Astronomy & Astrophysics manuscript no.

sgrA_flares_AA ©ESO 2024

July 22, 2024

Flares from plasmoids and current sheets around Sgr A*

I. Dimitropoulos⋆1, 2 , A. Nathanail⋆⋆2 , M. Petropoulou3, 4 , I. Contopoulos2 and C. M. Fromm5, 6, 7

1
Department of Physics, University of Patras, Rio 26504, Greece
2
Research Center for Astronomy and Applied Mathematics, Academy of Athens, Athens 11527, Greece
3
Department of Physics, National and Kapodistrian University of Athens, University Campus Zografos, GR 15784, Athens, Greece
4
Institute of Accelerating Systems & Applications, University Campus Zografos, GR 15784, Athens, Greece
5
Institut für Theoretische Physik und Astrophysik, Universität Würzburg, Emil-Fischer-Strasse 31, 97074 Würzburg,Germany
6
Institut für Theoretische Physik, Goethe Universität Frankfurt, Max-von-Laue-Str.1, 60438 Frankfurt am Main, Germany
7
Max-Planck-Institut für Radioastronomie, Auf dem Hügel 69,D-53121 Bonn, Germany

July 22, 2024

arXiv:2407.14312v1 [astro-ph.HE] 19 Jul 2024

ABSTRACT

Context. The supermassive black hole Sgr A* at the center of our galaxy produces repeating near-infrared flares that are observed
by ground and space based instruments. This activity has been simulated in the past with Magnetically Arrested Disk (MAD) models
which include stable jet formations. The present study uses a different approach in that it considers a Standard and Normal Evolution
(SANE) multi-loop model that lacks a stable jet structure.
Aims. The main objective of this research is to identify regions that contain current sheets and high magnetic turbulence, and to
subsequently generate a 2.2 µm light curve generated from non-thermal particles. These aims require the identification of areas which
contain current sheets and high magnetic turbulence, and the averaging of the magnetization in the regions surrounding these areas.
Subsequently, Particle-in-Cell (PIC) fitting formulas are applied to determine the non-thermal particle distribution and to obtain the
sought after light curve. Additionally, we investigate the properties of the flares, in particular, their evolution during flare events, and
the similarity of flare characteristics between the generated and observed light curves.
Methods. 2D GRMHD simulation data from a SANE multi-loop model is employed, and thermal radiation is introduced to generate
a 230 GHz light curve. Physical variables are calibrated to align with the 230 GHz observations. Current sheets are identified by
analyzing toroidal currents, magnetization, plasma β, density, and dimensionless temperatures. The evolution of current sheets during
flare events is studied, and higher-energy non-thermal light curves are calculated, focusing on the 2.2 µm near-infrared range.
Results. We obtain promising 2.2 µm lightcurves whose flare duration and spectral index behavior align well with observations. Our
findings support the association of flares with particle acceleration and nonthermal emission in current sheet plasmoid chains and in
the boundary of the disk inside the funnel above and below the central black hole.
Key words. black hole physics – GRMHD - Sgr A* - current sheets - magnetic reconnection - NIR flares

1. Introduction that are formed in the innermost region of the black hole and
produce synchrotron radiation (Dodds-Eden et al. 2009; Boehle
In the center of the Milky Way at a distance of D ≃ 8.3 kpc et al. 2016; Ponti et al. 2017; Chatterjee et al. 2021; Scepi et al.
lies Sagittarius A* (Sgr A*), a 4.2 × 106 M⊙ supermassive black 2022).
hole (hereafter SMBH; GRAVITY Collaboration et al. 2022).
The accretion disk around it is optically thin, with collisionless Analytical models have been applied to study the trajectory
high-temperature low-density plasma. Its bolometric luminosity of the hot spots in the resulting lightcurve and showed the im-
is in the sub-Eddington range of L ∼ 1036 erg s−1 (Genzel et al. portance of the effect of gravitational lensing (GRAVITY Col-
2010). The emission from Sgr A* is systematically monitored laboration & Loeb 2005; Younsi & Wu 2015; Bauböck et al.
over a very wide range of frequencies, allowing us to investigate 2020; Ball et al. 2021). As shown by Matsumoto et al. (2020),
the dynamics of the disk in the vicinity of a SMBH. assuming that its motion is in the equatorial plane, the GRAV-
Observations from the GRAVITY Collaboration have shown ITY hot spot follows a circular path at super-Keplerian velocity.
a hot spot rotating around Sgr A* with a period of ∼ 60 min- However, an outflowing conical sub-Keplerian orbit can fit these
utes at a distance of a few gravitational radii (hereafter rg ) from observations equally well (Antonopoulou & Nathanail 2024).
the black hole (GRAVITY Collaboration et al. 2018). Sgr A*
is known to exhibit variability in the near infrared which mani- Lately, the theoretical astrophysics community has increas-
fests itself in the form of several flares over a single day (Ghez ingly turned to numerical simulations, specifically to gen-
et al. 2005a; Do et al. 2019). For the first time, GRAVITY ob- eral relativistic magnetohydrodynamic simulations (hereafter
servations revealed that a flare in Sgr A* coincided with a hot GRMHD), in order to gain a comprehensive understanding of
spot moving around the central black hole. Several investiga- the dynamics involved in accreting black holes. Observations of
tions suggest that flares originate in highly magnetized structures flares and moving hot spots further emphasizes the significance
of considering both magnetic fields and fluid dynamics within
⋆
up1098378@upatras.gr the extreme gravitational environment near a black hole (Dexter
⋆⋆
anathanail@academyofathens.gr et al. 2020; Porth et al. 2021; Chatterjee et al. 2021; Čemeljić
Article number, page 1 of 12
Dimitropoulos et al. A&A proofs: manuscript no. sgrA_flares_AA Sgr A* flares

