Petr Heinzel

Physics of Solar Prominences: I—Spectral Diagnostics and Non-LTE Modelling

This review paper outlines background information and covers recent advances made via the analysis of spectra and images of prominence plasma and the increased sophistication of non-LTE (i.e. when there is a departure from Local Thermodynamic Equilibrium) radiative transfer models. We first describe the spectral inversion techniques that have been used to infer the plasma parameters important for the general properties of the prominence plasma in both its cool core and the hotter prominence-corona transition region. We also review studies devoted to the observation of bulk motions of the prominence plasma and to the determination of prominence mass. However, a simple inversion of spectroscopic data usually fails when the lines become optically thick at certain wavelengths. Therefore, complex non-LTE models become necessary. We thus present the basics of non-LTE radiative transfer theory and the associated multi-level radiative transfer problems. The main results of one- and two-dimensional models of the prominences and their fine-structures are presented. We then discuss the energy balance in various prominence models. Finally, we outline the outstanding observational and theoretical questions, and the directions for future progress in our understanding of solar prominences.

Hinode, TRACE, SOHO, and Ground-based Observations of a Quiescent Prominence

A quiescent prominence was observed by several instruments on 2007 April 25. The temporal evolution was recorded in Hα by the Hinode SOT, in X-rays by the Hinode XRT, and in the 195 A channel by TRACE. Moreover, ground-based observatories (GBOs) provided calibrated Hα intensities. Simultaneous extreme-UV (EUV) data were also taken by the Hinode EIS and SOHO SUMER and CDS instruments. Here we have selected the SOT Hα image taken at 13:19 UT, which nicely shows the prominence fine structure. We compare this image with cotemporaneous ones taken by the XRT and TRACE and show the intensity variations along several cuts parallel to the solar limb. EIS spectra were obtained about half an hour later. Dark prominence structure clearly seen in the TRACE and EIS 195 A images is due to the prominence absorption in H I, He I, and He II resonance continua plus the coronal emissivity blocking due to the prominence void (cavity). The void clearly visible in the XRT images is entirely due to X-ray emissivity blocking. We use TRACE, EIS, and XRT data to estimate the amount of absorption and blocking. The Hα integrated intensities independently provide us with an estimate of the Hα opacity, which is related to the opacity of resonance continua as follows from the non-LTE radiative-transfer modeling. However, spatial averaging of the Hα and EUV data have quite different natures, which must be taken into account when evaluating the true opacities. We demonstrate this important effect here for the first time. Finally, based on this multiwavelength analysis, we discuss the determination of the column densities and the ionization degree of hydrogen in the prominence.

Why Are Solar Filaments More Extended in Extreme-Ultraviolet Lines than in Hα?

A long solar filament was observed simultaneously in the Hα line by THEMIS/MSDP and in selected EUV lines by the Coronal Diagnostic Spectrometer on SOHO. Co-alignment of optical and EUV images reveals that the dark EUV filament is much more extended than the Hα filament. Assuming that the EUV filament represents Lyman continuum absorption of the background EUV-line radiation, a straightforward explanation of this effect is suggested. Based on non-LTE filament models, we demonstrate that the ratio of the Lyman continuum to Hα opacity can reach a factor of 50-100, and thus the EUV filament is still well visible while the Hα line contrast diminishes below the detection limit. This kind of interpretation leads to an important conclusion that the cool filament material in which the Lyman continuum absorption takes place is more abundant than one would expect from Hα disk observations. This then may have significant consequences on the filament structure and on formation models, as well as on mass considerations related to coronal mass ejections.

ON THE NATURE OF PROMINENCE EMISSION OBSERVED BY SDO/AIA

The prominence-corona transition region (PCTR) plays a key role in the thermal and pressure equilibrium of solar prominences. Our knowledge of this interface is limited and several major issues remain open, including the thermal structure and, in particular, the maximum temperature of the detectable plasma. The high signal-to-noise ratio of images obtained by the Atmospheric Imaging Assembly (AIA) on NASAs Solar Dynamics Observatory clearly shows that prominences are often seen in emission in the 171 and 131 bands. We investigate the temperature sensitivity of these AIA bands for prominence observations, in order to infer the temperature content in an effort to explain the emission. Using the CHIANTI atomic database and previously determined prominence differential emission measure distributions, we build synthetic spectra to establish the main emission-line contributors in the AIA bands. We find that the Fe IX line always dominates the 171 band, even in the absence of plasma at >106 K temperatures, while the 131 band is dominated by Fe VIII. We conclude that the PCTR has sufficient plasma emitting at >4 × 105 K to be detected by AIA.

