THE HYDRATION AND DEHYDRATION OF HYDROUS FERRIC IRON SULFATES. E. A. Hasenmueller

Introduction: Data collected by Viking, Pathfinder, and the Mars Exploration Rovers and by orbiters (e.g., Odyssey) provide multiple lines of evidence for the historical and present-day existence of water on Mars. For example, Mars Odyssey detected up to ~10 wt% equivalent H2O in equatorial regions of Mars where water ice is not stable [1]. It has been theorized that sulfate minerals, including hydrated ferric sulfates, may be part of the inventory of hydrous phases that account for water on the martian surface [2, 3, 4]. Several Fe-bearing minerals, such as jarosite and goethite, as well as Mg and Ca sulfates have already been identified in the martian regolith [5, 6, 7]. Knowledge of the martian regolith mineralogy is essential to understanding Mars’ hydrogeologic history, and hydrous minerals may serve as useful records of past aqueous alteration events. Indeed, important inferences about past conditions have been made from the presence of jarosite and goethite [5, 6], and it has been determined that jarosite is thermodynamically stable under assumed martian surface conditions [8]. Laboratory data measured under simulated martian surface conditions are crucial to provide constraints on hydrous mineral stability [9], and guided by current martian soil chemistry data we selected hydrated ferric sulfates for study (in lieu of ferrous sulfate minerals). Methods: Samples were chosen based on their H2O contents, the presence of independent H2O molecules that might evolve in a step-wise manner, and availability. Jarosite does not contain H2O but is OHbearing and was studied because it has been identified on Mars [5]. Jarosite, (KFe3(SO4)2(OH)6), from Bisbee, AZ; kornelite (Fe2(SO4)3·7H2O), from Napa Co., CA; botryogen, (MgFe(SO4)2(OH)·7H2O), from Coso Hot Springs, CA; and coquimbite, (Fe2(SO4)3·9H2O), from Alcaparrosa, Chile, were analyzed. Samples were ground dry and were mounted on an Anton-Paar TTK 450 heating stage on a Bruker D8 diffractometer with a VANTEC-1 position-sensitive detector (Cu radiation). Data collection ranges were tailored to each mineral to encompass the strongest peaks and collection times were typically 1-2 hrs. The heating stage was programmed to range from 25°C to 300°C in 50oC increments (25o, 50o, 100o, ...) and humidities ranged from room humidity (25-48%RH) to ~0% (roughingpump vacuum). Initial diffraction experiments evaluated sample purity, and reconnaissance heating measurements were made to determine the gross thermal behavior. Based on initial heating measurements, subsequent heating experiments were conducted on a much finer temperature scale and for longer times. Results: Jarosite. Diffraction data for jarosite changed little from 25o to 300oC, with only minor peak shifts occurring as a function of temperature. Exposure of the sample to vacuum (~20 mtorr) had no effect on the diffraction pattern. These results are consistent with the absence of H2O molecules in the structure. Coquimbite. Initial heating experiments from 25o300oC revealed gradual decomposition beginning at 50oC, evidenced by continuous decrease in peak intensities; no peak shifts or new phases were observed. This reaction was not reversible, which was surprising given the presence of independent H2O molecules in coquimbite (Fig. 1). The structure of coquimbite was destroyed between 200o and 250oC (Fig. 2), leaving an amorphous material that recrystallized to kornelite and copiapite (FeFe4(SO4)6(OH)2·20H2O) after reexposure to room temperature and humidity. Finer-scale heating at 30o, 40o, and 45oC showed a gradual decrease in peak intensities even at 30oC. Thus, coquimbite is unstable at temperatures as low as 30oC and may be unstable at 25oC in a vacuum; further experiments will evaluate the lower-temperature stability limits.