Electron acceleration in solar flares

For the period September 1978 to December 1982 we have identified 55 solar flare particle events for which our instruments on board the ISEE-3 (ICE) spacecraft detected electrons above 10 MeV. Combining our data with those from the ULEWAT spectrometer (MPI Garching and University of Maryland) electron spectra in the range from 0.1 to 100 MeV were obtained. The observed spectral shapes can be divided into two classes. The spectra of the one class can be fit by a single power law in rigidity over the entire observed range. The spectra of the other class deviate from a power law, instead exhibiting a steepening at low rigidities and a flattening at high rigidities. Events with power-law spectra are associated with impulsive (<1 hr duration) soft X-ray emission, whereas events with hardening spectra are associated with long-duration (<1 hr) soft X-ray emission. The characteristics of long-duration events are consistent with diffusive shock acceleration taking place high in the corona. Electron spectra of short-duration flares are well reproduced by the distribution functions derived from a model assuming simultaneous second-order Fermi acceleration and Coulomb losses operating in closed flare loops.     

Energy release in solar flares

We summarize key problems in our understanding of energy release in solar flares, as addressed by participants in a recent workshop. These problems fall into three broad areas: (i) Transport and thermalization of energy, (ii) acceleration of particles, and (iii) origin and effects of mass motions. We then describe how suitably coordinated collaborative observing sequences during the forthcoming Solar Maximum Year are potentially capable of resolving some of these issues.     

Electron acceleration in solar flares

We present observations of an intense solar flare hard X-ray burst on 1980 June 27, made with a balloon-borne array of liquid nitrogen-cooled germanium detectors which provided unprecedented spectral resolution (≲1 keV FWHM). The hard X-ray spectra throughout the impulsive phase burst fitted well to a double power-law form, and emission from an isothermal 10⁸–10⁹K plasma can be specifically excluded. The temporal variations of the spectrum indicate that the hard X-ray burst is made up of two superposed components: individual spikes lasting ∼3–15 s, whch have a hard spectrum and a break energy of 30–65 keV; and a slowly varying component characterized by a soft spectrum with a constant low-energy slope and a break energy which increases from 25 keV to ≳100 keV through the event. The double power-law shape indicates that acceleration by DC electric fields parallel to the magnetic field, similar to that occurring in the Earth's auroral zone, may be the source of the energetic electrons which produce the hard X-ray emission. The total potential drop required for flares is typically ∼10² kV compared to ∼10 kV for auroral substorms.

     

Electron acceleration in solar flares

We present observations of an intense solar flare hard X-ray burst on 1980 June 27, made with a balloon-borne array of liquid nitrogen-cooled germanium detectors which provided unprecedented spectral resolution (1 keV FWHM). The hard X-ray spectra throughout the impulsive phase burst fitted well to a double power-law form, and emission from an isothermal 10⁸–10⁹K plasma can be specifically excluded. The temporal variations of the spectrum indicate that the hard X-ray burst is made up of two superposed components: individual spikes lasting 3–15 s, whch have a hard spectrum and a break energy of 30–65 keV; and a slowly varying component characterized by a soft spectrum with a constant low-energy slope and a break energy which increases from 25 keV to 100 keV through the event. The double power-law shape indicates that acceleration by DC electric fields parallel to the magnetic field, similar to that occurring in the Earth's auroral zone, may be the source of the energetic electrons which produce the hard X-ray emission. The total potential drop required for flares is typically 10² kV compared to 10 kV for auroral substorms.     

Particle acceleration in solar flares

Particle acceleration during solar flares is a complex process where the main actors (Direct (D.C.) or turbulent electric fields) are hidden from us. It is easy to construct a successful particle accelertion model if we are allowed to impose on the flaring region arbitrary conditions (e.g., strength and scale length of the D.C. or turbulent electric fields), but then we have not solved the acceleration problem; we have simply re-defined it. We outline in this review three recent observations which indicate that the following physical processes may happen during solar flares: (1) Release of energy in a large number of microflares; (2) short time-scales; (3) small length scales; and (4) coherent radiation and acceleration sources. We propose that these new findings force us to reformulate the acceleration process inside a flaring active region assuming that a large number of reconnection sites will burst almost simultaneously. All the well-known acceleration mechanisms (electric fields, turbulent fields, shock waves, etc.) reviewed briefly here, can be used in a statistical model where each particle is gaining energy through its interaction with many small reconnection sites.     

