Stephen A. Drake

Coronal Structure and Abundances of Capella from Simultaneous EUVE and ASCA Spectroscopy

We report analysis of the simultaneous 1996 March EUVE and ASCA observations of the spectroscopic binary Capella. The EUVE spectrum is dominated by lines of highly ionized Fe, requiring a continuous emission-measure distribution over a wide range of temperatures. The ASCA spectrum shows He-like line emission features of S, Si, and Mg, as well as unresolved L-shell emission lines of Fe and Ni and H-like and He-like Ne lines. The flux in these line features cannot be determined independently from the continuum flux. The ASCA spectrum is relatively soft, with few counts above 4 keV. The emission-measure distribution determined by Line-Based Analysis of the EUV Fe line intensities is well constrained from Te ~ 6 × 105 to 2 × 107 K, but it is not constrained above this range since Fe XXIV is the highest temperature line observed with EUVE. Since repeated observations of Capella by EUVE have shown that emission-line intensities of the hottest EUV-emitting material (Fe XXI to XXIV) vary by factors up to 4, the ASCA spectrum is important for extending the temperature coverage. Thus, the high-energy cut-off of the ASCA spectrum provides a constraint on the highest temperature emission measures. In principle, elemental abundances are determined from global fits to the ASCA spectrum; however, no well-fitting model has been found for the high signal-to-noise ASCA performance verification spectrum of Capella (1993 September 2). The newer ASCA spectrum of Capella (1996 March 3-4) shows a similar pattern of fitting difficulties. Using the EUVE measurements (1996 March 3-7) to constrain models, we have conducted sensitivity studies of the atomic data, source physics, and instrument calibration. The plasma spectral emission models (Raymond-Smith, MEKAL, SPEX) around 1.2 keV appear to have flux deficits relative to the observed ASCA count spectrum. New atomic models by Liedahl and Brickhouse, calculated with the HULLAC code, provide a set of lines—missing from the existing plasma codes—to fill in this flux deficit. Incorporating these additional lines dramatically improves the spectral model fits to the data, allowing reliable determination of elemental abundances. The successful application of the new atomic models to the Capella problem can have widespread implications, affecting spectral models of galaxies, cluster cooling flows, and supernova remnants, as well as other stellar coronae. Analysis with the new atomic models of the simultaneous ASCA and EUVE data confirms the previous EUVE results that the continuous emission-measure distribution of Capella has a strong enhancement at Te ~ 6 × 106 K. While a two-temperature model actually provides a better fit to the ASCA spectrum than the EUVE-derived continuous model, the EUVE data are not well fitted with only two temperatures. We find that the abundances of Mg, Si, S, and Fe are consistent with solar photospheric values, while Ne appears to be underabundant by a factor of ~3 to 4.

The evolution of chromospheric activity of cool giant and subgiant stars

IUE spectra for a large sample of cool subgiant stars are examined, and evidence is found that subgiants in the mass range 1.2-1.6 solar masses undergo a sudden decline in UV transition region emission near B - V = 0.6, which corresponds to spectral type G0 IV. The decline in UV emission coincides with a sharp decrease in stellar rotation rates, and it is suggested that this decay in activity and rotation marks a transformation from acoustic heating in the early F stars to magnetic dynamo-driven activity in the cooler stars, resulting in a strong rotational braking action by stellar wind. For more massive giant stars, there is a similar transformation in the nature of chromospheric activity near B - V = 0.7, or spectral type G0 III, from acoustic heating in the F-type giants to a solarlike dynamo mechanism in the cooler giants. No sign of an abrupt drop in activity near spectral type G5 III at the location of Grays proposed rotational boundary line is seen. 102 refs.

