RHESSI and EUV observations as
diagnostic of accelerated electrons and
atmospheric response in solar flares
Marina Battaglia
[email protected]
University of Applied Sciences Northwestern Switzerland
RHESSI 15, Graz, 27.7.2016
2
Overview
1.  Some open questions in solar flare physics
2.  X-ray and EUV emission in the standard solar flare model
3.  Diagnostic of flaring processes at EUV and X-ray wavelengths
3.1 Differential emission measures
3.2 EUV line spectroscopy
4.  Selected recent examples
4.1 Location and drivers of chromospheric evaporation
4.2 Electron distribution functions
4.3 Acceleration region diagnostic
5.  Summary and Conclusions
RHESSI 15, Graz, 27.7.2016
3
1. Some open question in solar flare physics
Particle acceleration and
transport
Where and how are
electrons accelerated?
How much energy goes
into accelerated
electrons?
How are electrons
transported close to the
Sun and away from the
Sun?
RHESSI 15, Graz, 27.7.2016
Atmospheric response
What is the chromospheric
response to electron beam
heating?
What is the total flare energy?
Where and how is white light
emission generated?
How is chromospheric
evaporation triggered and
how does it evolve?
4
1. Some open question in solar flare physics
Particle acceleration and
transport
Where and how are
electrons accelerated?
How much energy goes
into accelerated
electrons?
How are electrons
transported close to the
Sun and away from the
Sun?
RHESSI 15, Graz, 27.7.2016
Atmospheric response
What is the chromospheric
response to electron beam
heating?
What is the total flare energy?
Where and how is white light
emission generated?
How is chromospheric
evaporation triggered and
how does it evolve?
5
2. X-ray and EUV emission in the standard solar flare model
1)  Release of magnetic energy
by magnetic reconnection
(non-thermal)
above-the-looptop
source
2)  Particle acceleration and
heating
3)  Accelerated electrons
produce HXR emission and
heat chromosphere
4)  “Chromospheric
evaporation”, hot plasma fills
coronal loop
(thermal)
loop-top source
hot flare loop
HXR footpoint
UV and visible light
chromospheric and TR
emission
RHESSI 15, Graz, 27.7.2016
6
3. Diagnostic of flaring processes at X-ray and EUV wavelengths
RHESSI imaging spectroscopy:
Krucker & Battaglia 2014
• 
• 
• 
• 
EUV:
•  Different wavelengths
for different
temperatures
•  Multi-thermal
diagnostic of whole
atmosphere
•  Flows (evaporation,
coronal rain, ….)
RHESSI 15, Graz, 27.7.2016
Location of energy deposition
Acceleration region
Hot flare loop
Electron flux distributions
The Astrophysical Journal Letters, 755:L16 (6pp), 2012 August 10
AIA images @ 4 WL
Milligan
Milligan et al. 2012
7
3.1 EUV diagnostic of plasma heating and energetic electrons
using differential emission measures
Differential emission measure along line of sight:
AIA temperature
response
Fletcher et al. 2013
The Astrophysical Journal, 771:104 (13pp), 2013 July 10
The Astrophysical Journal, 771:104 (13pp), 2013 July 10
Fletcher et al.
