Influence of Magnetic Fields on Precipitating Solar Wind Hydrogen at Mars.

Solar wind protons can interact directly with the hydrogen corona of Mars through charge exchange, resulting in energetic neutral atoms (ENAs) able to penetrate deep into the upper atmosphere of Mars. ENAs can undergo multiple charge changing interactions, leading to an observable beam of penetrating protons in the upper atmosphere. We seek to characterize the behavior of these protons in the presence of magnetic fields using data collected by the Mars Atmosphere and Volatile EvolutioN spacecraft. We find that backscattered penetrating proton flux is enhanced in regions where the magnetic field strength is greater than 200 nT. We also find a strong correlation at CO2 column densities less than 5.5 × 1014 cm-2 between magnetic field strength and the observed backscattered and downward flux. We do not see significant changes in penetrating proton flux with magnetic field strengths on the order of 10 nT.

Introduction

The magnetosphere and atmosphere of Mars provide us with a unique environment in our solar system. Unlike Earth, Mars does not have an intrinsic global magnetic field, which leaves the Martian atmosphere directly exposed to the solar wind. Photoionization of the upper Martian atmosphere by X‐rays and extreme ultraviolet solar radiation creates a highly conductive ionosphere, which acts as an obstacle for the incoming solar wind. The interaction between the magnetized solar wind plasma and ionosphere of Mars results in currents that create significant magnetic pressure, which slows and deflects the solar wind about the ionosphere, creating a dynamic induced magnetosphere (Intrilligator & Smith, 1979; Michel, 1971; Ramstad et al., 2020; Riedler et al., 1989). In addition to an induced magnetosphere, Mars also has strong crustal magnetic fields primarily located in the southern hemisphere with strengths reaching 1,000 nT at ionospheric altitudes (Acuña et al., 1998). These crustal fields are thought to have arisen from an ancient dynamo, and the disappearance of this intrinsic field is believed to have played a pivotal role in Mars's history (Lillis et al., 2008). Understanding the effects of both these remanent crustal fields and the induced magnetospheric environment on the Martian atmosphere is crucial as we seek to determine how the planet evolved to the cold, dry environment we observe today.

The present‐day atmosphere of Mars is primarily composed of CO2 along with small percentages of other gases, including atomic hydrogen (Mahaffy et al., 2015; Nier & McElroy, 1977). Due to the planet's weak gravity, hydrogen in the upper atmosphere can often reach thermal velocities greater than the planet's escape velocity. Hydrogen atoms on escaping and/or ballistic trajectories lead to the formation of an extended exosphere, or corona, about the planet. This neutral hydrogen corona extends past the planet's bow shock, allowing for direct interaction between the Martian atmosphere and the solar wind (Anderson, 1974; Anderson & Hord, 1971; Chaufrey et al., 2008). Solar wind protons can interact with neutral hydrogen in the corona through charge exchange, resulting in energetic neutral atoms (ENAs) that retain the solar wind velocity. These ENAs bypass plasma boundaries and penetrate down to altitudes of 120 km into the collisional atmosphere. Multiple charge‐changing reactions occur as these solar wind hydrogen atoms precipitate into the atmosphere and interact with other neutrals, resulting in a beam of protons (i.e., penetrating protons) measurable at altitudes between 120 and 400 km. Hydrogen ENAs and their charged byproducts have been observed and characterized by Mars Express (MEX) and the Mars Atmosphere and Volatile EvolutioN (MAVEN) missions. Previous studies have examined the behavior of precipitating hydrogen as a function of various temporal and global parameters (Brinkfeldt et al., 2006; Futaana et al., 2006; Girazian & Halekas, 2021; Gunell et al., 2006; Halekas, 2017; Halekas et al., 2015; Henderson et al., 2021). It has not been determined through observations; however, how the local magnetic fields in the Martian environment affect the behavior of these particles.

