Mars science and telecommunications orbiter

Exosphere and Atmospheric Escape

The Near-Space Environment

Boundaries

To answer questions concerning the erosion of the Martian atmosphere over time, we must have a detailed understanding of the physical mechanisms that are involved, their variability, and the causes of this variability.

It was realized early on that Mars’ ionosphere is too thin to routinely stand off the solar wind. Thus the effectiveness of the solar wind in scavenging the upper atmosphere hinged on the question of whether or not Mars possesses an Earth-like, dynamo-generated, planetary magnetic field to protect the atmosphere from erosion by the solar wind. Only very recently was it shown conclusively (by Mars Global Surveyor) that Mars does not posses such a dynamo field. However, large remanent magnetic fields were detected in localized regions on the planet, indicating that a dynamo field did exist, probably until some 4 billion years ago (Acuna, et al., 1999; Acuna, et al., 1998; Connerney, et al., 2004). The magnetic remanents, larger than any seen on Earth, add a dynamic character to the interaction as they rotate into view of the solar wind every Mars rotation, periodically adding magnetic pressure to help stand off the solar wind. They also indicate that during the early history of Mars a global magnetic field of significant strength must have influenced solar wind interaction-related loss processes.

Measurements by the Mars Global Surveyor (MGS) and by the Phobos-2 spacecraft led to a significant improvement in our knowledge of the regions and boundaries surrounding Mars (e.g., Winterhalter, et al., 2004). Figure 1, based primarily on recent insights gained from analyses of MGS and Phobos-2 data, depicts our current understanding of the Mars near-space environment. The main signatures include a bow shock which reflects the size of the Martian obstacle to the solar wind, a magnetosheath wherein the solar wind picks up planetary ions in a process known as mass-loading, and variously named features (such as "Magnetic Pileup Boundary") of the solar wind boundary (see Figure 2).

The boundary layer of the solar wind interaction, at altitudes between the exobase (~200 km) and the middle magnetosheath (~ 0.5 Rm) is a key region for characterizing the physics of solar wind interaction-related atmosphere escape. Neither Phobos-2 nor Mars Express has had the orbit or measurements to achieve this characterization. MGS did not carry thermal ion and electron instruments, and so it is impossible to determine whether the boundary layer at sites removed from the strong remanent magnetic fields is an "ionopause" in the sense of that observed at Venus on PVO. It has long been known from Viking mission measurements, that the ionospheric thermal pressure at Mars is frequently insufficient to balance the total incident solar wind pressure (e.g. Zhang, et al., 1991). Thus it is expected that the Martian ionosphere is magnetized, much like Venus' ionosphere during times of high solar wind dynamic pressure and/or low fluxes of ionizing solar radiation. The Viking Landers did not carry magnetometers, but the MGS magnetometer supports this expectation down to the altitudes sampled by that orbiter during its aerobraking phase (Acuna, et al., 1998). The presence of this magnetization is an essential factor in the erosion of the upper ionosphere by the solar wind interaction. Locally, crustal magnetic fields vastly complicate the topology of the boundary layer, and thus ionosphere escape, in as yet undetermined ways (e.g. Brain, et al., 2005; Mitchell, et al., 2001).

Both Phobos-2 and Mars Express established the existence of significant present-day escape of atmospheric constituents from the boundary layer in the form of ions observed in Mars’ solar wind wake (Lundin and al., 1990, 1993; Zakharov, 1992; Verigin, et al., 1991). However, the total flux, composition, and solar cycle variability of this escaping ion population is only roughly estimated, as is the pickup ion influx that sputters neutral atmosphere constituents at the exobase rather than escaping (Luhmann, et al., 1992; Luhmann and Kozyra, 1991). The lack of a magnetometer and other supporting contextual information on Mars Express seriously compromises that mission’s ability to obtain definitive results from its continuing ion escape measurements.

