Lunar and Planetary Science XLVIII (2017) 1786.pdf

LOW-CO2 ATMOSPHERE ON EARLY MARS? AN INTERPRETATION OF MANGANESE OXIDE ON GALE CRATER BY LABORATORY EXPERIMENTS. N. Noda1, S. Imamura1, Y. Sekine1, H. Tabata1, S. Uesugi1, T. Murakami1, and Y. Takahashi1, 1Department of Earth and Planetary Science, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033 Japan ([email protected])

Introduction: Both CO2 and O2 are key atmos- Table 1. Experimental conditions of the present pheric components for climatic and chemical evolution study. The flow rate of gas mixtures was adjusted using a of early Mars. However, their ancient partial pressures rotameter and mass flow controller. The mixing ratio of have been poorly constrained. Multiple lines of geolog- O2/CO2 in the gas mixture was determined by calibration ical and geochemical evidence support the existence of using a quadrupole mass spectrometer (QMS) liquid water on Mars, at least episodically, during the Run Air flux CO2 flux Mixing ratio Duration # [L/min] [L/min] O2/CO2 [min] late Noachian and early Hesperian [1]. This implies the 1 0.5 0 30 presence of greater amounts of greenhouse gas, repre- 2 0 0.5 180 sentatively CO2, than that of today [2]. On the other 3 0.5 0.5 0.6 330 hand, in-situ analyses for sandstones in the Kimberley 4 1 0.05 5 35 region on Gale crater by the Curiosity rover have found 5 1 0.005 40 240 Mn enrichments in fracture-filling materials, i.e., veins, crosscutting the surrounding rocks [3]. The Mn en- prepared in three steps. First, solutions of ~0.1 M N- richments were highly likely caused by deposition of Tris-methyl-3-aminopropanesulfonic acid (TAPS) and Mn oxides from reducing, soluble Mn (Mn2+) within 0.1 M NaOH were made in order to buffer solution pH ancient groundwater, rather than depositions of evapo- around 8–9 during the experiments. Then, the solutions rites such as Mn carbonate [3]. Since oxidation of Mn were deaerated by bubbling pure Ar gas in an Ar- requires high levels of redox potential (Eh) (Fig. 1), purged glovebox for more than 6 hours. Finally, these findings indicate a possible coexistence of a high- 99.0+% MnCl2･4H2O was introduced as a source of ly oxidizing atmosphere and wet conditions on early Mn2+ in the glovebox. The initial concentrations of dis- Mars [3]. solved Mn were 20 mM. Fig. 1: Eh-pH The gas mixture of artificial air and CO2 was then diagrams of Mn- bubbled into the starting solution within the reaction O-C system at vessel for 0.5–5.5 hours of reaction time (Table 1). The pCO2=0.5 bar total pressure of the gas mixture in the reaction vessel Equilibrium calcu- was ~1 bar. Solution samples were collected several lations were per- times through the valve during the reaction. These solu- formed based on Henry’s low and tion samples were collected by a syringe with 0.22-μm Nernst equation [4]. filter and added into ~5 ml of 0.1 M HNO3 to prevent oxidation by air during the analysis. After the reaction, The present study aims to further constrain the solid precipitates were recovered by filtering the rest of composition of early Mars’ atmosphere, especially the the solutions through 0.22-μm membranes. All these O2/CO2 ratio, that allows the formation of Mn oxides procedures were conducted in the Ar-purged glovebox, -12 found in Gale crater. Even under high CO2 conditions, in which low-O2 conditions (pO2 < 10 bar) were Mn oxide is thermochemically stable for high pO2 (Fig. maintained out of the reaction vessel by a circulation 2+ 1). However, the oxidation of Mn2+ is known to be slow pump and O2 removal system. Mn concentrations of [5]. Thus, Mn-oxide formation could be kinetically collected solution samples were measured using induc- inhibited under high CO2 conditions due to efficient tively coupled plasma atomic emission spectroscopy 2+ formation of Mn carbonate and subsequent depletion of (ICP-AES) to examine loss of Mn due to precipitation. Mn2+ in the solution. We performed laboratory experi- Recovered solid precipitates were analyzed with X-ray ments, in which gas mixtures of artificial air (N2/O2 = 2+ 4) and pure CO2 were introduced into Mn solutions. We conducted five experiments (run #1–5: Table 1) by Fig. 2. varying the mixing ratio of air and CO2 and analyzed Schematic solid precipitations. diagram of Materials & Methods: The experiments were per- experimental formed in a flow system using a Pyrex-glass reaction system. vessel (Fig. 2). Starting solutions of dissolved Mn were Lunar and Planetary Science XLVIII (2017) 1786.pdf