A survey of the radio continuum emission of RS Canum Venaticorum and related active binary systems

Observations of 77 RS CVn and related active binary systems were made with VLA at a continuum frequency of 4.86 GHz. The results are combined with other VLA observations to produce a total number of observed sources of 122. A search for correlations of radio properties with various systemic and stellar parameters is performed. Significant correlations are found between the normalized radio luminosity L(6)/L(bol) and the normalized soft X-ray luminosity L(x)/L(bol). It is suggested that the close correlation between low-level 6-cm emission and the hot component of the soft X-ray emission observed by the Einstein solid state spectrometer is caused by both emissions being produced by the same thermal electrons.

Corona(e) of AR Lacertae. I. The Temperature and Abundance Distribution

X-ray spectra of the RS CVn binary AR Lac, obtained from simultaneous observations with the ROSAT PSPC and the ASCA SIS and GIS detectors, have been analyzed to study the coronal temperature and abundance distribution. The spectra were jointly fitted with plasma emission models to the following possible temperature distributions: (a) one with discrete multitemperature emission components, (b) a continuous emission measure with a power-law dependence on temperature, and (c) a continuous emission distribution parameterized by the sum of a sixth-order Chebyshev polynomial. We find that (i) solar abundance plasma models with either discrete or continuous emission measure (CEM) distributions are rejected, (ii) the best fit is obtained with a two-temperature (2T) plasma emission model with an underabundance of the elements 0, Mg, Si, S, Ar, Ca, and Fe by a factor of 3-4 relative to the solar photospheric values, and (iii) the best-fit CEM distribution also has similarly reduced abundances but fits the data less well than the 2T model. These results are confirmed even when the Fe-L region, which is subject to uncertainties in the atomic physics, is excluded from the fit. We consider optical depth effects as unlikely to be the explanation for the observed weakness of the line complexes relative to the continuum. Analysis of the spectral data during the primary and secondary eclipses shows that the emission measure of the high-temperature component in the 2T models appears to be more affected by the primary eclipse than the low-temperature component, suggesting that part of the former is concentrated in structures that are spatially more compact.

Nonthermal Hard X-Ray Emission and Iron Kα Emission from a Superflare on II Pegasi

We report on an X-ray flare detected on the active binary system II Pegasi with the Swift telescope. The event triggered the Burst Alert Telescope (BAT) in the hard X-ray band on 2005 December 16 at 11:21:52 UT with a 10-200 keV luminosity of 2.2 ? 1032 ergs s-1?a superflare, by comparison with energies of typical stellar flares on active binary systems. The trigger spectrum indicates a hot thermal plasma with T ~ 180 ? 106 K. X-ray spectral analysis from 0.8 to 200 keV with the X-Ray Telescope and BAT in the next two orbits reveals evidence for a thermal component (T > 80 ? 106 K) and Fe K 6.4 keV emission. A tail of emission out to 200 keV can be fit with either an extremely high temperature thermal plasma (T ~ 3 ? 108 K) or power-law emission. Based on analogies with solar flares, we attribute the excess continuum emission to nonthermal thick-target bremsstrahlung emission from a population of accelerated electrons. We estimate the radiated energy from 0.01 to 200 keV to be ~6 ? 1036 ergs, the total radiated energy over all wavelengths ~10 38 ergs, the energy in nonthermal electrons above 20 keV ~3 ? 1040 ergs, and conducted energy 20 keV when compared to the upper and lower bounds on the thermal energy content of the flare. This marks the first occasion in which evidence exists for nonthermal hard X-ray emission from a stellar flare. We investigate the emission mechanism responsible for producing the 6.4 keV feature, and find that collisional ionization from nonthermal electrons appears to be more plausible than the photoionization mechanism usually invoked on the Sun and pre-main-sequence stars.

