutonium (Pu) and its role as a source of fast neutrons, criticality conditions for nuclear fission to occur could be met at much smaller local concentrations of U and Th, but an additional concentration factor of ~20 is necessary to enable sustained operation of georeactor [Meijer & van Westrenen, 2008]. No such factors have been identified at the lowermost mantle conditions if we leave aside unproven speculations.

Thus, the geochemical fate of actinides provides no room for deep-Earth georeactor to emerge neither in the core nor within the mantle. The only geochemical reservoir where naturally occuring nuclear fission reaction could be initiated is the upper layer of the Earth’s crust composed of sedimentary rocks. The differentiation of mantle-derived melts within magma chambers followed by hydrothermal transport of U from enriched residual melts to the upper crust resulted in the formation of U ore deposits.

Spontaneous nuclear fission is permitted to occur only in high-grade ores with U content more than 10 wt%. In addition to this requirement, a number of additional conditions should be met to reach criticality in the rich zone of ore. As for the sandstone-type ore deposits, this zone must be quite thick (> 0.5 m) and chopped up by tectonic faults in order to provide the entry of water acting as a neutron moderator [Naudet, 1991]. All these conditions were satisfied in Oklo deposits (Gabon, South Africa) where the only known natural nuclear reactor on Earth operated ~2.0 Ga ago for ~100 Ma. Since then, there was no chance for such reactors to initiate and operate as the ²³⁵U/²³⁸U ratio became too low to support criticality [Gauthier-Lafaye & Weber, 2003].

Neither deep-Earth georeactors nor Oklo-like natural reactors, even if they existed, wouldn’t be able to affect the Earth’s heat balance. The power output of the latter (~100 kW) was estimated to be quite negligible compared to the total heat input from the Earth’s interior (43-49 TW). According to theoretical estimates, the operation of hypothetical core georeactor would contribute up to 30 TW to geothermal heat flow (GHF) if the radiogenic heat was transferred by mantle convection to the Earth’s surface over geologically relevant timescales.

Nevertheless, convective flow velocities correspond to plate tectonic ones and do not exceed 1-10 cm·year-1, although mantle plumes could ascend from the core-mantle boundary two or more orders of magnitude faster (100 cm·year-1) [Bercovici, 2010]. Anyway, these upwelling processes are so slow that excessive radiogenic heat (ERC) must be dispersed into the surrounding mantle rocks instead of being transferred to the Earth’s crust. Even if ERC reached the Earth’s surface without losses, ERC-induced GHF alterations wouldn’t be synchronized with georeactor operating cycles due to the different timescales of heat production and heat transfer.

Moreover, warming of the Earth’s crust by enhanced GHF wouldn’t have been resulted in the release of GHG from their natural reservoirs into the atmosphere, i.e., carbon dioxide from the ocean water column and methane from deep sea gas hydrates (DSGH). As for the latter, it has been shown that DSGH destabilization requires an extreme warming of seafloor (e.g., from 1-2 to 14-15 ºC at the depth of 1200 m!) [Ruppel & Kessler, 2017] that couldn’t be provided by geothermal activity. Finally, one could only imagine the occurrence of such unrealistic event, but even in this case methane released from DSGH would be oxidized in bottom sediments and the overlying water column before it entered the atmosphere [Reeburgh, 2007].

Thus, the hypothesis of deep-Earth georeactor and its impact on the global climate is based on a multitude of poor assumptions that contradict a number of studies from several scientific disciplines, from geochemistry to climatology.