The solar wind pressure varies over the solar cycle, as does the upper atmosphere and ionosphere-controlling solar EUV flux, providing a changing view of the solar wind interaction that can be used to extrapolate related atmosphere escape into the past. As a whole, the measurements from the Viking, Phobos, MGS and Mars Express missions hint at the complexity and potential importance of the solar wind interaction with Mars in atmosphere escape, but no mission has had the necessary instrument complement, orbital coverage, and duration to sufficiently characterize this aspect of the evolution of the Martian atmosphere.

Escape Mechanisms

Mass loading and ion pickup are intimately related by the transfer of solar wind momentum and energy to planetary ions by the convective electric field, their common underlying physical cause. Mass loading is a manifestation of the replacement of solar wind protons by planetary ions or of the addition of significant masses of planetary ions to the solar wind. Ion pickup is the energization of planetary ions by the penetrating solar wind convection electric field, which can occur with or without the local penetration of the solar wind protons themselves. Whether one refers to mass loading or ion pickup, the end result can be removal of planetary ions if enough solar wind energy is transferred to accelerate them beyond the escape velocity of ~5 km/s. How the solar wind-related erosion varies with time, and on what external and/or internal parameters the variations depend, is unclear. Measurements of the boundary positions, and the escaping particle composition and fluxes are crucial aspects in the understanding of the impacts of the atmospheric erosion by the solar wind.

In addition to mass loading /ion pickup, there are other escape processes that are occurring at Mars, but have not been measured sufficiently to allow their overall influences to be determined. These include: photochemical escape; sputtering (by pick-up ions); possible other ion outflows in the wake and from the crustal field cusps; and bulk ionospheric plasma removal.

Photochemical escape relies on ionospheric photochemistry to generate suprathermal neutral populations above the escape velocity (e.g. by dissociative recombination). Nagy, et al. (1990) and Zhang, et al. (1993) estimated photochemical escape rates for Mars neutral oxygen based on the few Viking Lander altitude profiles of ionospheric composition. The observation of the hot oxygen corona of Mars (e.g. Paxton and Anderson, 1991), inferred from UV spectrometer measurements on Mariner 9, provides one way to study photochemical escape rates for oxygen by modeling the observed neutral exosphere altitude profile. Other photochemically escaping constituents, such as carbon, can also be indirectly detected by this method (Kim, et al., 1998), which remains unexploited. The sputtering of the neutral upper atmosphere at the exobase level by impacting atmospheric pickup ions, which models suggest is significant for atmosphere escape over time (Kass and Yung, 1995; Luhmann, et al., 1992) has yet to be established by any observation. It is difficult to observe the low energy sputtered neutrals except by indirect means. In addition, poorly characterized upper atmosphere properties play a significant role by determining what species are available to be ionized and removed. Nevertheless, impacting pickup ions can be detected by an ion spectrometer at the appropriate location, and upper atmosphere neutral particle density profiles can show evidence of some nonthermal heating by impacting pickup ions. In a similar manner model calculations (e.g. Ma, et al., 2004) indicate that ionospheric plasma outflow through the tail may be another escape channel, but no measurements exist to either confirm or contradict these suggestions. Study of special ion outflows that may be associated with the wake structure or with crustal magnetic fields are a subject of current analysis in the Mars Express community, but the lack of local magnetic field measurements represents major difficulties to those efforts. Similarly, the idea of bulk ionospheric plasma removal processes, such as proposed shear instabilities in the solar wind/ionosphere boundary layer (Penz, et al., 2005), also remain to be tested observationally.

Photochemical escape

Various investigators have modeled photochemical escape rates of heavy atoms from Mars, such as N, O, and C (e.g., Fox, 1993a, 1993b; Fox and Bakalian, 2001; Fox and Hac, 1997a, 1997b, 1999; Hodges, 2000; Jakosky, et al., 1994; Kim, et al., 1998; McElroy, et al., 1977; Nagy, et al., 2001). Planetary escape processes have been reviewed, for example by Chamberlain and Hunten (1987), Shizgal and Arkos (1996), and Hunten (2002).

