Energetic Electron Observations by Parker Solar Probe/IS⊙IS during the First Widespread SEP Event of Solar Cycle 25 on 2020 November 29

At the end of 2020 November, two coronal mass ejections (CMEs) erupted from the Sun and propagated through the interplanetary medium in the direction of Parker Solar Probe while the spacecraft was located at ∼0.81 au. The passage of these interplanetary CMEs (ICMEs) starting on November 29 (DOY 334) produced the largest enhancement of energetic ions and electrons observed by the Integrated Science Investigation of the Sun (IS⊙IS) energetic particle instrument suite on board Parker Solar Probe during the mission’s first eight orbits. This was also the first spatially widespread solar energetic particle event observed in solar cycle 25. We investigate several key characteristics of the energetic electron event including the time profile and anisotropy distribution of near-relativistic electrons as measured by IS⊙IS’s low-energy Energetic Particle Instrument (EPI-Lo) and compare these observations with contextual data from the Parker Solar Probe Fields Experiment magnetometer. These are the first electron anisotropy measurements from IS⊙IS/EPI-Lo, demonstrating that the instrument can successfully produce these measurements. We find that the electron count rate peaks at the time of the shock driven by the faster of the two ICMEs, implying that the shock parameters of this ICME are conducive to the acceleration of electrons. Additionally, the angular distribution of the electrons during the passage of the magnetic clouds associated with the ICMEs shows significant anisotropy, with electrons moving primarily parallel and antiparallel to the local magnetic field as well as bidirectionally, providing an indication of the ICME’s magnetic topology and connectivity to the Sun or magnetic structures in the inner heliosphere.


Introduction
Coronal mass ejections (CMEs) are a class of highly energetic eruptive solar events observed by coronographs near the Sun; fast CMEs drive shocks that can accelerate solar energetic particles (SEPs) to very high energies (Tsurutani & Lin 1985;Desai & Giacalone 2016;Bruno et al. 2018). Interplanetary coronal mass ejections (ICMEs; e.g., Cane & Richardson 2003), the manifestations in the solar wind of CMEs, have a variety of characteristic plasma and magnetic signatures that aid in the identification of ICMEs in in situ observations (Zurbuchen & Richardson 2006). CME-driven "gradual" SEP events (Reames 2013) do not tend to be as electron-rich as smaller "impulsive" events (Wang et al. 2012). The acceleration of electrons at shocks is not well understood; however, there are examples showing efficient electron acceleration (e.g., Guo & Giacalone 2010). A number of studies report significant enhancements of energetic electrons in the regions of shocks including the heliospheric termination shock (e.g., Decker et al. 2005), the Earth's bow shock (e.g., Gosling et al. 1989), Saturn's bow shock (e.g., Masters et al. 2013;Richardson 2013), and shocks driven by ICMEs (e.g., Tsurutani & Lin 1985;Lario et al. 2003;Simnett et al. 2005;Ho et al. 2008;Dresing et al. 2016b).
Shock spikes, where the local particle intensity peaks at the time of the shock crossing, are interpreted as enhancements of SEPs accelerated locally at the shock (e.g., Fraschetti & Giacalone 2015). Tsurutani & Lin (1985), Lario et al. (2003), and Dresing et al. (2016b) performed statistical studies of the electron measurements associated with ICME-driven shocks. Those studies found that shock spike electron enhancements are rare and may only occur with specific shock parameters and plasma conditions.
Recent studies have demonstrated that certain conditions, including shock-normal orientation, upstream magnetic turbulence, and a pre-event enhanced background can lead to efficient energization of electrons at collisionless shocks driven by CMEs (Burgess 2006;Guo & Giacalone 2015). Studies have found that quasi-perpendicular shocks are significantly more efficient accelerators of electrons compared with quasiparallel shocks (Holman & Pesses 1983;Wu 1984;Krauss-Varban et al. 1989;Jokipii & Giacalone 2007;Guo & Giacalone 2010. Shock-drift acceleration, in which electrons drift along the E = −V × B/c electric field of the shock, is believed to result in significantly higher electron acceleration efficiency than in quasi-parallel shocks. This is also consistent with diffusive shock acceleration theory in which quasi-perpendicular shocks have extremely high acceleration rates (e.g., Jokipii 1987;Zank et al. 2006;Schwadron et al. 2015). Numerical simulations in Guo & Giacalone (2015) indicate that large-scale magnetic fluctuations upstream of a propagating shock may significantly increase electron acceleration efficiency at any shock-normal orientation.