Over the past two decades, various groups have modeled the behavior of precipitating solar wind hydrogen in the Martian atmosphere (Bisikalo et al., 2018; Gérard et al., 2018; Holmström et al., 2002; Kallio & Barabash, 2001; Kallio et al., 1997; Shematovich & Bisikalo, 2021; Shematovich et al., 2011; Wang et al., 2018). Recent modeling studies have focused on how induced magnetic fields affect the precipitation of hydrogen ENAs in the Martian environment. These studies have found a strong dependence of both the backscattered and downward precipitating solar wind hydrogen ENA flux on magnetic field strength (Bisikalo et al., 2018; Shematovich et al., 2011). Shematovich et al. (2011) found that an induced horizontal magnetic field of 50 nT at an altitude of 437 km almost completely screened all downward proton precipitation. Bisikalo et al. (2018) found a similar shielding effect at 160 km; backscattered H+ flux as a percentage of the incident proton flux increased from 4.3% to 25% when a 30 nT horizontal magnetic field was included in their simulation.

Data from selected MAVEN orbits have been compared to modeling results (Bisikalo et al., 2018; Shematovich & Bisikalo, 2021), but a global comparison over the duration of the mission has yet to be done. Using in situ data collected over a period of 7 years by the Solar Wind Ion Analyzer (SWIA), magnetometer, and Neutral Gas and Ion Mass Spectrometer onboard MAVEN, we seek to determine how penetrating protons are affected by the Martian‐induced magnetosphere and crustal fields. We examine the behavior of downward and backscattered penetrating proton populations in the presence of magnetized regions at altitudes between 120 and 250 km and compare their observed behavior to previous modeling results.

Instrumentation

The primary data analyzed in this study were collected by SWIA onboard MAVEN. SWIA is a top hat electrostatic analyzer that measures ion energies and fluxes within the Martian environment (Halekas et al., 2013). SWIA sweeps through a range of applied voltages, filtering incoming ions based on their energies and entrance angles. In its coarse mode, SWIA collects data at a 4‐s cadence over 48 energies × 16 azimuthal angles × 4 polar angles. The coarse L2 data product is collected in a toroidal field of view and spans energies between 5.1 eV and 25 keV, resulting in thorough temporal and spatial coverage of the Martian plasma environment (Halekas et al., 2013).

We also utilize CO2 density measurements taken by the Neutral Gas and Ion Mass Spectrometer (NGIMS). NGIMS uses a quadrupole mass analyzer to measure mass to charge ratios of ions and neutrals with masses falling between 2 and 150 Da (Mahaffy et al., 2014). NGIMS measurements are acquired near MAVEN's periapsis between altitudes of 150 and 500 km (Jakosky et al., 2015; Stone et al., 2018).

In addition to data collected by SWIA and NGIMS, we utilize magnetic field measurements collected by the magnetometer (MAG) (Connerney et al., 2015). MAG consists of two independent triaxial fluxgate magnetometers that sample at a rate of 32 vector samples per second. Each axis covers ranges of ±512 nT or ±2,048 nT with a resolution of 0.008 nT, allowing for detailed characterization of magnetic fields in the Martian environment (Connerney et al., 2015).

Methods

We examine penetrating proton data collected by SWIA in conjunction with simultaneous MAG and NGIMS measurements between October 2014 and February 2021. We constrain our data set to orbits during which periapsis occurred at a solar zenith angle less than 90° for altitudes below 250 km, corresponding to the locations where penetrating protons are most clearly observed. With these constraints, we use SWIA coarse data to analyze penetrating proton spectra from 3,753 periapses with typical durations of ∼9 min.

During nominal solar wind conditions, only 4%–15% of upstream ENAs are converted to penetrating protons, leading to low count statistics (Halekas, 2017). To remove background counts, we determine an average background rate for each periapsis by finding the average number of counts over the periapsis in the three highest energy bins and dividing by the total duration of the periapsis in seconds. These bins were chosen since there are very few naturally occurring particles in the Martian environment at these energies as well as to quantify varying cosmic ray fluxes over the mission (Girazian & Halekas, 2021). We then examine each 4‐s penetrating proton energy spectrum. We first separate the proton signals into upward (i.e., backscattered) and downward propagating populations using the same methods employed in Girazian and Halekas (2021); we consider particles downward propagating if the angle between their velocity vector and a vector normal to the Martian surface at that observation point is greater than 90°. Once the backscattered and downward signatures are isolated, we apply a background correction to the 4‐s count spectrum by subtracting the average background count rate calculated for the periapsis from each energy‐anode bin.