AIA 131 18:03:09
200
22
10
21
10
20
10
6.0
18:05
K−1]
1023
0.8
RHESSI
6.5
log 10 T [K]
7.0
7.5
15, Graz, 27.7.2016
10
21
10
20
2
1023
1.0
3: southern loops
150
22
10
150
2
1
100
100
21
10
50
10
6.0
6.5
log 10 T [K]
7.0
7.5
5.5
1023
0.8
6.0
log
6.5
T [K]
7.0
7.5
-600
-550
-500
-450
-600
10
200
0.9
50
3
20
5.5
K−1]
5.5
10
22
10
0.9
Y (arcsec)
1 northern ribbons
10
200
23
DEM, ξ(T) [cm−5 K−1]
0.8
DEM, ξ(T) [cm−5 K−1]
DEM, ξ(T) [cm−5 K−1]
10
23
K−1]
23
AIA 131 18:06:09
200
8
Inferring DEMs from observations
Challenge:
Observed
intensity
Error of observed
intensity
DEM
Temperature
response
•  Several methods exist
-  Regularized inversion (Hannah & Kontar 2012)
-  Forward fitting model DEM
(e.g. Aschwanden & Boerner 2011, Ryan et al. 2014)
-  Monte-Carlo Markov Chain
(e.g. Testa et a. 2012)
-  ……
DEM-maps of an erupting CME seen in
SDO/AIA (Hannah & Kontar 2013)
•  Aschwanden 2015: Benchmark test of 11 methods for
active regions using response functions for SDO/AIA, SDO/
EVE, RHESSI, and GOES
RHESSI 15, Graz, 27.7.2016
9
)⟩ from thin kappa (red) and ξκ (T ) (blue). The grey areas give the confidence
DEMs from simultaneous X-ray and EUV observations
AIA is only sensitive to temperatures ~< 16 MK
RHESSI is sensitive to temperatures ~> 8 MK
Different approaches for simultaneous data
exploitation exist
& 6Christe
2014
mperature Inglis
response for
AIA wavelength-channels
physical Journal, 789:116 (12pp), 2014 July 10
Caspi et al. 2014
Simultaneous fits of full Sun EVE and
RHESSI spectra with multiple Gaussian
DEMs
Inglis & Christe
(solid lines, left axis) and
ure response at 7 keV, 15 keV, and 24 keV (dashed lines, right axis), for an
of 1049 cm−3 .
Gaussian and uniform DEM model (Epstein function)
hematic of the joint differential emission measure fitting procedure. AIA flux data from the six optically thin EUV channels is compared to the reconstructed
model DEM function folded through the AIA temperature response function. Simultaneously, the RHESSI count spectrum is compared to a model count
btained by calculating the expected photon spectrum that results from the model DEM function, and utilizing the RHESSI response function to convert
on spectrum to a count
spectrum. 15, Graz, 27.7.2016
RHESSI
sion of this figure is available in the online journal.)
10
–7–
– 17
Simultaneous forward fits to RHESSI and AIA data
ribution expressed via differential emission measure
Motorina & Kontar 2015:
!
flux spectrum ⟨nV F (E)⟩ =
n(r)F (E, r)dV in the emitting
•  Treat RHESSI and AIA data as one dataset
V
servable that can be directly inferred from the X-ray spectrum
the density or emitting volume. It can be related to the DEM
electron distribution at temperature T (Brown & Emslie 1988;
):
•  Generate one temperature
" ∞
ξ(T )
response
expmatrix
(−E/kB T )dT [electrons keV−1 s−1 cm−2 ].
3/2
(kB T )
0
(3)
– 10mean
–
ifferential emission measure, one can compute the
electron
Forward-fit
model DEM
ting•  volume.
We introduce
a DEM ξ(T ) of the shape:
$
#
T
κ
ξ(T ) ∝ T −(κ+0.5) exp − (κ − 1.5)
T
creasing T since ξ(T ) ∝ T −(κ+0.5)
This DEM has a
to RHESSI data (light-blue dashed); DEM fro
(4) physical meaning:
two
and green dashed
κ (T )s (blue dashed
It ξrepresents
a line
kappafits.
AIA loci-curves are indicated near the t
distribution!
for T ≫ Tκ and it falls off
Fig. 6.— Left: Comparison of DEMs from diff
DEM (with confidence range) from AIA data,
o exp(−T
). Graz,
The27.7.2016
DEM hasFig.a1.—single
maximum, dξ(T )/dT = 0
fit
κ /T 15,
Differential emission measure, ξ(T ), for plasma with emission measure⟨nV
EM F=(E)⟩ obtained from the simultaneous
11
RHESSI
1049 cm−3 and two sets of parameters: κ = 6.5, T
= 20 MK (solid line); κ = 8.5,
Line spectroscopy
The Astrophysical Journal, 813:113 (8pp), 2015 November 10
Battaglia et al.
-87 km/s
Fe
XXI
Si II
Fe II
H2
Si II
Fe II
CI
Fe II Fe II
Figure 6. A selection of spectral regions from IRIS (from left to right): the two C II lines; Fe XII which is generally faint and not visible here (rest wavelength indicated
by dashed line); Fe XXI which is well visible as inverse C-shape; several O IV lines (1399.8, 1401.2, 1404.8 Å) and the strong Si IV line at 1402.8 Å; part of the nearUV (NUV) window with the Mg k and h lines (2796.4, 2803.5 Å). The colors are inverted and the images are contrast-enhanced. Downflows in Si IV, Mg II and C II
are seen just north and south of the maximum blueshifts of Fe XXI (Y ≈ 272″ and 263″).