The ¹⁵N/¹⁴N ratio in the Martian atmosphere was measured as about 1.62 times the terrestrial value (Nier and McElroy, 1976, 1977). The enhancement is presumably due to preferential escape of ¹⁴N. There are several sources of escaping nitrogen, including photodissociation (Brinkmann, 1971), photodissociative ionization and electron impact dissociative ionization of N₂, dissociative recombination of N₂⁺, and the reaction of N₂⁺ with O. Quantifying the sources involving N₂ requires in situ measurements of the number density profiles of N₂ as a function of altitude in the thermosphere from about 100 km to 350 km at both low and high solar activities, and as a function of solar zenith angle and season. The only solid information we have about thermospheric density profiles of N₂ comprises the measurements made by the Vikings 1 and 2 entry probe neutral mass spectrometers, each in one pass at one location, one solar zenith angle, and low solar activity through the atmosphere. The altitude profiles cannot be simply determined from the measured N₂ mixing ratio of about 2.5 %, because the vertical distribution depends on the diffusion coefficients, the strength of mixing due to turbulence and convection. Usually this distribution is modeled in 1D atmospheres with the use of an eddy diffusion coefficient. The MTGCM’s of Bougher (2000) and Bougher, et al. (2004b) do not contain N₂ as a separate species, and the altitude profiles derived from these models are therefore only estimates.

For the contribution to escape of N from dissociative recombination of N₂⁺, density profiles of the ions and electrons are required from below the F1 peak to altitudes above the exobase. Estimates of the O densities over a small altitude range have been derived from the Viking RPA measurements of CO₂⁺ and O₂⁺ at low solar activity. There is no direct information about the density profiles of either N₂⁺ or any of the other minor ions, although several models have been constructed of many major and minor ions (e.g., Fox, 1993a, 1993b, 2004; Krasnopolsky, 2002), which have allowed estimates to be made of the source due to dissociative recombination.

Quantifying the important source of escaping N due to reaction of N₂⁺ with O requires knowledge also of the density profiles of O. There is no direct information about either the mixing ratio of O or the altitude profiles. Estimates over a small altitude range at one altitude have been made based on the low solar activity Viking 1 O₂⁺/CO₂⁺ peak ratio (Fox and Dalgarno, 1979; Hanson, et al., 1977), and from remote sensing of the moderate solar activity Mariner 9 1304A emission (Stewart, et al., 1992). The inferred mixing ratios range by about a factor of 3 from about 0.7 % to 2% at 130 km. There is no in situ information about the high solar activity mixing ratios or density profiles. Experience with Venus tells us that there may be large variations in the photochemically produced species, such as O, CO, C and N. This is a major uncertainty in important quantities, which requires in situ measurement by an (open source) mass spectrometer capable of measuring O, C, and N.

The most recent calculations by Fox and Hac (1997b) have shown that the presence of a dense, early atmosphere of 100 to 500 mb at 3.8 Gyr before present reduces the isotopic fractionation of nitrogen to the observed value. Thus quantifying the present escape rates of N is necessary to estimate the escape rates as a function of time before present, and thus to determine whether a dense, early atmosphere is needed to reduce the isotope fractionation.

The largest source of photochemical escape of C is photodissociation of CO, followed by dissociative recombination of CO⁺ Other sources of escaping C have been found to be negligible compared to these two mechanisms (Fox and Bakalian, 2001; Nagy, et al., 2001). Quantifying the sources of these mechanisms requires knowledge of the density profiles of CO and CO⁺, the total electron density, and their variations with location, solar zenith angle, season and solar activity.

The Viking 1 neutral mass spectrometer measured a CO mixing ratio of about 0.42% at 120 km, and a density profile over the lower thermosphere for ~45 degrees solar zenith angle and low solar activity. The only measurements available are those from the Vikings 1 and 2 descent modules, at only two locations and for low solar activity. Models have been constructed by, for example, Krasnopolsky (1993) of the CO profile in the lower thermosphere. There is no direct information available about the solar activity variation of CO. The CO airglow emissions arise from both CO and CO₂, and so are not indicative of the CO mixing ratio, with the possible exception of the (14v) series of the CO fourth positive bands, which are produced by scattering of Lyman alpha by CO. The CO variations have been computed by the MTGCM of Bougher (2000), but high accuracy is not expected from this model.

