Evidence for Calcium Carbonate at the Mars Phoenix Landing Site

See allHide authors and affiliations

Science  03 Jul 2009:
Vol. 325, Issue 5936, pp. 61-64
DOI: 10.1126/science.1172768


Carbonates are generally products of aqueous processes and may hold important clues about the history of liquid water on the surface of Mars. Calcium carbonate (approximately 3 to 5 weight percent) has been identified in the soils around the Phoenix landing site by scanning calorimetry showing an endothermic transition beginning around 725°C accompanied by evolution of carbon dioxide and by the ability of the soil to buffer pH against acid addition. Based on empirical kinetics, the amount of calcium carbonate is most consistent with formation in the past by the interaction of atmospheric carbon dioxide with liquid water films on particle surfaces.

The key to understanding Mars’ past climate is the study of secondary minerals that have formed by reaction with volatile compounds such as H2O and CO2. A wet and warmer climate during the early history of Mars, coupled with a much denser CO2 atmosphere, are ideal conditions for the aqueous alteration of basaltic materials and the subsequent formation of carbonates (1, 2). Although Mg- and Ca-rich carbonates are expected to be thermodynamically stable in the present martian environment (3, 4), there have been few detections of carbonates on the surface by orbiting or landed missions to Mars. Earth-based thermal emission observations of Mars were the first to indicate the presence of carbonates in the martian dust (1 to 3 volume percent) (5). Mg-rich carbonates have been suggested to be a component [2 to 5 weight percent (wt %)] of the martian global dust based on the presence of a 1480 cm−1 absorption feature in orbital thermal emission spectroscopy (6). A similar feature was observed in brighter, undisturbed soils by the Miniature Thermal Emission Spectrometer on the Gusev plains (7). Mg-rich carbonates have recently been identified in the Nili Fossae region by the Compact Reconnaissance Imaging Spectrometer for Mars instrument onboard the Mars Reconnaissance Orbiter (8), prompting the idea that the Mg-rich carbonates in the dust might be due to the eolian redistribution of surface carbonates. Carbonates have also been confirmed as aqueous alteration phases in martian meteorites (911).

Here, we describe observations made by instruments on the Phoenix Lander relevant to the identification of calcium carbonate in the soil: the Thermal and Evolved-Gas Analyzer (TEGA), which is similar to the TEGA instrument flown on the Mars Polar Lander (12), and the Microscopy, Electrochemical, and Conductivity Analyzer (MECA) (13). TEGA consisted of eight thermal analyzer cells that each contained an oven in which samples were heated up to 1000°C at a controlled ramp rate, with the necessary power recorded for calorimetry. Any gases that were generated from the heated samples were carried to the evolved-gas analyzer by a N2 carrier gas maintained at 12 mbar oven pressure (14). The MECA contained a Wet Chemistry Laboratory (WCL) that consisted of four analyzer cells, each of which added an aqueous solution to the soil and measured electrochemical parameters such as pH and the concentrations of various ions before and after the addition of various chemical reagents.

Soils analyzed by TEGA have similar thermal and evolved gas behaviors. Here, we focus on a subsurface sample dubbed Wicked Witch (15). We looked for an endothermic phase transition by fitting a curve to the oven power as a function of temperature and looked for differences from the smooth baseline fit. For the thermal analysis of a sample we routinely heated the oven containing the sample a second time with an identical temperature ramp within a day or two after the initial heating and compared the results to check for background transitions that occurred in the oven itself.

A clear endothermic peak is seen between 725°C and 820°C, and there is another one between about 860°C and 980°C (Fig. 1). The phase responsible for the 860°C peak is unidentified at this time. The most likely phase candidate for the 725°C endothermic reaction is a calcium-rich carbonate mineral phase (e.g., calcite, ikaite, aragonite, or ankerite) (16). The area of the endothermic peak is about 6 J. Based on the calcite decomposition enthalpy of 2550 J/g (17) we estimate that there is 2.4 mg of CaCO3 in the sample. For a TEGA oven volume of 0.052 cm3, full with a 1 g/cm3 sample, the concentration of CaCO3 is about 4.5 wt % (18).