The angular distribution of charged particles in the vicinity of an ICME can serve as a probe to investigate the acceleration processes that energized those particles, the connectivity of the ICME to magnetic structures in the heliosphere or the Sun, and the magnetic configuration of the ICME. In particular whether the particles are beamed parallel or antiparallel, are counterstreaming, or have some other pitch angle distribution (PAD) relative to the magnetic field direction can provide valuable information about the acceleration mechanisms and structure of the ICME. For example, particles may be injected along one or both of the legs of the ICME (Kahler & Reames 1991;Richardson et al. 1991;Richardson & Cane 1996;Larson et al. 1997;Dresing et al. 2016a;Gómez-Herrero et al. 2017) resulting in a direct connection of those particles to an observer within the ICME. Additionally, magnetic mirrors, either from connections of the ICME loop legs to magnetic structures in the interplanetary magnetic field or corona, can cause energetic charged particles to reflect back and forth inside an ICME flux tube (Marsden et al. 1987;Bieber et al. 2002;Malandraki et al. 2002;Roelof 2008). Examination of electron PADs throughout the structure of an ICME provides valuable insight into the transport of energetic electrons within the ICME.
Near the end of 2020 November, two successive ICMEs passed over Parker Solar Probe (Fox et al. 2016) while the spacecraft was near aphelion at 0.81 au. Parker Solar Probe was uniquely positioned close to the longitude of the solar eruptions associated with these events and directly encountered the related ICMEs. The second of these ICMEs produced the largest SEP event observed by the Integrated Science Investigation of the Sun (ISeIS) energetic particle instrument suite (McComas et al. 2016) during Parker Solar Probe's first eight orbits (Cohen et al. 2021a). This is the first spatially widespread SEP event of solar cycle 25 with SEPs detected throughout the inner heliosphere including at the Solar Terrestrial Relations Observatory A (STEREO-A), Solar Orbiter, the Earth, and Parker Solar Probe (A. Kollhoff et al. 2021). During these events, ISeIS observed significant enhancements in energetic ions and electrons. Prior to these events, ISeIS had only observed a handful of small energetic electron events (Mitchell et al. 2020;Wiedenbeck et al. 2020;Cohen et al. 2021b) with low statistics such that the anisotropy of the events could not be meaningfully analyzed. These 2020 November ICMEs produced high electron intensities, allowing for the study of the energetic electron distributions throughout the event.
In this study, we examine the electron intensity-time profiles, anisotropies, and PADs of the 2020 November ICMEs observed by ISeIS with respect to the local magnetic field as measured by the Parker Solar Probe Fields Experiment (FIELDS) magnetometers . Special attention is paid to electron measurements during the ICME intervals when EPI-Lo had good pitch angle coverage and observed significant anisotropy.

Instrumentation
The ISeIS instrument suite includes two energetic particle instruments designed to measure high-energy (EPI-Hi) and lower-energy (EPI-Lo) SEPs (McComas et al. 2016). This work focuses primarily on measurements from the EPI-Lo instrument (Hill et al. 2017). EPI-Lo consists of eight instrumental segments (wedges) organized in an octagonal configuration, which provides a ∼2π sr field of view (see Figure 1; McComas et al. 2016;Hill et al. 2017). EPI-Lo measures electrons from ∼17 keV-1 MeV using 700 μm thick isosceles trapezoidal solid-state detectors (SSDs) located at the rear of each of EPI-Lo's eight wedges. Each wedge includes three SSDs referred to as the particle SSD, ion SSD, and anticoincidence SSD. The particle SSD (labeled with an "E" in Figure 2) is designed specifically to measure electrons, with a ∼3.2 μm thick aluminum flashing to suppress low-energy ions. The anticoincidence SSD (highlighted in green in Figure 2) is located behind the particle SSD such that particles that penetrate the particle SSD can be identified and rejected. The combination of the particle and anticoincidence SSDs can also be used to make "dE/dx versus total energy" measurements (e.g., Stone et al. 1998bStone et al. , 1998c. The ion SSD (labeled with an "I" in Figure 2) is mounted alongside the particle SSD and does not include the aluminum flashing, allowing it to measure lower-energy ions than the particle SSD. The ion SSD is used primarily for triple-coincidence ion measurements incorporating both time-of-flight through the instrument and energy deposit in the SSD. It can also be used to measure electrons and give a measure of the contamination of ions to the electron measurements from the flashed particle SSD. Each EPI-Lo wedge has 10 individual apertures (for a total of 80 look directions across the entire instrument), which provide comprehensive anisotropy information for both ion and electron measurements. Since each wedge only has a single SSD assembly, the electron anisotropy can only be determined on a wedge-by-wedge basis with no information on which aperture each electron entered. The eight fields of view available for electrons provide unprecedented angular coverage for a three-axis stabilized configuration and contribute detailed information on electron anisotropy and PADs.
The higher-energy detector EPI-Hi measures ions with energies ∼1-20 MeV/nuc and electrons with energies ∼0.5-2 MeV using a "dE/dx versus total energy" technique (McComas et al. 2016;Wiedenbeck et al. 2017). Each EPI-Hi telescope is composed of stacks of SSDs providing measurements of the energy and species of incident particles.
