Atmospheric and water loss from early Venus

Abstract

Previous interpretations of the Pioneer Venus mass spectrometer data of the deuterium to hydrogen $(D/H)$ ratio of $1.9\times {10}^{-2}$ or $120\pm 40$ times the terrestrial value indicate that Venus may have had at least an ${\mathrm{H}}_2\mathrm{O}$ content of the order of about 0.3% of a terrestrial ocean (TO), and even much more during and shortly after the accretion period of $⩽300\mathrm{Myr}$, depending on the unknown ratio of a continuous supply of ${\mathrm{H}}_2\mathrm{O}$ by comets to a hydrogen blow-off loss and impact erosion of the early atmosphere. In view of the low ${\mathrm{H}}_2\mathrm{O}$ abundance in the present atmosphere, several studies suggest that the planet should have lost most of its ${\mathrm{H}}_2\mathrm{O}$ during the early high X-ray, EUV and solar wind period of the active young Sun. Because oxygen did not accumulate in Venus’ atmosphere it is commonly believed that a part of the oxygen from dissociated ${\mathrm{H}}_2\mathrm{O}$ vapor was dragged off to space along with the escaping hydrogen during a blow-off period, or could have oxidized the surface minerals to produce FeO and ${\mathrm{Fe}}_2{\mathrm{O}}_3$ to the depths of a few kilometers to tens of kilometers depending on the initial amount of ${\mathrm{H}}_2\mathrm{O}$. We use in the present study, for the first time, multi-wavelength X-ray and EUV (XUV) observations by the ASCA, ROSAT, EUVE, FUSE and IUE satellites and stellar winds inferred from mass loss observations by the Hubble Space Telescope of solar proxies with ages $<4.6\mathrm{Gyr}$ for the investigation of how efficiently the radiation and particle environment of the young Sun could have influenced the evolution of the early Venusian atmosphere and its ${\mathrm{H}}_2\mathrm{O}$ inventory due to the removal of oxygen picked up by the solar wind. For modelling the Venusian thermosphere over the planetary history we apply a diffusive-gravitational equilibrium and thermal balance model and investigate the heating of the early thermosphere by photodissociation and ionization processes, due to exothermic chemical reactions and cooling by ${\mathrm{CO}}_2$ IR emission in the $15\mathrm{\mu }\mathrm{m}$ band. Our model simulations result in expanded thermospheres with exobase levels between about 200 km at present and about 2200 km 4.5 Gyr ago. Moreover, our results yield high exospheric temperatures during the active phase of the young Sun even if we assume a “dry” ${\mathrm{CO}}_2$ atmosphere with similar composition that is observed on present Venus of more than 8000 K after the Sun arrived at the zero-age-main-sequence (ZAMS). Exospheric temperatures above about 4000 K lead to diffusion-limited escape and high loss rates for atomic hydrogen. The duration of this blow-off phase for atomic hydrogen essentially depends on the mixing ratios of ${\mathrm{CO}}_2$, ${\mathrm{N}}_2$ and ${\mathrm{H}}_2\mathrm{O}$ in the early Venusian atmosphere and could last between about 150 to several hundred Myr, which could result in a large thermal loss of hydrogen from Venus. For studying how much of the ${\mathrm{H}}_2\mathrm{O}$-related oxygen could have been lost to space by the ion pick up process due to the stronger solar wind and higher XUV fluxes of the young Sun we used our modelled atmospheric density profiles and studied the loss of ${\mathrm{O}}^{+}$ ion pick up from the upper atmosphere of Venus over the planet's history by applying a numerical test particle model. Depending on the used solar wind parameters, our model simulations show that ion pick up by a strong early solar wind on a non-magnetized Venus could erode during 4.6 Gyr more than about 250 bar of ${\mathrm{O}}^{+}$ ions, that corresponds to an equivalent amount of one terrestrial ocean. Finally, we discuss the implications of our findings for the formation of the Venusian atmosphere and discuss our results in the frame of previous studies.

Introduction

Present Venus is an extremely dry planet containing very little ${\mathrm{H}}_2\mathrm{O}$ vapor of about 200–300 ppm (Hoffman et al., 1980; Moroz et al., 1979; Johnson and Fegley, 2000) in its atmosphere. On the basis of the analysis of the Pioneer Venus large-probe neutral mass spectrometer (LNMS) data it was found that Venus’ atmosphere is enriched in D over H relative to Earth by a factor of $\approx 120\pm 40$. From this one can conclude that the planet was once more “wet”. McElroy et al. (1982) used the present hydrogen escape rates and showed that these loss rates over 4.5 Gyr would imply a lower limit of ${\mathrm{H}}_2\mathrm{O}$ on Venus of $\approx 0.3\%$ of a terrestrial ocean (TO¹) (e.g., see also Donahue et al., 1982; Lammer and Bauer, 2003). Donahue and Hartle (1992) and Hartle et al. (1996) deduced the amount of ${\mathrm{H}}_2\mathrm{O}$ lost from Venus from the measured $D/H$ ratio based on ion mass spectrometer measurements in a range of an equivalent TO-depth of several meters to tens of meters.

