Subway Rider Crystal Chemistry

The "Subway Rider" model applied to potassium-rich clinopyroxenes

High-Pressure Crystal Chemistry of Potassium in Clinopyroxene and Other Phases

The experimental examination of K-enrichment in clinopyroxene (Cpx) at high pressure uses multianvils in the laboratory of David Walker at Columbia University's Lamont-Doherty Earth Observatory. A structure refinement of a natural K-rich diopside appeared recently (Harlow, 1996), the first encompassing paper on the high-pressure experiments has been finished (Harlow, in press), refinement of a experimentally grown K-rich clinopyroxene (a crystal with 4.5 % K2O in Di40Ko39.5KCrcpx20.5 ) has been completed, further high-pressure experiments into different K-rich assemblages is underway. Several insights are:

1) The "Subway Rider" model of crystal chemistry at high pressure: Like a subway car jammed with passengers, where it is easier for a large guy to get a seat if he can squeeze in adjacent to small people, a large cation can enter a site in the constrained structure of a pyroxene if there are smaller adjacent cations so the average size is not too large. A principal cation to balance K in the M2 site of clinopyroxene is Mg2. The effect appears important at high pressure because temperature is high too, permitting the necessary disorder in the cation sites.

2) Experiments to yield a barium clinopyroxene component (BaMgSi2O6) indicate that the "Subway Rider" effect is necessary for the Ba-Cpx solution to occur. When Ba-Cpx component enters diopside, there is always concomitant solid solution with enstatite component (MgSi2O6), manifesting that small Mg2 balances the large Ba2 ions in the M2 site in clinopyroxene, the essence of the "Subway Rider" effect.

3) It is hypothesized that sodium must be highly compressible (more than usually predicted) because it acts as the small cation compensator that allows potassium to enter Cpx at high pressure as described in the "Subway Rider" effect. Model calculations based on polyhedral compressibility predicts that Na is 2 to 3 times more compressible than Ca2 in a clinopyroxene, much more than the 10% difference calculated from theoretical linear bond compressibilities.

4) K uptake in Cpx is controlled by K-cpx component solubility in melts rather than K2O content, which is consistent with observations for other Cpx components.

5) The role of cation vacancies in diopsidic Cpx is not inconsistent with KCpx-uptake. Moreover vacancy abundance, as Ca-Eskola component (Ca0.5Vac0.5AlSi2O6), is highly correlated with pressure and the coexistence of an aluminous phase, such as kyanite, rather than the bulk composition of the assemblage. This understanding has important implications for the composition of clinopyroxene from mantle eclogites, grospydites, etc.


Harlow, G. E. (1996) Structure refinement of a natural K-rich diopside: The effect of potassium on the average structure. American Mineralogist, 81, 632-638..

Harlow, G. E. (1997) Potassium in clinopyroxene at high pressure and temperature: An experimental study. American Mineralogist, 82, 259-269.

Harlow, G.E. (1998) Interpretation of Kcpx and CaEs in clinopyroxene from diamond inclusions and mantle samples. Proceedings of the Seventh International Kimberlite Convention (in press).