During the time period studied in this work, the Parker Solar Probe spacecraft was oriented with the spacecraft Z-axis (which points toward the Sun when the spacecraft is in nominal orientation 12 ) 45°off-pointing (alternating between westward and eastward of the Sun) within the ecliptic plane for communications, thermal, and power management purposes. As a result, the EPI-Lo wedges sampled particle populations primarily in the transverse (T) and normal (N) directions in the radial-tangential-normal (RTN) coordinate system (Fränz & Harper 2002). Figure 3 illustrates the look directions of the nominal geometric center of each of the EPI-Lo wedges as the larger labeled circles in the two orientations of the spacecraft during this time period in the RTN coordinate system. The smaller circles show the look direction of each of the EPI-Lo apertures color-coded based on wedge.
While the aluminum flashing of the particle SSD is effective at rejecting low-energy ions, the electron channels still suffer from contamination from higher-energy ions. Preliminary modeling with a newly developed Geant4 (Agostinelli et al. 2003) model of EPI-Lo suggests that only ions in the relatively narrow energy range of ∼300-600 keV contribute to <1 MeV measurements in the electron channels. Above 1 MeV, it is not possible to conclusively distinguish electrons from ions using the EPI-Lo electron channels alone. Thus, for energies 1 MeV, the EPI-Hi electron measurements are more reliable.
As an example of this effect, the EPI-Lo energetic electron spectrogram for the 2020 November SEP event shows two distinct populations, separated at ∼1 MeV as shown in Figure 4. The lower-energy population is believed to be electron dominated, while the higher-energy population is likely a mix of ions and electrons; we confirmed that highenergy electrons and ions were observed by EPI-Hi during this event (Cohen et al. 2021a). For this reason, we focus on EPI-Lo measurements of <1 MeV electrons and use comparisons in the relevant proton energies to confirm that these measurements are dominated by electrons. This is the first SEP electron event observed by ISeIS that is sufficiently large to enable the instrumental electron response to be calibrated using comparisons between the data collected and instrument simulations. This calibration work is ongoing and will be the subject of a future publication.
This work utilizes both the high-time-cadence and highenergy-resolution EPI-Lo electron data from all eight wedges. The high-time-cadence data uses wider energy bins resulting in lower energy resolution whereas the high-energy-resolution data has smaller energy bins but longer integration times. Throughout this work, the high-time-cadence data used has energies in the range ∼130-870 keV in wedges 3 and 7, ∼140-730 keV in wedge 5, and ∼180-790 keV in all other wedges. Similarly, the high-energy-resolution data used throughout this study are in the energy range ∼110-870 keV in wedges 3 and 7, ∼80-730 keV in wedge 5, and ∼100-790 keV in all other wedges. The slight differences in energy range between the wedges are due to the different detector gains and thresholds used in each of the wedges. In the course of this study, we found that electron energies were not being binned optimally in the flight software between the different wedges. The energy bins were re-calibrated after this discovery and these are the closest possible matches in the energy range of interest.

Observations
Between 2020 November 29 (DOY 334) and December 1 (DOY 336), two ICMEs passed over Parker Solar Probe, with the second, faster ICME observed by Parker Solar Probe instruments shortly after the first. The arrival times of the shocks 13 driven by the two ICMEs at Parker Solar Probe are indicated by the abrupt increases in the magnetic field intensity in panel (c) of Figure 5 marked by the black and blue vertical dashed lines. Both ISeIS energetic particle instruments observed significant ion and electron enhancements associated with the 2020 November ICMEs (Cohen et al. 2021a). Panels (a) and (b) of Figure 5 show the time profiles of the energetic electrons observed by EPI-Hi and EPI-Lo, respectively. The EPI-Hi data is the sum of the shortest energy range in all five EPI-Hi apertures (range 1 in the high energy telescope (HET) and range 3 in the low energy telescope (LET)), corresponding to ∼0.5-2 MeV electrons, with a 1 hr integration time (Wiedenbeck et al. 2017). The apparent depression in electron count rates observed by EPI-Hi in the top panel of Figure 5 during the passage of the shock and sheath regions of the Figure 2. CAD model image of the EPI-Lo SSDs. The particle SSD is labeled with an "E," the ion SSD is labeled with an "I," and the anticoincidence SSD is located behind the particle SSD. The outer region of the anticoincidence SSD is highlighted in green.