Example of an IRIS flare spectrum (Battaglia et al. 2015)
265–270 arcsec) does not show significant Fe XXI velocities,
only a small redshift. Northward of Y ≈ 270 and southward of
Y ≈ 265 Fe XXI shows a clear blueshift on the order of 0.5 Å,
corresponding to ≈110 km s−1. The other spectral lines do not
show clear shifts at these locations, rather some line-broadening. However, Si IV, Mg II, and C II show prominent downflows just north and south of these locations, which correspond
to leading edges of flare ribbons. It is possible that Fe XXI has
not formed yet there, which would indicate densities of the
order of 1010 cm−3 according to the equilibration time (see
Section 4.3). Or, Fe XXI could simply be too faint to be visible
at these locations. The flare was also observed by Hinode/EIS
and twice (∼17:46:28 UT and ∼17:48:52 UT) the EIS slit was
at the same location as the IRIS slit. The line selection was
rather sparse and does not provide full temperature coverage.
As expected, the upflow velocities in EIS Fe XXIII (12.5 MK)
are higher than those
seen in Fe
XXI Graz,
in IRIS. However,
analysis
RHESSI
15,
27.7.2016
of the other available data suggests redshifts in Fe XVI (2.8 MK)
HXR source. On the other hand, the EIS data could be
interpreted as showing explosive evaporation even at times
when no HXR emission was observed.
Future combined high-spatial resolution flare observations,
including lower temperature lines such as O IV and Si IV and in
combination with Hinode/EIS, should help shed some light on
the matter.
Different lines are formed under different conditions (temperature, density)
à diagnostic of photosphere up to corona
à tracing of flows (chromospheric
evaporation,
coronal rain, filament
This work was supported by the Swiss
National Science
Foundation (200021-140308) and through NASA contract
NAS 5-98033 for RHESSI. L.K. was supported by a Marie
eruptions, ….)
Curie Fellowship. D.G. acknowledges support by the European
Community’s Seventh Framework Programme (FP7/20072013) under grant agreement No. 606862 (F-CHROMA).
We would like to acknowledge the International Space
Science Institute (ISSI) for facilitating the collaboration on
this work.
12
4 Selected recent examples
4.1: Location and drivers of chromospheric evaporation
Milligan et al. 2009: VelocitiesNo.of
evaporating
plasma observed with Hinode/EIS
2, 2009
VELOCITY CHARACTERISTICS OF EVAPORATED PLASMA USING HINODE/EIS
Figure 5. Plasma velocity from a flare footpoint at ∼14:14:51 UT as a function of temperature for each of the emission lines used in this study
represent a weighted least-squares fit to the data points from 0.5 to 1.5 MK and 2.0 to 16 MK.
omitted from any further calculations. Taking the mean and
standard deviation of each fit parameter for each component
across the five individual detectors currently provides the best
ree rows: intensity maps in each of the 15 lines used in this study, ranging from 0.05 to 16 MK.estimate
Two footpoints
are clearly
visible, with
of the
parameter
andtheitssoutheastern
uncertainties. The isothermal
ighter of the two. Overlaid are the 20–25 keV emission contours (at 60% and 80% of the maximum) as observed by RHESSI from 14:14:28–14:15:00
fits
yielded
a
temperature
(T)
and
emission
measure (EM) of
marked with an “×” within the HXR contour was the focus of a more detailed spectral analysis. Bottom three rows: except for the Fe xxiii and Fe
xxiv
46 represent
sponding velocity maps for each of the above intensity maps. Red pixels denote material moving
observer,
17 away
± 1from
MKtheand
7 ±while
2 ×blue
10pixels
cm−3 , respectively. The low−1 . The Fe xxiii and Fe xxiv
g toward the observer. The same RHESSI 20–25 keV contours are overlaid. All velocity mapsenergy
are scaled
to ±150
cutoff
to km
thes assumed
power-law electron spectrum was
s formed over the enhanced blue wing of each line with the blue color scaled with the flux.
found to be Ec ! 13 ± 2 keV with a spectral index, δ, of
7.6 ± 0.7.