et al. 2022; Scepi et al. 2022; Mellah et al. 2023; Lin & Yuan 2. Numerical setup
2024).
The numerical methods employed for the GRMHD simulation
in this study closely mirror those used in previous research
(Nathanail et al. 2020). We initialize the simulation with mag-
Numerical investigations explore the formation of current netic field configurations designed to generate multiple current
sheets in the vicinity of a black hole and the subsequent produc- sheets as the accretion system evolves. The formation of cur-
tion of plasmoids and plasmoid chains (Nathanail et al. 2020; rent sheets occurs in the vicinity of the black hole, and their
Nathanail et al. 2022; Ripperda et al. 2022). Current sheets are subsequent reconnection gives rise to the development of mul-
responsible for particle acceleration and the generation of vari- tiple plasmoids and plasmoid chains. In order to assess the ra-
able non-thermal radiation. These models need to be supported diation characteristics of the current sheets formed and the sub-
by numerical studies of turbulence and magnetic reconnection in sequent generation of plasmoids, we employ a radiation proxy
collisionless plasmas that self-consistently capture the dynami- model. This model enables the calculation of both thermal and
cal interplay between particles and fields on the kinetic plasma non-thermal radiation at 230 GHz and 2.2 µm, respectively.
scales. Such investigation can be performed with Particle-in-Cell
(PIC) simulations (Drake et al. 2012; Guo et al. 2014; 2015;
Sironi & Spitkovsky 2014; Shay et al. 2014; Dahlin et al. 2014; 2.1. GRMHD simulation
Li et al. 2015; 2023; Petropoulou & Sironi 2018; Werner et al.
The accretion disk surrounding a black hole can be conceptual-
2018; Ball et al. 2018; Comisso & Sironi 2019). In particular,
ized as a hydrodynamic system in the context of curved space-
these investigations help us study the acceleration processes in
time, where the magnetic field significantly influences its dy-
detail by tracking individual particles, but also characterize the
namics. A GRMHD code integrates these physical properties
properties of the particle distribution function, e.g. shape of dis-
into a unified numerical simulation and allows a robust evolu-
tribution (power law, log-parabolic), fraction of energy carried
tion in time where the accretion and accumulation of magnetic
by non-thermal particles, maximum particle energy and others
flux produce energetic phenomena. We employ the BHAC code
(e.g., Werner et al. 2018; Ball et al. 2018; Petropoulou et al.
(Porth et al. 2017) which solves the ideal MHD equations in gen-
2019; Zhang et al. 2023). Some of the particle properties, like
eral relativity, namely
the slope of the power-law distribution, can be related, through
empirical expressions, to the local plasma properties, such as ∇µ (ρuµ ) = 0 (1)
magnetization σ B B2 /ρ ( where B is the magnetic field strength
and ρ is the density) and plasma β B 2P/B2 (where P is the pres-
sure). These prescriptions can then be combined with GRMHD ∇µ T µν = 0 (2)
simulations to anticipate the generation of non-thermal, high-
energy electron distributions with properties depending on the
local physical conditions (Chatterjee et al. 2021; Scepi et al.
2022; Aimar et al. 2023; Lin et al. 2023; Mellah et al. 2023). ∇µ ∗F µν = 0 (3)
The way to connect GRMHD simulations and PIC results is µν ∗ µν
where T and F are the energy momentum tensor and
the following: after the plasma accretion process has reached the dual of the Faraday tensor, respectively. Here, we de-
an inflow equilibrium (Dexter et al. 2020), the thermal radia- note with ρ the rest-mass density and with uµ the fluid four-
tion is attributed to the hot electrons with a parameterized tem- velocity. Some important characteristics of the code are the
perature derived from the simulation ion temperature (Mości- following: it implements second-order high-resolution shock-
brodzka et al. 2016). In order to include non-thermal electrons capturing finite-volume methods and adaptive mesh-refinement
to radiate, formulas from PIC simulations are employed. Non- (AMR) wherever needed. It also implements constrained-
thermal electron distributions are assumed in each cell depend- transport (Del Zanna et al. 2007) to guarantee a divergence-free
ing on magnetization (for σ < 5) and plasma β (Davelaar et al. magnetic field (Olivares et al. 2019).
2019; Fromm et al. 2022; Cruz-Osorio et al. 2022). This study Furthermore, the model is axisymmetric (2D), and the co-
proposes a novel approach for incorporating non-thermal parti- ordinates of the initial torus-like plasma distribution are spher-
cles into simulation results. Specifically, current sheets are first ical (r × θ × ϕ). We used a logarithmic radial grid and the do-
identified (see also Vos et al. 2024), followed by characterizing main extends out to 2500 rg . The resolution of the model is
their environment in terms of averaged quantities of magnetiza- 4096 × 2048 × 1 cells. The torus at the initial equilibrium state
tion σ and plasma β. Finally, non-thermal particles are sourced has constant specific angular momentum l = 4.28 (Fishbone &
from the plasma within the current sheet. In this paper, we utilize Moncrief 1976). The inner radius was set to rin = 6 rg and the
the 2D GRMHD simulations of Nathanail et al. (2020) to inves- pressure maximum radius to rmax = 12 rg . All quantities are cal-
tigate the ability of the current sheet and plasmoids to produce culated in geometrized units (G = c = 1) in which the gravita-
the observed non-thermal flaring activity of Sgr A*. A radia- tional radius is equal to rg = M. We considered a Kerr black hole
tion model is applied to produce lightcurves at 230 GHz (ther- with dimensionless spin a = J/M 2 = 0.93, where J and M are
mal radiation from the disk) and 2.2 µm (non-thermal radiation the angular momentum and mass of the black hole respectively.
from current sheets; see also Scepi et al. 2022). The structure of The radius of the event horizon is rh = 1.341 rg , and there are 29
the paper is as follows: Section 2 describes the methodology of grid cells inside the horizon.
our investigation divided in three subsections in which we intro- The initial conditions for the magnetic field inside the
duce the GRMHD simulation setup (2.1), the modeling of the torus consists of several loops of interchangeable clockwise-
radiation (2.2), and the method with which we determine cur- anticlockwise orientation between neighboring loops (see
rent sheets and their environment from the GRMHD simulation Fig. 1). In order to generate a magnetic field topology as de-
(2.3). In Section 3 we present our results for the thermal (3.1) scribed above we used a vector potential Aϕ :
and non-thermal radiation (3.2). Finally, in Section 4 we present
our conclusions and discuss prospects for future work. Aϕ ∝ A × B , (4)
Article number, page 2 of 12
Dimitropoulos et al. Dimitropoulos et al.: Sgr A* flares Sgr A* flares