We then sum over all anode bins to generate a count‐energy profile for each 4‐s spectrum. Focusing on energy bins above 100 eV in order to exclude potential low energy ions and spacecraft charging signatures, we locate the energy at which the peak number of counts occurs (E peak). We generate an angle‐averaged flux‐energy profile by converting the background‐corrected counts to differential energy flux and summing over all anode bins. We implement a quadratic spline method to interpolate flux values for energies between 0.25 and 1.75 E peak to encompass the core points of the flux distribution while excluding potential outliers in the low energy tail. This interpolated profile is then fit with a Gaussian to determine a characteristic height and full width at half maximum (FWHM) for each 4‐s profile. Figure 1 shows an example of ion data collected at periapsis by SWIA with downward and backscattered profiles. This process is repeated for each 4‐s spectrum collected during the periapsis, resulting in peak fluxes and energy widths collected at a multitude of regions within the Martian dayside environment.

Figure 1

(a) Example differential energy flux Solar Wind Ion Analyzer (SWIA) spectrum collected at periapsis on 6 November 2016. Characteristic penetrating proton signature can be seen between 10:01 and 10:17 a.m. (b) Example of angular‐averaged profile of downward‐propagating protons. Pink points represent background‐corrected, angle‐averaged flux. Gray points correspond to points that fall between 0.25 and 1.75 E peak. Blue points are a result of quadratic spline interpolation. Gaussian fit overplotted in black, which is used to find characteristic heights and full widths at half maximum (FWHMs) for each profile (c) Corresponding backscattered proton profile for timestamp in (b) (note different scale).

ENAs produced in the magnetosheath, precipitating magnetosheath protons, and/or precipitating pickup ions can occur in the same altitude range we are examining (Diéval et al., 2012; Gunell et al., 2006; Hara et al., 2017). In order to isolate precipitating solar wind hydrogen, we restrict the range for the energy of the spectral peak as well as the FWHM. Only ions whose E peak falls between 540 and 1,720 eV with FWHM/E peak < 5 are kept in this study. This energy range was chosen to ensure nominal solar wind conditions during data collection (Girazian & Halekas, 2021). We also employ an outlier rejection within each periapsis, omitting profiles that have peak fluxes greater than two times the median peak flux of all profiles within that periapsis.

Previous studies have demonstrated seasonal variability of precipitating solar wind hydrogen flux, in addition to changes in flux with respect to solar wind conditions (Brecht, 1997; Diéval et al., 2012; Halekas, 2017; Halekas et al., 2015; Henderson et al., 2021; Wang et al., 2013, 2018). To mitigate these temporal effects, we normalize the penetrating proton fluxes to gain unbiased measurements of these particles within the Martian ionosphere. We normalize the peak proton fluxes on a monthly basis for both the backscattered and downward populations by computing a mean downward flux for each calendar month and dividing all downward and backscattered peak fluxes measured within that time period by the mean flux. This largely eliminates seasonal variability within our data as shown in Figure S1 in Supporting Information S1, but there could still be variations associated with solar wind conditions and solar zenith angle. These factors should largely average out within our data since neither parameter should be strongly correlated with crustal magnetic fields as is confirmed Figure S2 in Supporting Information S1, which shows data normalized in the same fashion as described above except over each MAVEN orbit. We do observe a slight enhancement in the albedo for higher solar zenith angles as indicated in Figure S3 in Supporting Information S1, which is consistent with Girazian and Halekas (2021). However, we do not observe a strong change in trends with respect to magnetic field strength for different solar zenith angles. The monthly normalized flux values for the downward and backscattered populations will be the focus for the rest of this analysis.

Results

In order to better understand the physical processes governing the precipitation of penetrating protons in the Martian atmosphere, we examine the behavior of the observed backscattered and downward fluxes of these particles in the presence of magnetic fields. We investigate the influence of both magnetic field configuration and strength on these particles as well as the effects of collisional processes in the neutral atmosphere.

