In parts per million in the Earth's crust Uranium is around 1.8ppm and Gold 0.003ppm. Given that it takes far more energy to create Uranium than Gold, why is this?

  • 55
    $\begingroup$ The universe was shaken, not stirred -- seriously: neither the Earth nor much of anything in the universe is perfectly homogenized. BTW, as a matter of form, you should (1) provide a reference for your data, (2) specify whether that's ppm by atom count or mass or volume, etc. $\endgroup$ Commented Nov 4, 2014 at 13:39
  • 2
    $\begingroup$ Related: physics.stackexchange.com/q/141839 and physics.stackexchange.com/q/1918. $\endgroup$ Commented Nov 4, 2014 at 14:49
  • 4
    $\begingroup$ Shouldn't this be in earthscience.SE? $\endgroup$
    – Gimelist
    Commented Nov 5, 2014 at 7:11
  • 3
    $\begingroup$ Related: earthscience.stackexchange.com/q/102/725 $\endgroup$
    – Gimelist
    Commented Nov 5, 2014 at 8:44

2 Answers 2


Since gold is much more abundant in the universe than is uranium (by a factor of about 20:1)1, why is the situation reversed in the Earth's crust (by a factor of about 1:600)2? The answer lies in chemistry.

Uranium is chemically active. It readily oxidizes (pitchblende) and it readily combines with silicates. Uranium is a lithophile (literally, rock-loving) element3. It does not dissolve all that well in molten iron, and thus tended not migrate to the center of the Earth when the Earth differentiated. Uranium is a "high field strength element", one of the two classes of trace elements that are incompatible with the minerals that form the upper mantle4. When upper mantle rock undergoes a partial melt, incompatible elements such as uranium preferentially join the silicate melt rather than staying with the solid minerals. Over time, this magnifies the amount of uranium in the crust compared to that in the upper mantle5.

Gold on the other hand is rather inert chemically. It has little affinity to oxygen or sulfur. It does however readily dissolve in molten iron. Gold is a siderophile (literally, iron-loving) element3. Of the tiny bit of gold currently found in the crust, hardly any is primordial. Almost all of the primordial gold sank to the Earth's core when the planet differentiated. The gold currently found in the crust instead arrived in meteors that hit the Earth after the Earth had finished forming6.

The above assumes that the Bulk Silicate Earth (BSE) models of the Earth is basically correct, that the Earth formed from protoplanets and planetary embryos that had formed from material in the inner solar system, and that the proto-Earth differentiated into a core and primitive mantle. One prediction of these models is that the differentiation that created the Earth's core made the core strongly enhanced in siderophile elements and strongly depleted of lithophile elements, particularly so with regard to high refractory lithophile elements such as thorium and uranium. An opposing (not well accepted) model says that rather than being depleted of uranium, the Earth's core is uranium-enhanced, and to such an extent that there is a large georeactor at the very center of the Earth. These are testable hypotheses. Recent studies of geo-neutrinos are consistent with the BSE hypothesis, and simultaneously reject the possibility of a large georeactor at the center of the Earth7.


  1. Based on Lodders, "Solar system abundances of the elements." Principles and Perspectives in Cosmochemistry, Springer Berlin Heidelberg, 379-417 (2010), the abundance of gold to uranium by mass in chondritic meteorites is 18.1:1, 25:1 for the sun's photosphere. To one significant digit, this ratio becomes 20:1.

  2. From Lide, editor, CRC Handbook of Chemistry and Physics, 88th edition, the crustal ratio of uranium to gold is 675:1. From online resources such as web elements.com, I get ratios ranging from over 400:1 to over 600:1. I used 600:1.

  3. Victor M. Goldschmidt developed the concept of classifying elements as siderophile ("iron loving"), lithophile ("rock loving"), chalcophile (literally "ore loving", but Goldschmidt implied "sulphur loving"), and atmophiles ("air loving") in the 1920s. While Goldschmidt's initial concept of a siderophilic core surrounded by a chalcophilic layer surrounded in turn by a lithophilic outer layer didn't pan out, his classification scheme lives on. That uranium is a lithophile and gold is a siderophile is basic chemistry.

  4. There are two key classes of "incompatible elements": Those with an abnormally large ionic radius, and those with an abnormally large field strength. Uranium and thorium fall into the latter class.

  5. While the "incompatible elements" are lithophiles based on chemistry, they don't fit nicely in the crystalline structures that comprise typical rock. In rock undergoing a partial melt, incompatible elements such as uranium tend to migrate to the melt because of this structural incompatibility. Over time, plate tectonics has made the incompatible elements migrate to the Earth's crust.

  6. This is the conclusion of Willbold, et al., "The tungsten isotopic composition of the Earth/'s mantle before the terminal bombardment." Nature 477.7363: 195-198 (2011). Others disagree. One thing is certain: Gold is an extremely rare element in the Earth's crust.