For the amount of primordial ${\mathrm{H}}_2\mathrm{O}$ on Venus one has to consider two possibilities. The first hypotheses assumes that Venus was formed from condensates in the solar nebula that contained little ${\mathrm{H}}_2\mathrm{O}$ (Holland, 1963, Lewis, 1970, Lewis, 1972, Lewis, 1974, Lewis and Prinn, 1984), while the second hypothesis argues in agreement with previous theories (Dayhoff et al., 1967, Walker, 1975) for an water abundance more comparable to that on Earth and Mars (e.g., Wetherill, 1981, Donahue and Pollack, 1983, Kasting and Pollack, 1983, Morbidelli et al., 2000, Raymond et al., 2004).

The supply of ${\mathrm{H}}_2\mathrm{O}$ to the Venus’ atmosphere by comets was studied by Lewis (1974), Grinspoon and Lewis (1988) and more recently by Chyba et al. (1990). However, Grinspoon and Lewis (1988) have also argued that present Venus’ water content may be in a steady state where the loss of hydrogen to space is balanced by a continuous input of ${\mathrm{H}}_2\mathrm{O}$ from comets or from delayed juvenile outgassing. In case the external water delivery occurs, then no increase of Venus’ past ${\mathrm{H}}_2\mathrm{O}$ inventory is required to explain the observed $D/H$ ratio. The enrichment of D could conceivably have started out of more or less “dry” condition, as originally was suggested by Lewis (1972).

On the other hand, Walker (1975) pointed out that if the present escape flux of ${\mathrm{H}}_2\mathrm{O}$ is in a steady state with the expected input of juvenile water and is augmented by an additional source like comets, the outgassing rate would be by a factor of about $5\times {10}^3$ too weak to reproduce the equivalent amount of a TO in 4.5 Gyr. However, the initial water content on early Venus may have been larger, because a substantial amount of ${\mathrm{H}}_2\mathrm{O}$ is required to explain the onset of the large greenhouse effect observed at present (e.g., Shimazu and Urabe, 1968, Rasool and DeBergh, 1970, Donahue et al., 1982, Kasting and Pollack, 1983; Chassefière, 1996a, Chassefière, 1996b). From these considerations one may expect that the early outgassing of ${\mathrm{H}}_2\mathrm{O}$ from Venus should have been much more efficient than it is at present to generate an equivalent amount of a TO (e.g., Hunten et al., 1987). However, we note that it is difficult to estimate the initial ${\mathrm{H}}_2\mathrm{O}$ reservoir on early Venus from the present $D/H$ ratio, because all estimations depend on the “non-steady” or “steady” state Venus’ water content and possible unknown $D/H$ fractionation processes which could have had different rates over the planet's past compared to the present time.

Kasting and Pollack (1983), Kasting et al. (1984) and Kasting (1988) argue that early Venus may have had once a warm water ocean stable on the surface, but the runaway greenhouse effect shifted the climate to an extremely hot and dry planet. Another alternative would be a humid greenhouse effect where the water on Venus was, because of the high temperatures, mainly in a vapor phase. These authors argue that in both cases the hydrogen could have been lost by diffusion-limited hydrodynamic escape driven by the high X-ray and EUV (XUV) radiation expected during the early active period of the young Sun (Zahnle and Walker, 1982, Ayres, 1997, Guinan and Ribas, 2002, Ribas et al., 2005).

Watson et al. (1981) studied hydrodynamic loss from early Venus under the assumption that the upper atmosphere consisted of pure molecular hydrogen. They found that XUV-driven energy-limited hydrodynamic conditions would have resulted in the upper limit of the hydrogen escape flux of about ${10}^{12}{\mathrm{cm}}^{-2}{\mathrm{s}}^{-1}$ ($\approx 4.5\times {10}^{30}{\mathrm{s}}^{-1}$). Such an escape flux would evacuate the amount of a terrestrial water ocean in about 280 Myr (Watson et al., 1981).

Kasting and Pollack (1983) argued that the Watson et al. (1981) study neglected the so-called “cold trap” at the mesopause which determines the number density of hydrogen at about 90–130 km altitude, so that their energy-limited escape rate may be an overestimation of the hydrogen loss rate. By applying a more accurate atmospheric model which included a maximum assumed hydrogen mixing ratio of about 0.46 at the mesopause during a runaway greenhouse period they found maximum hydrodynamic hydrogen escape fluxes of about $2.7\times {10}^{11}{\mathrm{cm}}^{-2}{\mathrm{s}}^{-1}$ ($\approx 1.2\times {10}^{30}{\mathrm{s}}^{-1}$), $2.3\times {10}^{12}{\mathrm{cm}}^{-2}{\mathrm{s}}^{-1}$ ($\approx 1\times {10}^{31}{\mathrm{s}}^{-1}$) and $3.\times {10}^{12}{\mathrm{cm}}^{-2}{\mathrm{s}}^{-1}$ ($\approx 1.4\times {10}^{31}{\mathrm{s}}^{-1}$) for 1, 8 and 16 times higher XUV fluxes compared to that of the present Sun. However, their estimated atomic hydrogen escape fluxes are even higher at XUV fluxes more than one, than those of Watson et al. (1981).