MAGNETIC TOPOLOGY OF BUBBLES IN QUIESCENT PROMINENCES

We study a polar-crown prominence with a bubble and its plume observed in several coronal filters by the SDO/AIA and in Hα by the MSDP spectrograph in Bialkow (Poland) to address the following questions: what is the brightness of prominence bubbles in EUV with respect to the corona outside of the prominence and the prominence coronal cavity? What is the geometry and topology of the magnetic field in the bubble? What is the nature of the vertical threads seen within prominences? We find that the brightness of the bubble and plume is lower than the brightness of the corona outside of the prominence, and is similar to that of the coronal cavity. We constructed linear force-free models of prominences with bubbles, where the flux rope is perturbed by inclusion of parasitic bipoles. The arcade field lines of the bipole create the bubble, which is thus devoid of magnetic dips. Shearing the bipole or adding a second one can lead to cusp-shaped prominences with bubbles similar to the observed ones. The bubbles have complex magnetic topology, with a pair of coronal magnetic null points linked by a separator outlining the boundary between the bubble and the prominence body. We conjecture that plume formation involves magnetic reconnection at the separator. Depending on the viewing angle, the prominence can appear either anvil-shaped with predominantly horizontal structures, or cusp-shaped with predominantly vertical structuring. The latter is an artifact of the alignment of magnetic dips with respect to the prominence axis and the line of sight.

Relation between cool and hot post-flare loops of 26 June 1992 derived from optical and X-ray (SXT-Yohkoh) observations

We have analyzed the physical conditions of the plasma in post-flare loops with special emphasis on dynamics and energy transport using SXT-data (hot plasma) and optical ground-based data from Pic du Midi, Wrocław, and Ondřejov (cool plasma). By combining the Hα observations with the SXT images we can understand the relationship between cool and hot plasmas, the process of cooling post-flare loops and the mechanism which maintains the long duration of these loops. Using recent results of NLTE modeling of prominence-like plasmas, we derive the emission measure of cool Hα loops and this gives us a realistic estimate of the electron density (2.2 × 1010 cm−3). Then, by comparing this emission measure with that of hot loops derived from SXT data, we are able to estimate the ratio between electron densities in hot and cool loops taking into account the effect of geometrical filling factors. This leads to the electron density in hot loops 7 × 109 cm−3. We also derive the temperature of hot X-ray loops (≃ 5.5 × 106 K), which, together with the electron density, provides the initial values for solving the time-dependent energy balance equation. We obtain the cooling times which are compared to a typical growth-time of the whole loop system (∼ 2000 s). In the legs of cool Hα loops, we observe an excess of the emission measure which we attribute to the effect of Doppler brightening (due to large downflow velocities).

Spectroscopic diagnostics of an Hα and EUV filament observed with THEMIS and SOHO

A long filament has been observed with THEMIS/MSDP and SOHO/CDS - SUMER, during a coordinated campaign (JOPs 131/95) on May 5, 2000. The data were (a) 2-D Hα spectra, observed using THEMIS, (b) Lyman series spectra and Lyman continuum, observed using SOHO/SUMER, and (c) EUV spectra (in O  629 A, Mg  624 A, Si  520 A, Ca  557 A and He  584 A) observed using SOHO/CDS. A large depression of the line emissions in CDS images represents the EUV filament. A computed model shows that the EUV filament consists of an extended in height cloud of low gas pressure at an altitude lower than the top of the Hα filament, volume-blocking and absorbing coronal emission and absorbing transition region line emission. The optical thickness of the Lyman continuum is estimated by using the ratio of O  intensity inside and outside the EUV filament, while the optical thickness of Hα is computed from the Hα line profile by using an inversion technique. Using simultaneous Hα, Lyman lines and Lyman continuum spectroscopic data, we performed detailed, non-LTE radiative transfer diagnostics of the filament plasma conditions. The optical thickness of the Lyman continuum is larger than that of the Hα line by one to two orders of magnitude. This could be of a great importance for filament formation modeling, if we consider that more cool material exists in filament channels but is optically too thin to be visible in Hα images.

