Energetic Neutral Atoms Detected in the 2022 February 15 Solar Energetic Particle Event

Energetic neutral atoms (ENAs) are expected to be produced near the Sun during large solar energetic particle (SEP) events. However, their detection by SEP instruments near 1 au has been limited. The clearest reported measurement has been from the Solar Terrestrial Relations Observatory (STEREO) during the 2006 December 5 SEP event. Additional evidence of ENAs has been found through reanalysis of observations by the Solar Anomalous and Magnetospheric Particle Explorer obtained near the equator in low Earth orbit and associated with several large X-ray flares and fast coronal mass ejections (CMEs). Here we describe another detection of ENAs from the STEREO Low Energy Telescope associated with the large 2022 February 15 SEP event. Given the timing and spectrum of the ENAs and the location of the source region (behind the east limb from STEREO’s viewpoint), these ENAs are most likely a result of acceleration by a CME-driven shock when the CME was at approximately 2–3 R S. The possibility of a postflare loop origin is considered unlikely.


Introduction
As ions (solar energetic particles, SEPs) are accelerated near the Sun through either reconnection processes associated with solar flares or shocks driven by coronal mass ejections (CMEs), they can interact with the surrounding medium and, via charge exchange, create energetic neutral atoms (ENAs).Unlike their SEP counterparts, the ENAs are not affected by the interplanetary magnetic field, and thus, their energy spectrum remains unchanged over their ballistic path from escape near the Sun to observers in the heliosphere.Therefore, they can be a direct probe of the acceleration processes and conditions at the acceleration site.
The possibility of measuring solar ENAs in interplanetary space was studied by Hsieh et al. (1992), and it was a substantial discovery when they were detected by the Low Energy Telescope (LET; Mewaldt et al. 2008) on the Solar Terrestrial Relations Observatory (STEREO; Kaiser et al. 2008) as reported by Mewaldt et al. (2009).Recent reanalysis of data from the Low-energy Ion Composition Analyzer (Mason et al. 1993) on the Solar Anomalous and Magnetospheric Particle Explorer (SAMPEX; Baker et al. 1993) revealed additional detections of ENAs through signatures in the near-equatorial trapped proton population in the Earth's magnetic field following 19 large SEP events (Mason et al. 2021).Unfortunately, aside from these papers, no observations of solar ENAs have been reported.
This may partly be due to the lack of instruments optimized to detect solar ENAs.LET is fundamentally an SEP instrument, and ENAs are stripped as they pass through the instrument's front window to become protons.The challenge is to differentiate these "new" protons from the SEP protons that are generated in copious amounts by the same solar acceleration processes.In the case of the SAMPEX detections, SEP protons were excluded by the Earth's magnetic field, which did not affect the ENAs.The event reported by Mewaldt et al. (2009) was the result of fortuitous relative positioning of the STEREO spacecraft and the solar source region on the Sun (and, relatedly, the propagation direction of the likely CME).Over the years, the STEREO data have been periodically examined for a similar serendipitous event, without success.However, these searches were not particularly systematic.We have begun to remedy this and have found a similar, albeit smaller, ENA event.

