Heavy ion escape from Martian wake enhanced by magnetic reconnection

Lei Wang Institute of Geology and Geophysics, Chinese Academy of Sciences Can Huang (  huangcan@mail.iggcas.ac.cn ) Department of Geophysics and Planetary Science, University of Science and Technology of China Yasong Ge Institute of Geology and Geophysics, Chinese Academy of Sciences A. M. Du Rongsheng Wang University of Science and Technology of China https://orcid.org/0000-0002-9511-7660 Tielong Zhang Institute of Space Science and Applied Technology, Harbin Institute of Technology, Shenzhen Jinqiao Fan Institute of Geology and Geophysics, Chinese Academy of Sciences


Background
Mars has no global magnetic eld 8 . As a consequence, the solar wind interacts directly with its upper atmosphere and ionosphere. The solar wind erosion of the Martian atmosphere may explain the dehydration of present-day Mars. Ion escape, in particular, through magnetotail or plasma wake, is a signi cant part of Martian atmospheric escape 9,10 . The main ion escape channels for Mars involve polar wind, boundary layer, ion pick-up, and plasma sheet 2,11 . Ion escape through the tail or wake region, however, should be ine cient because of Mars' obstruction. It has been reported that bursty and e cient ion escape processes exist in tail regions 4 . A similar process occurs in the Venusian tail 7 , suggesting this process should be a common characteristic of unmagnetized planets.
Magnetic reconnection is a fundamental process that explosively dissipates magnetic energy 12 and depletes celestial bodies' charged particles, resulting in phenomena such as solar coronal mass ejection 13 , disconnection of the comet tail 14 , and atmospheric ion loss in planets 15 . Magnetic ux ropes often are identi ed in reconnection exhaust and diffusion region 16,17 , with scales ranging from electron inertial length 18 to ion inertial length 19 . They wrap substantial plasma and cause large-scale ion escape during their release from the planetary tail 7 . Such ux rope structures can also form in the Martian dayside ionosphere from macroscopic instabilities because of the plasma ow shear and subsequently can be dragged into the tail 20,21 . Thus, to evaluate the role of magnetic reconnection for ion escape, it is necessary to assess the ion content within ux ropes of different origins. We rst report the two types of ux ropes observed simultaneously during a single crossing of the Martian magnetotail current sheet ( Fig. 1). The oxygen ion escape ux within the ux rope generated by the reconnection was more than twice that formed by boundary instabilities.