second ICME is an artifact due to the EPI-Hi instrument going into higher dynamic threshold modes due to high particle intensities. The apparently flat profile with time on DOY 334-335 is also an instrumental artifact. For more information on this, please refer to Cohen et al. (2021a). The EPI-Lo data is the sum of ∼57-870 keV electrons from Wedges 3 and 7, which point sunward and anti-sunward, respectively, during periods of nominal orientation with a 1 minute integration time. This energy range is believed to have minimal ion contamination. Figure 6 shows the onset of the event summed over energies ∼130-870 keV in wedges 2, 3, and 4, which are the wedges with pitch angle coverage closest to 0°at the onset of the event. The onset of the EPI-Lo electron event was determined to be 13:53 UTC±1 minute on November 29 (DOY 334) as the first instance of three consecutive time bins with a >3σ increase above the background from the previous 1.5 days in which no electron event was observed (marked by the red dashed vertical line in Figure 6). As discussed by A. Kollhoff et al. (2021), this particle event was associated with an M4.4 soft X-ray flare in NOAA active region (AR) 12790, located at E98S23, just behind the east limb as seen from Earth. The flare started at 12:34 UTC and peaked at 13:11 UTC. Type II and III radio emissions were also observed, together with a fast (∼1500 km s −1 ) CME preceded by a faster (∼1800 km s −1 ) shock. As discussed further below, this is the likely origin of the second ICME (hereafter referred to as ICME2). It is likely that the first several hours of the electron enhancement resulted from impulsive electrons accelerated at the solar event that produced ICME2. In this scenario, one would expect to observe energy dispersion on the leading edge of the event (Krucker et al. 1999). We compared time-intensity data from each of the EPI-Lo electron energy bins and identified no clear energy dispersion signature. Indeed, the onset of the electron event appears approximately concurrent across all energies perhaps implying a change in connectivity resulting in an abrupt "turn on" of the energetic electrons. Energetic particles associated with the eruption of ICME2 were observed by spacecraft throughout the heliosphere (A. Kollhoff et al. 2021) including Solar Orbiter (Mueller et al. 2013), ACE (Stone et al. 1998a and Solar and Heliospheric Observatory (Domingo et al. 1995). As CME1 moved significantly slower than CME2 (∼550 km s −1 ), it is unlikely to have locally accelerated electrons that contributed to this event. The passage of the shock driven by ICME1 (black dashed vertical line in Figure 5) does not show a further concurrent electron enhancement in either EPI-Hi or EPI-Lo. CME modeling efforts obtained from the Space Weather Database of Notifications, Knowledge, Information (DONKI 14 ) compiled by the NASA Community Coordinated Modeling Center, suggest that the first ICME to reach Parker Solar Probe erupted at 21:24 UTC on 2020 November 26 from NOAA AR 12,787 at a longitude of −100°in Heliocentric Earth Equatorial (HEEQ; Thompson 2006) coordinates with a speed of ∼550 km s −1 . From Wang-Sheeley-Arge (WSA)-ENLIL-Cone modeling (e.g., Odstrcil et al. 2020), ICME1 was expected to reach Parker Solar Probe on 2020 November 29 (DOY 334) at 14:12 UTC. The possible ICME1-driven shock was observed to reach Parker Solar Probe later that day, around 23:10 UTC. On 2020 November 29 (DOY 334), the second ICME that passed over Parker Solar Probe during this time period erupted at 13:24 UTC from AR 12790 with a longitude −75°HEEQ traveling ∼1336 km s −1 according to DONKI estimates. WSA-ENLIL-Cone modeling predicts that this much faster ICME would reach Parker Solar Probe at 19:37 UTC on 2020 November 30 (DOY 335). This predicted arrival time is ∼1 hr later than the observed arrival time of the ICME2-driven shock. During this time, Parker Solar Probe was located around −95°, well in the path of the CMEs.
The structure of the two ICMEs can be inferred from the magnetic field direction and magnitude as measured by the FIELDS magnetometers. Panel (c) of Figure 5 shows the magnetic field components in RTN coordinates throughout the passages of ICME1 and ICME2, as well as the magnitude of the magnetic field strength from 12:00 UTC on 2020 November 29 (DOY 334) to 00:00 UTC on 2020 December 2 (DOY 337). Panels (d) and (e) show the azimuthal and polar angles, respectively, of the RTN magnetic field vector throughout this time period. As discussed above, the possible shock driven by ICME1 arrives at Parker Solar Probe at ∼23:10 UTC on November 29 (DOY 334) as indicated by the sharp increase in the magnetic field strength (black dashed vertical line in Figure 5). Magnetic signatures of the sheath followed by a region of smoothly rotating magnetic field, suggestive of a magnetic cloud (Klein & Burlaga 1982), entered at ∼03:30 UTC on November 30 (DOY 335) and extending to ∼16:00 UTC, are observed by FIELDS. The smoothly rotating flux rope-like structure of ICME1 is shown by the first shaded region in Figure 5. The shock of ICME2 is detected at ∼18:35 UTC, as evidenced by the large magnetic discontinuity at this time (blue dashed vertical line in Figure 5). The ICME2 sheath then passes over Parker Solar Probe until ∼02:30 UTC on December 1. At this point, FIELDS detects what appears to be a magnetic cloud based on the smoothly rotating magnetic field observed (second shaded region in Figure 5). Interestingly, shortly after Parker Solar Probe passed into the magnetic cloud of ICME2, a strong depression in the magnetic field strength began at ∼03:00 UTC and lasted for ∼15 minutes. During this time, the spacecraft may have briefly left the magnetic cloud and reentered a flux tube with properties similar to that of the preceding sheath based on the observed SEP enhancement within this structure. Further interpretation and confirmation of the various features identified in Figure 5 require more detailed study of the plasma and magnetic field data.
