Dipolarization fronts (DFs) are a commonly observed phenomenon in the tail of Earth’s magnetosphere. They are observed as a sudden increase in the z-component of the magnetic field (e.g., Russell and McPherron, 1973; Angelopoulos et al., 1992; Nakamura et al., 2002). This increased z-component is a result of the fact that DFs carry a more dipolar field than the stretched tail field around it. Accompanying reconnection in the tail, there is often a high-speed earthward flow such as a bursty bulk flow (BBF) (e.g., Angelopoulos et al., 1992). Dipolarizing flux bundles (DFBs) are smaller flux tubes embedded in BBFs that carry a more dipolar field than the surrounding plasma (e.g., Liu et al., 2014). DFs are the kinetic-scale boundaries between DFBs and are often considered a tangential discontinuity between the dipolar field and the stretched tail field (e.g., Sergeev et al., 2009; Fu et al., 2012), although not always (Balikhin et al., 2014). This process is associated with substorms (e.g., Baumjohann et al., 1999; Fu et al., 2020, and references therein) with evidence that they are more common with higher geomagnetic activity (e.g., Liu et al., 2014), a link furthered by the occurrence rate of DFs being about five events per day (Liu et al., 2013; Xiao et al., 2017), which is comparable to substorms.

An increase in the flux of energetic (a few 10 s of keV) ions has been found in many cases at DFs (e.g., Runov et al., 2011; Pan et al., 2014; Birn et al., 2015; Malykhin et al., 2018), most prominently at DFs with the strongest increase in B z (Malykhin et al., 2018). At lower energies, below around 20 keV, the proton flux can decrease due to those ions coming from less dense sources (Birn et al., 2015). The energetic ions at ∼ 20 keV- ∼ 80 keV are found earthward of the DF, while they are found closer to the DF above that energy (Birn et al., 2015). This is because the ions are often accelerated when reflected by the DF, so they end up in front of it (Zhou et al., 2010; Zhou et al., 2019). These ions are typically anisotropic, sometimes with pancake-like (i.e., trapped) pitch-angle distributions, likely as a result of betatron-type acceleration (Birn et al., 2017; Zhou et al., 2018). Zhou et al. (2018) tested the anisotropy against an adiabaticity parameter and found that the anisotropy was stronger when the plasma was more adiabatic. Although these energetic ions are primarily H + , O + can be efficiently accelerated at DFs as well, and are accelerated even further when the DFs are followed by turbulence (Panasyuk et al., 2021).

The acceleration of ions at DFs is much more complicated than electrons at the same location because in general the thickness of the front is smaller than the ion gyroradius, so fully adiabatic acceleration is not possible (e.g., Malykhin et al., 2018), producing orbits that are “partially adiabatic and weakly chaotic” (Büchner and Zelenyi, 1989). For example, a typical DF thickness is ∼ 1,000 km (e.g., Runov et al., 2011) while the gyroradius of a 50 keV proton in a 10 nT field can be up to > 3,000 km but a 50 keV electron in the same field has a gyroradius of < 75 km, so nonadiabatic effects that are not relevant for electrons are for protons and heavier ions. An example of the type of particle motion occurring in this situation is a Speiser orbit (Speiser, 1965) with half orbits around B z in the equatorial plane followed by quasi-adiabatic motion out of the plane (Birn et al., 2015). Ions can also be reflected at DFs multiple times (e.g., Zhou et al., 2010; Zhou et al., 2012; Birn et al., 2015; Lu et al., 2016; Zhou et al., 2018). Another way in which the ions can be not fully adiabatic is when they originate near the reconnection site and initially gain energy nonadiabatically before adiabatic or quasi-adiabatic acceleration once they reach the DFs (Pan et al., 2014). There are other fully non-adiabatic ways in which ions can be accelerated as well however. They can be trapped if the DF has a negative B z dip or quasi-trapped if B z never drops below 0. Quasi-trapping can accelerate particle over 40 keV depending on the number of encounters with the front before getting magnetized and trapping can accelerate particles up to 100 keV depending on how much time the particle is in phase with the front. Additionally wave-particle interactions can accelerate ions. Some of these include waves in turbulent magnetic structures (Grigorenko et al., 2015) and lower hybrid drift (LHD) waves, especially in slower DFs (Greco et al., 2017).

