Goldschmidt's classification of the elements

Saturday, May 24, 1997

Eric Reardon

Within the first two weeks of an introductory geochemistry course, a student usually learns the meaning of the words 'lithophile', 'siderophile', 'chalcophile', and 'atmophile'. These terms were introduced in the early 1920's by a remarkable scientist. His name was V. M. Goldschmidt and it was his way of classifying the elements according to their affinities (philos - love) for various earth materials. Thus a siderophile element would be one with an affinity for iron metal. Before discussing this classification scheme in detail, let's learn more about its developer.

Victor Moritz Goldschmidt was born in Zurich, 1888. By the time his family moved to Oslo when he was thirteen, he had also lived in Amsterdam and Heidelburg. The move to Oslo was so that his father could take up the Chair of Chemistry at what is now the University of Oslo. Up until 1905, Oslo was part of Sweden and called `Kristiania'. From his own interest in Earth Sciences and likely because of his father's chemistry influence on him, Goldschmidt carved out a research area that fused together both of these disciplines. Even though the word 'Geochemistry' was coined a century earlier by C. F. Shonbein, another Swiss-born chemist and inventor of gun cotton, most geochemists today recognize Goldschmidt as the Founder of this modern field of Earth Sciences.

Goldschmidt completed his Ph.D. thesis in 1911 at the age of 23; was appointed Director of the Mineralogic Institute of the University of Oslo three years later; and at the age of 29, was appointed Chair of the Norwegian government's 'Commission on Raw Materials' as well as the Director of its laboratory. This meteoric rise to prominence within the Norwegian scientific community simply reflected Goldschmidt's unbounded passion, ability and prodigiousness in his research discipline.

Goldschmidt made many important scientific contributions during his lifetime. Perhaps the most important was in providing an understanding of the geochemical behaviour of rare earth elements and the factors which control their distribution in mineral phases during magma crystallisation. It was Goldschmidt who first estimated the ionic sizes of the rare earth elements and discovered the somewhat paradoxical decrease in their ionic size with increasing atomic number -- the now familiar 'Lanthanide Contraction', a term which he coined. However, as intimated in the opening line of this article, the most noted of Goldschmidt's contributions was his division of the periodic table into siderophile (iron-loving), lithophile (rock-loving), chalcophile (chalco indicates copper but Goldschmidt wanted to imply sulphur-loving), and atmophile (atmosphere-loving) elements. This classification scheme, first suggested by Goldschmidt in 1922, is illustrated in the accompanying periodic table. It represents the results of careful study of the elemental compositions of phases within two seemingly dissimilar materials: meteorites and sulphide ore smelting products.

In most meteorites, there are three principle phases present: an iron-nickel metal alloy, a troilite or iron sulphide phase, and a silicate phase. The silicate phase has a mineral assemblage that closely resembles mafic igneous rocks on earth. In some meteorites, there is evidence that all three phases originally existed as immiscible liquids at a higher temperature. Goldschmidt thus reasoned that an analysis of these phases should indicate the relative preference of these phases for individual elements, which would provide a general basis for classifying elements. Goldschmidt also had recourse to a large amount of analytical data on solid material produced from the smelting of the copper shale ore (Kupferschiefer) at Mansfeld, Germany. During the smelting operation, three immiscible liquids develop and solidify: the iron melt; the copper sulphide matte; and the silicate slag. Detailed analyses of the composition of these materials by Goldschmidt, and also by Ida and Walter Noddack from Berlin, confirmed the similar affinities of meteorite and smelter product phases for a wide range of elements.

Goldschmidt had a modern view of meteorites. He considered them derived from the asteroid belt and represented material that had failed to coalesce into a planet, rather than as material produced from the breakup of a larger planet. The close association of iron, sulphur and silicate phases in meteorites, unlike any terrestrial rock, was clear evidence to him of this. He ascribed the fact that similar rocks are not found on Earth to a period of heating, partial melting, and elemental fractionation that occurred soon after the Earth was formed. He envisioned the Earth to be originally composed of meteoritic material with a uniform composition. Heating would first result in the melting and sinking of the lowest melting point and highest density phase (iron-nickel alloy). This process formed the core of the earth. With continued heating, the sulphide phase would melt and sink. Thus a sulphide layer would mantle the core, leaving behind a residual outer silicate layer (slag).

Today, there is no geophysical evidence for a sulphide layer as Goldschmidt proposed. At the time, he overestimated the abundance of sulphur in his conjectured primordial Earth material. Evidently, the actual lower amounts of sulphur present were easily accommodated in the iron and silicate phases during heating such that a separate immiscible sulphide phase did not form.

Interpretation of Goldschmidt's classification scheme is straight forward. Given a system where all three melts are in equilibrium (i.e. molten Fe-Ni, sulphide and silica magma), a lithophile element, such as potassium, will be in highest concentration in the silica magma; whereas a siderophile element, such as one of the precious metals: platinum, palladium or gold; will be most enriched in the Fe-Ni melt. We would thus expect potassium to be enriched in the crust of the Earth and precious metals in the core of the earth. The strong clustering of elements with particular affinity evident in the accompanying table is no coincidence. It merely reflects an element's bonding characteristics and thus its electron configuration. Siderophile elements are mostly confined to Group VIIIB of the periodic table - elements characterised by a mostly complete outer d electron shell and by this shell's strong attraction for any outer s electrons. This makes the outer bonding shell of electrons of these elements generally unavailable for any type of bonding except metallic. Lithophile elements tend to have an outer bonding shell that contains s and/or p subshells, but no d; whereas the outer bonding shell of chalcophile elements contains a d subshell.

As with all attempts to divide up a continuum of behaviour into several simple categories, there will be exceptions, crossovers and examples of dual behaviour. The chief property not factored into Goldschmidt's classification is elemental valence. He was aware that many elements can occur in two or more valence states and when they do, they may exhibit different affinities. For example, in reducing environments, Cr3+ is strongly chalcophile; whereas under oxidizing conditions, Cr6+ is distinctly lithophile. Another example is that phosphorus only has an affinity for iron metal in the reduced form. In the oxidized form, it has lithophile character and occurs as phosphates in crustal rocks. Different valence forms of the same element can behave chemically as differently as different elements. Although not shown in the periodic table in this article, Goldschmidt indicated the elements with such secondary affinities.

Anyone interested in reading other treatments of this topic can consult any of the following references. There is also a very informative write-up of the life and accomplishments of Goldschmidt by K. H. Wedepohl. For example, you can find out about his 1942 arrest in occupied Norway and his rescue by the Norwegian resistance. This can be accessed on the Internet at


  • Brownlow A.H. Geochemistry; Prentice-Hall, Inc.: New Jersey, 1979.
  • Faure G. Principles and Applications of Inorganic Geochemistry; MacMillian Publishing Co.: New York, 1991.
  • Goldschmidt V.M. Geochemistry (A. Muir, ed); Oxford University Press: London, 1954.
  • Mason B.; Moore C.B. Principles of Geochemistry, 4th ed; John Wiley & Sons, Inc.: New York, 1982.
  • Wedepohl K. H. 1995, Geochim. Cosmochim. Acta 59, 1217