The EPI-Lo electron time series shows a shock spike in which the count rates reach a peak concurrent with the passage of the ICME2-driven shock arrival at 18:35 UTC on November 30 (DOY 335; blue vertical dashed line in Figure 5). At the crossing of the shock, the electron count rates measured by EPI-Lo are roughly two orders of magnitude higher than within ICME1 and more than an order-of-magnitude higher than the electron count rates within the sheath of ICME2. Energetic electron shock spikes are rare and noteworthy features. The significance of this feature will be discussed in Section 4. Neither a shock spike nor a further enhancement timed with the arrival of the shock are observed in connection with ICME1. A clear drop in electron count rate is observed in both instruments through the magnetic cloud associated with ICME2 on December 1 (DOY 336) from ∼03:30-08:30 UTC (identified as the second gray shaded region in Figure 5). The implications of these features outlined above are discussed in Section 4.

EPI-Lo Measurements of Electron Anisotropies
The following subsections outline four periods during which EPI-Lo measured electron enhancements with significant anisotropy throughout these events.  (a) of Figure 7 shows the EPI-Lo electron count rate 15 in the range ∼130-870 keV in wedges 3 and 7, ∼140-730 in wedge 5, and ∼180-790 keV in all other wedges using high-timecadence (1 minute) electron data. Figure 7 also shows the magnetic field vector components and magnitude (panel (b)), the azimuthal angle of the magnetic field vector (where 180°is sunward-panel (c)), the polar angle of the magnetic field vector (panel (d)), the pitch angle relative to the magnetic field direction of the geometric center of each of the EPI-Lo wedges in the same colors as panel (a) (panel (e)), and the PAD relative to the magnetic field vector over a 5 minute interval as a function of time determined by calculating the fraction of the total count rate for a given time interval contributed by a particular wedge (panel (f)). Panel (f) forgoes the high-timecadence data in favor of the high-energy-resolution data to allow for direct comparison between each of the wedges using closer energy ranges as discussed above. In this panel, wedges 3 and 7 use the energy range ∼110-870 keV, wedge 5 is in the energy range ∼80-730 keV, and all other wedges show energy range ∼100-790 keV. The observation that the EPI-Lo wedges are measuring different count rates during this time interval indicates that EPI-Lo is measuring energetic electron anisotropies. It is also evident from Figure 7 that the count rates of each of the wedges change with time, sometimes abruptly. This could be caused by changes in the magnetic field direction, also changing the pitch angle coverage, or changes in the electron anisotropy. The changes in the relative count rates of individual wedges also indicate that the slight differences in their energy ranges do not account for variations in count rates.
The change in the magnetic field orientation starting around 07:00 UTC suggests that the electron enhancement in wedge 1 (red data in panel (a) of Figure 7) is due to the spacecraft entering a magnetic structure with a different electron distribution than the surrounding plasma (this change is most obvious in panel (d) of Figure 7) as the field turns farther northward. This rotation in the magnetic field vector also results in a change in the pitch angle coverages of different EPI-Lo wedges. At this time, wedge 1 (red data in panel (a) of Figure 7) covered the pitch angle range closest to 0°(panel (e)). At roughly 07:50 UTC, the spacecraft performed a roll maneuver, marked by the vertical red dashed line in   Figure 7) into the approximate position formerly held by wedge 1, resulting in wedges 5 and 6 having the pitch angle coverage closest to 0°(see Figure 3). At this point, the highest count rate moves from wedge 1 to wedges 5 and 6, demonstrating a clear anisotropy signature during this time period. Throughout the next ∼7 hr, the wedges with pitch angle coverage closest to 0°have the highest count rates as shown in panel (f) of Figure 7, demonstrating a field-aligned distribution of energetic electrons during ICME1. That the peak intensity at ∼0°persists across the time of the spacecraft roll clearly illustrates the capability of EPI-Lo to observe energetic electron anisotropies. Unfortunately, during this time, the pitch angle coverage toward 180°is poor, so it is not possible to determine whether a counterstreaming flow, such as is often found in ICMEs, was in fact present in this ICME.
The average PAD during the time period marked by the black vertical dashed lines in Figure 7 is shown as the average fraction of the total count rate for each wedge versus pitch angle cosine (μ) in panel (a) of Figure 11. This clearly shows unidirectional field-aligned flow of energetic electrons. Since the minimum intensity appears to be at ∼90°, this hints at the possibility of a bidirectional flow, though it is not possible to confidently draw that conclusion without larger pitch angle coverage.
Close examination of Figure 7 shows other features including several magnetic field discontinuities at which the pitch angle coverage changes abruptly and which in several cases appear to bound structures within the ICME. An example is that at ∼09:25-10:25 UTC (blue vertical dashed lines in Figure 7) during which the particles appear to be more isotropic than outside the structure. The widths of these magnetic structures are discussed in Section 4.
Overall, during the period in Figure 7 within ICME1, the higher rates around pitch angle 0°appear to be present throughout the interval, although the distributions appear to be more isotropic until ∼07:00 UTC, just before the spacecraft roll, when a possible bidirectional flow, with minimum rates near ∼90°develops and appears to extend to the end of this interval (with the exception of the structure at ∼09:25-10:25 UTC).
