fl205.aschwanden02 Posted: 25-Apr-95 Updated: 26-Nov-95, 08-Apr-96 Events specified: N/A
M.J.Aschwanden (UMd), B.R.Dennis (GSFC), R.A.Schwartz (GSFC), Kosugi (NAOJ)
An average time delay of 17 ms has been recently discovered between the energies of 25-50 keV and 50-100 keV for subsecond hard X-ray pulses, observed by BATSE on Compton GRO, using high-time resolution (64 ms) data from 640 solar flares (Aschwanden, Schwartz, and Alt, 1995, ApJ 447, July 10 issue). This energy-dependent time delay is believed to be a kinematic effect that is produced by the time difference of simultaneously accelerated electrons propagating from the coronal injection site to the chromospheric (thick-target) hard X-ray emission site. The time delay corresponds to an average propagation distance of 13,000 km, (or a loop height of 8,300 km if the particles are injected near the loop-top). We propose to compare these electron time-of-flight distances measured by CGRO with flare loop geometries seen in Yohkoh/SXT and HXT images. From the 160 flares simultaneously observed with Yohkoh/HXT and BATSE/CGRO (with high time-resolution in the burst trigger mode) we will mostly concentrate on flares that show footpoint and loop-top sources in HXT images. We anticipate that this study will provide information on the localization of acceleration sources in solar flares.
Data requirements: A subset of the HXT and SXT images of 160 flares that have been observed simultaneously with COMPTON GRO during 1991-1993
Update 08-Apr-96
Actually this is a final report that this project has now been completed to the extent outlined in the original proposal of May 95. The results are described in 3 new papers since the last update, Paper (3) has been submitted to the AIP proceedings, Paper (4) has been submitted to ApJ on January 26, and Paper (5) will be submitted to ApJ during March. The complete list of publications resulting from this proposal and the abstracts of the last 3 new papers are listed below. I will also send of the submitted and not yet submitted preprints to the Yohoh DUCs (Keith Strong and Kaz Shibata) as requested by the team bulletin board managers. All 5 papers are also available in form of PostScript files from the anonymous ftp account of UMd,
%ftp ftp.astro.umd.edu
Name: anonymous
Password: your@full.email.address
ftp>cd /pub/markus
ftp>mget *
ftp>quit
List of publications resulting from Yohoh/DUC proposal
fl205.aschwanden02
1) Aschwanden,M.J., Hudson,H.S., Kosugi,T., and Schwartz,R.A. 1996,
``Electron Time-of-Flight Measurements During the Masuda Flare
1992 Jan 13'',
Astrophys.J., 464, issue 1996 June 20, in press.
2) Aschwanden,M.J., and Schwartz,R.A. 1996, ``The Inversion of
Electron Time-of-Flight Distances from Hard X-Ray Delay
Measurements'',
Astrophys.J., 464, issue 1996 June 20, in press.
3) Aschwanden,M.J. 1996, in Proc. of Workshop on "High Energy Solar
Physics", NASA/GSFC, Greenbelt, Aug.16-18, 1995, AIP,
(eds. R.Ramaty, Mandzhavidze,N., and Hua,X.-M.),
``Hard X-Ray Timing'', in press.
4) Aschwanden,M.J., Wills,M.J., Hudson,H.S., Kosugi,T., and Schwartz,R.A.
1996, ``Electron Time-of-Flight Distances and Flare Loop Geometries
Compared from CGRO and Yohkoh Observations'',
Astrophys.J., submitted, 1996 Jan 24
5) Aschwanden,M.J., Kosugi,T., Hudson,H.S., Wills,M.J., and
Schwartz,R.A. 1996, ``The Scaling Law between Electron
Time-of-Flight Distances and Loop Lengths in Solar Flares'',
Astrophys. J., to be submitted, March 1996
ABSTRACT OF PAPER 3)
\title{Hard X-Ray Timing}
\author{Markus J. Aschwanden}
High-time resolution (64 ms) hard X-ray (HXR) data from {\sl BATSE/ CGRO} allow us to study the energy-dependent timing of acceleration, propagation, energy loss, and trapping of $\gapprox 20$ keV HXR-emitting electrons during solar flares. In many flares two different HXR flux components can be distinguished: (1) the fine structure of a HXR time profile (containing sequences of subsecond pulses) exhibits delays of $\approx 10-100$ ms for the low-energy electrons, while (2) the unmodulated HXR time profile (a smooth lower envelope to the fine structure) shows a delay of opposite sign and much larger magnitude, of typically 1-10 s. We model the timing of various acceleration mechanisms and find that the delay of the HXR pulses is dominated by time-of-flight differences rather than by acceleration time scales, while the timing of the unmodulated HXR flux is governed by trapping and collisional time scales.