where jν is the emissivity of thermal or non-thermal processes at

a specific frequency1 and τν is the √opacity of the fluid to a given
process at frequency ν. The terms −g dϕdθdr represents the in-
finitesimal volume element dV for each grid point. Close to the
black hole radiation in specific is suppressed by the gravitational
√
redshift term (g3redshift ; Viergutz 1993), where gredshift = Σ∆/A.
Here, A = (r2 + a2 )2 − a2 ∆ sin2 θ, Σ = r2 + a2 cos2 θ and
∆ = r2 − 2rM + a2 . The parameter a corresponds to the spin
of the BH, while M represents its mass, which is fixed at 1 in
these forms (an estimation of gredshift for r = 2rg , θ = π/4 and
a = 0.93 is approximately on the order of 10−1 ). The observ-
able radiation flux is given by Fν = Lν /4πD2 , where the distance
Fig. 1. Initial magnetic field (lines) and density (color) configuration in from the black hole is D = 8.277 kpc (GRAVITY Collaboration
the initial torus. Axis coordinates calculated in rg . Red and blue colors et al. 2022). The integration of the luminosity works best when
denote clockwise and anticlockwise direction of the magnetic field.
the viewing angle is nearly face on, as we are not taking into ac-
count the effects of Doppler beaming and gravitational lensing.
with The difference of such a method compared with a full GRRT
calculation is of order unity (Scepi et al. 2022).
At 230 GHz, the thermal electron synchrotron radiation
A ≡ max(ρ/ρmax − 0.2, 0) , (5)
comes predominantly from the disk region, namely π/3 ∼< θ ∼<
2π/3. In order to calculate the synchrotron emission of a thermal
distribution of plasma, we utilize the formula developed by Le-
ung et al. (2011) (see also Pandya et al. 2016). At 230 GHz, the
B ≡ cos [(N − 1)θ] sin [2π(r − rin )/λr ] , (6) emissivity of a fluid element in the disk is given by the equation
where N is the number of poloidal loops and λr their character- jν ∝ ne e2 ν s Θ−2 −1 −X 1/3
(8)
e c Xe
istic length scale, and ρmax is the maximum rest mass density of
the disk. ne is the electron number density and X ≡ ν/ν s , where ν s ≡
The above model is characterized as SANE multi-loop and (2/9)νL Θ2e sin λ (λ is the particle pitch angle with respect to the
has important differences from a classic Magnetically Arrested direction of the magnetic field), and νL ≡ eB/(2πme c) is the Lar-
Disk (MAD) model. In a classic MAD there is a steady jet, mor frequency. We use Θe to denote the dimensionless tempera-
whereas in the SANE multi-loop model no steady polar outflow ture of the electrons defined as
develops. We have not yet observed a jet from Sgr A* and it is an mp Te P
open question in which direction our models should go for the Θe = (9)
interpretation of phenomena near the black hole. As discussed in me T p ρ
Nathanail et al. (2022), images at 43 and 86 GHz will produce where m p , me are the masses of the proton and electron respec-
an extended stable jet base in the case of MAD model, whereas tively, it is important to mention that the composition of the
in the case of the SANE multi-loop model, the images will pos- fluid is assumed to be electron-proton plasma. The ratio T p /T e
sibly have an extended jet-like structure only in the flaring state. is calculated from the prescription of Mościbrodzka et al. (2016)
We expect that new images of the Sgr A* black hole will further which describes the electron-to-proton coupling in low (disk)
clarify the landscape of pertinent models. and high magnetized regions (funnel), namely
Tp β2 1
2.2. Radiation model = Rhigh + Rlow , (10)
Te 1 + β2 1 + β2
In this subsection we discuss and present in detail the model
that we used for the thermal and non-thermal radiation expected where Rhigh and Rlow are free parameters that determine the heat-
from plasma in the accretion flow. We employ an approximate ing ratio of electrons. In our calculations we set Rhigh = 20 and
"proxy" model for radiation which integrates the thermal emis- Rlow = 1, since we do not focus on a large parametric inves-
sivity across the entire disk. The validity of this method has been tigation of flares from such configurations. In Dihingia et al.
tested in the past against General Relativistic Radiative Transfer (2023b), a comparison was made regarding the prescription of
(hereafter GRRT) calculations (Porth et al. 2019). These compar- temperature ratios proposed by various studies (Mościbrodzka
isons demonstrated a close correspondence between the proxy et al. 2016; Dihingia et al. 2023a; Meringolo et al. 2023). The
method’s results, the detailed GRRT calculations, and the ob- results indicate that the model presented by Mościbrodzka et al.
served variability in the lightcurve. Scepi et al. (2022) employed (2016) offers the most accurate approximation of temperature
a similar strategy to investigate flares originating from current ratios within the disk region. The electron temperature in a
sheets and plasmoids for MAD models. GRMHD simulation can be determined through two temperature
The total luminosity Lν emitted by the fluid at a specific fre- 1
There seems to be some confusion in the literature in the definition of
quency ν (in units of ergs per second per Hz), is calculated based ‘emissivity’. Emissivity usually expresses the effectiveness of the sur-
on the following equation: face of a material in emitting energy as thermal radiation. In the recent
astrophysical literature, however, emissivity jν is defined as the energy
Z 2π Z π Z rmax loss per unit time per frequency per unit volume (e.g. Leung et al. 2011;
Lν = e−τν × jν (B, p, ρ) × g3redshift Ghisellini 2013; Scepi et al. 2022). Note that Rybicki & Lightman 1986
0 0 rh define jν as the ‘emission coefficient’, i.e. the energy loss per unit time
√
−g dϕdθdr (7) per frequency per unit volume per steradian.

Article number, page 3 of 12

Dimitropoulos et al. A&A proofs: manuscript no. sgrA_flares_AA Sgr A* flares

Fig. 2. Upper panel (from left to right): magnetic field Bz component, toroidal current and density. Lower panel (from left to right): magnetization,
plasma β and dimensionless temperature Θ p . By setting limits in these parameters, we define the red regions in Fig. 3.

simulations that independently evolve this quantity (Jiang et al. where

2023). Such studies highlight the importance of self-consistently
evolving the electron temperature to accurately explore the ther- Bν = (2hν3 /c2 )[exp(hν/kT e ) − 1]−1 (12)
mal synchrotron radiation (Jiang et al. 2024).
The thermal radiation absorption term e−τν is important only is the Planck function. So, the total luminosity for thermal radi-
R θ0
inside the disk. The optical depth is defined as τν = π/3 aν,th ds ation at each grid point (r0 , θ0 ) is calculated as follows: First,
√ we determine the emissivity at the point. Next, we compute
where ds = gθθ dθ (π/3 ≤ θ0 ≤ 2π/3 are the latitudinal limits
of the disk). In the case of local thermal equilibrium, the source the absorption term at all grid points (r, θ) where r = r0 and
function of the emitting material is given by the Planck func- π/3 ≤ θ ≤ θ0 . The sum of absorption coefficients (aν,th ) at
tion Bν . Then, the absorption coefficient of the material can be these points provides us with the value of the optical depth τν at
expressed as (r0 , θ0 ). In this way, it is like integrating along light rays perpen-
dicular to the equatorial plane.
The non-thermal emissivity at 2.2 µm, which is attributed
aν,th = jν /Bν , (11) to the synchrotron radiation of relativistic electrons, is given by
Article number, page 4 of 12
Dimitropoulos et al. Dimitropoulos et al.: Sgr A* flares Sgr A* flares
√
(Rybicki & Lightman 1986,Ghisellini 2013) where A p = 1.8 + 0.7/ σ, B p = 3.7σ−0.19 , and C p = 23.4σ0.26 .
Moreover, the electron non-thermal efficiency with respect to the
σT cKU B ν
!− p−1
2 magnetization (σ) and plasma β is
jν ∝ (13)
νL νL ϵ = Aϵ + Bϵ tanh(Cϵ β) , (18)

where U B = B2 /(8π) is the energy density of the magnetic field. where Aϵ = 1 − 1/(4.2σ 0.55
+ 1), Bϵ = 0.64σ , and Cϵ =
0.07

