Atmosphere-Surface Interactions on Mars: Δ17O Measurements of Carbonate from ALH 84001

See allHide authors and affiliations

Science  05 Jun 1998:
Vol. 280, Issue 5369, pp. 1580-1582
DOI: 10.1126/science.280.5369.1580


Oxygen isotope measurements of carbonate from martian meteorite ALH 84001 (δ18O = 18.3 ± 0.4 per mil, δ17O = 10.3 ± 0.2 per mil, and Δ17O = 0.8 ± 0.05 per mil) are fractionated with respect to those of silicate minerals. These measurements support the existence of two oxygen isotope reservoirs (the atmosphere and the silicate planet) on Mars at the time of carbonate growth. The cause of the atmospheric oxygen isotope anomaly may be exchange between CO2 and O(1D) produced by the photodecomposition of ozone. Atmospheric oxygen isotope compositions may be transferred to carbonate minerals by CO2-H2O exchange and mineral growth. A sink of17O-depleted oxygen, as required by mass balance, may exist in the planetary regolith.

Chemical and isotopic measurements of the present-day martian atmosphere indicate that many elements and isotopes are fractionated from their initial compositions (1). Modeling suggests that the composition of the martian atmosphere has evolved on planetary time scales as a result of escape processes, impact erosion, outgassing, and photochemical reactions (2-6). Morphologic evidence points to times early in martian history when greenhouse heating supported warmer climates and liquid water shaped Mars's surface (7, 8). Measurements of trapped gases and mineral phases in martian meteorites indicate that atmospheric evolution occurred before the excavation of these meteorites from the martian surface (9–15).

The present martian atmosphere interacts with the hydrosphere and regolith through diurnal, seasonal, and longer term exchange of water between the atmosphere and regolith (16, 17). In the past, liquid water acted as a medium for atmosphere-surface exchange and new mineral precipitation. High rates of oxygen exchange between H2O and CO2 facilitated the transfer of atmospheric oxygen isotopic characteristics to H2O and to minerals that precipitated from it. Because oxygen is present in CO2, H2O, dust, and other atmospheric constituents such as O2, CO, and O3, the oxygen isotope compositions of minerals formed in the regolith serves as a tracer for atmospheric chemistry and evolution and its interactions with the martian surface. Here we report measurements (18) of 16O, 17O, and 18O in carbonate minerals from martian meteorite ALH 84001 [δ18O = 18.3 ± 0.4‰, δ17O = 10.3 ± 0.2‰, and Δ17O = 0.8 ± 0.05‰ (23)]. High-temperature silicate phases in ALH 84001 give δ18O = 4.53 to 4.64‰, δ17O = 2.58 to 2.74‰, and Δ17O = 0.22 to 0.327‰ (12). Secondary ion mass spectrometry analysis of SiO2 yielded δ18O = 20.4 ± 0.9‰ (24), and analyses of carbonate have yielded δ18O = –9 to 26‰. The mean of these measurements is 15 ± 5‰ (one 3σ outlier is omitted) (15,24-26). Our δ18O measurements fall within the range of previous measurements for carbonates but differ from the Δ17O of the high-temperature silicate phases.

Karlsson et al. (27) extracted H2O from six martian meteorites by stepwise pyrolysis and found that the Δ17O of this water ranged from 0.1 to 0.9‰. H2O extracted from Nakhla, Lafayette, Chassigny, and Zagami meteorites during 600° and 1000°C pyrolysis steps (Δ17O = 0.4 to 0.9‰) was fractionated relative to martian meteorite silicates [Δ17O ∼ 0.3‰ (12)]. The same pyrolysis steps for Shergotty produced H2O unfractionated with respect to martian meteorite silicates, and for meteorite EETA 79001 they produced H2O with Δ17O = 0.1‰, which is intermediate between terrestrial waters (Δ17O = 0) (11) and martian meteorite silicates. These results (27) were interpreted to indicate that H2O extracted from Nakhla, Lafayette, Chassigny, Zagami, and Shergotty had a martian origin and that H2O extracted from EETA 79001 was a mixture of martian and terrestrial waters. Because water extracted from Nakhla, Lafayette, Chassigny, and Zagami was not in oxygen isotopic equilibrium with martian meteorite silicate minerals, two distinct oxygen reservoirs were inferred to be present on Mars—the silicate planet (the crust and mantle) and the atmosphere (including H2O that exchanged oxygen with it).

Our measurements of carbonates (Δ17O = 0.8‰) are fractionated relative to measurements of martian meteorite silicates and indicate that ALH 84001 also preserves evidence of oxygen isotope disequilibrium between the atmosphere and the silicate planet. Carbonate nodules in ALH 84001 preserve chemical and isotopic gradients (24, 25, 28) that would have been homogenized by diffusion and exchange if post-growth reequilibration occurred. Our oxygen isotope data therefore indicate that the atmosphere and lithosphere were out of isotopic equilibrium at the time of carbonate growth.