Results
One Martian current sheet crossing event. On 1 March 2021, the Mars Atmosphere and Volatile EvolutioN (MAVEN) orbiter 22 experienced a southward crossing of the tail current sheet. It successively grazed three ux ropes, as depicted in Fig. 1a 23 observed a reversal in the B x component (Fig. 2b) and simultaneously a dip of |B| ( Fig. 2a) at 10:14:33 UT, indicating that MAVEN crossed the tail current sheet. The crustal magnetic eld obtained from the spherical harmonic model 24 was vanishingly small throughout the interval (Fig. 2a).
The energetic ion (up to a few hundreds of eV) population shown in Fig. 2e also implies that MAVEN crossed the current sheet 9,25 . Figure 2c shows that the current sheet was embedded in a relatively wide channel of plasma ow mainly in the anti-sunward direction. The Suprathermal and Thermal Ion Composition (STATIC) 26 instrument recorded multiple ion species, including , and across the current sheet (Fig. 2f). Heavy ions were more abundant than protons (Fig. 2d).
Referring to Fig. 2a, three evident local magnetic eld enhancements were detected during the current sheet crossing. Meanwhile, distinct bipolar variations in the B z component appeared around |B| peaks ( Fig. 2b), showing the characteristic features of ux ropes. For convenience, we name the three ux ropes FR1, FR2, and FR3. Note that the axial core eld of FR1 and FR3 was mainly in the x-direction, while FR2 mainly aligned with the y-direction, implying that FR1 and FR3 may have different origins with FR2.
Magnetic reconnection and ux rope embedded in current sheet. As shown in Fig. 3, we investigate the MAVEN measurements within the overall current sheet (CS) coordinates. Figure 3a  to 100 km/s (Fig. 3b). This speed was comparable to the Alfvén velocity in the adjacent lobe region (Fig.   3c), signaling that the proton ow was the out ow in the reconnection exhaust 27,28 . Such tailward proton ow and the negative B NCS suggest that the reconnection site was on the sunward side of the MAVEN path (as indicated in Fig. 1b). In collisionless magnetic reconnection, Hall currents usually are directed toward the X-line along the magnetic eld lines just inside the separatrices and away from the X-line along the separatrices 29,30 . Such a current system leads to a quadrupole Hall magnetic eld. During 10:14:51-10:15:26 UT (shadowed region), FR2 traversed MAVEN quickly along the -L CS and -M CS directions. Meanwhile, a distinct polarity reversal was found in the N CS direction, while a unimodal peak was found in the M CS direction (Fig. 3e). According to our method of determining the axial direction of a ux rope, the M FR2 axis is adopted as the FR2 axial direction. The angle between M FR2 and M CS axis was about 162°, that is, the axial core eld of FR2 was approximately antiparallel to the M CS direction or was along the reconnection guide eld. These observations are consistent with a ux rope formed by magnetotail reconnection 31,32 . Figure 3f exhibits the electron differential energy ux from the Solar Wind Electron Analyzer (SWEA) 33 . Enhancements of energetic electron ux above 500 eV can be seen during the crossing of the current sheet. Time slices of the electron energy spectrum near core regions of three ux ropes are provided in Fig. 3g-i. The spectrum shapes in these regions are different. Electron spectra within FR1 and FR3 displayed approximate Maxwellian distributions. For FR2, however, there was a at-top population between 150 eV and 320 eV superimposed on a Maxwellian distribution, which is a typical feature of magnetic reconnection exhaust 34,35 . Moreover, suprathermal (300-800 eV) electrons substantially increased within FR2. This combined evidence indicates that FR2 was generated by reconnection and expelled quickly from the X-line in the exhaust.
Flux ropes on the edge of current sheet. We also perform an MVA analysis on the magnetic eld measurements of FR1 and FR3 (Fig. 4) (Fig. 4a, c). Both axes were quasi-perpendicular (128° and 119°) to the cross-tail direction of the current sheet. Hodograms on the L FR -M FR plane show that the magnetic eld variations of FR1 occured mainly in quadrants 1 and 4 and rotated clockwise (Fig. 4b), whereas that of FR3 occured mainly in quadrants 1 and 2 and rotated counterclockwise (Fig. 4d). Because the signs of are positive for both of them, the helicity of FR1 and FR3 are left-handed and right-handed, respectively.
The quasi-perpendicularity between the axis and M CS , as well as the opposite helicities of the ux ropes on either side of the current sheet, suggesting that they may be generated by dayside ionospheric instabilities and dragged into the tail by the solar wind 36,37 .
To compare the MAVEN observations with the theoretical prediction, we study the con guration of magnetic eld lines hanging on the dayside ionosphere. In this event, MAVEN traveled from the nominal bow shock and magnetic pile-up boundary to the site adjacent to the current sheet during 09:05:00-10:10:00 UT (see Supplementary Fig. 1). Four snapshots of the magnetic eld clock angle in the M CS -N CS plane are shown along the orbit. During this time, the clock angle changed less than ±15° from the Martian magnetosheath to the tail current sheet, suggesting that the orientation of the incoming and hemisphere. As shown in the schematic (Fig. 1c), the velocity shear between the magnetosheath and the ionosphere could lead to eld line twisting and rolling up to form a ux rope. In this scenario, the twisted eld lines on the two sides of the current sheet rotate in opposite directions. The observed opposite helicities of FR1 and FR3 are in good agreement with this theoretical prediction, given that FR1 and FR3 were observed on both sides of the current sheet.
Note that the observed densities (about 0.7 cm -3 for and 1.2 cm -3 for ) of oxygen ions in this event are lower than that in the typical plasma sheet 39 . The oxygen ion escape caused by the reconnection in this study may even have been underestimated. The mean oxygen ion ux within FR1, FR2, and FR3 were 4.8×10 6 cm -2 s -1 , 1.1×10 7 cm -2 s -1 , and 4.0×10 6 cm -2 s -1 , respectively. The heavy ion escape rate in the ux rope generated by reconnection was also greater than that generated by boundary instabilities. In the current sheet normal direction, the spacecraft velocity was about -2.9 km/s, and the average proton velocity was about 3.6 km/s, so the half-width of the current sheet was about 720 km, given the crossing time was about 220 s. We also derive ion inertial lengths of ~620 km for (d H ), ~1180 km for and ~1250 km for from the average ion densities during the crossing of the current sheet (see Fig. 2d). Thus, this current sheet was as thin as the characteristic proton length that recently has been detected frequently [40][41][42] . Theoretically, magnetic reconnection can be triggered only after a current sheet becomes thinner than d H 43 , implying that magnetic reconnection in the Martian tail may occur more frequently and thus be critical to the Martian ion escape process. This enhancement of ion escape due to reconnection should also occur on other non-magnetized planets like Venus.