Observations
In an effort to prioritize our search for signatures of ENAs in the nearly 18 yr long STEREO data set, we focused on time periods corresponding to the launch of fast CMEs originating from the solar east (as viewed by STEREO), working backwards in time from late 2023.The LET data are divided into 16 sectors identified through the combination of triggered segments for the L1 and L2 detectors (see Appendix and description in Mewaldt et al. 2008).On the STEREO-A spacecraft, nominally sectors 3 and 4 include directions of a standard Parker Spiral for solar wind speeds of 400 km s −1 ; however, in late 2015, in order to maintain communications with the spacecraft, STEREO-A was rotated 180°about the spacecraft-Sun line and left in this position until 2023 August when it was rotated back to its nominal orientation.This did not change which sector included the radial direction to the Sun, sector 7, but did change the viewing orientation of the rest of the sectors, leaving the nominal Parker spiral direction outside the telescope's field of view.Our search consisted of examining the 1.8-3.6MeV proton-sectored intensities from Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence.Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
LET on STEREO-A5 and searching for an increase in the intensities from sector 7 with no corresponding increase in the other sectors.Since the STEREO spacecraft are stabilized, each sector's viewing direction is fixed throughout the event, in contrast to spinning spacecraft such as Wind and ACE where each "sector" is a time-dependent portion of a spacecraft rotation.The 1.8-3.6MeV energy range was selected as it is the lowest energy interval available (since late 2010 November) in the LET sectored data, and we expected any signature to be strongest at low energies due to a falling energy spectrum (e.g., see Figure 5 of Mewaldt et al. 2009).
The choice to focus on eastern CMEs was to maximize the time difference between radially traveling particles (e.g., ENAs) and SEPs that are constrained to travel along interplanetary magnetic field (IMF) lines.SEP events with eastern source regions typically have more gradual and later onsets than their western counterparts (see, e.g., Cane et al. 1988).As ENAs will be stripped as they pass through the instrument window, their signature in LET will be that of a charged ion, e.g., a proton in the case of neutral hydrogen atoms and instrumentally indistinguishable from SEP protons except for their arrival direction and timing.
A candidate ENA signal was identified in this manner for the fast CME on 2022 February 15. Figure 1 shows the intensities measured in sectors 0, 4, and 7 as a function of time.The onset of the SEP event is evident in the latter half of February 16 (day 47), but an increase is seen in only sector 7 (covering angles −17.5°to 7.5°relative to the spacecraft-Sun line) at the end of February 15 and shown on an expanded timescale in the right panel.The CME speed was ∼2200 km s −1 (Mierla et al. 2022), and the SEP event was large enough to be observed by multiple spacecraft, including Maven in orbit around Mars (Khoo et al. 2024).Unfortunately, from the perspective of STEREO-A, Solar Orbiter (SolO), and near-Earth spacecraft the source region was behind the east limb of the Sun, so it is unknown to what X-ray class the flare belonged (Mierla et al. 2022).
To investigate the candidate event further, we examined the pulse height analysis (PHA) data.During quiet times, LET telemeters all measured PHA events (each one corresponding to an individual ion detection), which have more detailed information regarding which detector segments were triggered and the associated energies deposited.Fourteen PHA events occurred in the candidate ENA time period of days 46.95-47.05(2022 February 15 22:48 UT-2022 February 16 01:12 UT) that were identified on board as protons arriving in sector 7. The energy deposits in the L1 and L2 detectors were compared to those of all PHA events in a larger time period, which includes the start of the SEP event, and were found to lie on the proton track as expected (see Appendix, Figure A1).
The distribution in the ecliptic arrival directions (relative to the spacecraft-Sun line) for all protons detected during the ENA time period is compared to the distribution for a prior time period 5 times longer (to increase statistical accuracy) in Figure 2 (see also Appendix, Figure A2).The ENA time period is strongly peaked near the solar radial direction (0°) as compared to a flatter distribution for the earlier time period, indicating most of the particles during this ENA time period are traveling radially outward from the Sun.We also examined the direction of the magnetic field as measured by STEREO/MAG (Acuña et al. 2008) during the ENA time period.The range of angles was 21.8°-78°and clearly not radial, showing that the peak near 0°is mostly ENAs unaffected by the IMF during this time period.
Next we examined the arrival times of the candidate ENAs (hereafter referred to simply as ENAs) as a function of their inverse velocity.Clear velocity dispersion is evident with a determined path length very close to 1 au, in contrast to the dispersion of the onset of the SEP event later on 2022 February 16 with a substantially longer path length (see Appendix, Figure A3).Additionally, the solar release time of the ENAs determined from the velocity dispersion analysis was 21:55 UT, close to the 21:50 UT time of eruption as reported by Mierla et al. (2022).The solar release time profile can be calculated by using the velocities of the individual particles and the locations of the STEREO-A spacecraft at the observation times.Light-travel time has been added, so comparisons can be made to STEREO-A remote-sensing observations (Figure 3).The profile is not well defined due to the low counting statistics and cannot be compared to an X-ray or extreme-ultraviolet (EUV) time profile of the flare emission as it was not observed by any spacecraft.However, the CME was well observed by STEREO-A/SECCHI (Howard et al. 2008), and the heighttime profile suggests the CME was at ∼2-3 R S when the ENAs were released.
Finally, we have calculated an average intensity spectrum from the 14 detected ENAs.As the incoming particles are not isotropic, we use the effective area of the combined L1A0 and L1A1 detectors, assuming a beam perpendicular to L1A0 (L1A1 is tilted 45°relative to L1A0; see Figure A4) instead of the standard geometry factor for sector 7.For comparison, we have made the same calculation for the ENA time period on 2006 December 5 identified by Mewaldt et al. (2009), where we have similarly selected particles arriving in Sun-pointed sectors from the STEREO-A and STEREO-B spacecraft.Both spectra are plotted in Figure 4; only the lowest two energy points are well defined for our spectrum, but it is clear that the Mewaldt et al. (2009) ENA event was larger by approximately a factor of 2.5.Given the lack of higher-energy measurements, it is difficult to determine how the spectral slopes compare, but they appear roughly similar and close to the −2.46 power law reported by Mewaldt et al. (2009).