These ions present in the plasma sheet are a mix of ions from the ionosphere like H + , O + , H e + , and N + and ions from the solar wind like H + and H e + + (e.g., Young et al., 1982; Daglis, 2006; Kistler, 2020). H + is present at both sources, so it is not useful for distinguishing between the two populations. The solar wind is highly ionized, with the most common species after H + being H e + + at ∼ 1%–4% of the plasma depending on the solar cycle and whether it is in the fast or slow solar wind. Meanwhile, the ionosphere is mostly singly ionized, with O + as the most common species after H + , followed by H e + , and N + (e.g., Kistler, 2020). O + can enter the plasma sheet from either the nightside auroral region or dayside cusp, especially during storms (e.g., Kistler et al., 2010; Kistler et al., 2016; Kistler et al., 2019). During storms, O + in the plasma sheet is enhanced (Kistler et al., 2005) and the source evolves from being primarily solar wind dominated to primarily ionosphere dominated (Kistler et al., 2023). O + is also enhanced in the plasma sheet during substorms (Pandya et al., 2018) and more specifically the levels of O + increase with higher Kp index (Young et al., 1982; Mouikis et al., 2010; Pandya et al., 2018). These cold ionospheric ions can be found behind a DF (Wang et al., 2017; Xu et al., 2019) and can even affect the DF itself, such as by making the DF slower (Liang et al., 2016). O + density is enhanced at DFs like H + density is, but it occurs more gradually (Zhao et al., 2018).

O + can reach the plasma sheet from two origins: either the cusp on the dayside or the auroral region on the nightside (e.g., Yau et al., 1985). Of these two, both can be relevant but generally the auroral zone is a larger contributor (e.g., Yau et al., 1985; Winglee, 2003). The tail is generally more disturbed when more O + is from the cusp than the auroral zone (Yu and Ridley, 2013). These ions are most commonly deposited into the plasma sheet at a distance of ∼ 20–40 R E downtail (Li et al., 2013; Artemyev et al., 2020), which is often tailward of a DF (Liang et al., 2017). In fact, increased levels of O + can cause the nightside reconnection site to move earthward, so O + is often deposited near the reconnection site or sometimes even downtail of it (Wiltberger et al., 2010). Polar wind plasma, which is primarily H + , reaches the plasma sheet first, followed by the hot cusp plasma and auroral outflow that have more O + (Kistler et al., 2023). The distance downtail ions are transported into the plasma sheet depends on the particle velocity, not energy, so for the same energy O + will enter closer to Earth than H + will (Kistler, 2020). There is also a dawn-dusk asymmetry in where ions are transported, with more on the dusk side despite there not being a corresponding asymmetry in the source region (Li et al., 2013), with this asymmetry also present for energetic O + in addition to cold ions (Kronberg et al., 2015). Once the ions are in the plasma sheet, their flow is primarily earthward, but there is a very small dawnward component on the dawnside and a more significant duskward component on the duskside (Hori et al., 2000).

This work builds off of the study we previously undertook in Chepuri et al. (2023). In that study we studied how well adiabatic acceleration equations explained energetic electron observations at DFs. We found that betatron acceleration overestimated the observed electron acceleration while a combined betatron and Fermi acceleration equation underestimated it while having a high error. Two potential explanations are that there are non-adiabatic processes occurring or that the implicit assumption in this method that the quiet plasma sheet before the DF is similar to the source population is not generally valid. In this work, first we attempt to study ion acceleration to see if they exhibit similar behavior to electrons. Then, we study the composition of the plasma as a proxy for what the source population is to try to answer the open questions from the previous study.

The Magnetospheric Multiscale (MMS) mission consists of four spacecraft in tight formation launched in 2015 (Burch et al., 2016). Starting in 2017, the orbit had an apogee of ∼ 25 R E to spend the most amount of time possible in the nightside reconnection region (Fuselier et al., 2016). The orbit is such that MMS has a tail season when apogee is on the nightside in the Northern hemisphere summer, meaning we can focus our search for events in that time range.