Magnetic Field Strength and Elevation Angle

To gain a better understanding of how the magnetic fields at Mars affect the precipitation of hydrogen ENAs, we examine the behavior of penetrating protons as a function of magnetic field elevation angle and strength. The magnetic field elevation angle is measured with respect to a vector tangent to the Martian surface at each penetrating proton measurement collected throughout a periapsis. An elevation angle of 90° (0°) corresponds to a radial (horizontal) magnetic field configuration (Brain et al., 2006; Hara et al., 2018). A previous study has shown that at altitudes between 200 and 350 km, directly precipitating ion fluxes are 2–3 times greater in strongly magnetized regions when the magnetic field is radially oriented and significantly lessened when the magnetic field is in a parallel configuration (Hara et al., 2018).

We do not observe these behaviors for penetrating protons at altitudes below 250 km as indicated in Figure 2. For magnetic field strengths less than 200 nT, both the downward and backscattered penetrating proton fluxes are nearly uniform with respect to the magnetic field elevation angle. Figure 2a shows that on average, the highest downward penetrating proton fluxes are observed at these lower magnetic field strengths, regardless of the elevation angle. The enhancement in flux shown in Figure 2a at ∼100 nT for B θ < 10° is apparently an artifact of solar wind variations; this feature is nearly eliminated when using an orbital normalization process as seen in Figure S2 in Supporting Information S1. We also see in Figure 2b that the lowest backscattered fluxes are measured at magnetic field strengths less than 200 nT at all elevation angles. The uniform distribution of both the backscattered and downward fluxes for different magnetic field geometries indicates that at these weaker field strengths, the penetrating protons are not primarily governed by electromagnetic forces. Above 200 nT, we see in Figure 2c that the backscattered flux reaches up to ∼80%–85% of the downward flux as well as an increase in the backscattered flux compared to lower field strengths in Figure 2b. These trends indicate that the particles are deflected/reflected by magnetic fields at lower altitudes in these regions. We also see in Figure 2a that the downward flux at higher magnetic field strengths decreases, which implies that the penetrating protons are reflected/deflected locally and/or at higher altitudes.

Figure 2

(a) Median, normalized downward penetrating proton flux as a function of magnetic field strength and elevation angle. (b) Median, normalized backscattered proton flux. (c) Ratio of median normalized backscattered and downward proton flux. (d) Total number of data points per bin.

Protons propagating perpendicular to the local magnetic field have gyroradii of ∼450 and ∼5 km for nominal solar wind conditions and magnetic field strengths of 10 and 1,000 nT, respectively. At the average altitude where these data are collected (∼170 km), the mean free path between H+‐CO2 charge exchange interactions is ∼29 km (Lindsay et al., 2005). It is therefore not surprising to see little variation in the observed fluxes in Figures 2a and 2b at magnetic field strengths below 200 nT. The gyroradii of penetrating protons at these magnetic field strengths would exceed the mean free path scale, making it difficult for electromagnetic forces to significantly affect their trajectories.

We believe we do not see the stark, monotonic trends of increased flux with increasing magnetic field elevation angles and magnetic field strength reported by Hara et al. (2018), most likely due to both different observation locations and the source of the protons analyzed in this study. Hara et al. (2018) examined the flux of solar wind protons and hydrogen pickup ions at altitudes between 200 and 350 km, whereas we focus on protons generated from ENA charge exchange interactions and examine their behavior below 250 km. Penetrating protons spend significantly less time in a charged state compared to solar wind protons and hydrogen pickup ions, reducing the effect of electromagnetic forces. This would result in a less apparent effect of magnetic fields compared to solar wind protons, which spend their lifetime in a charged state.