{{\sl THE ASTROPHYSICAL JOURNAL} \hfill{Submitted, 1996 Jan 24}}
\begin{abstract} The distance between the coronal acceleration site and the photospheric hard X-ray (HXR) emission site can be determined from velocity-dependent electron time-of-flight (TOF) differences in the framework of the thick-target model. We determine these electron TOF distances $l$ with relative time delay measurements in the 30-300 keV energy range, using 16-channel data from {\sl BATSE/CGRO} for the 8 largest flares simultaneously observed with {\sl Yohkoh}. We filter the HXR fine structure from the smoothly-varying HXR flux with a Fourier filter, in order to separate competing time delays. In the {\sl Yohkoh/HXT} images we identify the corresponding flare loops that show $\ge 30$ keV HXR footpoint emission and project the electron TOF distances into the loop plane, assuming a semi-circular shape (with radius $r$). The flare loop radii vary in the range of $r=5,600-17,000$ km. In all 8 flares we find that the {\sl projected electron TOF distance $l'$} exceeds the loop half length $s=r(\pi /2)$, with a scale-invariant ratio of $l'/s=1.3\pm 0.2$. Projecting the electron TOF distances onto an open field line that extends to the cusp region above the flare loop, we find an average ratio of $h/r=1.6 \pm 0.4$ for the height $h$ of the acceleration site. This geometry is compatible with acceleration mechanisms operating in the cusp region, perhaps associated with magnetic reconnection processes above the flare loop. Alternatively, acceleration sites inside the flare loop cannot be ruled out (since $l'/s < 2$), but do not provide a natural explanation for the observed length ratio $l'/s$. Large-scale electric DC field acceleration mechanisms are found to be less suitable to explain the observed HXR timing and pulse durations. \end{abstract}
ABSTRACT OF PAPER 5)
{{\sl THE ASTROPHYSICAL JOURNAL} \hfill{Manuscript, 1996 March 10}}
\begin{abstract} >From the complete dataset of solar flares simultaneously observed with the {\sl Burst and Transient Source Experiment (BATSE)} onboard the {\sl Compton Gamma Ray Observatory (CGRO)} in the high-time resolution mode (64 ms) and the {\sl Hard X-Ray Telescope (HXT)} onboard {\sl Yohkoh} we were able to determine the electron time-of-flight (TOF) distance and the flare loop geometry in 42 events. The electron TOF distances $l'$ were determined from hard X-ray (HXR) time delays ($\approx 10-100$ ms) occurring in the 16-channel spectra (at $\approx 20-200$ keV), produced by the velocity difference of the HXR-producing electrons. The flare loops were mostly identified from double footpoint sources in $\gapprox 30$ keV HXT images, with radii ranging from $r=3000$ to $r=25,000$ km. We find a scaling law between the electron TOF distance $l'$ and the flare loop half length $s=r(\pi /2)$, having a mean ratio (and standard deviation) of $l'/s=1.41\pm 0.29$. In 5 flares we observe coronal $\gapprox 30$ keV HXR sources of the Masuda-type in the cusp region above the flare loop, and find that their heights are consistent with the electron TOF distance to the footpoints. These results provide strong evidence that particle acceleration in solar flares occurs in the cusp region above the flare loop and that the coronal HXR sources discovered by Masuda are a signature of the acceleration site, probably controlled by a magnetic reconnection process. \end{abstract}
Update 26-Nov-95
this is a progress report on the project fl205.schwanden02. Two studies on this project have been completed and were submitted to ApJ for publication:
1) Aschwanden,M.J., Hudson,H.S., Kosugi,T., and Schwartz,R.A. 1996,
``Electron Time-of-Flight Measurements During the Masuda Flare
1992 Jan 13'', submitted (1995 Aug 15)
2) Aschwanden,M.J., and Schwartz,R.A. 1996, ``The Inversion of
Electron Time-of-Flight Distances from Hard X-Ray Delay
Measurements'', {\sl Astrophys.J.}, submitted (1995 Oct 10)
We include the abstracts of these 2 papers below.
We will make the papers publicly available with preprints
and in electronic form on the WWW page.
A third study is in progress. Together with a summer student (Meredith Wills CfA) we analyzed 100 solar flare events simultaneously recorded with Yohkoh and CGRO. We produced for each flare 3 SXT maps, 3 EM maps, 3 Te maps, and 3 HXT maps, from the rise time, peak time, and decay time of the flare. We analyzed for each flare the timing of the hard X-ray fine structure from the MER (Medium Energy Resolution) 16-channel data from BATSE/CGRO with 16 ms and 64 ms time resolution. In many events we were able to fit a 1-parameter timing model that describes the relativistic electron time-of-flight differences and determined the electron flight distances. We are in the process now to do geometric projections of these flight distances onto the flare loop geometries of the Yohkoh SXT and HXT images. We plan to write up the results in near future. A tentative reference of this study is,
3) Aschwanden,M.J., Hudson,H.S., Kosugi,T., Schwartz,R.A., and Wills,M.,
1996, ``Electron time-of-Flight Distances and Flare Loop Geometries
Compared from CGRO and Yohkoh data'', ApJ, in preparation.