Methods
Local coordinate system for the current sheet and ux ropes. We obtain the local (LMN) coordinate system from the minimum variance analysis 44 . To obtain the overall current sheet coordinate system, we apply MVA on B CS observed between 10:13:00 and 10: Calculation of axial orientation and handedness of a ux rope. This method is valid regardless of whether or not the ux rope is force-free and is based only on single-spacecraft measurements. According to characteristics of azimuthal and axial elds (bipolar for the azimuthal and unipolar for the axial) of ux ropes, we calculate in the ux rope LMN coordinate system, where ξ represents L FR or M FR, and j is the temporal index. The larger (smaller) one corresponds to the axial component (azimuthal component). We de ne the sign of the helicity as . The positive (negative) helicity means that the handedness of a ux rope was right-handed (left-handed). The term is the directional area on the B L -B M plane. Its magnitude denotes the integral area, and the positive (negative) sign denotes the +N FR (-N FR ) direction. The term means the sign of observed helicity is also associated with the relative trajectory of crossing. V p and V S are the velocities of the protons and the satellite, respectively.
Estimation of oxygen ions' escape rate. In the overall exhaust, oxygen ions mainly moved along the -L direction of the current sheet. Thus, we obtain uxes of oxygen ions in the exhaust and diffusion regions by integrating the product of densities and V L . To compare the ux of oxygen ions within the three ux ropes, it is reasonable to take the mean value of the product as an indicator. Estimating the cross-section of the exhaust is critical for calculating the ion escape rate. Observations have found that the length of the cross-tail current of Mars is even larger than the diameter of the planet 46 . We assume that the exhaust region extended to the length of the diameter of Mars. In the current sheet normal direction, we obtain the  The purple tube represents a magnetic ux tube, and the green line means the magnetic eld line   rolled up by the shear ow (the blue arrow). As a result, the handedness of the two sides of the twisted ux tube is opposite. The ux rope is dragged by the solar wind to the magnetotail, which corresponds to FR1 and FR3 on both sides of the current sheet. Note that we cannot tell if FR1 and FR3 were connected to a same magnetic eld line or combined in a same ux tube, as shown in Fig. 1a with the dashed segment. calculation. The X component of background magnetic eld exhibited a tangential shape, consistent with the singal of a current sheet crossing. The cyan dashed line in Fig. 2b shows the modeled crustal eld, which is much smaller than the measured magnetic eld. E ux in Fig. 2e Fig. 1a.

Figure 3
Magnetic and plasma measurements in the reconnection region and ux ropes. (a) Magnetic eld data after removing the data within ux rope regions, as a representation of background magnetic eld contributed by current sheet, (b) proton bulk velocities, (c) Alfvén velocity , where is the total ion mass density, (d) tailward uxes (-nVX) of and , (e) original magnetic eld data in the current sheet LMN coordinates. The current sheet-based minimum variance analysis is performed by the interval between two dotted lines (10:13:00-10:16:40 UT). (f) The SWEA measurements of differential energy uxes (in units of eV/cm2/str/s/eV) of electrons with energies of 199-568 eV. (g-i) Electron distributions extracted at the time denoted by the arrows below Fig. 3f. The asymmetric Hall eld, the Alfvénic out ow and the at-top electron distribution within FR2 (Fig. 3h) suggest that FR2 was generated by local reconnection.  Magnetic measurements of ux ropes generated by boundary instabilities. (a and c) Time pro les and (b and d) hodograms of the vector magnetic eld components transformed into the ux rope-based minimum variance coordinate frame after removing the current sheet signal. Black dots in Fig. 4b and 4d are the start points of the hodograms; h represents the sign of the helicity of the ux ropes. Axial orientations (relative to the MSO coordinate) of the ux ropes FR1 and FR2 were [0.68, -0.43, -0.59] and