s not unique; many of the neighboring pixels within
of the lines listed inThe
Table
1. Assuming
linear relationship
combination
of ahigh-resolution
images and spectra from
show similar profiles with a dominant stationary
between velocity RHESSI
(v up and vallows
temperature
(T) of
of the
the flux of nonthermal
down ) and
a
measurement
RHESSI
n the two hottest
lines. 15, Graz, 27.7.2016
form v = A + BT , where A and B are constants, a leastelectrons responsible for driving chromospheric evaporation to
hows the derived line-of-sight velocities from this
squares fit was applied to both the blueshifted and redshifted
Temporal and spatial correlation of HXR emission with upflows and downflows
à explosive evaporation driven by non-thermal electron beam
13
The Astrophysical Journal Letters, 807:L22 (5pp), 2015 July 10
Graham & Cauzzi
Graham & Cauzzi 2015: high resolution observationsMultiple
of evaporating flare sources
evaporation and condensation with IRIS
100
40
Fe XXI Velocity
0
30
km s-1
-100
km s-1
- very high evaporation velocities (up to
300 km/s)
-  time-lag between condensations and
evaporation
-  Same behavior for each pixel à
elementary flare kernels
Mg II Velocity (30% bisector)
20
-200
10
MgII
FeXXI
-300
-400
0
0
240
480
720
Time (s)
960
0
30
60
90
120
150
180
Figure 3. Mg II intensity spacetime map (inverted B/W color table) with
Time (s)
overlays of Mg II downflows at the 30% bisector level at 15 and 30 km s 1 (red
contours).
The
Fe
XXI upflow velocities above 270, 200, and 100 km s 1 are
Figure 5. Superposed epoch analysis of Fe xxi and Mg ii flows for every slit pixel in Figure 3 (negative velocities showing rising
indicated
with yellow,
blue,
lightdetail).
blue crosses, respectively.
The greyscale darkens with increasing occurrence within a given velocity
interval
(seedark
text
forandfull
Battaglia et al. 2015: spatial and temporal evolution of chromospheric evaporation
the flare, and we find that the coronal line is always
that the characteristics of the energy release a
with IRIS and RHESSI à sustainedofentirely
evaporation
impulsive
phase
due
blue-shifted. This is after
consistent with
other recent spatial
remarkably
uniform,
each
time
occurring
in a
andthermal
Fe
detectors was
offset between
the Mgto
by aligning or
the have
spectrographʼs
fiducial marks.
IRIS results and seems to imply that IRIS observations corrected
environment,
little influence
over the sub
shows a sudden enhancement due to the
The
Mg intensity
conduction
fully resolve single flaring kernels. Still, upflows persist
plasma
evolution.
flare, and the development of the ribbon clearly appears as a
II
XXI
II
RHESSI 15, Graz, 27.7.2016
in any given position for a remarkable length of time,
and should be readily visible even at the coarser resolu2. Leftor
panels:
of theFig.
Fe XXI 3),
1354.1
Å line (green)
for slit
position
tion Figure
of CDS
EISfits(cf.
contrary
the
commonly
y = 122″. The observed spectrum is shown by the stepped black line, with the
reported
a dominant
component
total fit occurrence
in black on top; of
cooler
components arestationary
shown by a dotted
line. The
in such
observations.
A shown.
possible
might
Right explanation
panels: Mg II subordinate
line renon-thermal
FWHM, vnt, is also
bisector
positions
at 10% intensity
increments
marked byof
black
squares,
side with
in
the
very
simple
linear
progression
the
newly
and the rest wavelength (2791.56 Å) by the vertical dashed line.
activated kernels analysed. However, we cannot state if
these observed characteristics are representative of the
(Canfield
et of
al. 1990;
et al. 1995): almost all the flaring
entire
flare or
otherDing
events.
pixels initially display a red asymmetry of
the line, which
2) disappears
All of therapidly;
flaringthepixels
(at ∼ 0.3′′ resolution)
also
peak of emission is only slightly
display
sudden
Mgbisector
ii condensation
redshifted;
andand
the strong
shift of the
with respect downflows,
to the rest
withwavelength
values in appears
agreement
with from
earlier
from the
both
to increase
the results
center toward
visible
andperhaps
EUV signifying
observations.
