Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://star.arm.ac.uk/~ambn/pcuvflare.dvi-ps Äàòà èçìåíåíèÿ: Fri May 31 18:25:41 1996 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 00:21:36 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: ecliptic

Stellar flares -- UV
P. Brendan Byrne
1 Armagh Observatory, Armagh BT61 9DG, N. Ireland
2 eMail: pbb@star.arm.ac.uk
1 Introduction
Early studies of stellar flares were made entirely in the optical regime. It was
recognised that flares arose from the generation of hot plasma within the stellar
chromosphere at whose temperature (indicated, for instance, by the presence of
a strong, blue optical continuum) a substantial emission in the ultraviolet would
be expected. It was not until the advent of space­borne instruments of adequate
sensitivity, however, that direct confirmation of this prediction was forthcoming.
In this review I examine some results of more than a decade of observation of
stellar flares.
1.1 Sources of UV data
The major source of ultraviolet data on stellar flares has been the International
Ultraviolet Explorer (IUE) satellite (Boggess et al. 1979). In order to understand
the limitations of our current understanding in this area it is important to ap­
preciate some of the characteristics of its instrumentation. IUE`s telescope is of
40cm aperture and it is equipped with a spectrograph which can operate at two
resolutions, i.e. \Delta–=– ¸350 (LORES) and ¸17 000 (HIRES). Its detectors are
optimised for operation in two wavebands, i.e. ¸1150--1950 š A (SW) and ¸1950--
3200 š A (LW). IUE`s small aperture results in a limited sensitivity, a consequence
of which is a modest time resolution when studying stellar flares (a long expo­
sure time is needed to gain adequate signal­to­noise). IUE`s elliptical 24­hour
quasi­geosynchronous orbit and its resulting interactive mode of operation make
continuous monitoring feasible, a feature suiting flare star work.
The recent advent of the Hubble Space Telescope (HST) (Giacconi et al.
1982) with its 2.4 m aperture represents in some ways such a huge gain in light­
gathering power that it might be expected to render IUE obselete. Its prime UV
spectroscopic instrument is the Goddard High Resolution Spectrograph (GHRS)
whose detectors span essentially the same spectral bands as IUE. This offers
spectral resolutions of \Delta–=– ¸2 000, ¸25 000, and ¸80 000. However, the spec­
tral ranges available at each of these resolutions are ¸285š A, ¸30š A, and ¸8š A

2 P. Brendan Byrne
respectively. Thus while HST offers a giant leap in sensitivity, it fails to offer the
wide spectral coverage of IUE. The orbit of HST imposes a further restriction on
its use. Its low orbit means that, for objects at low ecliptic latitude, the Earth
occults many targets for up to half the observing time.
Thus most of the ultraviolet flare data available to date has come from IUE.
Observations from HST are beginning to appear, however, and will, no doubt,
increase in the near future.
2 Flares on dMe stars
2.1 dMe quiescent emission
Quiescent emission from dMe flare stars is characterised by strong line emis­
sion from neutral and ionised atomic species arising in temperatures charac­
teristic of the solar chromosphere (6 10 3o KŸT e Ÿ10 4o K) and transition region
(3 10 4o KŸT e Ÿ2 10 5o K). A weak continuum is sometimes present in the long
wavelength portion of the LW spectra but is entirely absent in SW. An example
of a LORES IUE spectrum of dMe flare star in quiesence will be found in Fig. 1
and the corresponding mean surface fluxes in Table 1.
Fig. 1. An example IUE SW LORES spectrum of the dMe flare star, YZ CMi,
in quiescence and while flaring (J.G. Doyle, private communication). Prominent
emission lines are indicated. The quiescent spectrum has been displaced to aid
visibility of the two spectra. The horizontal line indicates the zero level for
each spectrum. Note the strong continuum in the flare spectrum, which is not
detectible in quiescence.