THE MOUSE THAT ROARED: A SUPERFLARE FROM THE dMe FLARE STAR EV LAC DETECTED BY SWIFT AND KONUS-WIND

We report on a large stellar flare from the nearby dMe flare star EV Lac observed by the Swift and Konus-Wind satellites and the Liverpool Telescope. It is the first large stellar flare from a dMe flare star to result in a Swift trigger based on its hard X-ray intensity. Its peak fX from 0.3 to 100 keV of 5.3 ? 10?8 erg cm?2?s?1 is nearly 7000 times larger than the stars quiescent coronal flux, and the change in magnitude in the white filter is ?4.7. This flare also caused a transient increase in EV Lacs bolometric luminosity (L bol) during the early stages of the flare, with a peak estimated L X /L bol ~ 3.1. We apply flare loop hydrodynamic modeling to the plasma parameter temporal changes to derive a loop semi-length of l/R = 0.37 ? 0.07. The soft X-ray spectrum of the flare reveals evidence of iron K? emission at 6.4 keV. We model the K? emission as fluorescence from the hot flare source irradiating the photospheric iron, and derive loop heights of h/R = 0.1, consistent within factors of a few with the heights inferred from hydrodynamic modeling. The K? emission feature shows variability on timescales of ~200 s which is difficult to interpret using the pure fluorescence hypothesis. We examine K? emission produced by collisional ionization from accelerated particles, and find parameter values for the spectrum of accelerated particles which can accommodate the increased amount of K? flux and the lack of observed nonthermal emission in the 20-50 keV spectral region.

X-Ray and Radio Observations of the Cygnus OB2 Association

Several OB stars in the Cygnus OB2 association are among the strongest stellar X-ray and radio sources in the Galaxy. The radio emission is particularly unusual, displaying a high level of variability and nonthermal behavior. We obtained two ROSAT PSPC observations, a ROSAT HRI observation, and three VLA observations of the association during a 2 yr time span. Our study will focus on four stars, Cyg OB2 No. 5, No. 8A, No. 9, and No. 12. Three of the four (Cyg No. 5, No. 9, and No. 12) were relatively constant in their X-ray emission over the 2 yr time frame. The fourth, Cyg OB2 No. 8A, increased in intensity by ~34%. No short-term (hourly) variability was detected. The observed X-ray characteristics (e.g., luminosity, temperature) are found to be consistent with the X-ray properties of other OB stars. The exception is Cyg OB2 No. 12, whose X-ray characteristics are found to be inconsistent with its spectral classification. Detailed spectral analyses of the PSPC data are presented for two absorption models: (1) ISM (cold absorber) and (2) Wind + ISM (warm absorber). The spectral fits suggest that the X-ray sources are located within the stellar wind, and estimates of the X-ray locations are presented. Adopting the radio-derived mass-loss rates, these X-ray locations are found to be consistent with the shock scenario proposed for OB stars. As expected, the radio emission has continued to be highly variable. Nonthermal characteristics are observed in Cyg OB2 No. 8A and No. 12. One of the most unusual nonthermal radio sources, Cyg OB2 No. 9, was found to be thermal in one of our observations. An observation of Cyg OB2 No. 5 also displayed a thermal radio spectrum. A comparison of the observed and intrinsic X-ray fluxes with the observed radio fluxes suggests that these quantities are anticorrelated; the strongest X-ray source is the weakest radio source. This is contrary to normal expectations for a wind-generated model of X-ray and radio emission. We investigate the long-term temporal behavior of both the X-ray and radio emission by comparing our newly acquired data with the previous X-ray (IPC) and radio data over the past 15 yr. Except for three events observed in Cyg OB2 No. 5, which displayed significant increases in its X-ray emission, the X-ray emission has remained relatively constant with a variability level less than 20% over this time span, whereas the radio emission has stayed highly variable with various levels of nonthermal behavior. We present a model to investigate the case in which the X-ray and radio emission are controlled by stellar wind properties and find that the predicted variability should be comparable in both emission processes. This is not observed. It is very intriguing that every time we observe the radio emission, it is different, whereas the X-ray emission always appears to be constant. If the stellar wind is as variable as suggested by the radio data, we believe it is highly implausible that we just happened to miss all periods of X-ray variability. We also investigate the implications of the observed nonthermal radio spectrum of Cyg OB2 No. 8A. Using the synchrotron emission model of White, we find that this nonthermal radio spectrum predicts a mass-loss rate almost 2 orders of magnitude less than that expected for a thermal radio spectrum. This lower mass-loss rate is consistent with an X-ray source located at the base of the stellar wind, contrary to the basic shock scenario. Since these stars show evidence of changing from thermal to nonthermal radio characteristics, it is difficult to understand how such a large change in mass loss did not produce a significant change in the observed X-rays.