Fig. 1

Plot of differential power (relative to a degree 4 polynomial fit) for the subsurface sample dubbed Wicked Witch.

The amount of CO2 generated during the heating is shown in Fig. 2. There is a low-temperature release between 400°C and 680°C, which may be due to Mg or Fe carbonate, adsorbed CO2 contained in a zeolite type phase, or organic molecules that are converted to CO2 by oxidants in the soil. We made an independent estimate of the amount of CaCO3 from the amount of CO2 generated at higher temperatures. The majority of the high-temperature gas release appears to happen at a temperature higher than that of the oven phase transition (Fig. 2). There is a time lag of about 3 min or 60° for the gas to reach the EGA, but the observed time lag is greater than expected (19). Pending laboratory studies, this time lag may represent an interaction between the CO2 and the walls of the plumbing. The area under the peak corresponds to 0.66 mg of CO2, which implies a concentration of 3 wt % CaCO3 in the sample. Based on the combined calorimetry and evolved gas results, the Wicked Witch sample has on the order of 3 to 5 wt % CaCO3.

Fig. 2

Plot of Mass 44 (CO2) count rate and temperature versus run time. The first peak is atmospheric CO2 that diffuses into the oven and plumbing overnight; the second may be due to carbonates with a low decomposition temperature (e.g., FeCO3 or MgCO3). The high temperature peak is due to decomposition of CaCO3.

The WCL was designed to perform an assay for carbonate using classical wet chemical methods based on sample acidification. After 25 mL of water and a calibrant were added to the WCL beaker, soil was added and the pH was measured for a period of time. After a 4-sol freeze and subsequent thaw, 2.5 × 10−5 mole of 2-nitrobenzoic acid was dispensed into the soil solution (20). The pH was monitored by two ion selective polymer membrane pH electrodes and an iridium oxide pH electrode (12). Based on the acid dissociation constant (pKa) of 2-nitrobenzoic acid, the pH of the WCL solution should be 3.2 after acid addition in the absence of buffering.

After a surface sample dubbed Rosy Red (15) was added, the pH of the WCL leaching solution rose to 7.7 ± 0.5, a value consistent with the pH of a carbonate buffered solution and a partial pressure of CO2 (Pco2) of ~0.3 mbar in the WCL headspace (Fig. 3A). Over the course of sols 30 to 34, the pH reading decreased slightly to pH 7.5 ± 0.5, possibly due to changes in headspace Pco2, temperature, or intrinsic sensor drift. The addition of acid to the sample solution did not change the pH, confirming that the system is buffered by carbonate (Fig. 3B). The addition of the acid was confirmed by WCL cyclic voltammetry measurements, which detect the presence of 2-nitrobenzoic acid in solution. If calcium carbonate is the only saturated carbonate species present and thermodynamically independent of other equilibrium reactions, its solubility product Ksp will determine the upper limit of Ca2+ concentration in solution. The measured Ca2+ concentration in the Rosy Red soil solution of 5.5(±3.0) × 10−4 M is consistent with a Pco2 of ~0.3 mbar and a saturated calcium carbonate solution (21).

Fig. 3

Plot of pH versus local Mars time (A) upon soil delivery and (B) upon addition of 2-nitrobenzoic acid.

Assuming that a 1 cm3 sample with a density of 1 g/cm3 was added to the WCL cell, equilibrium calculations based on the measured ion concentrations in solution and the quantity of acid added show that the amount of calcium carbonate needed to neutralize the acid is 0.3% by weight. Lacking a full titration of the amount of carbonate, this quantity is only a lower limit to the concentration of calcium carbonate in the soil. Two other samples, both subsurface samples from the Sorceress location (similar to Wicked Witch), were shown also to contain ≥ 0.3% of calcium carbonate by weight. The WCL results are entirely consistent with the TEGA calorimetry and evolved-gas data, which strongly suggest the presence of calcium carbonate in the soils around the Phoenix landing site.