One needs also to determine the non-thermal electron distribu- −68σ0.13 . On the other hand Eq. (19) approximates the slope p
tion n(γ) = Kγ−p , so K is the electron number density multiplied (for Eq. (19) it holds that p = k − 1) with respect to the same
by the efficiency ϵ of the radiation mechanism, and p is the slope parameters based on Meringolo et al. (2023).
of the distribution. k1
In the case of non-thermal radiation the integration of Eq. (7) k = k0 + √ + k2 σ−6/10 tanh (k3 βσ1/3 ) , (19)
σ
is performed along the radial direction. In the funnel region (see
e.g. colored region in Fig. 3) , this is like integrating along rays where k0 = 2.8, k1 = 0.2, k2 = 1.6, and k3 = 2.25. The efficiency
perpendicularRto the equatorial plane. The optical depth in this of particle acceleration in turbulent plasmas is equal to
rmax √
case is τν = r aν,nth ds where ds = grr dr. aν,nth is the di- e1
mensionless absorptivity which for the non-thermal process is ϵ = e0 + √ + e2 σ0.1 tanh[e3 βσ0.1 ] , (20)
σ
defined as
√ 2 where e0 = 1, e1 = −0.23, e2 = 0.5 and e3 = −10.18.
πe K p+2 p+4
In the upcoming section 2.3, we will elaborate on how we
aν,nth = ν 2 ν− 2 fa (p) (14)
8me c L determine current sheets and their associated parameters, such
as magnetization and plasma β in their vicinity. The reason we
(Ghisellini 2013; Pandya et al. 2016), where the function fa (p) are interested in the latter is because dissipative current sheets
is a product of Γ functions, approximated by are fed with plasma from their environment. These conditions
determine the strength of the non- ideal electric fields that accel-
p2
!
p+1 1.8 erate particles and the rate of energy dissipation. All empirical
fa (p) ∼ 3 2 + , (15)
p0.7 40 relations that describe properties of the accelerated particle pop-
ulations refer to the σ and plasma β parameters in the upstream
also in this case, the total luminosity for non-thermal radiation at regions of the current sheets. Once these areas of interest are
each grid point is calculated as follows: First, we determine the identified, these formulas can then be applied to determine the
emissivity at the point (r0 , θ0 ). Next, we compute the absorption slope of the distribution of accelerated electrons within the cur-
term at all grid points (r, θ) where r0 ≤ r ≤ rmax and θ = θ0 . The rent sheets or turbulent regions at the disk boundary.
sum of absorption coefficients (aν,nth ) at these points provides us
with the value of the optical depth τν at (r0 , θ0 ).
2.3. Determination of current sheets and their environment
A GRMHD simulation does not have the necessary micro-
physics to follow the acceleration of particles and extract the PIC simulations that investigate current sheets start with a spe-
electron distribution function or the efficiency for microscopic cific magnetization, a specific plasma β and an initial current
processes occurring within the plasma. In order to incorporate layer in the middle. As the magnetic field evolves magnetic re-
how energy is transferred from the magnetic field to the plasma, connection events take place, magnetic energy is dissipated, and
we used two different post-processing formulas from results of particles are accelerated from the thermal pool to higher ener-
PIC simulations. The first approach is from a local investigation gies. Eqs. (17) and (19) are fitting models of the results of such
of idealized current sheets (Ball et al. 2018). The second one PIC simulations which can give us the slope of the spectrum of
is from an investigation of microphysical properties of special- the non-thermal electrons as functions of GRMHD quantities. It
relativistic turbulence (Meringolo et al. 2023). The first approach is important to emphasize that the parameter values that must be
was applied to the current sheets in the funnel region above the used in these formulas are roughly those set as initial conditions
disk boundary, and the second one in the vicinity of the disk in a PIC simulation. As it turns out, these values are maintained
boundary. The results of such local investigations give us the during the calculation only in the environment surrounding the
opportunity to approximate the properties of the accelerated par- current sheets.
ticles that will eventually radiate through the local plasma char- So, in order to use eqs. (17) and (19) it is necessary to con-
acteristics. Let us first define the non-thermal acceleration ef- strain the current sheet and its surrounding environment. First,
ficiency ϵ as the fraction of the kinetic energy carried by the it is useful to distinguish the regions where current sheets form
non-thermal particles to the kinetic energy of the total electron from the polarity reversal of the magnetic fields (Fig. 2, reversal
population, of Bz ). We utilize five variables to determine the current sheets,
R∞ which are as follows: toroidal current Jϕ := (∇ × B)ϕ , density ρ,
γ pc
(γ − 1)[ dNdγ − f MB (γ, Θe )]dγ
magnetization σ, plasma β and dimensionless temperature Θ p
ϵ= R inf , (16) (normalized to mc2 which is equal with P/ρ). The remaining
γ
(γ − 1) dN
dγ dγ plots in Fig. 2 provide limits for each parameter that defines the
pc
exact region where a current sheet occurs. As we see in Fig. 2, a
where γ pc denotes the peak of the spectrum and f MB is a rela- current sheet is characterized by high values of toroidal current,
tivistic Maxwellian fitting function. high values of density, low values of magnetization, high val-
According to Ball et al. (2018), Eq. (17) gives us the slope of ues of plasma β and high values of dimensionless temperature.
the spectrum with respect to the magnetization σ and plasma β. Thus, we set the following limits for each parameter: |Jϕ | > 10−4 ,
a density floor (cutoff) at ρcut = 2 × 10−5 , a magnetization ceil-
p = A p + B p tanh (C p β) (17) ing cutoff at σcut = 10, a plasma β floor at βcut = 10−3 and
Article number, page 5 of 12
Dimitropoulos et al. A&A proofs: manuscript no. sgrA_flares_AA Sgr A* flares

Fig. 4. Cumulative 230 GHz flux from all angles θ for radii less than or
equal to r.

Fig. 3. Determination of plasma (red color) in the reconnection region

and places of high magnetic turbulence. They are selected after applying
limits on toroidal current, magnetization, plasma β, density and dimen-
sionless temperature. The green grid corresponds to the discretization
for the purpose of parameter averaging. Same snapshot as in Fig 2.