Discussion/Summary
Based on the following key signatures, we have identified the 14 particles as ENAs: 1.All the particles arrived from a narrow direction consistent with radial from the Sun and not the measured magnetic field directions during the time period.2. The particles exhibit velocity dispersion with a path length consistent with the radial distance of STEREO-A and a release time near that of the solar eruption.It is distinctly different from the velocity dispersion evident in the onset of the ensuing SEP event.3. The energy deposits in the LET detectors are consistent with protons as would be expected from hydrogen ENAs being stripped after passing through the LET window.This is the first reported detection of ENAs from the STEREO spacecraft since the seminal observation and paper by Mewaldt et al. (2009).Although it is smaller than the Mewaldt et al. (2009) event, it shows similar characteristics, including narrow radial arrival directions well before the onset of an SEP event; velocity dispersion consistent with 1 au path length and release times near that of the solar flare/eruption; and a powerlaw spectral index close to the previously reported value of −2.46.It should be noted that these features are also inconsistent with those expected from neutron-decay protons, which would travel along the IMF direction and have a substantially different spectrum (see discussion in Mewaldt et al. 2009).
As discussed and modeled by Li et al. (2023), there are two primary ways in which ENAs can be produced in association with solar eruptions.The first is from CME-driven shocks, where ENAs can be converted from energetic protons accelerated both upstream and downstream of the propagating shock.ENAs can also be created from energetic protons trapped in large postflare loops.The height at which the ENAs are created is important in both scenarios because, if the surrounding material is too dense, the ENAs will be reionized to become SEP protons.
In their CME-driven shock simulations, Li et al. used a shock speed of 1500 km s −1 and calculated the time profiles of 0.1-20 MeV ENAs for observers at 1 au, with three different locations in longitude relative to the travel direction of the shock front.The integrated fluence spectrum was also calculated.While there are some differences, all three cases show ∼1-2 MeV ENA onsets occurring ∼10 minutes after the initiation of the shock and spectra above 1 MeV that are approximately power laws with indices of ∼−2.5.
Li et al. found a strong dependence on the final loop height for the ENA time profiles and fluence spectra in their postflare loop simulations.Although the onset times for ∼2 MeV ENAs were not that different than for the CME shock scenario, the fluence spectra were quite different.For smaller loops (e.g., final heights of 0.22 R S ), no ENAs below 1 MeV reached 1 au, and for larger loops (e.g., final heights of 0.6 R S ), below 2 MeV, the spectrum decreased with decreasing energy, while above this energy, the spectrum was consistent with a power law of index ∼−2.Mewaldt et al.ʼs (2009) calculation led them to conclude that the ENAs they observed were generated at heights above 1.6 R S and thus must have been associated with a CME-driven shock (although they did not consider the idea of large postflare loops as a potential source).Our ENA detection appears to be of similar origin, having a solar release time corresponding to when the leading edge of the CME was ∼2-3 R S (see Figure 3).Our spectrum is also more consistent with that resulting from Li et al.ʼs simulations for a shock source than for a postflare loop source although we do not have measurements below 1.8 MeV, where the spectral shape is most distinctive between the shock and loop scenarios.Another relevant argument against the postflare loop origin is the observation, reported by Mierla et al. (2022), that postflare loops were first observed by SolO's Extreme Ultraviolet Imager (EUI; Rochus et al. 2020) at 23:50 UT on 2022 February 15, well past the peak of our calculated ENA solar release time profile (Figure 3).
It should be mentioned that this solar eruption was unusual in the amount of prominence material ejected and in the fact that SolO/EUI was able to track it in EUV wavelengths to distances of 6 R S (Mierla et al. 2022).The SOHO LASCO CDAW CME catalog estimated the CME mass as >10 16 gm; since the solar activity occurred far from the limb, that mass must be considered "uncertain," but it places the CME in the top 2% of over 20,000 LASCO CMEs with mass determinations.The solar event gained significant press6 and attention; whether the presence of the cooler prominence material had an impact on the generation of ENAs (possibly increasing their likelihood of being detected at 1 au) is not known.The heighttime profile of the leading edge of the prominence is shown in Figure 3, and it was below ∼2 R S when the ENAs left the Sun.
Given the paucity of ENA detections since the original report of Mewaldt et al. (2009), it is tempting to consider their production or their probability of escape from the Sun to be extremely low.However, the simulations of Li et al. suggest significant numbers can be produced from CME-driven shocks and possibly in large postflare loops.The recent reanalysis of  quasi-trapped, equatorial protons observed by SAMPEX has shown the presence of ENAs in several large SEP events (Mason et al. 2021).The difficulty is that these detections are being made by instruments not designed to measure ENAs specifically; thus, unique conditions must exist to separate the ENA signal from that of SEPs.Unfortunately, this is also the case for Parker Solar Probe (and other spacecraft, such as SolO, with similar SEP instrumentation) where the SEP suite, the Integrated Science Investigation of the Sun (McComas et al. 2016), is capable of similar observations to those presented here and in Mewaldt et al. (2009) but would require unusual positioning relative to the solar source We plan to continue our search for ENA signatures in the STEREO data set, but to really understand the ENAs expected from large solar events, a dedicated ENA mission is required.cooperation between ESA and NASA.Measurements for Figure 3 and imagery for Figure 5 were made using JHelioviewer (Müller et al. 2017).