On Earth, plate tectonics facilitates oxygen exchange between the hydrosphere and lithosphere through global-scale hydrothermal circulation systems at plate margins. Because the amount of oxygen in the silicate Earth is greater than that in the hydrosphere and atmosphere, this exchange buffers the oxygen isotope composition of the hydrosphere (29). The oxygen isotope composition of water in Earth's oceans is buffered to a steady-state value on a time scale with a mean life (1/rate constant) of ∼100 to 250 million years (29). The lack of plate tectonics on Mars is seen as a means of maintaining an atmosphere that is isotopically distinct from the solid planet (27). It is unclear whether other types of hydrothermal activity may have caused sufficient atmosphere-surface oxygen exchange to affect atmospheric δ18O, δ17O, and Δ17O. Massive hydrothermal systems have been invoked to explain surface features associated with the Oceanus Borealis and spectroscopic features in the Valles Marineris canyon system (7, 30) and also as a means of generating catastrophic floods and release of CO2 of sufficient quantity to stabilize episodic, greenhouse-state, atmospheric conditions with higher CO2 pressure (7). If these processes released CO2 and H2O with juvenile composition in amounts comparable to the atmospheric burden, they might affect the oxygen isotopic composition of the atmosphere.

Hydrothermal interactions have been invoked to explain differences between the Δ17O data of (27) and the H/D data of martian meteorite waters extracted in a second set of pyrolysis experiments (14). It has been suggested that hydrogen and oxygen isotope systematics may have been decoupled during martian hydrothermal interactions in the same way that they are decoupled in meteoric hydrothermal systems (14, 31). Because the relative H/O ratio of hydrothermal fluids is much greater than the H/O ratio in silicate rock, the hydrothermal fluid Δ17O will approach the value of the rock for much lower fluid/rock than does the fluid D/H ratio. For a 17O- and D-enriched atmosphere, these interactions should produce positively correlated D/H ratios and Δ17O for both fluid and rock. The pyrolysis data (14, 27) form a negatively correlated array with a low-D/H, high-Δ17O end member and a high-D/H, low-Δ17O end member (Fig.1). Although hydrothermal decoupling of D/H from Δ17O is an inevitable consequence of martian hyrothermal interactions, the negatively correlated array cannot be interpreted solely in this context. We suggest that this negatively correlated array reflects the operation of additional processes—secular changes that occurred in martian atmospheric D/H and Δ17O as a result of the mixing of evolved and juvenile oxygen-bearing and hydrogen-bearing reservoirs.

Figure 1

Plot of maximum values of D/H versus Δ17O for martian meteorite H2O (▪) extracted by pyrolysis (14, 27) and carbonate (⧫) Δ17O data versus pyrolysis D/H data from (14). Note that D/H and Δ17O were collected in different sets of experiments. The H2O and the carbonate data form a negatively sloped array. Assuming the martian atmosphere has elevated D/H and Δ17O, positively sloped arrays would be predicted by simple mixing with terrestrial waters (terrestrial contamination arrays), hydrothermal exchange, or mixing of juvenile and evolved martian reservoirs. Variable atmospheric D/H ratios would produce horizontal arrays, and variable atmospheric Δ17O would produce vertical arrays. The primordial D/H field is from (13).

Yung and others (2) have modeled the evolution of the D/H ratio by Jeans escape. The present atmospheric burden of H2O is sufficiently low to accommodate evolution from D/H ∼ 0 to the present atmospheric D/H in ∼105years (2). To extend this time scale to the 4.5-billion-year planetary time scale requires an initial column of 3.6 m and a present column of 0.2 m of exchangeable water (2), an amount that is less than the total water budget inferred on the basis of geomorphological features (7, 8, 17). The amount of exchangeable water has changed on planetary time scales (7,8) and may change episodically on time scales of 105to 107 years as a result of orbital changes and concomitant mobilization of water trapped in the polar caps and of injection of juvenile waters by floods and volcanic activity (32). The atmospheric D/H ratio would undergo secular variations if the size of the exchangeable H2O reservoir changed by an equal or greater amount than itself on time scales <105 years. Because there is more oxygen in the atmosphere than hydrogen, atmospheric Δ17O is more difficult to change than the D/H ratio but may also have been influenced by changes in the size of the exchangeable oxygen-bearing reservoir.

Suggestions to explain the evolution of atmospheric Δ17O include an atmosphere that originated as a late cometary or meteoritic veneer (heterogeneous accretion), an atmosphere whose composition evolved as a result of escape processes, and an atmosphere that evolved as a result of photochemical reactions (5, 6, 27). No direct tests exist to support the possibility that atmospheric Δ17O was produced by heterogeneous accretion.