finally a temperature floor at Θ p = 10−3 . By combining these

constraints, we obtain the red regions in Fig. 3 which are clearly
associated with the current sheets we identified. Additionally,
the plasma in these red-colored regions has undergone the re-
connection process. Efficient particle acceleration occurs in re-
gions with high magnetization and low plasma β Ball et al. 2018.
Such regions develop only in the funnel region, so we applied the
above methodology exclusively to the plasma within this region.
Following this procedure we have identified the current
Fig. 5. Top panel: the lightcurve at 230 GHz (red line), the moving
sheets developed in the simulation. The next step involves char-
average of the lightcurve (m3 , in a time window of 3 hours, black line)
acterizing the current sheet environment in terms of averaged and the 1s3 standard deviation (grey area). Lower panel: The measure
quantities in order to use Eq. (17) and (19) for the plasma within of variability the lightcurve M3 defined as M3 = s3 /m3 .
the current sheet. To achieve this, a discretization method is em-
ployed, wherein the region is divided into cells with a radial
length of 1 rg and a meridional extent of 10 degrees in the θ Subsequently, we present results from the non-thermal radiation
direction. In each cell that contains part of a current sheet (red calculations at 2.2 µm, including the resulting light curve, an
region) we calculate the mean values of magnetization σ and analysis of flares, and a comparison with the GRMHD simula-
plasma β of their environment (white region) over 2 cells to the tion results.
right and 2 to the left in the θ direction, namely
cell=+2 3.1. Radiation at 230 GHz
1 X
σ= σ (21)
4 cell=−2 Observations of Sgr A* at 230 GHz have been conducted by var-
ious telescopes, including the Submillimeter Array (SMA), the
Atacama Large Millimeter/submillimeter Array (ALMA), and
cell=+2 notably, the Event Horizon Telescope (EHT). The latter, in par-
1 X
β= β (22) ticular, has provided a wealth of high-resolution images from
4 cell=−2
the innermost region of the accretion disk onto Sgr A* itself,
presenting an invaluable resource for the scientific community
Subsequently, the slope of the non-thermal electrons inside the
(EHT collaboration 2022a-f). Thus there exists a plethora of data
current sheet or in the turbulent plasma in each cell is com-
on the properties of the lightcurve at that particular frequency.
puted using equations Eq. (17) and (19) with parameters values
Therefore, utilizing much of this data, our main goals are to re-
as σ = σ and β = β. The plasma assumed to be accelerated produce the main features of the light curve at 230 GHz (am-
corresponds to the region depicted in red in Fig. 3. plitude and variability) by calibrating the various physical pa-
rameters from dimensionless code units to actual physical units.
3. Results When we combine how the flux changes with distance and angle
(as illustrated in Fig. 4), it becomes clear that the majority of the
In the next two subsections, we present our results on the ther- radiation is emanating from a distance of approximately 10 rg .
mal and non-thermal radiation. The first part covers calculations This underscores that the events we are studying occur in ex-
at 230 GHz (thermal component), which help calibrate the phys- tremely close proximity to the black hole, and consequently, the
ical parameters for Sgr A*. Further calculations provide infor- critical data points we are interested in are highly concentrated
mation about the model’s variability compared to observations. in this area.
Article number, page 6 of 12
Dimitropoulos et al. Dimitropoulos et al.: Sgr A* flares Sgr A* flares

Fig. 7. Lightcurve at 2.2µm. Our model can produce one very bright
flare around 7 mJy and another three that pass the threshold of 1.5 mJy.

Fig. 6. Distribution of M3 of the lightcurve at 230 GHZ (blue) compared

with historical data from observations of Sgr A* (grey). We have also
cut the lightcurve and respectively M3 in three time intervals, first for
t ∈ [0, 370] min (green), second for t ∈ [370, 746] min (red) and third
fort ∈ [746, 1143] min (yellow).

As we know from observations (Wielgus et al. 2022) the ra-

diation flux of Sgr A* at 230 GHz is approximately stable. In
order to turn the parameters of the GRMHD simulation from
code units to physical cgs units a conversion and calibration of
variables is necessary. We define a scaling parameter S, that ap-
propriately multiplies physical quantities such as magnetic field
(B), pressure (P) and density (ρ) to convert values of physical
quantities from geometric (code) units to cgs units. Obviously,
the conversion leaves unchanged dimensionless quantities such
as Θe , σ, and β
√
Bcgs = 4πcS 1/2 Bgeometric (23)

ρcgs = S ρgeometric (24)

Pcgs = c2 S Pgeometric (25) Fig. 8. Histogram comparison between our model and observations
from Do et al. (2019).
Through an iterative procedure applied in various snapshots of
the GRMHD calculation, the model limited the radiation flux in
a range around 4 Jy and finally produced the value of the scaling a downward trend, suggesting that if we keep running the same
parameter S . After calculating the parameter S we can calcu- simulation for further time, the flare variability will eventually
late the radiation flux at all snapshots and finally produce the stabilize at lower levels. Fig. 6 is a histogram of M3 that provides
lightcurve at 230 GHz. The upper panel of Fig. 5 presents the a broader view derived from observing Sgr A*, the historical
lightcurve at 230 GHz (red line). Noticeably, the radiation flux data is represented in gray (see Wielgus et al. 2022). The other
remains almost constant throughout the calculation period. The distributions are derived from the simulation results for different
higher flux values observed in the beginning of the light curve time windows. It is important to keep in mind that the yellow
gradually decrease towards the end of the evolution. line distribution represents the last part of the M3 curve inside
To investigate the variability of the lightcurve we define the the time interval t ∈ [746, 1143] min. This closely aligns with
moving average radiation flux over a 3-hour period (m3 , black the observations and shows that a new calculation of the same
line in Fig. 5) and the standard deviation over the same dura- model for longer times will probably yield the desired variabil-
tion (s3 , grey area in Fig. 5). The lower panel of Fig. 5 shows ity in the lightcurve.
the spread of M3 where M3 = s3 /m3 . M3 is a measure of the
variability of the lightcurve and can be compared with historical
observations from Sgr A* (Fig. 6). 3.2. Results at 2.2 µm (NIR flares)
Sgr A* lightcurve shifts and changes over time due to a 3.2.1. Lightcurve and statistics
combination of stellar winds and turbulence on various scales.
Longer changes come from variations in the stellar winds, espe- In Fig. 3 we carefully marked the specific areas where non-
cially near the S stars. Shorter changes happen because of tur- thermal radiation processes occur. These dynamic areas reside
bulence closer to the center. The spread of M3 in Fig. 5 exhibits within the funnel region and in the limit of the disk. There is
Article number, page 7 of 12
Dimitropoulos et al. A&A proofs: manuscript no. sgrA_flares_AA Sgr A* flares