The in
chromospheric
conwing,
a gradient
the condensation
velocityin(Cauzzi
et al. 1996).
deriveinthe
densation
each flaring
kernelTostops
∼ condensation
50 − 60 s, at
weof
interpret
the bisector
at the 30% intensity
leastvelocity,
a factor
two faster
thanposition
any previously
reported
level
in
terms
of
Doppler
shifts.
While
this
might
underestimate
value, but consistent with predictions of 1-D
hydrodythe actual condensation velocity (e.g., Canfield et al. 1990), it
namical
simulations of flares affecting ‘undisturbed’ chrois the best compromise between assessing the flows while
mosphere.
avoiding contamination from possible blends in the far wings,
3) although
Surprisingly,
pixels show
these areonly
mostlya offew
photospheric
origin.a(T.simultanePereira,
ous onset
coronal andWe
chromospheric
flows,shifts
while
private of
communication)
confirm that similar
are for
when
the coronal
other triplet
lines, albeit lags
with more
mostderived
of them
theusing
initial
evaporation
behind
We conclude by remarking that the large nu
diagonal
strip across flaring
the diagram.
The observed,
chromospheric
downindependent
pixels
and
the c
flows are co-spatial and co-temporal with each of the new
temporal
coverage
of
their
dynamical
evolution
intensity enhancements, as shown by the red contours at the 15
allowThese
us to
derive
common flows
character
values
of condensation
are
and cadence
30 km s 1 levels.
whatwith
we can
define
as ‘prototypical’
with
consistent
results
from many
earlier ground-flares,
and spaceby the
actual resolution
of o
basedextension
studies, but limited
Figure 3 offers
unprecedented
detail on their
spatial
evolution.
i.e.and≤temporal
0.5′′ . In
principle our results can be imm
The
corresponding
velocities,
as derived from
compared
withevaporation
the output
of numerical
simula
fits, areFor
overlaid
as coloredwith
crosses
on
the centroid
the Fe XXI
single offlaring
loops.
example,
respec
the same figure. For about 70% of the pixels, we find flows
third point above, one could attempt to derive
reaching nearly 300 km s 1 (yellow crosses identify flows
the actual coronal emission during the early p
above 270 km s 1), these encompass the fastest, early Fe XXI
flare pertaining
chromospheric
heating,
as predictedstrip
with
emission,
to a very
thin spatio-temporal
collisional
thick-target
(Allred
et
al.
2005)
or
approaching our 0″. 5 resolution, and lasting only one to two co
(Longcope
2014)
temporal
steps. Such
valuesmodels.
of evaporation flows are among the
strongest ever reported (300–400 km s−1; e.g., Antonucci et
al.1982) and the largest documented yet from IRIS. Within the
This theresearch
has
received
funding
from
same pixels,
flows decay
to 200
blue region)
km s 1 (dark
ropean
Community’s
Seventh
Framework
and below rather uniformly in time. For pixels northward ofPro
124″,(FP7/2007-2013)
we detect only slowerunder
flows, grant
up to agreement
150 km s 1. no. 60
From
Figure 3, weIRIS
see that
flaring small
pixels display
clearmi
CHROMA).
is aallNASA
explorer
14 yet
signatures
of both
evaporation
andby
condensation,
the onset
veloped
and
operated
LMSAL
with
missio
of opposite
flows is co-temporal
(within
one Research
to two time steps)
tions executed
at NASA
Ames
center
!
The mean electron flux spectrum ⟨nV F (E)⟩ = VV n(r)F (E, r)dV in the emitting
!
volume V is the only observable that can
be directly
inferred from the X-ray spectrum
n(r)E
23/2
dV
dT ,
exp (−E/kB T (r))
1/2 (k T (r))3/2
(πm
)
dT
e functions
B
T
without
on the
density
or emitting
volume.
It can be related to the DEM
4.2:assumptions
Electron
distribution
(3)
=
n(r)
ξ(T ) via the Maxwellian electron distribution at temperature T (Brown & Emslie 1988;
DEM
is directly
mean
flux spectrum <nVF>
where nrelated
dV /dT =toξ (T
) is the electron
DEM and thus
Battaglia &
Kontar
2013):
" ∞
23/2 E
ξ(T ) 23/2 E ! ∞ ξ (T )
⟨nV F (E)⟩ =
exp (−E/kB T )dT [electrons keV−1 s−1 cm−2 ]. (3)
1/2
3/2
⟨nV
exp (−E/kB T ) dT . (4)
(πme )
(kFB⟩T=
)
0
1/2
3/2
2
(πme )
(kB T )
0
Therefore, knowing the differential emission measure,
– 8 –one can compute the mean electron
Therefore, knowing the DEM, one can compute the electron flux
flux spectrum in the emitting volume. We introduce a DEM ξ(T ) of the shape:
−2
Battaglia et spectrum
al. 2015:
function
fits
in Electron
the emittingdistribution
volume (electrons
keV−1from
s−1 cmsimultaneous
).