Altitude profiles of CO⁺ have not been measured, but they have been modeled by several investigators (e.g., Fox, 1993a, 1997, 2004). The model altitude profiles of CO+ should be verified by in situ spacecraft measurements from high altitudes down to the F1 and E peaks (~130 and ~105 km at low solar activity). The altitudes of the peaks are predicted to be higher at high solar activity than at low solar activity ({Fox, 1995 #144; Fox, et al., 1995; Krasnopolsky, 2002). As it is now, it is not possible to distinguish the models, since there are no measurements of minor ions in the Martian ionosphere, other than the O⁺ “measurements” from the Viking RPA (e.g., Hanson, et al., 1977).

Quantifying the present escape rates of O is necessary for determining the history of water on Mars. The sources of photochemical O escape include dissociative recombination of O2⁺, and charge transfer of H⁺ to O in the Martian exosphere. The major source is dissociative recombination, determination of which source requires measurements of the density profiles of O₂⁺ and electrons. Information from the RPA on Viking and models suggests that, since O₂⁺ is the major ion, its density profile is close to that of the electrons. Dissociative recombination reactions tend to produce very energetic neutrals and their rate coefficients are dependent on the electron temperatures. Generally the dissociative recombination coefficients are of the form k=A (300/Te) ^(0.2-0.7). Thus determining the magnitude of the escape fluxes due to dissociative recombination requires knowledge of the altitude profiles of the electron temperature, and their variation with solar zenith angle and solar activity, for which comprehensive (e.g. time, season, solar cycle variations) information has not been available from either measurements or models. Estimates from the Viking RPA indicate values of Te of 2000-3000 K above 200 km, at one solar zenith angle and low solar activity. The values below 200 km have only been modeled. Modeling the Viking measured electron and ion temperature profiles has required imposing an ad hoc energy source at the top of the atmosphere, the origin of which is unknown. The ion temperature is also important in determining the high altitude density profiles of ions, and the rate coefficients for some ion-molecule reactions. An instrument that is capable of measuring the plasma temperatures and their variation with solar zenith angle, local time, season and solar activity, such as an RPA, would be necessary. Such an instrument would also be important for determining the source of energy at high altitudes that is necessary to model the plasma temperatures.

The computed values of the O escape rate have been compared to the thermal escape rate of McElroy (1972). McElroy et al. (1977) suggested that H and O escape in the stoichiometric proportions of water, (2:1). If this is true it implies that the oxidation state of the atmosphere has remained unchanged over time. Hydrogen escapes thermally mostly as H with a small additional component due to H₂ escape. There are also some non-thermal mechanisms for H escape, such as the charge transfer reaction of H⁺ with atomic H in the thermosphere and corona. These non-thermal escape rates are more important for D than for H. Recently, many investigators have modeled the rate of O escape due to dissociative recombination and found that the rates are too small to balance the H escape rate, and have thus questioned this “balanced escape hypothesis”. (We note here that this conclusion has been debated by Hodges (2000). Thus it is possible that the atmosphere of Mars may have been more reducing in the past, leading to larger densities of H₂ in the atmosphere. Fox (2003) modeled the ionosphere of Mars, for various H₂ mixing ratios from 4 to 100 ppm. Since H₂ reacts with N₂⁺, CO⁺, CO₂⁺ and O⁺, the nonthermal escape rates of heavy atoms arising from these species will be greatly reduced as the mixing ratio of H₂ increases, i.e. for a more reducing atmosphere. The implications of a more reducing atmosphere for escape rates of C and N due to dissociative recombination of CO⁺ and N₂⁺ are that the escape rates are reduced by a factor of ~6 and 2.3, respectively, as the H₂ mixing ratio in the thermosphere increases from 10 ppm, approximately the current value (Krasnopolsky and Feldman, 2001) to 100 ppm. The escape rates of N and C due to photodissociation of N₂ and CO are reduced by factors of 2-3 as the H₂ mixing ratio increases from 10 to 100 ppm. Since the modeled rates of N escape seem to be too large at present, and integrating backwards in time actually over-fractionates the N isotopes, a more reducing state of the atmosphere in the past would decrease the escape rates and provide a possible other or additional mechanism to that of a dense early atmosphere for inhibiting the escape of Nitrogen.