With best regards,
Markus Aschwanden
Electron Time-of-Flight Measurements During the Masuda Flare 1992 Jan 13
Markus J. Aschwanden Department of Astronomy, University of Maryland, College Park, MD 20742, USA internet: markus@astro.umd.edu
Hugh Hudson Institute for Astronomy, University of Hawaii, Honolulu, Hawaii 96822, USA
Kosugi Takeo National Astronomical Observatory, Mitaka, Tokyo 181
Richard A. Schwartz Hughes STX \& LASP, NASA/GSFC, Code 682, Greenbelt, MD 20771, USA
ABSTRACT
The solar flare of 1992 Jan 13, 1729 UT, has become famous for Masuda's discovery of a hard X-ray looptop source (Masuda 1994). Here we analyze energy-dependent time delays occurring in 30-120 keV hard X-ray (HXR) emission during this flare, observed by BATSE onboard CGRO with a time resolution of 64 ms. The purpose of this study is to reconstruct the kinematics of HXR-producing electrons from energy-dependent HXR delays and to relate the inferred time-of-flight distance to the spatial geometry of the flare loop as observed by SXT and HXT onboard Yohkoh. The findings are:
(1) The HXR flux can be decomposed into a sequence of pulses with ca. 2-3 s duration and into a smoothly-varying envelope that accounts for 90\% of the >30 keV flux. Cross-correlating the pulses between 5 different energy channels in the 30-120 keV range we find that the HXR pulses are delayed t_P = 40-220 ms) at the lower energies with respect to the higher energies. For the HXR envelopes we find much larger delays (-t_E=2.1-6.6 s) of opposite sign.
(2) We fit kinematic models that quantify electron acceleration and propagation times to the observed HXR timing, for small-scale and large-scale accelerating fields, in semi-circular and cusp-like flare loop geometries. We find that the acceleration site is most likely located in an altitude of h=44,000+6000 km in the cusp region above the SXR-emitting flare loop (h=12,500 km), and also significantly above Masuda's coronal HXR source (h=22,100 km). This finding offers an interpretation of Masuda's HXR source in terms of nonthermal bremsstrahlung by electrons partially confined in the cusp region either by magnetic mirroring or by wave turbulence in the reconnection outflow.
(3) The delay of the smoothly-varying HXR flux is found to be consistent with trapping time differences in terms of collisional deflection, based on measurements of the electron density (n_e ~ 2 10^11 cm^-3) from SXT emission measure maps.
This study provides the first quantitative localization of the electron acceleration site in a solar flare and demonstrates that energy-dependent HXR delays offer a sensitive diagnostics for electron acceleration, propagation, and trapping in solar flares.
The Inversion of Electron Time-of-Flight Distances from Hard X-Ray Time Delay Measurements
Markus J. Aschwanden Department of Astronomy, University of Maryland, College Park, MD 20742, USA internet: markus@astro.umd.edu
Richard A. Schwartz Hughes STX & LASP, NASA/GSFC, Code 682, Greenbelt, MD 20771, USA
ABSTRACT
The electron time-of-flight distance L between the acceleration site and the chromosphere can be measured during solar flares from energy-dependent hard X-ray (HXR) time delays t(e), based on the applicability of the thick-target model. The determination of the path length L represents an inversion problem because the time-dependent electron injection spectrum at the acceleration site, N(E,t,x=0), is retarded by the propagation time t^prop(E)=L/v(E) at the thick-target site, i.e. N(E,t,x=L)=N(E,t-t^prop[E],x=0), and has to be convolved with the bremsstrahlung cross-section s(e,E) and the instrumental detector response function R_i(e) to reproduce the observed HXR time profiles I(e_i,t) (in different detector channels i), from which the time delay differences t(e_i)-t(e_j) can be measured.
In this study we solve this inversion problem by numerical forward integration of time-dependent electron injection spectra N(E,t) with gaussian pulse shapes to obtain the convolved time-dependent HXR spectra I(e,t), using specific detector response functions from BATSE/CGRO and HXRBS/SMM. We find that the timing of HXR pulses can be accurately represented with the (monoenergetic) photon energy e_i that corresponds to the median of the channel count spectra C_i(e)= I(e) R_i(e). We compute numerical conversion factors q_E(e,gamma,E_0) that permit to convert the timing of photon energies e_i(t) into electron energies E_i(t)=q_E e_i(t), from which kinematic parameters can be fitted to determine the electron time-of-flight path length L. We test the inversion procedure with numeric simulations and demonstrate that the inversion is accurate within s_L/L ~ 1\% for noise-free data. This inversion procedure is applied to the Masuda flare (in the accompanying paper) to localize the electron acceleration region.