Energy release in solar flares

We examine observational evidence concerning energy release in solar flares. We propose that different processes may be operative on four different time scales: (a) on the sub-second time scale of sub-bursts which are a prominent feature of mm-wave microwave records; (b) on the few-seconds time scale of elementary bursts which are a prominent feature of hard X-ray records; (c) on the few-minutes time scale of the impulsive phase; and (d) on the tens-of-minutes or longer time scale of the gradual phase.We propose that the concentration of magnetic field into magnetic knots at the photosphere has important consequences for the coronal magnetic-field structure such that the magnetic field in this region may be viewed as an array of elementary flux tubes. The release of the free energy of one such tube may produce an elementary burst. The development of magnetic islands during this process may be responsible for the sub-bursts. The impulsive phase may be simply the composite effect of many elementary bursts.We propose that the gradual phase of energy release, with which flares typically begin and with which many flares end, involves a steady process of reconnection, whereas the impulsive phase involves a more rapid stochastic process of reconnection which is a consequence of mode interaction.In the case of two-ribbon flares, the late part of the gradual phase may be attributed to reconnection of a large current sheet which is being produced as a result of filament eruption. A similar process may be operative in smaller flares.Also, Department of Applied Physics, Stanford University.     

Heat transfer in solar flares

Radiative cooling and heat conduction determine the temperature structure of flare plasmas along magnetic field. It is shown that both in the case of slow heating and of impulsive heating, temperatures are distributed in such a way that classical collisional heat conduction is valid.     

Energy release in solar flares

Team 2 of the Ottawa FLARES 22 Workshop dealt with observational and theoretical aspects of the characteristics and processes of energy release in flares. Main results summarized in this article stress the global character of the flaring phenomenon in active regions, the importance of discontinuities in magnetic connectivity, the role of field-aligned currents in free energy storage, and the fragmentation of energy release in time and space.Report of Team 2, Flares 22 Workshop, Ottawa, May 25–28, 1993.     

Stellar and solar flares

Some ideas are developed concerning solar flares which have been presented earlier by the author (Schatzman, 1966a). Emphasis is laid on the problem of energy transport; from the energy supply to the region of the optical flare, on the storage of low energy cosmic ray particles in a magnetic bottle before the beginning of the optical flare, and the mechanism which triggers both the optical flare, and the production of high-energy cosmic rays. The relation between solar and stellar flares is considered.Lecture given at Goddard Space Flight Center, November 4, 1966.     

Magnetic flux changes associated with the solar flares of August 1972

The active region associated with Mt. Wilson sunspot group 18 935 (McMath, 11 976) which had a central meridian passage on August 4 and 5, 1972 produced a number of flares during transit. These included two importance 3B flares on August 4 and 7 as well as several of importance 1 and 2. Calculations of the total magnetic flux in this region were made during the period July 31 through August 9 using data from six observatories. For the 3B flare on August 4, the total flux changed from about 7.2 × 10²² Mx just before onset to about 5.6 × 10²² Mx two hours after onset. For the 3B flare on August 7, the flux was about 6.4 × 10²² Mx three hours before onset and about 5.2 × 10²² Mx three hours after onset. An importance 2B flare on August 2 had no measurable effect on the flux nor did any of several 1N or 1B flares which also occurred in this region during the period. The flux changes measured for the 3B flares occurred in the umbral and penumbral fields and no significant changes were observed in facular fields.The Aerospace Corporation, P.O. Box 92957, Los Angeles, Calif. 90009, U.S.A.     

Fine structures in solar microwave flares

The pulsed electron acceleration and release from the energy release volume in solar flares implies that there is a possibility of interaction between a group of electrons reflected from the foot of a bipolar flux tube with a newly injected beam. It is shown that interaction can lead to the stoppage of the synchrotron maser instability caused by the loss cone distribution and hence can produce further millisecond fine structures in the solar microwave bursts.     