Column Density

In addition to electromagnetic forces, the collisional atmosphere of Mars also greatly affects the behavior of these particles as they precipitate. To better understand how neutrals within the Martian atmosphere affect precipitating solar wind hydrogen atoms, we examine how their flux changes as a function of column density. We integrate the CO2 number density measured by NGIMS along a line of sight between each proton observation point and the Sun (Henderson et al., 2021; Mahaffy et al., 2014). This traces the path of the precipitating solar wind hydrogen from the Sun to the point at which it is observed by SWIA, allowing us to quantify the column density through which the particle has passed. The specific details of our calculations are described in Henderson et al. (2021). Since CO2 comprises ∼95% of the Martian atmosphere within the altitude range we are examining, we determine that using column density values for this species suffices for a first‐order approximation. Only inbound verified NGIMS CO2 density values are implemented in this calculation since outbound measurements are known to have higher backgrounds due to residual gas left in the instrument at periapsis (Stone et al., 2018). Because of this data quality restriction, we are only able to calculate the column densities for 2,601 orbits.

We examine the normalized fluxes of the backscattered and downward penetrating protons at various column densities to determine the importance of the interaction of these particles with the neutral atmosphere versus magnetic field control. Figure 3 demonstrates how the backscattered and downward fluxes vary both with magnetic field strength and column density. There is a clear relationship between magnetic field strength and penetrating proton flux for column densities less than 5.5 × 1014 cm−2. In Figures 3b and 3e, we see that as magnetic field strength increases, the ratio of the backscattered and downward proton flux increases. This ratio increases linearly with magnetic field strength, indicating that penetrating protons are being significantly deflected by magnetic fields at these column densities. This linear relationship also steepens with decreasing column density as demonstrated by the best‐fit slopes, which shows that the larger magnetic fields are the dominant factor controlling the behavior of these particles in the upper atmosphere. We also observe a decrease in the downward flux as magnetic field strength increases in Figures 3a and 3d, and S4–S10 in Supporting Information S1, demonstrating that more protons are being deflected and backscattered by the Lorentz force in these regions of the atmosphere. In Figures 3a and 3d, and S4–S10 in Supporting Information S1, we also see the backscattered flux approach nearly the same value as the downward flux for magnetic field strengths greater than 200 nT at column densities less than 5.5 × 1014 cm−2, indicating that a significant fraction of the penetrating protons is deflected/reflected by the magnetic fields.

Figure 3

Behavior of downward and backscattered penetrating proton flux as a function of magnetic field strength for different column density ranges. The column density (altitude) increases (decreases) from Panels a to i. Slopes (m) and their standard errors (SEs) are displayed to the right of Panels c, f, and i. (a) Median normalized downward and backscattered penetrating proton flux with Q1 and Q3 displayed as lower and upper error bars, respectively, in 10 nT bins. These statistical quantities were chosen due to a nonnormal distribution of penetrating proton fluxes. Data were collected at column densities 1013–5.5 × 1013 cm−2. Panels d and g follow this format and are collected at column densities indicated in titles. (b) Ratio of median backscattered and downward fluxes for data collected at column densities 1013–5.5 × 1013 cm−2. Error bars displayed are Q1 and Q3 computed over 10 nT bins. Panels e and h follow this format and are collected at column densities indicated in titles. (c) Distribution of backscattered and downward data at column densities 5.5 × 1013–1014 cm−2. Panels f and i follow this format and are collected at column densities indicated in titles.

We see from Figures S4–S14 in Supporting Information S1 that generally both the backscattered and downward fluxes are lower than the monthly averages, suggesting that some fraction of the downward‐going particles has been scattered before reaching these column densities. After tedious examination of the data, the “bump” at ∼60 nT in Figures 3g and S11C–S14C in Supporting Information S1 appears to be an artifact caused by solar wind variations as previously discussed. Outside of this feature, Figures 3h and S11D–S14D in Supporting Information S1 show that there is not a strong dependence of the ratio of the backscattered to downward flux on magnetic field strength for column densities above 5.5 × 1014 cm−2. This is in contrast to column densities less than 5.5 × 1014 cm−2, for which we see a clear dependence of this ratio on the magnetic field strength. At higher column densities, it appears that collisions with atmospheric neutrals dominate as indicated by lower downward flux values and little dependence of the backscattered‐downward flux ratios on the magnetic field strength.