3.1.2. December 1 ∼03:00-12:00 UTC ICME2 Magnetic Cloud At roughly 03:30 UTC on 2020 December 1, Parker Solar Probe passed from the sheath region into the magnetic cloud of ICME2. This time period is shown in Figure 8 in the same format as Figure 7. This magnetic cloud was identified based on the smooth rotation of the magnetic field lasting until ∼09:00 UTC on that day shown in panel (b) of Figure 8 (approximate boundaries of the magnetic cloud are marked by the blue vertical dashed lines in Figure 8). During the beginning of this time period, EPI-Lo had good pitch angle coverage up until the spacecraft roll at 05:00 UTC (red dashed vertical line in Figure 8), and observed the dominant electron count rates smoothly rotating with the magnetic field in panel (a) of Figure 8. Throughout this time, the wedges with pitch angles closest to 0°and 180°had the highest count rates, with the wedges nearer to 90°measuring significantly fewer electrons, indicating a bidirectional energetic electron distribution. After the spacecraft roll, the pitch angle coverage changes, resulting in poorer coverage at pitch angles less than 50°. After this maneuver, the wedges with pitch angles closest to 180°c ontinued to have higher count rates, and the pitch angles closer to 90°had lower count rates. Since there are no measurements with pitch angles close to 0°, we cannot conclusively state that the bidirectional flow continued. However, it is not unreasonable to assume that the bidirectional electron flow continued in the smoothly rotating magnetic field during the remainder of the magnetic cloud passage. From ∼08:00 UTC, at the trailing edge of the magnetic cloud, the distribution appears to change to become more isotropic, at least based on the limited pitch angle coverage.
The average PAD during the time period marked by the vertical black dashed lines in Figure 8  Due to limited pitch angle coverage, it is challenging to draw conclusions about the electron PAD at the onset of the SEP event upstream of ICME1 shown in Figure 9 in the same format as Figure 7. This figure focuses on the period from just before the onset of the electron event to shortly following the ICME1 shock (vertical blue dashed line). From ∼15:00-16:45 UTC during the early stages of the event, it is possible that the PAD is consistent with a unidirectional flow that peaks at small pitch angles, indicative of antisolar flow based on higher rates observed by wedges with pitch angles <90°and low counts observed by wedges with pitch angle coverage closer to 180°. However, this cannot be fully confirmed due to a lack of coverage at small pitch angles during the onset of the event.
The electron anisotropy discussed so far has focused on periods of long-lasting anisotropy during ICME1 and ICME2. Anisotropy observations can also be used to help identify small flux tube structures through which the spacecraft passes. These structures can have significantly different energetic electron distributions than the surrounding plasma. Upstream of ICME1, around 20:00 UTC on November 29 (DOY 334), Parker Solar Probe appears to have passed through a small flux tube with a highly beamed energetic electron distribution as shown in Figure 9. Around this time, the configuration of the magnetic field changes dramatically, turning more northward temporarily and fortuitously increasing the pitch angle coverage, coincident with a spike in the electron count rate observed by wedges 4 and 5 (orange and purple curves in panel (a)), which are the wedges with the pitch angle coverage closest to 0°. The spike lasts for ∼15 minutes. The average PAD during the time period marked by the black dashed lines in Figure 9 is shown in panel (c) of Figure 11 showing the unidirectional field-aligned electron flow. This change in the magnetic configuration is short-lived, and the electron distribution returns to a roughly isotropic distribution when the field rotates back to near its original configuration. Unfortunately, EPI-Lo has poor pitch angle coverage before and after this rotation, as shown in panel (e) of Figure 9, so it is not possible to draw firm conclusions about the anisotropy of this time period apart from within this flux tube. This also raises the possibility that this spike is observed as a result of improved pitch angle coverage instead of a separate magnetic structure with a different electron distribution. In any case, this scenario still implies that the electron distribution is highly field-aligned during the onset of the energetic electron event. Figure 10 shows observations for a 3 hr period from 17:00-20:00 UTC on November 30 (DOY 335) around the shock of ICME2 (vertical red dashed line). At the crossing of the shock driven by ICME2, the dramatically fluctuating magnetic field direction shown in panel (b) of Figure 10 and limited pitch angle coverage make it challenging to characterize the PAD. However, at the point of the shock crossing (shown by the vertical dashed red line in Figure 10), the highest count rates are in those wedges with pitch angles closest to and slightly below 90°. Limited pitch angle coverage toward 180°m akes it challenging to draw firm conclusions regarding the PAD at the shock crossing. The average PAD during the time period around the shock (marked by the two black dashed vertical lines in Figure 10) is shown in panel (d) of Figure 11. The relative count rates during this time indicate that the . Panel (f) shows the pitch angle time series for each EPI-Lo wedge in the energy range ∼80-870 keV (see the text for a full description of energy ranges) calculated as the fraction of the total count rate in each time bin in 5 minute data averaging intervals. The vertical red dashed line indicates a spacecraft maneuver. The two dashed black lines mark the time period used to calculate the average PAD shown in panel (a) of Figure 11. The two vertical dashed blue lines indicate a magnetic structure in which the electron distribution is more isotropic. highest rates are in the wedges with μ « 0.3-0.4. The error bars on this plot are large due to the widely varying pitch angle coverage and count rates in each wedge.