Chassefière (1996b) showed that oxygen produced by photodissociation of ${\mathrm{H}}_2\mathrm{O}$ vapor in an early stage of the evolution of terrestrial planets may have been lost by hydrodynamic escape only in relatively modest amounts. It was found that if dynamic escape of hydrogen contained in an ocean equivalent to a few TO occurred at a slow rate over the first Gyr, only about 2% of atomic oxygen could be lost from Venus.

On the other hand, the same study indicates that a short period of escape due to XUV-driven energy-limited hydrodynamic conditions could remove a primitive ocean on Venus of about 0.45 TO ($\approx 120\mathrm{bar}$) with an average depth of about 1300 m, and $\approx 30\%$ of atomic oxygen initially contained in the ocean could be dragged away by a planetary hydrogen wind. In such a case Venus’ early atmosphere would have been left with an oxygen-rich atmosphere (Chassefière, 1996b).

Chassefière (1997) argued that the problem of the fate of oxygen atoms left behind ($\approx 85\mathrm{bar}$) by hydrodynamically escaping hydrogen can only be solved if a primitive solar wind enhanced to ${10}^3$–${10}^4$ times bombarded and heated an extended dense upper atmosphere. Even if the hydrogen escape was diffusion-limited, oxygen atoms could have been efficiently dragged off, so that ${\mathrm{H}}_2\mathrm{O}$ molecules could have been lost in a flux of about $6\times {10}^{13}{\mathrm{cm}}^{-2}{\mathrm{s}}^{-1}$. With such escape rates about one TO could completely be lost in about 10 Myr (Chassefière, 1997).

Due to the lack of astrophysical observations of solar wind parameters of young solar-like stars, Bauer (1983) estimated the solar wind-induced loss of ${\mathrm{H}}_2\mathrm{O}$ from Venus of about 6% of the atmospheric mass by assuming a T-Tauri wind, which is about ${10}^6$ times larger over about 1 Myr (e.g., Cameron and Pollock, 1976). This loss rate corresponds to about 2% of a TO. However, T-Tauri winds develop in the early stage of the stellar evolution when planetary formation has not yet completed.

Moreover, Bauer (1983) used the simple idealized mass loading model of Michel (1971) which is based only on the charge exchange process between the solar wind and the atmosphere, neglecting finite ion gyroradii and non-spherical ionopause geometry, atmospheric expansion due to higher XUV radiation, and changes in photoionization and electron impact ionization rates. Because solar radiation and particle fluxes play a major role in all atmospheric processes, the evolution of the Venusian atmosphere can only be understood within the context of the evolving solar energy and particle fluxes.

The aim of the present work is, therefore, to apply for the first time a thermospheric balance and diffusive equilibrium model to the study of the early Venusian thermosphere–exosphere region and the efficiency of the ${\mathrm{O}}^{+}$ ion loss by the pick up process due to the evolving solar radiation and particle environment over the planet's history. We consider in Section 2 the effects of higher solar XUV fluxes as obtained by the Sun in Time multiwavelength program (X-rays to UV) of solar proxies with ages covering 0.13–7 Gyr (Guinan and Ribas, 2002, Ribas et al., 2005). To investigate the solar wind mass flux of the young Sun we use minimum and maximum values inferred from stellar wind particle collisions with the interstellar medium which surrounds the solar proxies with different ages (Wood et al., 2002, Wood et al., 2005, Linsky and Wood, 2004).

In Section 3 we discuss the heating of the present “dry” Venus atmosphere, apply a thermospheric model, and simulate the thermosphere–exosphere temperature and number density distributions as a function of altitude for a “dry” 96% ${\mathrm{CO}}_2$ atmosphere, also for ${\mathrm{N}}_2$-rich atmospheres and various XUV flux values. Further, we simulate ionospheric profiles as a function of XUV flux values using the ionospheric model of Shinagawa et al. (1987). Hot oxygen atoms which are generated by dissociative recombination of ${\mathrm{O}}_2^{+}$ molecular ions are calculated by applying the two-stream Monte Carlo model of Lammer and Bauer (1991) and Lammer et al. (2000a). By using the calculated exospheric temperatures, “hot” and “cold” exobase number densities, and XUV-induced expanded exobase altitudes we apply Chamberlain's equations (e.g.,, Chamberlain, 1963) for the calculation of the exospheric number density profiles as a function of altitude over large planetary distances.

In Section 4 we apply a numerical test particle model which includes ionization by solar wind charge exchange, electron impact and XUV radiation for the calculation of the ${\mathrm{O}}^{+}$ pick up ion loss rates over the planet's history. Finally, we discuss in Section 5 the results of our study and the implications for the evolution of a “dry” or “wet” early Venus atmosphere and its ${\mathrm{H}}_2\mathrm{O}$ inventory.