Oxygen isotope fractionations produced by gravitational separation above the homopause, followed by escape at the exobase, raise atmospheric δ18O, δ17O, and Δ17O (5). Given the range of atmospheric parameters determined for Mars (1, 5), isotopic fractionations produced by this process predict an increase in δ18O of ∼15 to 50‰ for a Δ17O enrichment of 0.5‰. The δ18O of H2O and carbonates in ALH 84001 do not support an enrichment of ∼50‰ in atmospheric CO2. Smaller δ18O enrichments are not supported unless the balance of exchangeable H2O is much smaller than even the H2O/CO2 ratio in the present-day martian atmosphere and polar caps, because H2O-CO2 exchange can account for the δ18O enrichments observed in ALH 84001 carbonates.

Oxygen isotope fractionations produced by atmospheric gas-phase mass-independent chemistry on Earth produce positive Δ17O values in stratospheric CO2 (6, 33). For the martian atmosphere, the requirements for production of positive Δ17O values are the presence of ultraviolet (UV) radiation, O3 chemistry, CO2-O(1D) chemistry, and a negative Δ17O sink. Because of the thinness of the martian atmosphere, UV radiation has been sufficient to cause O3 chemistry throughout the martian atmospheric column throughout martian history. CO2-O(1D) chemistry follows from photodecomposition of O3 and CO2, with net production of 17O-enriched CO2 and 17O-depleted O2.

Many reactions may have contributed to the observed Δ17O values of martian meteorite waters and carbonates. We suggest that the predominant reactions responsible for producing the observed anomaly are as follows: a1: CO2 + hυ ⇀ CO + O(3P), wavelength (λ) < 227.5 nm; a2: ⇀ CO + O(1D), λ < 167 nm; b1: O(3P) + O(3P) + M ⇀ M + O2; b2: O(3P) + OH ⇀ H + O2; b3: O(3P) + NO2 ⇀ NO + O2; c: O2 + O(3P) + M ⇀ O# 3+ M; d: O# 3 + hυ ⇀ O#(1D) + O2, λ < 310 nm; e: O#(1D) + CO2 ⇋ CO#*3 ⇋ CO# 2 + O(3P), exchange reaction; f: CO# 2 + H2O ⇋ H2CO# 3 ⇋ CO2 + H2O#, exchange reaction, where hυ denotes the energy of a photon of frequency υ, # denotes isotopically anomalous oxygen, and ∗ denotes electronically excited states. Reactions a1 and a2 are CO2 photodissociation reactions. Reactions b1 through b3 are responsible for production of O2 in the martian atmosphere. Reaction c, an ozone formation reaction, has an associated mass-independent fractionation of ∼100‰ in δ17O and δ18O. Reaction d, an ozone photodissociation reaction, may transfer this mass-independent anomaly to O(1D). Reaction e is the exchange reaction between O(1D) and CO2 but may have a mass-independent fractionation associated with the CO*3transition state, which would only require the presence of O(1D), such as in reaction a2. Reaction f is the CO2-H2O oxygen exchange reaction. Two possible sources of the isotopic anomaly in CO2 include the combined action of reactions c and d, or a2, and reaction e.

Photochemical production of O(1D) by ozone or CO2 photolysis and subsequent exchange with CO2provides a mechanism for generating a 0.5‰ increase in Δ17O in the martian atmosphere with only a ∼1‰ increase in δ18O. A requirement of this model is that there be a corresponding sink for a negative Δ17O atmosphere-surface system. O2 is the atmospheric species that acquires negative Δ17O, but in the martian atmosphere there is only sufficient O2 to accommodate a Δ17O enrichment of 0.1‰ (assuming a CO2-O2 exchange value for Δ17O of ∼50‰). Another possibility is that a sink for O2 with negative Δ17O exists. An oxidized sink of ∼100 mg/cm2 in the martian regolith (ferric? oxides and hydroxides) with Δ17O ∼ –0.5‰ would accommodate atmospheric Δ17O of ∼ 0.5‰.

We suggest that mass-independent chemistry and oxidation of the martian regolith may be the principal reason for the positive Δ17O of the martian atmosphere at the time of carbonate growth in ALH 84001. A 0.5‰ contribution to positive atmospheric Δ17O by escape processes is not supported by the δ18O of ALH 84001 carbonates. Models that explain positive atmospheric Δ17O by heterogeneous accretion are not required and cannot be tested directly. Operation of the former processes permits temporal variations of atmospheric Δ17O if the size of exchangeable O reservoirs varied. Temporal changes in martian atmospheric Δ17O and D/H may explain the negative correlation between the D/H ratio and Δ17O of martian meteorite carbonate and H2O extracted by (14,27). Tests of these hypotheses include oxygen isotope and D/H analyses of additional secondary phases in martian meteorites and samples of atmosphere, soil, and ice to be returned from Mars.


View Abstract

Stay Connected to Science

Navigate This Article