Discussion
This study demonstrates that EPI-Lo observed near-relativistic electrons during the first widespread SEP event of solar cycle 25, and successfully measured their anisotropies, including during the passage of two ICMEs, at the crossing of the ICME2-driven shock, and within an isolated magnetic structure upstream of ICME1.
The observation of energetic electron count rates peaking at the arrival of an ICME-driven shock as found at the shock upstream of ICME2 is relatively rare and may have important implications regarding the shock parameters that result in efficient electron acceleration (Tsurutani & Lin 1985;Lario et al. 2003;Dresing et al. 2016b). Dresing et al. (2016b) found that quasi-perpendicular shocks may be more efficient accelerators of electrons than quasi-parallel, based on a statistical study examining the time profiles of SEP increases for 475 interplanetary shocks measured by STEREO. The higher electron acceleration efficiency of quasi-perpendicular shocks than quasi-parallel is supported by numerical modeling showing that quasi-perpendicular shocks may accelerate electrons through shock-drift acceleration (e.g., Wu 1984;Guo & Giacalone 2010. J. Giacalone et al. (2021) determined that the shock driven by ICME2 was quasiperpendicular with an angle between the shock normal and the ambient magnetic field, θ BN , of 65°. As shown in Figure 10, at the time of the ICME2 shock crossing, the wedges with the highest count rates were those with pitch angle coverage between 40°and 90°. "Pancake" distributions in which the intensities peak at pitch angles of 90°are a classic signature of shock-drift acceleration at a quasi-perpendicular shock (e.g., Sarris & Van Allen 1974;Marhavilas et al. 2003;Richardson & Cane 2010). In this case, the peak of the distribution is slightly below 90°, perhaps indicating that the electrons are streaming along the shock front around 65°relative to the ambient field or perhaps indicating a "loss cone" distribution in which particles with low pitch angles are missing (e.g., Brain et al. 2007). This could result if the electrons encounter a barrier of higher magnetic field such that the smaller pitch angle particles can pass through and be lost while the higher pitch angle particles are reflected. The electron shock spike observed by EPI-Lo is another example of a growing list of observations and simulations supporting the idea that electrons may be efficiently accelerated by quasi-perpendicular ICME-driven shocks.
Factors other than θ BN may also contribute to the presence of an electron shock spike observed with ICME2. Dresing et al. (2016b) found that an existing pre-event electron enhancement was present in all of the cases they studied in which the peak electron intensity was observed coincident with the shock crossing. In the case of the shock spike observed by EPI-Lo associated with ICME2, a pre-shock background that could serve as a seed population of energetic electrons is clearly present in the electron enhancement beginning at 13:53 UTC on November 29 (DOY 334). The presence of ICME1 preceding ICME2 also suggests the possibility that the energetic electrons are trapped between ICME1 and the shock driven by ICME2, similar to the mechanism proposed in Dresing et al. (2018). Guo & Giacalone (2015) found that magnetic fluctuations upstream of a shock can significantly increase electron acceleration efficiency via shocks regardless of the shock-normal angle. In the case of ICME2, an energetic electron seed population is clearly present, and ICME1 may have provided increased turbulence ahead of the shock driven by ICME2, which may provide more efficient electron acceleration and result in the observed energetic electron shock spike (Li & Zank 2005;Li et al. 2012).
Despite the limited pitch angle coverage, EPI-Lo observations show energetic electron anisotropy measured in each of the ICMEs, which may be indicative of the magnetic connection of the ICME to the Sun or magnetic structures in the heliosphere. The prolonged bidirectional anisotropy observed throughout the magnetic cloud of ICME2 is a signature of electron trapping within the magnetic structure (Malandraki et al. 2005). These observations may imply that the magnetic cloud of ICME2 has a closed topology with legs still connected to the Sun or another magnetic structure resulting in the reflection of energetic electrons back and forth between the closed field boundaries (Richardson & Cane 1996;Torsti et al. 2004;Leske et al. 2012;Dresing et al. 2016a). ICME1 appears to have a prolonged electron enhancement beamed along the direction of the magnetic field, though we cannot rule out the presence of a bidirectional flow due to poor pitch angle coverage toward 180°. The flux rope of ICME1 passes over Parker Solar Probe roughly half a day after the solar event in which ICME2 erupted, likely ruling out the beamed energetic electrons as being due to impulsive electron acceleration from the event that ejected ICME2. Examination of the FIELDS radio frequency spectrometer radio data on November 30 (DOY 335) shows that FIELDS observed a single type III radio burst at ∼13:00 on November 30 (DOY 335) in addition to radio emissions associated with the eruption of ICME2 on November 29 (DOY 334). The electron anisotropy began long before this radio burst, likely precluding the possibility that the beamed electrons were due to a separate impulsive SEP event. The SEP electrons could become beamed if they entered the ICME where the magnetic fields are relatively smooth and hence there is little scattering, so they tend to be focused along the field direction. Due to nonoptimal pitch angle coverage, it is not possible to use the electron anisotropy during ICME1 to draw concrete conclusions regarding the magnetic topology of ICME1.