Appendix
Here we provide more details regarding the measurements of the ENAs detected by the STEREO-A spacecraft.
Figure A1 plots the energy deposited in the L1 detector (all segments) as a function of the energy deposited in the L2 detector (all segments) for all PHA events identified on board as protons for the time period 2022 February 10 00:00 UT to 2022 February 17 12:00 UT.The energies have not been corrected for the ions' angles of incidence.The distribution shows the pattern for protons stopping in the L2 detector (upper curve) and the "fold-back" portion (lower curve) due to protons passing through the L2 detector (possibly stopping in deeper detectors of the instrument).Also shown in blue are the 14 sector 7 ENAs identified during the time period 2022 February   The blue points are the sector 7 particles.They are clearly clustered around the direction radially toward the Sun, with a mean of −0.2°and standard deviation of 5°, and not spread throughout sector 7 angles (which includes angles of −17.5°to 7.5°).Over the period of 2022 February 15 22:00 UT to 2022 February 16 02:00 UT, the IMF direction varied in angle between 21.8°and 78°(as indicated by the vertical arrow to the left of the y-axis in Figure A2) and mostly between 40°and 70°, indicating the IMF was not radial during the period when the ENAs were observed.
Figure A3 gives the velocity dispersion of the particles for the same time period as Figure A2 in the left panel and a zoom into the few hours around the ENA time period in the right panel.The blue points are the detected ENAs, and they show a clear "edge" indicating velocity dispersion in the first arriving particles.This can be nicely fit with a line corresponding to a path length of 0.967 au (the distance of STEREO-A from the Sun at this time) and a release time of 21:55 UT.The SEP event starting a bit before day 47.5 does not have a nicely defined edge, as expected from a far eastern SEP event (as this one is), but it is clear that the velocity dispersion corresponds to a much longer path length than that for the ENAs.A line corresponding to a path length of 6 au and a release time of day 47.0 are not a terrible representation of the SEP dispersion, which is likely affected by the timing of the accelerating shock and particle scattering in the transport to 1 au.
As there are not many ENAs in our event, for completeness we provide their measured parameters in Table A1, including observed time, total energy, and which detector segments were triggered.A schematic showing the identification of the LET detectors is given in Figure A4 (adapted from Mewaldt et al. 2008).Note.
a This event is possibly not an ENA but is included for completeness.