(κ−0.5)
$
#
(κ−0.5)
i.e. the proportionality
constant
in Equation
isisEM
× Tκ
(κ
1.5)
Although
Equation
(4), which
an equivalent
of−the
Laplace/Γ(κ − 0.5).
T(4)
κ
−(κ+0.5)
(κ
−
1.5)
(4)
ξ(T
)
∝
T
exp
−
Model DEMtransform of a function f (t)
Tκ, EMκ, κ
T
2
For a uniform plasma density n, the emission measure can be written as EM = n V , where
−(κ+0.5)
This DEM drops with increasing T since ξ(T ) ∝!T ∞
for T ≫ Tκ and it falls off
V is the
emitting volume.
InsertingFEquation
(7)
into Equation
represents
the kappa-distribution:
(s) =
exp(−st)f
(t) dt, (3) one finds:(5)
quickly for T < Tκ due to exp(−Tκ /T ). The DEM 0has a single maximum, dξ(T )/dT = 0
Γ(κ + 1)
E/k T
23/2
. (8)
(πme )1/2 (kB Tκ )1/2 (κ − 1.5)1.5 Γ(κ − 1/2) (1 + E/kB Tκ (κ − 1.5))κ+1
numerical
integration
could bethe
rather
challenging due
to in
the
ξ(T ) can be solved analytically.
Moreover,
it represents
kappa-distribution
given
!
exponential
kerneldistribution
(e.g., Prato etfunction
al. 2006).
of kappa-distribution:
If we Advantages
introduce
an isotropic
electron
⟨f Following
(v)⟩, n = Rossberg
⟨f (v)⟩d3v, so that
Equations
(1) and (2).
This can be
shown as
follows:
B κ
at T⟨nV
Tκ (κ −
0.5) (see aFigure
1). It has the
advantageover
thattemperature,
the integral
over
max =
F (E)⟩
= 1.5)/(κ
n2 V is+formally
straightforward
integration
the
(2008), we
rewrite the
Laplace transform (Equation (5)) via
• 
Single
analytic
function
3
3
2 to describe whole spectrum
F (E)dE
= v⟨f (v)⟩d
v, and
dis vdefined
= 4πv
can write:
the
convolution
integral,
which
will
numerical
The
total emission
measure
EM
asdv
theweintegral
of ξ(Tallow
) overefficient
all temperatures
in
•  No cutoffcomputations
needed of ⟨nV F ⟩ via ξ (T ) and vice versa. Using the
3/2
the plasma
me n
2by
+exp(y)
1)
me v 2let
/(2k
TBian
•  Supported
stochastic
acceleration
(e.g.
et al 2014)
κ)
change
of variablesΓ(κ
s=
and t = models
exp(−x),
usBrewrite
⟨f (v)⟩ =
"2 ∞
"
1/2∞
3/2 Γ(κ − 1/2) [1 + m v 2 /(2k T (κ − 1.5))]κ+1
4πv
(πm
−the
1.5)
Tform:
e kB Tκ ) (5)(κ
e
B κ
Equation
in−(κ+0.5)
following
κ
−3
EM =
$
#
(κ − 1.5) dT [cm ].
(5)
T
0
0
Applied to
RHESSI data
by e.g.! Kasparova
& Karlicky 2009, Oka
∞
where E = me v 2 /2 was used. Or, simplifying:
!