an environment characterized by high magnetization and low

plasma β values which are important to activate the radiation
mechanisms with sufficient efficiency. In Section 2.3 we dis-
cussed in detail that it is this local environment that characterises
the process of acceleration for the plasma residing within current
sheets.
Using data from these regions, we calculate the light curve
at 2.2 µm, and our results are depicted in Fig. 7. Notably, during
quiescent states, the average flux is remarkably low, consistent
with observations. Within the simulation, several flares occur,
each typically exceeding a flux of > 1.5 mJy (Witzel et al. 2021;
von Fellenberg et al. 2023). Noteworthy, there are four flares dur-
ing the simulation time that surpass this threshold, with an addi-
tional three flares marginally meeting the criteria for acceptance
as flares. One particularly significant radiation flare, with a flux
of approximately 7 mJy and a longer duration than typical flares,
stands out as a potential candidate within the observed limits for
both flux and duration, resembling to the unprecedented flare re-
ported in Do et al. 2019.
In a theoretical comparison with observations we do not have
a flux limit that allows us to define what is considered as flare and
what not but it is obvious that the values of the flux in Fig. 7 rise
up several orders of magnitude between a quiescent and a flare
state (F ∼> 1.5 mJy). Sgr A* produces flares with lower flux but
also some very bright ones (F > 6 mJy; Do et al. 2019) like those
produced in our model. In Fig. 8 we see a comparison of the
historical flux distribution in histogram for Sgr A* from Witzel
et al. 2018 as presented in Do et al. (2019) compared to the high
flux flare observed in 2019. The probability of observing such a
flare (>6mJy) was computed to be less than 0.05% (Witzel et al.
2018). Our model is capable of reproducing high-flux flares that Fig. 9. Part of the lightcurve at 2.2 µm. In the upper plot, Flare A, for t ∈
cover the full range of observations. [86, 150] min (the first bright flare with approximately 7 mJy radiation
According to Fig. 8 our model appears to have fairly good flux) and in the bottom plot, Flare B, for t ∈ [879, 926] min (a bright
statistics in terms of the population of flares and in terms of their flare with two peaks). The duration of the flares is about 40-60 minutes,
intensity. This is not the case for Sgr A* flares generated from duration relevant to observations. We can divide the lightcurve in three
MAD models which seem to show a very large population of states: pre-flare, dim-state and bright-state. This separation is based on
flares and and tend to overproduce the quiescent state flux (Dex- the flux and spectral index
ter et al. 2020; Scepi et al. 2022; White & Quataert 2022). This
is probably due to the very strong magnetic field they possess,
3.2.2. Properties of flares
and to the fact that the main source of the flares is the current
sheet in the equator which feeds high values of absorptivity in In the upper panel of Fig. 9 we zoom in on the first major flare
the density due to the Ṁ (as seen characteristically in the paper within the time range t ∈ [86, 150], hereafter referred to as Flare
of Scepi et al. 2022). A. Whereas, on the lower panel, we zoom on a second, smaller
Another crucial variable to consider is the cooling time of flare within the time range t ∈ [879, 926], hereafter referred to as
particles responsible for producing non-thermal radiation. Our Flare B. We examine how the properties of each flare change
method for computing the light curve involves each simula- across three different stages: the pre-flare state, which occurs
tion snapshot radiating under the assumption of the fast cooling when the flux is several orders of magnitude smaller than 1 mJy;
regime. Additionally, no cooling effects are considered, treating the dim-state, which occurs when the flux begins to increase (as
all radiation as if emitted instantaneously. The accuracy of this shown in Fig. 11, where the spectral index also increases) and
approach depends on the estimation of the cooling parameter. finally, the bright-state, which occurs when the flux exceeds 1
The cooling time of a particle with Lorentz factor γ due to syn- mJy. In both flares, we can see the specific points and their cor-
chrotron radiation is given by the equation: responding colors that represent the three different states. Each
point is derived from a specific snapshot of the GRMHD sim-
γme c2 ulation. The first flare is characterized by a distinct peak, while
t syn = (26)
4/3σT cU B γ2 β2 the second flare has two peaks. Similar phenomena have been
observed (GRAVITY Collaboration et al. 2018). The classifica-
where the Lorentz factor γ is γ = (ν/νL )1/2 . Using typical values tion of points into the different flare states was based on flux
for magnetic field B from our model, approximately 60 G for the and spectral index. The rationale for this classification is clearly
region of interest calculating the cooling time from relation (26) illustrated in Fig. 11.
we obtain t syn ∼ 3 min. This suggests that particles have enough Following the same calculation methods used for the non-
time to radiate their energy in a snapshot of the GRMHD sim- thermal radiation, we estimate the radiation flux at three addi-
ulation, given that simulation outputs occur every 10 M, which tional frequencies: 6.66, 7.95, and 18.1 GHz, which correspond
translates to approximately 200 seconds for Sgr A*. Therefore, to the M, L, and H bands, respectively. The primary frequency of
cooling effects do not significantly impact the calculation. our calculation, 2.2 µm, corresponds to the K-band for Sgr A*.
Article number, page 8 of 12
Dimitropoulos et al. Dimitropoulos et al.: Sgr A* flares Sgr A* flares

Fig. 10. Radiation flux around 2.2 µm. Specific to the frequencies 6.66
(M- band), 7.95 (L-band), 13.8 (K-band) and 18.1 (H-band) GHz for
Flare A (stars) and Flare B (circles). The fitting (black and orange) lines
correspond to power-laws where the exponential gives us the spectral
index of the spectrum, in our case a = −0.44 for Flare A and a = −0.46
for Flare B.

Table 1. Radiation flux for Flare A and Flare B at frequencies 2.2µm,

1KeV and 10KeV.

Flare A Flare B
Flux (mJy) Flux (mJy)
2.2µm 7 3.2 Fig. 11. Spectral index vs Flux for the two flares (Flare A upper plot
1KeV 1 × 10−5 2.6 × 10−6 and Flare B bottom plot). The dots correspond to the snapshots that are
10KeV 3.7 × 10−6 9.2 × 10−7 marked also in Fig. 9: cyan color corresponds to pre-flare states, green
color represents the dim-state, red color represents the bright state of
each flare. The orange lines are a fitting power-law curve which can be
compared with observations
For this calculation we used two snapshots at times t = 136 min
and t = 920 min, one for each flare at the peak of the bright state.
We thus extract the spectrum of our model for this flare state the phenomena originated. Therefore, we plotted the parameters
(Fig. 10). The fitting curves in Fig. 10 can give us the spectral in- ρ, σ and plasma β, which play crucial roles in determining the
dex (which defined as the exponent of a power law like F ∝ νa ) current sheet (Fig. 12). The plots refer to the three states of the
for each flare. The two values that we obtained (a = −0.46 and second flare we investigated, specifically the snapshots at times
a = −0.44) are within the observed range (Ghez et al. 2005b,a; t = 886 min (pre-flare), t = 900 min (dim-state), t = 910 min
Bremer et al. 2011). (bright-state). The corresponding points of these snapshots can
Assuming that the particles can emit at higher energies via be seen in the light curve of Fig. 9.
the mechanism of magnetic reconnection, and using the spec- The plot of the density (ρ) (top row in Fig. 12) shows a clear
trum shown in Fig. 10, we extrapolated the radiation flux for X- structure in the upper funnel region that appears as the flare tran-
rays at 1 keV and 10 keV. The results are presented in Table 1. sitions from one state to the next. This structure is also character-
As shown, the flux of particles in the case of Flare A at 1 keV ized by low magnetization and high plasma β. According to the
reaches levels comparable to those observed (Eckart et al. 2012). definition given above, it constitutes a clear current sheet which,
In all other cases, the flux is at least an order of magnitude lower. in the bright-state, is surrounded by regions with high values of
magnetization (σ) (middle row in Fig. 12) and low plasma β
During a flare, both the flux and the spectral index simultane- (bottom row in Fig. 12). These are ideal conditions for radiat-
ously increase. We categorize the points into three groups, distin- ing non-thermal radiation. This current sheet in the bright state
guished by the colors used to represent them. Fig. 11 shows the can be characterized as a plasmoid chain, similar to those seen in
evolution of the spectral index as the flares transition from the localized PIC simulations (Ball et al. 2018; Sironi & Spitkovsky
pre-flare state to the dim-state and finally to the bright-state. The 2014; Petropoulou & Sironi 2018).
fitting power law curve (orange line) approximates how the spec- It is important to emphasize that this increase in magnetiza-
tral index evolves as the flare transitions from one state to the tion σ and decrease in plasma β in the bright state of the flares
next, which is comparable to observations (Bremer et al. 2011). (conditions ideal for the activation of flares) are also found at the
The same colors were used to distinguish the points in Fig. 9. boundaries of the disk. These are possible regions for flare gen-
eration through high magnetic turbulence in the plasma, as men-
3.2.3. Flare states compared with GRMHD simulation tioned in subsection 2.2. Fig. 12 clearly shows that the creation
of the current sheet inside the funnel region is a phenomenon
As a final property of the flares we wanted to see how these peaks perfectly connected to the organization of the magnetic field and
in the lightcurve relate to the GRMHD calculation from which results in the creation of conditions (magnetization and plasma
Article number, page 9 of 12
Dimitropoulos et al. A&A proofs: manuscript no. sgrA_flares_AA Sgr A* flares

Fig. 12. Top row: density of the second bright flare (t ∈ [886, 963] min) for its three states: pre-flare, dim-state, birght-state at snapshots t = 886 min,
t = 900 min and t = 910 min respectively. Middle row: magnetization (σ) at the same snapshots. Bottom row: plasma β at the same snapshots. All
panels refer to Flare B.