The observation of the short-lived unidirectional anisotropy with electrons beamed along the direction of the magnetic field on November 29 (DOY 334) demonstrates a brief change in connectivity, likely due to the spacecraft passing into a flux tube with a unique magnetic topology and electron distribution compared with the surrounding plasma. Though pitch angle coverage was incomplete throughout this time, prior to that spike, the enhanced energetic electron signal appeared roughly isotropic. This observation indicates that mesoscale structures in the solar wind can be studied by the ISeIS instruments and that PADs are another valuable tool in identifying and analyzing these structures.
Through this flux tube, the solar probe analyzer-ion instrument within the SWEAP instrument suite on Parker Solar Probe (Kasper et al. 2016) measured an average solar wind velocity of ∼390 km s −1 . Using Equations (1) and (2) from Borovsky (2008), the approximate wall-to-wall width of the flux tube encountered upstream of ICME1 is calculated to be 7.0 × 10 5 km using a time period of November 29 19:45-20:15 UTC (shown in Figure 9). This thickness is reasonable compared with the median thickness of 4.4 × 10 5 km found for the flux tubes analyzed at 1 au in Borovsky (2008). The same analysis was performed on the magnetic structure identified in the time period of ∼09:25-10:25 UTC on November 30 (DOY 335; identified by the vertical blue dashed lines in Figure 7). In this time period, the solar wind velocity is significantly slower, ∼230 km s −1 . Using this solar wind velocity, a flux tube thickness of 7.3 × 10 5 km was calculated for this structure. This is well within the range found by Borovsky (2008), suggesting that this structure may be an individual flux tube within the overall magnetic structure of ICME1.

Summary
The ISeIS instrument suite on Parker Solar Probe detected ions and electrons associated with the widespread SEP event commencing on 2020 November 29. Two CMEs erupted in the direction of the Parker Solar Probe spacecraft near the end of 2020 November with the second, faster, ICME passing Parker Solar Probe shortly after the first. This event produced the largest SEP enhancement observed by the ISeIS energetic particle suite in the mission's first eight orbits. In addition to a large enhancement in energetic ions, ISeIS also observed a significant enhancement in energetic electrons. The size of this event permitted the study of the anisotropy of energetic electrons observed by the ISeIS/EPI-Lo instrument for the first time during this mission. The results presented clearly demonstrate the ability of EPI-Lo to make electron anisotropy measurements in SEP events, even despite the limited pitch angle coverage when Parker Solar Probe is in suboptimal configuration far from the Sun. These include clear evidence of bidirectional electron flow in ICMEs as have been reported in previous studies.
The electron enhancements observed by EPI-Lo are noteworthy due to the presence of a shock spike, as well as the presence of significant anisotropy. The EPI-Lo electron count rate peaked at the same time as the passage of the shock driven by the second ICME. While not unheard of, this is rare due to the inefficiency of the acceleration of energetic electrons by Figure 11. Average PAD for each of the times of interest as a function of pitch angle cosine, μ. Panel (a) shows the average PAD of a 30 minute interval during ICME1 in which EPI-Lo measured electrons aligned with the magnetic field (see Figure 7). Panel (b) shows the average PAD of a 30 minute interval during ICME2 in which bidirectional electrons were measured (see Figure 8). Panel (c) shows the average PAD during a 15 minute interval through the brief flux tube encountered upstream of ICME1 showing a unidirectional flow (see Figure 9). Panel (d) shows the average PAD during a 15 minute interval around the crossing of the shock driven by ICME2 (see Figure 10). Error bars are calculated as the standard deviation of the pitch angle cosine (computed based on the center of each wedge) and count rate throughout each time period. The black vertical dashed lines in each plot show the minimum and maximum pitch angle observed during the time period by individual EPI-Lo apertures. CME-driven shocks. Several studies have found that ICMEdriven quasi-perpendicular shocks are significantly more efficient at accelerating electrons than quasi-parallel shocks. The results of the present study support the possibility of efficient electron acceleration by a quasi-perpendicular shock. Furthermore, the brief PAD with a peak between 40°and 90°supports the idea that shock-drift acceleration contributed to the electron acceleration efficiency of this event. The presence of an upstream electron enhancement providing a seed population for shock acceleration as well as the passage of a prior ICME perhaps resulting in increased turbulence may also contribute to a high efficiency of energetic electron acceleration from the shock.
These events were the first observed by the ISeIS/EPI-Lo instrument to have significant energetic electron anisotropy. The observed anisotropy may indicate a closed magnetic topology of the observed ICMEs, implying direct connection to the Sun or other magnetic structures in the heliosphere.
As the mission progresses and solar activity increases, we look forward to observing more ICME-associated SEP events and examining their characteristics including anisotropies, within the inner heliosphere.