We do not see a clear contrast between the column density trends for horizontal versus radial magnetic field configurations. Figures S4–S14 in Supporting Information S1 show that for each column density interval, the trends observed for penetrating proton data collected in horizontal (B θ < 45°) and radial (B θ ≥ 45°) magnetic field configurations are very similar to one another. We would anticipate nearly no dependence of the observed downward flux on magnetic field strength in radial magnetic field regions since these particles' velocities would be essentially parallel to the magnetic fields. This trend is not observed possibly due to a lack of observations in cusp regions (i.e., where B θ ∼ 90°). Further measurements at high magnetic field elevation angles are required to better clarify this behavior.

Model Comparison

Previous studies have suggested that magnetic field orientation and strength play a crucial role in governing the behavior of precipitating solar wind hydrogen in the Martian atmosphere. Some models have predicted that magnetic field strength variations of ∼10 nT would significantly affect the amount of backscattered and downward flux of penetrating protons born upstream of the Martian bow shock and within the magnetosheath (Bisikalo et al., 2018; Shematovich et al., 2011). Our results do not clearly support these predictions; our observations indicate that magnetic field variations of this magnitude do not significantly affect the fluxes of the backscattered or downward penetrating protons.

Bisikalo et al. (2018) modeled the behavior of precipitating solar wind hydrogen in the presence of induced, horizontal magnetic fields ranging from 0 to 30 nT at an altitude of 160 km (or a corresponding column density of ∼3 × 1015 cm−2 for their input CO2 density profile). They found that the downward and backscattered flux of penetrating protons changed by orders of magnitude with just 10 nT increases in the induced magnetic field strength. We do not observe clear signatures of such changes, particularly at low magnetic field strengths. In Figure 3g, we see little dependence on magnetic field strength and flux for the downward population. Bisikalo et al. (2018) found that the backscattered flux of penetrating protons decreased by three orders of magnitude with an increase in the induced horizontal magnetic field from 0 to 30 nT. In contrast, the backscattered flux in Figure 3g shows very little variation across all horizontally configured magnetic field strengths ranging from 0 to 750 nT. Furthermore, we observe an ∼30% increase in the downward penetrating proton flux between 0 and 60 nT due to solar wind variations, which is in contrast to the nearly three orders of magnitude decrease in downward flux found by Bisikalo et al. (2018) between 0 and 30 nT. At this region within the ionosphere, our observations suggest that the dominant process affecting the precipitation of these penetrating protons is collisions with neutral constituents rather than magnetic shielding.

The disagreement between the model predictions and observations is surprising, and further investigation is required to better understand what underlying assumptions need to be changed in order to reconcile our findings. It is possible that these discrepancies may partially stem from the assumption of uniform horizontal magnetic fields made by Bisikalo et al. (2018), which would reflect/shield protons more efficiently than fields in perpendicular configurations. Further complicating direct comparisons, our H+ observations only represent a fraction of the reflection rate of H+/H at a given altitude. The observed H+ may be biased toward the high energy fraction of hydrogen (due to the energy dependence of the charge exchange cross section) and thus on average may not have experienced as much energy loss from collisions. We are not able to directly compare our results to other models due to the altitude range in which they were taken as well as the fact that these models examined protons born in the magnetosheath rather than in the upstream solar wind (Shematovich et al., 2011).

Discussion

We find that magnetic fields in the Martian environment play an important role in governing the behavior of penetrating protons in the upper atmosphere. At column densities less than 5.5 × 1014 cm−2, our observations suggest that magnetic field deflection is the dominant process controlling the behavior of penetrating protons. We find both downward flux depletion and enhancement of the backscattered flux as magnetic field strength increases. This indicates that the Lorentz force causes the penetrating protons to deflect around the magnetic field and scatter back to higher altitudes. At column densities above 5.5 × 1014 cm−2, we find little dependence of the backscattered and downward penetrating proton fluxes on magnetic field strength. This, in conjunction with lower flux values, suggests that collisional interactions with the neutrals dominate in this region of the atmosphere. We also find that magnetic field variations on the order of 10 nT do not obviously affect the behavior of the downward or backscattered penetrating proton fluxes in contrast with some previous model results. Further investigation into the behavior of other ion species in the presence of magnetic fields at Mars will help us to better understand the role the magnetospheric environment has played in the planet's atmospheric evolution over the past billions of years.

Supporting Information S1

Click here for additional data file.