An interpretation of rapid changes in the magnetic field associated with solar flares

The energy source of a flare is the magnetic field in the corona. A topological model of the magnetic field is used here for interpreting the recently discovered drastic changes in magnetic field associated with solar flares. The following observational results are self‐consistently explained: (1) the transverse field strength decreases at outer part of active regions and increases significantly in their centers; (2) the center‐of‐mass positions of opposite magnetic polarities converge towards the magnetic neutral line just after flares onset; (3) the magnetic flux of active regions decreases steadily during the course of flares. For X‐class flares, almost 50% events show such changes. (© 2008 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Non-thermal processes in large solar flares

We analyze particle acceleration processes in large solar flares, using observations of the August, 1972, series of large events. The energetic particle populations are estimated from the hard X-ray and γ-ray emission, and from direct interplanetary particle observations. The collisional energy losses of these particles are computed as a function of height, assuming that the particles are accelerated high in the solar atmosphere and then precipitate down into denser layers. We compare the computed energy input with the flare energy output in radiation, heating, and mass ejection, and find for large proton event flares that:

1. The ～10–10² keV electrons accelerated during the flash phase constitute the bulk of the total flare energy.
2. The flare can be divided into two regions depending on whether the electron energy input goes into radiation or explosive heating. The computed energy input to the radiative quasi-equilibrium region agrees with the observed flare energy output in optical, UV, and EUV radiation.
3. The electron energy input to the explosive heating region can produce evaporation of the upper chromosphere needed to form the soft X-ray flare plasma.
4. Very intense energetic electron fluxes can provide the energy and mass for interplanetary shock wave by heating the atmospheric gas to energies sufficient to escape the solar gravitational and magnetic fields. The threshold for shock formation appears to be ～10³¹ ergs total energy in >20 keV electrons, and all of the shock energy can be supplied by electrons if their spectrum extends down to 5–10 keV.
5. High energy protons are accelerated later than the 10–10² keV electrons and most of them escape to the interplanetary medium. The energetic protons are not a significant contributor to the energization of flare phenomena. The observations are consistent with shock-wave acceleration of the protons and other nuclei, and also of electrons to relativistic energies.
6. The flare white-light continuum emission is consistent with a model of free-bound transitions in a plasma with strong non-thermal ionization produced in the lower solar chromosphere by energetic electrons. The white-light continuum is inconsistent with models of photospheric heating by the energetic particles. A threshold energy of ～5×10³⁰ ergs in >20 keV electrons is required for detectable white-light emission.

The highly efficient electron energization required in these flares suggests that the flare mechanism consists of rapid dissipation of chromospheric and coronal field-aligned or sheet currents, due to the onset of current-driven Buneman anomalous resistivity. Large proton flares then result when the energy input from accelerated electrons is sufficient to form a shock wave.     

Particle acceleration in impulsive solar flares

Second-step acceleration of nonrelativistic protons and ions in impulsive solar flares is discussed extending our earlier calculations for relativistic electrons. We derive the relevant particle transport equation, discussing in detail the influence of the particle's effective charge and mass number on the various momentum gain (stochastic acceleration, diffusive shock wave acceleration) and loss (Coulomb interactions, particle escape) processes. Analytical solutions for the ion-momentum spectra in the hard-sphere approximation are given. The inclusion of Coulomb losses modify the particle spectra significantly at kinetic energies smaller than E _B = 0.64( _e /5.0) MeV nucl.^–1 from the well-known Bessel function variation in long-duration flares. For equal injection conditions this modification explains the observed much smaller ion fluxes from impulsive flares at high energies as compared to long-duration flares. We also calculate the ³He/⁴He-isotope variation as a function of momentum in impulsive flares in the hard-sphere approximation and find significant variations near E _m = 0.38(T _e /2 × 10⁶ K) MeV nucl.^–1, where T _e is the electron temperature of the coronal medium.     