absorption fine structure (XAFS) and X-ray diffraction even for low-CO2, high-O2 gas mixtures (e.g., O2/CO2 = (XRD) to determine the mineral compositions and oxi- 4 (the run # 4) and 40 (the run #5), efficient precipita- dation state of Mn. tion of MnCO3 occurs before the formation of MnO2 Dissolved Mn in solution: Time evolution of Mn2+ (Fig. 4). In the solution, dissolved Mn is highly deplet- concentrations for each run is shown in Fig. 3. In all ed due to the precipitation of MnCO3 (Fig. 3). These runs, the Mn concentrations decreased from the initial results suggest that kinetics of these two competition concentrations of ~20 mM. Under high CO2 conditions reactions are the critical factor that determines the for- (i.e., the runs #3 and 4), the Mn concentrations declined mation of Mn oxides. rapidly within 50 mins of reaction time. After 310 mins As mentioned above, the Curiosity rover has found of reaction time for the run #3, the Mn concentration in Mn-oxide enrichments in veins of sandstones in the the solution was nearly zero within the error, which Kimberley region [3]. Precipitation of Mn carbonate indicates that most of initial Mn2+ was precipitated as cannot explain the Mn enrichments [3]. Mn carbonate solid particles. does not form in a solution for pH < ~4–5 (Fig. 1); however, oxidation of Mn in an acidic solution does not Fig. 3: Time evo- proceed effectively due to a strong pH dependence of lution of concen- the reaction rate [5]. In addition, the long-term presence trations of Mn of acidic groundwater may be inconsistent with less- in solution. chemically-altered surrounding rocks [3]. Thus, our The procedure error results suggest that the findings of Mn oxides by Curi- would be a few mM. osity indicate not only oxidizing conditions but low- Thus, apparent CO2 ones in the ancient groundwater of Gale crater. increase in Mn for One possible explanation of the low CO2 conditions the run #5 would is that the groundwater was not equilibrated with the be due to the procedure error. atmosphere. This could have happened if efficient pre- Dat a f or Mul t i pl ot cipitation of carbonate minerals occurred in an open DATA TYPE Xanes Dat a Of f set X = 0 Y = 0 Solid precipitates: The XANES spectra for the water (e.g., Gale’s crater lake [6]) and consequent CO2- collected solid precipitates are shown in Fig. 4. Com- depleted water transported into the subsurface, includ- paring with the results of the standard materials, all of ing fractures of the Kimberley region. However, a lack Dat a f or Mul t i pl ot DATA TYPE theXa nesolids Dat a precipitates collected from the experimental of sedimentary carbonates on Mt. Sharp of Gale crater Of f set X = 0 Y = 0 runs #2–5 are predominated by MnCO3. The results of by remote sensing might not be consistent with this idea. XRD analysis support this conclusion. The major peaks Further in-situ investigations of Mt. Sharp sediments

Fi l eName of Sa thempl eName XRDComment spectra are attributedFact or Co l o fromr MnCO3. The would be important to evaluate this possibility. sampl e1. xan sampl e1. dat 1. 000 sampl e2. xanXANESsampl e2. dat spectrum for run #11. 00 0 cannot be simply ex- Alternatively, if the groundwater was equilibrated sampl e3. xan sampl e3. dat 1. 000 sampl e4. xan sampl e4. dat 1. 000 sampl e5. xanplainedsampl e5 by the standard material1. 000 s of MnO2, Mn2O3, with the atmosphere, our results imply low pCO2 at the Mn2O3. xan Mn2O3. dat 1. 000 Mn3O4. xan MnMn3O4O. dat , and MnCO suggesting1. 000 that the solid precipi- time of deposition of the Mn oxides. For instance, as- MnCO3. xan MnC3O3. dat4 3 1. 000 MnO2. xan MnO2. dat 1. 000 tates would be a mixture of Mn (hydr)oxidants with suming pO2 of 0.2 bar, pCO2 should have been < 5 different valences. mbar in order to avoid efficient precipitation of MnCO3 ut (the run #5). Given ~0.1–1 bar of pO2 for formation of Fi l eName Sampl eName 2.0Comment Fact or Col or sampl e1. xan sampl e1. dat 1.8 1. 000 Run #1 MnO2, required pCO2 would be <~10 mbar. These val- sampl e2. xan sampl e2. dat 1. 000 Run #2 sampl e3. xan sampl e3. dat 1.6 1. 000 Run #3 ues are comparable to or less than pCO2 in today’s at- sampl e4. xan sampl e4. dat 1.4 1. 000 Run #4 sampl e5. xan sampl e5 1. 000 Run #5 mosphere (i.e., 6 mbar). This could imply that early Mn2O3. xan Mn2O3. dat 1.2 1. 000 Mn2O3 Mars’ atmosphere may have contained additional Mn3O4. xan Mn3O4. dat 1.0 1. 000 Mn3O4 MnCO3. xan MnCO3. dat 1. 000 0.8 MnCO3 greenhouse gases to keep the surface warm, or that MnO2. xan MnO2. dat 1. 000 MnO2 0.6 Mars had been basically cold and arid, similar to that of 0.4 today, at the time when groundwater was active within 0.2 Gale crater. ut 0.0 6530 6540 6550 6560 6570 6580 2.0 Energy(eV) References: [1] Ehlmann, B.L. et al. (2011). Nature 479, doi:10.1038/nature10582. [2] Wordsworth, R.D. (2016). 1.8 Fig. 4: XANES spectra of solid precipitates and Annu. Rev. Earth Planet. Sci. 44, 381-408 [3] Lanza, N.L. et standard materials. 1.6 al. (2016). Geophys. Res. Lett., 43, 7398-7407. [4] Usui, A.

1.4 et al. (2015). Geoscience of Marine Manganese Deposits, 1st Discussion: The Eh-pH diagrams for the experi- ed., University of Tokyo Press. [5] Stumm, W. and Morgan, 1.2 mental conditions indicate that Mn oxide (i.e., MnO2) is J.J. (1996). Aquatic Chemistry: Chemical Equilibria and 1.0 thermochemically stable for all of our experimental Rates in Natural Waters, 3rd Edition, Wiley-ii. [6] Grotzinger, J.P. et al. (2015). Science 350, doi:10.1126/science.aac7575. 0.8 conditions (Fig. 1). However, our results show that

0.6

0.4

0.2

0.0 6520 6540 6560 6580 6600 6620 6640 6660 6680 6700 6720 6740 Energy(eV)