In summary, quantifying the photochemical escape rates of heavy atoms and the thermal escape rates of H requires in situ measurements of both major and minor ions, and major and minor neutral species. The mass range required is from 1 to 44 amu for both ions and neutrals. Also required are the neutral, ion and electron temperatures as a function of altitude. The escape rates are expected to be dominated by those at high solar activity, where we have essentially no in situ data at present. Thus measurements at high solar activity are essential. The only in situ data we have at low solar activity is from the neutral mass spectrometers and the retarding potential analyzers of Vikings 1 and 2, which sampled the atmosphere at only two locations, and at low solar activity. We have no data on the minor ion profiles, since the Viking RPA could measure only O₂⁺, CO₂⁺ and O⁺.

Ion Escape

The escape of ions is by traditional ion-pickup processes (McElroy, et al., 1977 Luhmann, 1990), and by ion outflows (Ma, et al., 2004). The computed rates of traditional ion pickup are computed as the rates of ionization of the atoms above the ionopause, followed by “pickup” by the convection electric field of the flowing solar wind plasma. Fox (1997) suggested, however, that theoretically all the ions above the photochemical equilibrium region could be stripped away, rather than only those created above the ionopause. This theory has been confirmed by MHD (Ma, et al., 2002; Ma, et al., 2004) models in which the computed escape rates are more similar to those computed by Fox than to those computed as traditional ion pickup. On Mars, the major ion flowing away from the planet is predicted to be O₂⁺ (Fox, 1997), whereas the main ion observed in the planetary wake is O⁺ (e.g., Lundin, et al., 1989). The MHD model of Ma, et al. (2004) suggested the major escaping ion is O⁺ rather than O₂⁺, but their model ion profiles showed an F2 peak of O⁺, while other models have shown that O₂⁺ is the major ion over the whole ionosphere. Their model was computed with ad hoc plasma temperature profiles designed to produce an F2 peak of O⁺ at high altitudes. This is yet another reason for measuring the values of Te and Ti in the ionosphere at several values of the solar zenith angle and solar activity.

The escape of species as ions has the potential to be more important than those of the photochemical processes. For example, the predicted O⁺ escape rates are larger than the H nonthermal escape rates, and can more than make up for the smaller predicted values of the photochemical escape of O. H atoms will escape also as H⁺. Quantifying the escape rates of ions requires density profiles of minor ions from 1 amu (H⁺) to 44 amu (CO₂⁺).

The escape rates of ions can be estimated by determining the scale heights of the ions above the photochemical equilibrium region. This was done for O₂⁺ by Chen et al. (1978) and by Fox (1997). The altitude profiles could only be fitted if an upward flux or velocity boundary condition were imposed on the ions. Determining the scale height compared to that expected for zero flux requires altitude profiles of ions from below the peaks, i.e., from about 110 km to 350 km. Measurements that would provide only high altitude scale heights are of little value for the determination of escape rates of ions because comparisons to models with zero flux upper boundary conditions would not be possible without simultaneous measurements of the peaks. It should be noted here that an upward flux boundary condition in a one-dimensional model probably actually represents the divergence of the horizontal fluxes of ions (Shinagawa and Cravens, 1989).

In summary, determining the rates of escape of ions requires in situ measurements of major and minor ions from high altitudes, i.e., in the solar wind, down to below the F1 peaks (near 110 to 140 km), as well as measuring the fluxes of ions in the wake of the planet.