Electric field measurements in solar flares

Meaurements of solar flare spectra have allowed the electric field strengths in two flares to be determined, using the Inglis-Teller formula. Further, an independently estimated value for the electron density has allowed the two components of this field, that is, the interionic component and the external component that arises, for example, through plasma instabilities, to be separately extracted. External electric field strengths 0.5 kV cm^–1 for a limb flare and 1.3 kV cm^–1 for a white-light flare are found. Estimates of electric fields strengths generated by the resistive magnetic tearing instability indicate that this process could account for a significant part of the electric field if pre-existing magnetic field strengths in the flaring regions are characterized by a few kilogauss. Other plasma processes probably contribute measurably as well.Operated by the Association of Universities for Research in Astronomy, Inc., under contract NSF AST84-18716 with the National Science Foundation.     

Particle acceleration in impulsive solar flares

A model for second-step electron acceleration in impulsive solar flares is presented. We have extended the theory of stochastic particle acceleration to include Coulomb energy losses which become important at low coronal heights. This inclusion successfully explains the observed steepening of interplanetary electron spectra below 3 MeV following impulsive solar flares taking place at low coronal heights. It also explains the observed spectral differences of relativistic electrons in long-duration and impulsive flares.     

X-ray emission from solar flares

Solar X-ray Spectrometer (SOXS), the first space-borne solar astronomy experiment of India was designed to improve our current understanding of X-ray emission from the Sun in general and solar flares in particular. SOXS mission is composed of two solid state detectors, viz., Si and CZT semiconductors capable of observing the full disk Sun in X-ray energy range of 4–56 keV. The X-ray spectra of solar flares obtained by the Si detector in the 4–25 keV range show evidence of Fe and Fe/Ni line emission and multi-thermal plasma. The evolution of the break energy point that separates the thermal and non-thermal processes reveals increase with increasing flare plasma temperature. Small scale flare activities observed by both the detectors are found to be suitable to heat the active region corona; however their location appears to be in the transition region.     

Relativistic electrons from solar flares

Observations of interplanetary relativistic electrons from several solar-flare events monitored through 1964 to mid-1967 are presented. These are the first direct spectral measurements and time histories, made outside the magnetosphere, of solar-flare electrons having relativistic velocities. The 3- to 12-MeV electrons detected have kinetic energies about two orders of magnitude higher than those solar electrons previously studied in space, and measurements of both the time histories and energy spectra for a number of events in the present solar cycle were carried out. These measurements of interplanetary electrons are also directly compared with solar X-ray data and with measurements of related interplanetary solar protons.The time histories of at least four electron events show fits to the typical diffusion picture. A demonstrated similarity between the electron and the medium-energy proton fits for the event of 7 July, in particular, indicates that at these electron energies, but over several orders of magnitude of rigidity, whatever diffusion does take place is very nearly on a velocity, rather than a rigidity or an energy, basis. Diffusion-fit time histories varied as a function of T ₀ also indicate that the electrons in certain flare events originate at times near the X-ray and microwave burst, establishing their likely identity as the same electrons which cause the impulsive radiations. Also, the energy spectra and total numbers of the interplanetary electrons, compared with those of the flare-site electrons calculated from X-ray and microwave measurements, indicate that probably a small fraction of flare electrons escape into interplanetary space.     

Magnetic theories of solar flares

We have used the 512 channel diode array and vacuum telescope at KPNO to study the photospheric intensity distribution around sunspots, for comparison with isotherms predicted by convective blocking models of heat flow. Raster scan observations of 10 spots on 18 days were carried out in 1980 and 1981. Continuum passbands of 0.25 Å width were selected to avoid contamination by weak Fraunhofer lines, whose strength is sensitive to the presence of magnetic faculae often found near spots. Our observations show no evidence of extended bright rings around the spots at the level of 1–2%, as reported in one recent study using photographic photometry and much wider passbands. But 6 of the 10 spots we measured show marginally significant (2–3σ) bright rings of peak amplitude 0.1–0.3%. We are not able to explain these rings as a result of either residual facular signal, or instrumental effects. The excess radiative flux in these rings is small compared to the missing flux in the spot umbra and penumbra. We compare the brightness of the observed rings with peak brightnesses calculated from models of heat flow around spots of various depths and radii. Even if the spot is assumed to be unrealistically shallow, a detectable bright ring requires that the effective thermal conductivity (and/or its depth gradient) in layers surrounding the spot be significantly lower than the values indicated by mixing length models of the solar convection zone.