Radio emission from chemically peculiar stars

In five VLA observing runs the initial survey of radio emission from magnetic Bp-Ap stars by Drake et al. is extended to include a total of 16 sources detected at 6 cm out of 61 observed, giving a detection rate of 26 percent. Of these stars, three are also detected at 2 cm, four at 3.6 cm, and five at 20 cm. The 11 new stars detected as radio sources have spectral types B5-A0 and are He-weak and Si-strong. No classical (SrCrEu-type) Ap stars have yet been detected. The 16 detected sources show a wide range of radio luminosities with the early-B He-S stars on average 20 times more radio luminous than the late-B He-W stars and 1000 times more luminous than Theta Aurigae. Multifrequency observations indicate flat spectra in all cases. Four stars have a detectable degree of circular polarization at one or more frequencies. It is argued that the radio-emitting CP (chemically peculiar) stars form a distinct class of radio stars that differs from both the hot star wind sources and the active late-type stars. The observed properties of radio emission from these stars may be understood in terms of optically thick gyrosynchrotron emission from a nonthermal distribution of electrons produced in a current sheet far from the star. In this model the electrons travel along magnetic fields to smaller radii and higher magnetic latitudes where they mirror and radiate microwave radiation.

The discovery of nonthermal radio emission from magnetic Bp-Ap stars

In a VLA survey of chemically peculiar B- and A-type stars with strong magnetic fields, five of the 34 stars observed have been identified as 6 cm continuum sources. Three of the detections are helium-strong early Bp stars (Sigma Ori E, HR 1890, and Delta Ori C), and two are helium weak, silicon-strong stars with spectral types near A0p (IQ Aur = HD 34452, Babcocks star = HD 215441). The 6 cm luminosities L6 (ergs/s Hz) range from log L6 = 16.2 to 17.9, somewhat less than the OB supergiants and W-R stars. Three-frequency observations indicate that the helium-strong Bp stars are variable nonthermal sources.

Corona(e) of AR Lacertae. II. The spatial structure

The X-ray light curves in the 0.4-1.5 keV and 2-7 keV bands of the RS CVn binary AR Lacertae observed on 1993 June 1-3 over one full orbital cycle with the ASCA satellite have been used to map the spatial structure of AR Lacs coronae. We find that both stars are X-ray active, that the corona of the K-type secondary star appears to be hotter than that of the G-type primary star, that X-ray emission is concentrated on the sides of the stars facing each other, and that there are compact and well-localized regions of enhanced X-ray emission with heights much smaller than the stellar radii. In one class of solutions there are additional extended regions with dimensions similar to or greater then the radii of the underlying stars which may be structures that interconnect the two stars. There are also other acceptable models without extended structures, however our analysis indicates that solutions with extended sources are more probable. Also, about 50% of the X-ray emission is unmodulated and could come from either an extended halo region, from the poles of the larger K star, or from other symmetric or uneclipsed structures in the orbital plane. We compare the coronal structures inferred from the ASCA observations with those inferred using the same technique from an EXOSAT observation of AR Lac made in 1984 and find that there are substantial differences between the coronal structures at these two epochs. For the solution with extended material in the orbital plane, we have derived the rough physical parameters for the X-ray-emitting plasma, using the derived information on the spatial sizes of the various spatial components together with information about the emission measure and temperatures obtained from a simple spectral analysis of the ASCA data.