∞ x−1
RHESSI
15,be
Graz,
27.7.2016
F (ey )function
=
K(y ≡
− x)h(x)
(6)
This integral
can
solved
using the gamma
Γ(x)
y dx,
exp(−y)dy resulting
0
ξ(T )dT ∝
"
T
exp −
#3/2
−∞
"
#−κ−1
(9)
et. al. 2013/2015
15
17 –
<nVF>
RHESSI fit range
ig. 3.— AIA images in 4 wavelength-channels at the time for which the RHESSI spectrum
ξĸ(T) on RHESSI
data, only
Comparison of mean
Fig. 6.— Left: Comparison of DEMs from different meth
electron flux spectrum ξĸ(T)on RHESSI
AIA DEM
data from simulta
to RHESSI
datafit(light-blue and
dashed);
from
different
combined
methods
two ξκ (T )s (blue dashed line(low
and and
greenhigh
dashed line). Th
T component)
fits. AIA loci-curves are indicated near the top of the
AIA DEM from
regularized
DEM (with confidence range)
from AIA data, only, foun
inversion
⟨nV F (E)⟩ obtained from the simultaneous fit of AIA an
as fitted. 30%, 50%, and 70% contours from a RHESSI CLEAN image at 6-12 keV are
iven in red.
RHESSI 15, Graz, 27.7.2016
16
black line and dashed light-blue lines give ⟨nV F (E)⟩ from
fit-parameters. In addition to the electron number density we can a
energy density from Equation (8) as (see Appendix for full derivatio
Comparison of total energy
Total energy density
3
Uκ = kB nTκ ,
2
Total energy: UκV where V ≈1.5x1027 cm3
RHESSIκDEM
energy by multiplying Uκ with the volume (compare Table 1). Fitt
th ξκ (T ) reduces the total energy by a factor of ∼ 2.9 compared wi
Fig. 6.— Left: Comparison of DEMs from different methods: DEM from fit with one ξκ (T )
à Without
to RHESSI
data (light-blue
DEM
simultaneous
fit of low-energy
RHESSI
AIA 5
withl
aneously fitting RHESSI and
AIA
datadashed);
ξκ (T
) from
gives
a factor
ofand∼
AIA & RHESSI
constraint,
total
two ξκ (T )s (blue dashed line and green dashed line).
The red line gives
the sum of the two
RHESSI thin-κ
energies
derived
from the
fits. AIA loci-curves are indicated near the top of
the plot. The
grey area indicates
ppa. This shows that,
while
fitting
RHESSI
spectra
with
ξ
(T
)bealre
κ
x3
RHESSI
data
could
DEM (with
x5 confidence range) from AIA data, only, found by regularized inversion. Right:
over-estimated
byThe dotted
⟨nV F (E)⟩ obtained from the simultaneous fit of AIA
and RHESSI data (red).
nt reduction of the total electron
number,
the
lowest
electron
energi
factor
~5
black line and dashed light-blue lines give ⟨nV F (E)⟩ from thin kappa and from a single ξ (T )
κ
fitted to RHESSI data.
covered by simultaneously fitting RHESSI and AIA data, as can be
RHESSI 15, Graz, 27.7.2016
17
4.3 Acceleration region diagnostic
Masuda et al. 1994
Krucker & Battaglia 2014: Bulk acceleration in
above-the-looptop source
Krucker & Battaglia 2014
quiet corona
•  RHESSI imaging spectroscopy to infer density of accelerated electrons: nnt~109
cm-3
•  SDO/AIA differential emission measure analysis to determine ambient density n0
à  ratio nnt/n0 is close to 1
pre-flare
flare
Interpretation: Entire plasma is accelerated
(non-thermal) in bulk energization process
Above the loop-top-source is acceleration
region
RHESSI 15, Graz, 27.7.2016
ambient
f
v
accelerated
v
18
C4.5 flare
flare and initial ascent
y and EUV loop-top source
Liu et al. 2013: Particle acceleration in magnetic reconnection outflow regions
ray and EUV loop-top descent
slow to fast rise of the flux rope CME
0 km s−1 ) of upward ejections
8 km s−1 ) of downward loop shrinkages
to −23 km s−1 ) of overall
p-top descent
ase, hard X-ray burst
all X-ray and EUV loop-top descent
cond ascent
downward, low-altitude fast contractions
titude fast loop contractions
op shrinkages
18 km s−1 ) of downward fast contractions
Krucker & Battaglia 2014
Same event:
19 July 2012
pre-impulsive phase
The Astrophysical Journal, 767:168 (18pp), 2013 April 20
Liu, (18pp),
Chen,2013
& Petrosian
The Astrophysical Journal, 767:168
April 20
rather than the reconnection site
rview in Section 2, we present
and EUV emission in Section 3.
utflows in forms of plasmoids
ion 4. In Section 5, we analyze
rgy- and temperature-dependent
onal X-ray sources. We conclude
ndices A and B on supplementary
ns.