Article number, page 10 of 12

Dimitropoulos et al. Dimitropoulos et al.: Sgr A* flares Sgr A* flares

to have the specific slope and efficiency, producing the reported

flux. Two-thirds of the non-thermal radiation flux comes from
the disc boundary, where the turbulent plasma dominates, while
one-third comes from the current sheet.
Similar phenomena can be observed in Flare A at time t =
136 min. A clear current sheet has produced a plasmoid chain,
and the local environment is characterised by high magnetiza-
tion and low plasma β (see Fig. 15). Accordingly, the slope p
and the efficiency ϵ (middle and right panel of Fig. 14) peak at
the current sheet and produce the radiation flux (left panel of
Fig. 14), which strengthens our conclusions about how the flares
are generated in this particular calculation.

Fig. 13. Radiation flux at 2.2 microns (in mJy), slope p and efficiency 4. Conclusions
ϵ in each grid point of the GRMHD simulation are shown in the left,
middle and right panels, respectively. The plots illustrate the bright state We analyzed data from the GRMHD simulation of Nathanail
of Flare B. et al. (2020), specifically focusing on model D, which is charac-
terized as a SANE multi-loop model. Unlike MAD simulations,
this model describes the accretion disk of a black hole without
the production of a stable jet.
By applying a thermal radiation model to this particular sim-
ulation enabled us to reproduce the lightcurve at 230 GHz, from
which we were able to calibrate the simulation variables. En-
couragingly, the model’s variability demonstrated good agree-
ment with observations, prompting us to conduct larger simula-
tions with extended time evolution.
In order to calculate the non-thermal radiation from the
simulation, we introduced a novel method that identifies cur-
rent sheets and places with high magnetic turbulence. This
identification based on the magnetic field polarity reversals, is
done by constraining primarily the current density together with
the micro-physical plasma parameters, like magnetization and
Fig. 14. Same plots as Fig. 13 but for the bright state of Flare A. plasma β. After the identification is done, all quantities are av-
eraged in the local environment. The constraints we placed on
the parameters provided clear spatial boundaries within which to
apply the model that calculates the non-thermal radiation. This
association not only advances our comprehension of the funda-
mental processes at play, but also provides a crucial framework
for future investigations in similar contexts.
The results for the lightcurve at 2.2 µm are very encouraging
as they produce not only small flares up to 2 mJy but also a
couple of brighter ones that reach 7 mJy. The duration of such
flares is also consistent with observations. Further analysis of
the flare spectral index demonstrated the success of the model
in reproducing the observations. It is interesting that during the
evolution of the flare the spectrum follows a power law similar to
those given by the observations with spectral index a = −0.44.
In summary, we emphasize once again that a very important
Fig. 15. Density (left panel), magnetization σ (middle panel) and
plasma β (right panel) for Flare A in the bright state. result of this work is the identification of the source of flaring
non-thermal radiation in GRMHD simulations. We have clearly
shown that such flares most probably originate in current sheets
β) which will activate non-thermal particles in the current sheet and their associated plasmoid chains in the funnel area, as well
itself and along the disk boundary. as in the disk boundary due to magnetic turbulence.
Indeed, in accordance with the above, the areas that radiate
are the current sheet and the disk boundary as shown in Fig. 13. Acknowledgements
Fig. 13 illustrates the radiation flux (in mJy), the slope p and
the efficiency ϵ at time t = 910 min (bright state of Flare B that ID is supported by the Hellenic Foundation for Research and In-
we investigate). The visible boxes in each plot correspond to the novation (HFRI) under the 4th Call for HFRI PhD Fellowships
identification of current sheets and their environment as shown (Fellowship Number: 9239). Support comes from the ERC Ad-
in Fig. 3. Within each box, all plasma quantities are averaged in vanced Grant “JETSET: Launching, propagation and emission
order to characterize this specific reconnection region. However, of relativistic jets from binary mergers and across mass scales”
only the plasma within the current sheet layer - determined by (Grant No. 884631). This work was supported by computational
the constraints and limits discussed in Section 2.3 - is considered time granted from the National Infrastructures for Research and
Article number, page 11 of 12
Dimitropoulos et al. A&A proofs: manuscript no. sgrA_flares_AA Sgr A* flares