The primary instruments used to measure energetic particles for this study came from the Energetic Particle Detector (EPD) investigation (Mauk et al., 2016). These were the Fly’s Eye Energetic Particle Spectrometer (FEEPS) for electrons (Blake et al., 2016) and the Energetic Ion Spectrometer (EIS) for ions (Mauk et al., 2016). FEEPS measures electrons in the energy range of 25–650 keV while EIS can measure H + in an energy range of 20–500 keV and O + above ∼ 130 keV. FEEPS has 16 energy channels and observes nearly a full sky with 2.5 s time resolution in survey mode. EIS uses time-of-flight measurements to determine energies as well as basic differentiation between Hydrogen, Helium, and Oxygen. In addition to EPD data, we also used the Fast Plasma Investigation (FPI) to measure lower energies up to 30 keV (Pollock et al., 2016). Each spacecraft has four dual 180-degree that spectrometers for electrons, which allow for a 4 π -sr field of view. To study the composition of the thermal plasma, we used the Hot Plasma Composition Analyzer (HPCA) (Young et al., 2016). It measures ions from 1 eV to 40 keV and can differentiate between H + , H e + , H e + + , and O + . It does this all with a time resolution of around 10 s.

Data from the FIELDS instrument suite (Torbert et al., 2016), especially the fluxgate magnetometer (FGM) (Russell et al., 2016) was also necessary to measure magnetic fields and provide other context. Finally, we studied waves with the search coil magnetometer (Le Contel et al., 2016) and the electric field double probes (Lindqvist et al., 2016; Ergun et al., 2016).

We were not able to conclusively test this method of ion betatron acceleration at DFs, but we still were able to test some aspects of it. Adding a quasi-adiabatic correction did not improve the method’s results, which points to focusing on the source population as the main driver of error. However, the imprecise data introduces even more uncertainty. Each measurement of energetic protons covers nearly 20 s, while the changes in the magnetic field occur on the order of a few seconds. This could potentially lead to an underestimate of the acceleration since the measurement of energetic ions after the dipolarization includes some time from before the dipolarization as well. It could lead to other less predictable errors as well though, so this data is not generally reliable, meaning most of the conclusions we can draw from this work are related to the composition of the plasma.

Using HPCA data to study the plasma source provided some more useful results. We are using this data to determine the source of the plasma because that is potentially one of the biggest reasons why the implemented does not work. This method relies on the source population being similar to the quiet plasma sheet. Birn et al. (2014) showed that is not necessarily accurate. They found in simulations that the distance the source population traveled to the DF was dependent on energy and pitch angle. Parallel electrons above ∼ 100 keV are from the inner tail with x > − 22.5 , but below ∼ 10 keV they are from the plasma sheet boundary layer or lobes, with the intermediate energies being from the distant tail. For perpendicular electrons, the lower energy boundary is similar, but it goes up to a few 100 s of keV before the electrons are primarily from the near tail.

We were able to see that using betatron acceleration is more accurate when there are higher levels of O + . O + outflowing from the ionosphere tends to be deposited relatively closer to Earth, with those ions being transported to the plasma sheet around 85% of the time, typically Earthward of 40–55 R E during disturbed times, as many of our events are (Li et al., 2013). More precisely, O + typically reaches the plasma sheet at around 20–40 R E (Artemyev et al., 2020). This would put the bulk of O + ions near the tail reconnection site. As is shown in Figure 5 (panel (c)), this is around or not far downtail of our observations. There may be a slight trend towards having higher levels of O + at higher x distances, which could be a result of being closer to the location in the plasma sheet where O + is deposited. Therefore, if the plasma during an event has high levels of O + , it is likely from a more local location in the tail and therefore more likely to be similar to the plasma sheet prior to the dipolarization. The same, however, is not true of H e + + entering the tail from the solar wind (e.g., Winglee, 2003) so we do not have a comparable relationship between H e + + levels and accuracy of this method.

We have statistically studied dipolarization fronts in the tail with observations of energetic particles by MMS, following the work in (Chepuri et al., 2023). First, we attempted to test the accuracy of an equation for betatron acceleration of ions, including a quasi-adiabatic correction. However, because of the quality of the data, we could not determine a precise relationship beyond setting a bound for the acceleration level. This led us to test the composition of the plasma at these DFs to learn more about the source population. For the DFs with energetic electron acceleration, we found that this method of testing betatron acceleration was more accurate when more O + was present. We hypothesize that this is because O + that flows out of the ionosphere reaches the tail closer to our observations, so higher levels of O + are indicative of a source closer to the observed DF, which makes this method more accurate.