Low-temperature carbonate dissolution and precipitation are among the most important pedogenic processes in terrestrial semiarid to arid soils (22). Carbonates precipitate as coatings on soil particles and in soil pores that can result in cementation of particles. Calcium carbonate generally precipitates by reaction of CO2-charged water with Ca2+ released by dissolution of parent materials. Soils at the Phoenix landing site are in physical proximity to an ice-cemented soil layer several cm under the surface and are covered with a surface frost during the winter. These water reservoirs suggest the diffusion of water vapor through the soils (23) and the possibility of films of unfrozen water on particle surfaces. The latter can weather basaltic minerals and mobilize ions (24, 25). Given that atmospheric CO2 would dissolve in these thin films, calcium carbonate would precipitate if the water film becomes supersaturated. These precipitation events may be facilitated by the evaporation of the thin films during diurnal or seasonal cycles and may be much more frequent during periods of high obliquity (26). Another possibility for the formation of calcium carbonate is by hydrothermal aqueous alteration associated with an impact event or other thermal source (e.g., volcanic). An impact event into a volatile-rich target (i.e., ice and ice-cemented sediments) would rapidly melt and/or vaporize the ice, resulting in localized hydrothermal conditions. The Phoenix landing site lies on the ejecta of the Heimdall crater, which is located 20 km away. Consequently, calcium carbonate formed in the subsurface may have been transported to the Phoenix site by the excavation of subsurface material during the impact event that formed Heimdall. In this case, calcium carbonate has persisted in the soil since the impact event, ~0.5 billion years ago (27).

There are nonaqueous processes that can form carbonates, but the reaction rates appear to be far too small to account for the amount of calcium carbonate that we find (28). The presence of several percent by mass of calcium carbonate in the soil at the Phoenix site is of the expected magnitude if there has been periodically wet or damp martian soil in the geologic past, but it is difficult to produce under current martian conditions. This inference applies even if the calcium carbonate has been transported to the Phoenix site by wind, because the calcium carbonate would still have formed in damp or wet soil elsewhere on Mars. Two possible objections are (i) that the carbonate might be magmatic (e.g., carbonatites) and (ii) that the soil particles might have an abnormally large specific surface area, which would make a larger mass fraction of dry carbonate formation more feasible. There is, however, no evidence to support either objection. For the latter, preliminary analysis of microscopic imagery (29) indicates a modest specific surface area. Consequently, the simplest explanation is that the amount of carbonate provides geochemical evidence of past liquid water.

The presence of calcium carbonate in the soil has implications for our understanding of Mars. Calcium carbonate buffers an alkaline pH, which is similar to that of many habitable environments, notably terrestrial seawater. Another implication of the presence of calcium carbonate is that various ions may be adsorbed at exposed Ca2+ lattice sites of calcium carbonate and affect Mars’ soil geochemistry, and calcium carbonate can cement small soil grains and change the physical properties of the surface of Mars.

  • Present address: Department of Chemistry, University of New Hampshire, Durham, NH 03824, USA.