Technology S.A. (GRNET S.A.) in the National HPC facility Lin, X. & Yuan, F. 2024, arXiv e-prints, arXiv:2405.17408
- ARIS - . Simulations were performed also on the GOETHE- Matsumoto, T., Chan, C.-H., & Piran, T. 2020, Monthly Notices of the Royal
HLR cluster at CSC in Frankfurt Astronomical Society, 497, 2385
Mellah, I. E., Cerutti, B., & Crinquand, B. 2023, arXiv preprint
arXiv:2305.01689
Meringolo, C., Cruz-Osorio, A., Rezzolla, L., & Servidio, S. 2023, The Astro-
Data Availability physical Journal, 944, 122
Mościbrodzka, M., Falcke, H., & Shiokawa, H. 2016, Astronomy & Astro-
The data underlying this article will be shared on reasonable re- physics, 586, A38
quest to the corresponding author. Nathanail, A., Dhang, P., & Fromm, C. M. 2022, Monthly Notices of the Royal
Astronomical Society, 513, 5204
Nathanail, A., Fromm, C. M., Porth, O., et al. 2020, MNRAS, 495, 1549
References Nathanail, A., Mpisketzis, V., Porth, O., Fromm, C. M., & Rezzolla, L. 2022,
MNRAS, 513, 4267
Aimar, N., Dmytriiev, A., Vincent, F. H., et al. 2023, A&A, 672, A62 Olivares, H., Porth, O., Davelaar, J., et al. 2019, Astronomy & Astrophysics, 629,
Antonopoulou, E. & Nathanail, A. 2024, arXiv e-prints, arXiv:2405.10115 A61
Ball, D., Özel, F., Christian, P., Chan, C.-K., & Psaltis, D. 2021, ApJ, 917, 8 Pandya, A., Zhang, Z., Chandra, M., & Gammie, C. F. 2016, The Astrophysical
Ball, D., Sironi, L., & Özel, F. 2018, The Astrophysical Journal, 862, 80 Journal, 822, 34
Bauböck, M., Dexter, J., Abuter, R., et al. 2020, Astronomy & Astrophysics, 635, Petropoulou, M. & Sironi, L. 2018, Monthly Notices of the Royal Astronomical
A143 Society, 481, 5687
Boehle, A., Ghez, A., Schödel, R., et al. 2016, The Astrophysical Journal, 830, Petropoulou, M., Sironi, L., Spitkovsky, A., & Giannios, D. 2019, ApJ, 880, 37
17 Ponti, G., George, E., Scaringi, S., et al. 2017, Monthly Notices of the Royal
Bremer, M., Witzel, G., Eckart, A., et al. 2011, Astronomy & Astrophysics, 532, Astronomical Society, 468, 2447
A26 Porth, O., Chatterjee, K., Narayan, R., et al. 2019, ApJSS, 243, 26
Chatterjee, K., Markoff, S., Neilsen, J., et al. 2021, Monthly Notices of the Royal Porth, O., Mizuno, Y., Younsi, Z., & Fromm, C. 2021, Monthly Notices of the
Astronomical Society, 507, 5281 Royal Astronomical Society, 502, 2023
Comisso, L. & Sironi, L. 2019, The Astrophysical Journal, 886, 122 Porth, O., Olivares, H., Mizuno, Y., et al. 2017, MNRAS, 4, 1
Cruz-Osorio, A., Fromm, C. M., Mizuno, Y., et al. 2022, Nature Astronomy, 6, Ripperda, B., Liska, M., Chatterjee, K., et al. 2022, The Astrophysical Journal
103 Letters, 924, L32
Dahlin, J., Drake, J., & Swisdak, M. 2014, Physics of Plasmas, 21 Rybicki, G. B. & Lightman, A. P. 1986, Radiative Processes in Astrophysics
Davelaar, J., Olivares, H., Porth, O., et al. 2019, A&A, 632, A2 Scepi, N., Dexter, J., & Begelman, M. C. 2022, Monthly Notices of the Royal
Del Zanna, L., Zanotti, O., Bucciantini, N., & Londrillo, P. 2007, Astronomy & Astronomical Society, 511, 3536
Astrophysics, 473, 11 Shay, M., Haggerty, C., Phan, T., et al. 2014, Physics of Plasmas, 21
Dexter, J., Tchekhovskoy, A., Jiménez-Rosales, A., et al. 2020, Monthly Notices Sironi, L. & Spitkovsky, A. 2014, The Astrophysical Journal Letters, 783, L21
of the Royal Astronomical Society, 497, 4999 Čemeljić, M., Yang, H., Yuan, F., & Shang, H. 2022, ApJ, 933, 55
Dihingia, I. K., Mizuno, Y., Fromm, C. M., & Rezzolla, L. 2023a, Monthly No- Viergutz, S. 1993, Astronomy and Astrophysics, Vol. 272, p. 355 (1993), 272,
tices of the Royal Astronomical Society, 518, 405 355
Dihingia, I. K., Mizuno, Y., Fromm, C. M., & Younsi, Z. 2023b, arXiv preprint von Fellenberg, S. D., Witzel, G., Bauböck, M., et al. 2023, A&A, 669, L17
arXiv:2305.09698 Vos, J. T., Olivares, H., Cerutti, B., & Mościbrodzka, M. 2024, MN-
Do, T., Witzel, G., Gautam, A. K., et al. 2019, The Astrophysical Journal Letters, RAS[arXiv:2309.03267]
882, L27 Werner, G. R., Uzdensky, D. A., Begelman, M. C., Cerutti, B., & Nalewajko, K.
Dodds-Eden, K., Porquet, D., Trap, G., et al. 2009, The Astrophysical Journal, 2018, Monthly Notices of the Royal Astronomical Society, 473, 4840
698, 676 White, C. J. & Quataert, E. 2022, ApJ, 926, 136
Drake, J., Swisdak, M., & Fermo, R. 2012, The Astrophysical Journal Letters, Wielgus, M., Marchili, N., Martí-Vidal, I., et al. 2022, The Astrophysical Journal
763, L5 Letters, 930, L19
Eckart, A., García-Marín, M., Vogel, S., et al. 2012, Astronomy & Astrophysics, Witzel, G., Martinez, G., Hora, J., et al. 2018, ApJ, 863, 15
537, A52 Witzel, G., Martinez, G., Willner, S. P., et al. 2021, ApJ, 917, 73
EHT collaboration. 2022, The Astrophysical Journal Letters, 930, L12 Younsi, Z. & Wu, K. 2015, Monthly Notices of the Royal Astronomical Society,
Fishbone, L. G. & Moncrief, V. 1976, The Astrophysical Journal, 207, 962
454, 3283
Fromm, C. M., Cruz-Osorio, A., Mizuno, Y., et al. 2022, A&A, 660, A107
Zhang, H., Sironi, L., Giannios, D., & Petropoulou, M. 2023, ApJ, 956, L36
Genzel, R., Eisenhauer, F., & Gillessen, S. 2010, Reviews of Modern Physics,
82, 3121
Ghez, A., Hornstein, S., Lu, J., et al. 2005a, The Astrophysical Journal, 635,
1087
Ghez, A., Salim, S., Hornstein, S. D., et al. 2005b, The Astrophysical Journal,
620, 744
Ghisellini, G. 2013, Radiative processes in high energy astrophysics, Vol. 873
(Springer)
GRAVITY Collaboration, Abuter, R., Amorim, A., et al. 2018, A&A, 618, L10
GRAVITY Collaboration, Broderick, A. E. & Loeb, A. 2005, Monthly Notices
of the Royal Astronomical Society, 363, 353
GRAVITY Collaboration, Abuter, R., Aimar, N., Amorim, A., et al. 2022, As-
tronomy & Astrophysics, 657, L12
GRAVITY Collaboration, Abuter, R., Amorim, A., Anugu, N., et al. 2018, As-
tronomy & Astrophysics, 615, L15
Guo, F., Li, H., Daughton, W., & Liu, Y.-H. 2014, Physical Review Letters, 113,
155005
Guo, F., Liu, Y.-H., Daughton, W., & Li, H. 2015, The Astrophysical Journal,
806, 167
Jiang, H.-X., Mizuno, Y., Dihingia, I. K., et al. 2024, arXiv e-prints,
arXiv:2404.03237
Jiang, H.-X., Mizuno, Y., Fromm, C. M., & Nathanail, A. 2023, MNRAS, 522,
2307
Leung, P. K., Gammie, C. F., & Noble, S. C. 2011, The Astrophysical Journal,
737, 21
Li, X., Guo, F., Li, H., & Li, G. 2015, The Astrophysical Journal Letters, 811,
L24
Li, X., Guo, F., Liu, Y.-H., & Li, H. 2023, ApJ, 954, L37
Lin, X., Li, Y.-P., & Yuan, F. 2023, MNRAS, 520, 1271

Article number, page 12 of 12