Response of optical hydrogen lines to beam heating - I. Electron beams

Context. Observations of hydrogen Balmer lines in solar flares remain an important source of information on flare processes in the chromosphere during the impulsive phase of flares. The intensity profiles of optically thick hydrogen lines are determined by the temperature, density, and ionisation structure of the flaring atmosphere, by the plasma velocities and by the velocity distribution of particles in the line formation regions. Aims. We investigate the role of non-thermal electrons in the formation regions of H α , H β , and H γ lines in order to unfold their influence on the formation of these lines. We concentrate on pulse-beam heating varying on a subsecond timescale. Furthermore, we theoretically explore possibility that a new diagnostic tool exists indicating the presence of non-thermal electrons in the flaring chromosphere based on observations of optical hydrogen lines. Methods. To model the evolution of the flaring atmosphere and the time-dependent hydrogen excitation and ionisation, we used a 1-D radiative hydrodynamic code combined with a test-particle code that simulates the propagation, scattering, and thermalisation of a power-law electron beam in order to obtain the flare heating and the non-thermal collisional rates due to the interaction of the beam with the hydrogen atoms. To not bias the results by other effects, we calculate only short time evolutions of the flaring atmosphere and neglect the plasma velocities in the radiative transfer. Results. All calculated models have shown a time-correlated response of the modelled Balmer line intensities on a subsecond timescale, with a subsecond timelag behind the beam flux. Depending on the beam parameters, both line centres and wings can show pronounced intensity variations. The non-thermal collisional rates generally result in an increased emission from a secondary region formed in the chromosphere. Conclusions. Despite the clear influence of the non-thermal electron beams on the Balmer line intensity profiles, we were not able on the basis of our simulations to produce any unambiguous diagnostic of non-thermal electrons in the line-emitting region, which would be based on comparison of individual Balmer line intensity profiles. However, fast line intensity variations, well-correlated with the beam flux variations, represent an indirect indication of pulsating beams.

Multi-wavelength study of a high-latitude EUV filament

A large filament was observed during a multi-wavelength coordinated campaign on June 19, 1998 in the Hα line with the Swedish Vacuum Solar Telescope (SVST) at La Palma, in the coronal lines Fe ix/x 171 Å and Fe xi 195 Å with the Transition Region and Coronal Explorer (TRACE) and in EUV lines with the SOHO/CDS spectrometer and the hydrogen Lyman series with the SOHO/SUMER spectrometer. Because of its high-latitude location, it is possible to disentangle the physical properties of the Hα filament and the filament channel seen in EUV lines. TRACE images point out a dark region fitting the Hα fine-structure threads and a dark corridor (filament channel), well extended south of the magnetic inversion line. A similar pattern is observed in the CDS EUV-line images. The opacity of the hydrogen and helium resonance continua at 171 Å is almost two orders of magnitude lower than that at the Hi head (912 Å) and thus similar to the opacity of the Hα line. Since we do not see the filament channel in Hα, this would imply that it should also be invisible in TRACE lines. Thus, the diffuse dark corridor is interpreted as due to the coronal ‘volume blocking’ by a cool plasma which extends to large altitudes. Such extensions were also confirmed by computing the heights from the projection geometry and by simulations of the CDS and TRACE line intensities using the spectroscopic model of EUV filaments (Heinzel, Anzer, and Schmieder, 2003). Finally, our NLTE analysis of selected hydrogen Lyman lines observed by SUMER also leads to a conclusion that the dark filament channel is due to a presence of relatively cool plasma having low densities and being distributed at altitudes reaching the Hα filament.

Post-flare loops of 26 June 1992

We observed the large post-flare loop system, which developed after the X 3.9 flare of 25 June 1992 at 20∶11 UT, in Hα with the Multichannel Subtractive Double Pass Spectrograph at Pic-du-Midi and in X-rays with the it Yohkoh/SXT instrument. Following the long-term development of cool and hot plasmas, we have determined the emission measure of the cool plasma and, for the first time, the temporal evolution of the hot-loop emission measure and temperature during the entire gradual phase. Thus, it was possible to infer the temporal variation of electron densities, leading to estimates of cooling times. A gradual decrease of the hot-loop emission measure was observed, from 4 × 1030 cm−5 at 23∶00 UT on 25 June 1992 to 3 × 1028 cm−5 at 13∶10 UT on 26 June 1992. During the same period, the temperature decreased only slowly from 7.2 to 6.0 × 106 K. Using recent results of NLTE modeling of prominence-like plasmas, we also derive the emission measure of cool Hα loops and discuss their temperature and ionisation degree. During two hours of Hα observations (11–13 hours after the flare) the averaged emission measure does not show any significant change, though the amount of visible cool material decreases and the volume of the loops increases. The emission measure in Hα, after correction for the Doppler-brightening effect, is slightly lower than in soft X-rays. Since the hot plasma seems to be more spatially extended, we arrive at electron densities in the range ninfesupho≤ ninfesupcool∼ 2 × 1010 cm−3 at the time of the Hα observations.These results are consistent with the post-flare loop model proposed by Forbes, Malherbe, and Priest (1989). The observed slow decrease of the emission measure could be due to an increase of the volume of the loops and a gradual decrease of the chromospheric ablation driven by the reconnection, which seems to remain effective continuously for more than 16 hours. The cooling time for hot loops to cool down to 104 K and to appear in Hα would be only a few minutes at the beginning of the gradual phase but could be as long as 2 hours at the end, several hours later.