References and Notes

  1. R. E. Arvidson et al., Abstr. 1067, 40th Lunar and Planetary Science Conference, The Woodlands, TX, 23 to 27 March 2009.
  2. The onset temperature for this endothermic peak is similar to the calcite decomposition onset temperature of 738°C measured in the TEGA engineering qualification model (17). Other carbonates have decomposition temperatures that are lower than that of calcite.
  3. In a calibration run on the TEGA engineering qualification model with a known amount of calcite, we found the heat of transition to be 2550 J/g, which implies 2.4 mg of CaCO3 in the Wicked Witch sample.
  4. Estimation of the concentration of calcium carbonate in the sample is uncertain because the mass of sample in the oven is not tightly constrained. We estimate that this error is on the order of ± 25%.
  5. The cause of this delay will be explored in planned laboratory experiments in the future. The plumbing temperatures are heated to greater than 35°C, so one would not expect CO2 to condense out on any plumbing surfaces. The sol 70 run shows that this effect is not due to a background signal.
  6. A martian solar day has a mean period of 24 hours 39 min 35.244 s and is referred to as a sol to distinguish this from a ~3% shorter solar day on Earth.
  7. T. L. Heet, R. E. Arvidson, M. Mellon, Abstr. 1114, 40th Lunar and Planetary Science Conference, The Woodlands, TX, 23 to 27 March 2009.
  8. A previous experiment suggested that carbonate could form in relatively dry Mars-like conditions as submicrometer coatings on soil particles (5). However, the results from this study cannot be extrapolated over time because the experiments did not proceed beyond a monolayer of carbonate. A more comprehensive experimental study under Mars-like conditions showed that, in a given time span, carbonate forms on the surface of basaltic particles in an amount that increases with the thickness of films of water around the particles (30, 31). Growth of more than a monolayer of carbonate occurs with logarithmic reaction kinetics. For damp or wet conditions (meaning 0.1 to 0.5 g H2O per g soil, equivalent to 102 to 103 monolayers of water), the number of CO2-reacted monolayers of substrate, L, per particle is found empirically to follow the relationship (30, 32) L(t) = Dlog10(1 + t/t0) (Eq. 1). Here, the constant of proportionality D is ~1 monolayer CO2/log10t for powdered basalt, while t0, which represents a time scale for CO2 adsorption or dissolution in H2O, is ~10−2 to 10−3 days. Because the Phoenix site contains material that was likely ejected from the nearby Heimdall crater up to 0.6 billion years ago (27), or 2.2 × 1011 (Earth) days ago, the number of monolayers reacting with CO2 can be calculated from Eq. 1 as ~13 to 14. By stoichiometry, 1 mol of CO2 generates 1 mol of carbonate, so it follows that the mass fraction, F, of carbonate produced under these “damp” or “wet” soil conditions would be F = M × [(L As/A)/NA] (Eq. 2), where M is the molar mass of carbonate (100 g/mol for CaCO3), As is the specific surface area of the particles, A = 2 ×10−19 m−2 molecule−1 is the area taken up by a CO2 molecule in reaction with the surface (32), and NA is Avogadro’s number. Basalt glasses pulverized to 0.1 to 1 micrometer size have As ~1 to 10 m2/g (33), whereas 17 m2/g was estimated for Viking lander 1 soil (34). Substituting numerical values into Eq. 2 gives a carbonate mass fraction up to 20 wt % for continuously “wet” conditions. In contrast, under persistently dry or vapor conditions (meaning zero to a few monolayers of H2O), the fraction of carbonate produced is smaller because the kinetic coefficient D is smaller. D was measured up to 0.07 monolayers CO2/log10t at 6.6 mbar and –15°C (24), giving up to 0.8 wt % in Eq. 2. Other values of D for dry or vapor conditions ranged from ~0.01 to 0.2, but high values occurred only at higher CO2 pressures, up to 995 mbar (32).
  9. We acknowledge the work of the Robotic Arm team in delivering the samples to TEGA and WCL and the contributions of the engineers, scientists, and managers who made these instruments and the analysis of their data possible, including H. Enos, C. Fellows, K. Harshman, M. Finch, M. Williams, M. Fitzgibbon, G. Droege, J. M. Morookian, V. Lauer, D. C. Golden, R. Chaney, H. Hammack, L. Brooks, M. Mankey, K. Gospodinova, J. Kapit, C. Cable, P. Chang, E. Coombs, and S. Stroble. The Phoenix Mission was led by the University of Arizona, Tucson, on behalf of NASA and was managed by NASA’s Jet Propulsion Laboratory, California Institute of Technology, of Pasadena, CA. The spacecraft was developed by Lockheed Martin Space Systems, Denver.
View Abstract

Navigate This Article