APXS Composition Results


Detailed information on the APXS instrument

[Composition plot of martian rocks]

Volcanic Rock Classification

Igneous rocks are usually classified according to the minerals they contain. In the absence of mineralogic data, volcanic rocks can be classified using their chemical compositions. Shown here is a scheme which classifies volcanic rocks on the basis of their alkalis (Na2O and K2O) and silica (SiO2) contents. This commonly used chemical classification for lavas shows that Barnacle Bill and Yogi (corrected for adhering dust) are distinct from basaltic Martian meteorites (shown as red squares). The Pathfinder APXS analyses have been corrected for the presence of a small amount of salt, and sulfur is assumed to be present as sulfide. These rocks plot in or near the field of andesites, a type of lava common at continental margins on the Earth. The preliminary data for alkalis are likely to represent upper limits, so refinement of these analyses could shift them to slightly lower Na2O + K2O and higher SiO2. We do not presently know whether these are igneous (crystallized from a melt), sedimentary (grains/fragments deposited by wind or water or precipitates), or metamorphic rocks (deformed). If they are igneous rocks, this classification would indicate that they has been remelted and reprocessed, implying that Mars may have had a more active thermal history than previously thought. However, this measurement could also be a result of a physical mixture of particles of rocks such as granite and basalt. A spot reflectance spectra of Barnacle Bill shows that if the rock is made of a mixture, the particles are smaller than about 1 cm.


[MgSi/AlSi plot of martian rocks]

Magnesium/Silicon versus Aluminum/Silicon

This diagram (preliminary X-ray data) illustrates chemical differences between terrestrial rocks and meteorites inferred to have been derived from Mars. The Martian meteorites (as well as Viking soil analyses) all plot to the left of the fields for Earth rocks. Pathfinder APXS analyses of rocks (stars) and soils (yellow dots) appear to plot in the gap between these previously defined fields, although they are similar to at least one basaltic meteorite. The other two stars represent the compositions of Barnacle Bill and Yogi. The analysis of Yogi appears to be contaminated by dust adhering to the rock's surface. The rock composition can be estimated by subtracting a portion of dust; the resulting Yogi composition is very similar to that of Barnacle Bill (we have assumed 50% dust having the composition of drift analysis A-5 and used a linear mixing model to subtract the dust which is only strictly valid if the dust, where present, is thicker than the APXS penetration depth). Barnacle Bill is also contaminated by dust, but to a lesser extent.

[Piecharts of martian rocks]

Mineral Compositions

The Pathfinder APXS chemical analyses of Barnacle Bill and Yogi (estimated based on a correction for adhering dust) have been recast into plausible minerals using the CIPW norm calculation. If they are fully crystalline igneous rocks, both possibly consist of orthopyroxene (magnesium-iron silicate), feldspars (aluminum silicates of potassium, sodium, and calcium), quartz (silicon dioxide), and other minerals that include magnetite, ilmenite, iron sulfide, and calcium phosphate


[Na/Si vs. Fe/Mn plot of martian rocks]

Sodium/Silicon versus Iron/Silicon

Shown here are the analyses of Yogi (A-7) and Barnacle Bill (A-3) on a plot of Na/Si vs. Fe/Mn. Na/Si is not a good indicator of different planetary bodies (and the APXS analyses of Na have a large error), but the Fe/Mn ratio is a diagnostic feature that separates Martian rocks from all other rocks. As can be seen, Yogi and Barnacle Bill are quite Martian.


[Composition plot of Barnacle Bill]

Calcium/Silicon versus Iron/Silicon

These three elements are especially well suited for APXS analysis. The compositions of SNC meteorites, as well as Viking soils and Mars Pathfinder soils, have higher iron/silicon ratios than terrestrial rocks. Barnacle Bill's composition (A-3) plots to the left, because of its high silicon content. The analysis of Yogi appears to be contaminated by dust adhering to the rock's surface. The rock composition can be estimated by subtracting a portion of dust; the resulting Yogi composition is very similar to that of Barnacle Bill (we assumed 50% dust having the composition of drift analysis A-5 and used a linear mixing model to subtract the dust which is only strictly valid if the dust, where present, is thicker than the APXS penetration depth). Barnacle Bill is also contaminated by dust, but to a lesser extent.


[Plot of martian soils]

Martian Soil Composition

APXS analyses of Martian soils are compared with Viking soil analyses. Each element is normalized to silicon in this diagram. The yellow boxes representing Viking data include all analyses and their analytical uncertainties reported by B.C. Clark and others (1982) JGR, vol. 87, p. 10,064. Although the first APXS soil analysis (A-2) was reported to be almost identical to Viking soils, subsequent analyses demonstrate some variability and a few significant differences from Viking analyses. Specifically, soils at the Pathfinder site generally have higher aluminum and magnesium, and lower iron, chlorine, and sulfur. Scooby Doo, which appears to be a sedimentary rock composed primarily of compacted soil, also exhibits a few chemical differences form the surrounding soils. Analysis A-5 represents a deposit of windblown dust (called drift), whereas the other soil analyses may be cemented materials.


* Table of elemental compositions - Preliminary results for rocks and soils in wt %
* Detailed information on the APXS instrument
*APXS Mars surface composition results - Press Release 04 December 1997
* Mars Fact Sheet
* Mars Pathfinder home page at NSSDCA


Questions and comments about this page should be addressed to:
Dr. David R. Williams, david.r.williams@nasa.gov, (301) 286-1258
NSSDCA, Mail Code 690.1, NASA/Goddard Space Flight Center, Greenbelt, MD 20771

NASA Official: Dr. Ed Grayzeck, edwin.j.grayzeck@nasa.gov
Last Updated: 18 December 2001, DRW