  7. For example, see Bellini, et al., "Observation of geo-neutrinos." Physics Letters B 687.4:299-304 (2010), Fiorentini, et al., "Geo-neutrinos and earth's interior." Physics Reports 453.5:117-172 (2007), and a host of other recent papers on this topic.

  • 2
    $\begingroup$ I'm curious: regarding georeactors, is it possible at all that a planet with a georeactor could form naturally anywhere, even if Earth is not it? Or does this process rule this out on pretty much all planets? I'm guessing the answer is no, since silicon is so common (8th most common by mass in the universe) that any rocky planets are going to have a lot of it in them. $\endgroup$ Commented Sep 5, 2015 at 1:55
  • 2
    $\begingroup$ @mike4ty4 - en.wikipedia.org/wiki/Oklo . $\endgroup$ Commented Sep 5, 2015 at 7:45
  • 3
    $\begingroup$ Fun fact: The prefix "sidero-" is used in astronomy in two completely unrelated ways. Sometimes it comes from the Latin word "siderus," meaning "constellation," and sometimes from the Greek word "sideros (σίδηρος)," meaning "iron." It is not known whether these two words are related etymologically or whether their similarity is a coincidence. $\endgroup$
    – tparker
    Commented Jul 8, 2017 at 23:43
  • 1
    $\begingroup$ @Vorac - See table 2. It does not contain the word uranium, but it does list 92 U (i.e., uranium) as the very last entry in the table. Also see tables 6 and 8. $\endgroup$ Commented Aug 2, 2020 at 11:56
  • 1
    $\begingroup$ @tparker That's correct. $\endgroup$ Commented Dec 24, 2022 at 22:35

I'll be using the tabulated values from a Wikipedia article of abundance of elements in the Earth's crust. Gold has a tabulated value of 0.0031 ppm in mass for crustal abundance. Uranium has a tabulated value for 1.8 ppm in mass for crustal abundance. These figures are interesting in that uranium has an abundance of almost 500 times more ppm than that of gold even though uranium has an atomic mass 238.02891 and gold has an atomic mass of 196.96657.

One of the most prevalent theories for the low natural abundance of gold inside the crust yet it's easily accessible ores is that while the Earth was still molten, isotopes of naturally heavier elements naturally sank due gradually separating into the core. Dense material sinks depleting the crust of such elements.

A paper from Nature - "The tungsten isotopic composition of the Earth’s mantle before the terminal bombardment" - describes how ‘iron-loving’ elements such as gold should bring them into the core with them as well, yet there remains an over-abundance in the crust.

From a paper "The Cosmic Origins of Uranium" written by Professor Richard Arculus at the Australian National University, describes how multiple supernovae over a period from 6 billion to 200 million years ago, in addition to ten separate stellar sources are responsible for the relative abundance of Uranium in the solar system. Again, since the Earth was molten and at the melting point of nearly all elements, heavy elements including Uranium sank to the core of the Earth.


The present-day abundance of uranium in the 'depleted' mantle exposed on the ocean floor is about 0.004 ppm. The continental crust, on the other hand, is relatively enriched in uranium at some 1.4 ppm. This represents a 70-fold enrichment compared with the primitive mantle. In fact, the uranium lost from the 'depleted' oceanic mantle is mostly sequestered in the continental crust.

The extraction of Uranium from the mantle and upbringing into the crust has mutiple theories. Archus hypothesizes that over a period of 2 billion years,

1.) formation of oceanic crust and lithosphere through melting of the mantle at mid-ocean ridges, 2.) migration of this oceanic lithosphere laterally to a site of plate consumption (this is marked at the surface by a deep-sea trench), 3.) production of fluids and magmas from the downgoing (subducted) lithospheric plate and overriding mantle 'wedge' in these subduction zones, 4.) transfer of these fluids/melts to the surface in zones of 'island arcs' (such as the Pacific's Ring of Fire), 5.) production of continental crust from these island arc protoliths, through remelting, granite formation and intra-crustal recycling

these processes help explain why there is a larger than expected abundance of Uranium ore throughout the Earth's crust.

  • 2
    $\begingroup$ The Nature paper does not say that all heavy elements sank to the core during the planet's differentiation. The article specifically addresses the ‘iron-loving’ metals or siderophilic elements such as gold. Uranium is not siderophilic. The bulk silicate Earth model predicts that uranium is depleted rather than enhanced in the Earth's core. $\endgroup$ Commented Nov 4, 2014 at 16:22
  • 1
    $\begingroup$ @DavidHammen Thanks, I reflected that change. I like the new changes to your answer especially the part with the "high field strength element", and where it preferentially joins the silicate melt staying inside the upper mantle. I was unable to find such information. $\endgroup$
    – user43617
    Commented Nov 4, 2014 at 16:30