Figure 1 .
Figure 1.(Left) Intensities of 1.8-3.6MeV protons as a function of time for three sectors.Ranges of angles (relative to the spacecraft-Sun line) for each sector are given in the legend; note that sector 7 includes directions radially from the Sun.(Right) Zoom-in on the sector 7 increase around 2022 February 16 00:00 UT (day 47) with statistical uncertainties shown.

Figure 2 .
Figure2.Angular distribution of the ENAs and protons for the ENA time period (blue line) and an earlier time period (red line).To increase statistical accuracy, the selected earlier time is 5 times longer; thus, for quantitative comparison, the red histogram is scaled by a factor of 5.The range of the IMF angles for the ENA time period, as indicated by the black arrow, was 21.8°-78°( and primarily varying between 40°and 70°).

Figure 3 .
Figure 3. Height as a function of time for the leading edges of the CME (black points) and the prominence (red points).For comparison, the release time profile of the ENAs is overplotted.The time for light to travel from the Sun to STEREO-A is added to enable direct comparison to remote-sensing data.

Figure 4 .
Figure 4. Average intensity spectrum for the ENAs averaged over the time period 2022 February 15 22:48 UT to 2022 February 16 01:12 UT (green solid circles) and for the Mewaldt et al. (2009) event (blue open squares) averaged over the period 2006 December 5 11:30 UT to 13:50 UT.A power law with index of −2.46 is shown for comparison to that reported by Mewaldt et al. (2009).

Figure 5 .
Figure 5. (Left) Images from STEREO-A/EUVI, COR1, and COR2 at different times, showing the leading edge of the erupting CME.To enhance the visibility of the CME, a background frame a few minutes earlier was subtracted from each image.(Right) Positions of various spacecraft at the time of the eruption (Gieseler et al. 2023).The black arrow indicates the approximate direction of the eruption.

Figure A1 .
Figure A1.Energy deposited in detector L1 vs. energy deposited in detector L2 for all protons (red) and ENAs (blue).The upper portion of the trace is due to protons stopping in the L2 detector; the lower portion is from protons passing through both L1 and L2.
15 21:48 UT to 2022 February 16 01:12 UT.It is clear that they are part of the standard proton distribution as would be expected from ENAs entering the instrument (which at these energies would be stripped of their electrons to become protons as they passed through the outer window of LET).This shows the identified ENAs have the expected proton signature after passing through the telescope's entrance foil.Shown in FigureA2are the incident particle angles relative to the spacecraft-Sun line as calculated from the L1 and L2 detector segments for protons measured over the 2 days of 2022 February 15 and 16 (as compared to Figure2ofMewaldt et al. 2009).It should be noted that the STEREO-A spacecraft was rolled by 180°along the spacecraft-Sun line during this time for communication requirements.Thus, the A-side of LET spans angles from approximately 20°to −110°, and the B-side views angles from approximately −160°to −290°(where positive directions are west of the Sun and negative directions are east).The increase in density of points just before day 47.5 (February 16 12:00 UT) is due to the start of the SEP event.

Figure A2 .
Figure A2.Red circles: arrival angle from the spacecraft-Sun line for protons as a function of time; in blue filled circles are the sector 7 particles during the time period of 46.95-47.05.The variability of the magnetic field direction during the ENA time period is indicated by the vertical arrow to the left of the y-axis and extends to 78°.

Figure A4 .
Figure A4.Positions and labels of the detectors for STEREO/LET.Adapted from Mewaldt et al. (2008).

Figure A3 .
Figure A3.(Left) Inverse velocity vs. time for all protons (red points) and the ENAs (blue).Dashed lines show expected velocity dispersion for path lengths of 0.967 au (the distance to STEREO-A) and 6 au (the latter also has a later release time to better correspond to the apparent SEP onset).(Right) The same plot for 8 hr of 2022 February 15-16, illustrating the good agreement between the expected velocity dispersion for a release time of 21:55 UT and path length of 0.967 au and the leading edge of the ENAs.Energy equivalents are shown on the right y-axes.