SOL2017-09-04 (M5.5) 2017 as a Source of Relativistic Electrons and Protons

{{Infobox Nugget
|name = Nugget
|title = SOL2017-09-04 (M5.5) 2017 as a Source of Relativistic Electrons and Protons
|number = 373
|first_author = Alexei Struminsky
|second_author = (see Acknowledgments)
|publish_date = 16 March 2020
|next_nugget = Triumphal return of the overlappogram
|previous_nugget = {{#ask: [[Category:Nugget]] [[RHESSI Nugget Index::372]]}}
}}

== Introduction ==

As related in many of these Nuggets, 
[https://www.cosmos.esa.int/documents/1866264/3219248/MatthewsS_Solar_Particle_Acceleration_Radiation_Kinetics.pdf/e92c0bab-fbd1-1337-409a-1c7403958fea?t=1565184646167 particle acceleration] 
figures prominently in the physics of solar flares and coronal mass ejections.
We can observe these particles in the flare regions via their secondary
X-ray and &gamma;-ray emissions, and also in space more directly via the
solar energetic particles (SEPs; "solar cosmic rays").

The relation between the number of accelerated protons and electrons
is an important characteristic of these processes.
The RHESSI observations are consistent with two (possibly concurrent) 
acceleration processes: one that accelerates electrons above 50 keV 
but not above 0.3 MeV (the first phase), and a second that accelerates 
both ions and electrons proportionally to high energies 
(the second phase) (Ref. [1]). 
In order to choose a specific acceleration mechanism suggested
for solar flares (see the Table) we need to track electrons of ~10
MeV and protons of ~100 MeV. 
From the table it is clear that any acceleration
mechanism may operate during the first phase, but we have only two
possibilities for the second phase.

[[File:373f0.png|600px|thumb|center|
]]

To get information on 10 MeV electrons is possible via modeling
the results of radio, X-ray and &gamma;-ray observations, but 
this problem has not been solved unambiguously yet. 
It is impossible in principle to get information
on protons with energies between 30 and 300 MeV (between protons
generating nuclear &gamma;-ray lines and 
[https://en.wikipedia.org/wiki/Pion &pi;<sup>0</sup>-mesons]).  
Therefore let's consider both solar and interplanetary observations,
as described in Ref. [2].  
As a proxy of solar > 100 MeV proton flux we use the count rate of the 
anticoincidence shield of the spectrometer on the 
[https://www.isdc.unige.ch/integral/ INTEGRAL] spacecraft.
Information on relativistic solar electrons is available from the Earth via 
[https://en.wikipedia.org/wiki/Solar_and_Heliospheric_Observatory SOHO].

In studying a solar event, we must be clear about timing.
Since the acceleration phases differ by either the presence or the 
absence of relativistic electrons, we assume a time "zero" at the onset 
time of the 15.4 GHz radio emission, generated by relativistic
electrons via the
[http://astronomy.swin.edu.au/cosmos/S/Synchrotron+Emission gyrosynchrotron] 
mechanism. 
The chosen time zero can be considered as the time at which relativistic 
electrons first appear in the solar atmosphere, and 
the boundary between the first and the second phases.

== Observations of SOL2017-09-04 (M5.5)==

This flare is a brilliant example of two-phase acceleration.
It had distinct first and second phases separated by about ten
minutes. 
The M5.5 flare in the SXR radiation (GOES) began at 20:28
UT (- 15 min with respect to our zero time). 
However, RHESSI recorded the beginning of a gradual signal increase in its 
25-50 keV channel substantially earlier, approximately from
20:17 UT (-26 min).
The RHESSI data also revealed two separate episodes of HXR radiation
("spikes") as clearly seen in Figure 1. 
The first spike started at time -13 min. It was
characterized by a relatively soft spectrum (HXR below 100 keV,
radio waves at frequencies <8.8 GHz) and coincided with the maximum
of the SXR flare according to the GOES data. 
The small delay time of the emission-measure maximum relative to the 
maximum of the temperature characterizes the impulsive character of 
the flare at this time. 
The preflare and impulsive stages of this event lasted more 
than 20 min and were accompanied by the acceleration of electrons with 
energies of less than 100 keV.
The second spike was observed during the decay phase decay of the SXR 
radiation. 
It started simultaneously with the emission of radio waves at 15.4 GHz 
and was characterized by a harder spectrum (up to 100 keV), along with 
a gradual temporal profile. 
Note that both HXR spikes correspond to the post-eruption phase, because they
were detected after the inferred time of the CME position at a
distance of one solar radius, as judged from the 
[https://lasco-www.nrl.navy.mil LASCO] 
data shown in Figure 1.

[[File:373f1.png|600px|thumb|center|Figure 1:
The М5.5 flare of September 4, 2017.
Three vertical lines mark the RHESSI onset (-26 min relative to our "time zero"), the first spike onset (-13 min) and the second spike onset (0 min). 
Top to bottom: the SXR flux (GOES, 1–8 A, dashed line) and the count rate of RHESSI HXR detector (light grey - 25–50-keV; magenta 50–100-keV; green - 100–300-keV); temperature (red) and emission measure (blue) of flare plasma from GOES data, CME position (open circles - SOHO LASCO C2 data, long black arrow - approximation to one solar radius).
]]

==  Interplanetary observations ==

Comparing Figures 1 and 2 reveals that the second HXR spike was the source of the 
gradual electron and proton event.
These show energetic electrons in the 0.27-0.7 MeV channel (v/c 0.78-0.91) to be increasing 
at approximately +2.5 min; i.e., if the electrons were
injected at time zero, they had propagated the distance of 1.2 AU. 
Hence, electrons in the 0.67-3.0-MeV channel (v/c = 0.91-0.99) arrived at +7.5 min, they 
were injected at time +5.8 min, and those in the 2.64-10.4 MeV 
channel (v/c = 0.99) that arrived at +12.5 min were injected at time +10.8 min. 
These data are indicative of a prolonged and gradual acceleration of electrons.