F OBSERVATIONS
(a)
Reconnection
at X−Point
Outflows:
Plasmoids
Loop Contractions
Main HXR peak
(b)
Coronal X−ray Sources
Acceleration, Heating:
Turbulence
Shock
Collasping Trap
(c)
Pre−impulsive Pha
LT Source Ascent
Top of AIA FOV
X
X−Point X
Loop−top
(LT)
X−ray
Source
Cooler but Denser Loops
s an M7.7 flare that occurred at
NOAA active region (AR) 11520
Thick−Target Footpoints
s well observed by RHESSI and
Figure 12. (a) Schematic of the proposed flare model in which particle acceleration and plasma heating
Figure 10. Energy dependence of the double coronal X-ray sources likely located in the oppositely directed
outflows.
(a) and development
(b) RHESSIof contours
thereconnection
reconnection site.
(b)–(e) Temporal
reconnection and the up–down–up three-stage mo
ected
by
Fermi
and
its
impulsive
19 which each pair
RHESSI
15, Graz,
27.7.2016
overlaid on AIA
131 Å images.
The higher
energy sources (blue dotted contours at 43% and 95% of the imageportions
maximum)
are
closer
toward
one
another
the lower site, above and below
of reconnected magnetic field lines nearthan
the reconnection
relaxations.
that the higher
speeds (e.g.,
of the
green loop) from (c) to (d) are associated w
energy
sources (green
(c) Table
Contoured
along Note
the fiducial
Cut 0contraction
of the centroids
of the
two
ay
Telescope
oncontours).
Hinode.
1 6–8 keV image at the time of (b). (d) and (e) Projected heights
Sneak preview: Quantitative analysis of electron acceleration in the outflow
region during the pre-impulsive phase
Fitting a kappa-distribution to RHESSI and AIA data simultaneously we find a
hardening spectrum at ~ constant temperature à acceleration
The Astrophysical Journal, 767:168 (18pp), 2013 April 20
Liu, Chen, & Petrosian
Table 1
Event Time Line (2012 July 18–19)
Peak of the earlier C4.5 flare
Onset of the M7.7 flare and initial ascent
of the overall X-ray and EUV loop-top source
05:02–05:07
Onset of overall X-ray and EUV loop-top descent
and transition from slow to fast rise of the flux rope CME
Max. velocity (1050 km s−1 ) of upward ejections
Max. velocity (−58 km s−1 ) of downward loop shrinkages
Max. velocity (−7 to −23 km s−1 ) of overall
X-ray and EUV loop-top descent
Flare impulsive phase, hard X-ray burst
Min. height of overall X-ray and EUV loop-top descent
and onset of the second ascent
Upward ejections; downward, low-altitude fast contractions
Downward, high-altitude fast loop contractions
Downward slow loop shrinkages
Max. velocity (−918 km s−1 ) of downward fast contractions
05:15
05:16
05:15–05:20
05:16–05:43
05:21–05:31
04:17–05:20
05:20–16:00
04:17–16:00
06:52
<nVF> [cm-2keV-1s-1]
22:18, 07/18
04:17, 07/19
place in the outflow regions, rather than the reconnection site
itself.
After an observational overview in Section 2, we present
motions of the overall X-ray and EUV emission in Section 3.
We examine bi-directional outflows in forms of plasmoids
and contracting loops in Section 4. In Section 5, we analyze
the spatial distribution of energy- and temperature-dependent
emission, including double coronal X-ray sources. We conclude
in Section 6, followed by Appendices A and B on supplementary
RHESSI 15, Graz, 27.7.2016 AIA and STEREO observations.
Liu et al. 2013
20
Summary and Conclusions
•  Combining RHESSI observations with EUV measurements from
instruments such as SDO/AIA, SDO/EVE, IRIS, ……. provides
unprecedented diagnostic of accelerated electrons and
chromospheric response to flare energy input
•  Many different methods for combining those rich datasets exist
à Follow the presentations in the respective working groups and
talk to the relevant people
RHESSI 15, Graz, 27.7.2016
21