[[File:373f2.png|600px|thumb|center|Figure 2:
Top to bottom: electron flux in the 0.27–0.7(squares), 0.67–3.0(triangles), and 2.64–10.4 (circles) MeV channels from SOHO, 
the 
[https://heasarc.gsfc.nasa.gov/W3Browse/integral/intspiagrb.html Integral]
count rate per 50 ms (background equal to 5720 counts subtracted), smoothed over 1 min (1200 points) – red curve. CME position (blue circles - SOHO LASCO C2 data, blue arrows – possible approximations to one solar radius)
]]

The delay time relative to electromagnetic radiation for protons
with energies of 100 and 500 MeV (v/c 0.43 and 0.76) is ~14 and 5 min,
respectively. 
The arrival time of the first protons varied between
+15 (possibly) and +25 min (Fig.2, ACS SPI). 
Hence, the 100-MeV protons appeared in interplanetary space between 
0 and 10 min, while the 500-MeV protons appeared 
between 10 and 20 min (Fig. 2), which coincides
with the second HXR spike at the decay stage of the M5.5 flare.
Therefore, we identify the first HXR spike with the first phase of
the flare acceleration, while identifying the second spike with the
second phase of the flare acceleration. 
Apparently, electrons
exhibiting a softer spectrum that were accelerated in the first HXR
spike did not reach the threshold energies for their detection by
SOHO and/or their intensity was below the background (Fig. 2).

The similarity of the time profiles of the fluxes of electrons and
protons (Figure 2) suggests that the mechanisms of their acceleration
and propagation are the same. 
To explain the time profiles of proton fluxes in interplanetary space, 
it is necessary to assume their prolonged injection, which can take 
place either due to prolonged acceleration or trapping. 
The time profiles of the proton fluxes are similar to those of the
electrons, for which prolonged trapping is precluded by
synchrotron losses.
Accordingly, the hypothesis of prolonged acceleration is 
strongly preferred.
A prolonged acceleration of protons would be possible either at the 
post-eruptive phase of the flare or at the CME shock
front. 
Since relativistic electrons of the second phase could not
be accelerated at the shock wave, only one possibility remains,
namely, stochastic acceleration at the post-eruptive phase.  

== Scenario ==

The flare itself can consist of a number of elementary actions of
acceleration stochastically distributed in the chromosphere and
corona, both in the first and the second phases (Ref. [3]). 
Elementary acts can differ from one to another by their duration and energy.
The physical mechanisms of acceleration in these elementary actions can be
different and are unimportant in the proposed scenario. 
The energy of accelerated electrons in a flare at the first phase does not
reach the maximum possible value, and the proton energy is too low
to allow their detection by nuclear &gamma;-ray lines (the so-called
"silent protons"). 
Later the radiation losses and continued acceleration smooth out
the spectra of electrons accelerated in different elementary actions.

There is no role for CME/SEPs in this scenario of particle acceleration
in long duration flares, as explained in Figure 3.

[[File:373f3.png|400px|thumb|center|Figure 3:
In this sketch the index ''j'' refers to the flare region, and ''i'' to the CME region.
The outward motion of the CME successively extends its influence
over a wider and wider domain of closed magnetic fields.
In principle the CME can play the role of a mirror, returning
accelerated particles into the flare region and enabling further
acceleration.
The CME additionally accelerates particles that
cross its front many times, creating the conditions for the particles
to escape in a wide spatial angle. 
Protons trapped behind a CME may return back to the dense solar atmosphere and
thus produce the sustained &gamma;-ray -emission.
]]

== Conclusions ==

<ul>
<li>
The observational data do not contradict the scenario of
stochastic acceleration in which protons and electrons simultaneously
gain energy in elementary actions, whose duration is much shorter than
the duration of the flare itself.
</li>

<li>
The proposed scenario combines two, at first sight irreconcilable,
approaches, namely, acceleration both in the flares and in the CME shock
waves as well.
</li>

<li>
The acceleration rate at the flare site should be greater than that at
the shock front. 
Flare-accelerated protons, rather than  shock-accelerated SEPS (Ref. [4]) 
would produce the sustained &gamma;-ray -emission.
</li>
</ul>

== References ==

[1] [http://adsabs.harvard.edu/abs/2009ApJ...698L.152S "RHESSI Observations of the Proportional Acceleration of Relativistic >0.3 MeV Electrons and >30 MeV Protons in Solar Flares"] 

[2] [https://link.springer.com/article/10.1134/S1063780X20020130 "Solar Electrons and Protons in the Events of September 4-10, 2017 and Related Phenomena"]

[3] [https://ui.adsabs.harvard.edu/abs/1985SoPh...96..357B/abstract "Radio Spikes and the Fragmentation of Flare Energy Release"]

[4] [https://ui.adsabs.harvard.edu/abs/2020SoPh..295...18G "Source of Energetic Protons in the 2014 September 1 Sustained Gamma-ray Emission Event"]

== Acknowledgements ==
Irina Grigorieva, Yuri Logachov and Andrei Sadovskii are co-authors of this Nugget.