Stellar flares -- UV 3
Table 1. Typical mean emission line surface fluxes, in units of 10 4 erg cm \Gamma2 sec \Gamma1 , for
dMe stars in quiesence and during flare, compared to the mean solar quiescent flux
in these lines. Note that the flaring values are lower limits for the reasons discussed
in Section 2.2. Figures are also given for continuum surface fluxes in IUE`s SW region
(1150--1950 š A).
Star V1005 Ori 1 BYDra 2 YZCMi 3 HKAqr 4 Quiet
Line logfq logff logfq logff logfq logff logfq logff Sun
CII 7 32 16 24 6 22 12 19 0.5
CIV 11 229 26 33 13 75 22 98 0.6
NV ---­ ---­ 8 14 6 18 4 11 0.1
cont. ---­ 1200 ---­ ---­ ---­ 280 ---­ 1200
1 Mathioudakis et al. 1991; 2 Butler et al. 1987; 3 Doyle in prep.; 4 Byrne & McKay, 1990
2.2 dMe flaring emission
During flares all of the lines visible in the quiescent spectrum are seen to in­
crease, some by factors of up to ¸10 or greater. In general the degree of increase
is dependent on the excitation of the relevant line, up to mid­transition region
(T e ¸ 10 5 ffi K). So in the IUE wavebands the greatest flux enhancements are
seen in the resonance doublets of CIV(–1548/51 š A) and NV(–1238/42 š A). An ex­
ample flare spectrum is shown in Fig. 1 and some representative integrated flare
line surface fluxes at source are given in Table 1. It should be noted, however,
that, since areas of individual flares are unknown, it has been assumed in Table1
that they cover an entire hemisphere. Obviously, if a flare covers a fraction, 1
n
,
of a hemisphere, the fluxes in Table1 should be increased by the factor n. It
will be seen from Fig. 1 that a strong continuum is also present. Such continua
are recorded frequently in the more energetic dMe flares, but by no means uni­
versally. Lack of continuum detection in some flares may be a matter of IUE`s
limited sensitivity, however. The origin of dMe flare continua has been discussed
by Phillips et al. (1992).
Flaring emission in dMe`s is usually only detected in a single spectrum. This
is a result of the relatively long exposure times needed by IUE to detect flare stars
with adequate signal­to­noise compared with the typical duration of a flare as
gauged from other wavebands. In practise, exposure times in LW LORES need be
longer than ¸10min and in SW LORES longer than ¸20min for flare detection.
At the s/n levels attainable by IUE the effect of the flare on the spectrum is
usually undetectable in subsequent spectra. Thus flare parameters derived from
such observations should be judged with this fact in mind. In particular, mean
surface fluxes are calculated on the assumption that the flare lasts for the entire
exposure time.
Monitoring of active dMe stars for flares has revealed, apart from discrete
individual flares, continuous variability in these line fluxes (Butler et al. 1987,

4 P. Brendan Byrne
Byrne et al. 1991). This has led to the suggestion that, on the time scales re­
solvable with IUE, the most active dMe`s flare continuously at a low level. This
raises the question as to whether the measured ``quiescent'' fluxes of dMe`s are
``quiet'' in the same sense as the solar case.
3 Flares on RSCVn stars
3.1 RSCVn quiescent emission
Fig. 2. An example IUE SW LORES spectrum of the RSCVn star, II Peg, in
quiescence and while flaring (P.B. Byrne, private communication). The two flare
spectra were separated by ¸2 hr in time. Prominent emission lines are indicated.
The flare spectra have been displaced to aid their visibility. The horizontal line
indicates the zero level for each spectrum. Note the continuum in the flare spec­
tra, which is not detectible in quiescence.
Quiescent ultraviolet line emission from RSCVn systems is qualitatively the
same as that from the dMe`s except that the sources are more luminous. Some,
but not all, of the extra luminosity is accounted for by the greater surface areas
of these sub­giant or giant stars. The greater brightness of RSCVn`s has meant
that IUE`s HIRES mode can be used, in addition to SW LORES,to study their
spectra, at least in the strong MgII h&k resonance lines. An example of such
spectra will be found in Figs. 2 and 3.

Stellar flares -- UV 5
Table 2. Typical mean emission line surface fluxes, in units of 10 4 erg cm \Gamma2 sec \Gamma1 , for
RSCVn stars in quiesence and during flare, compared to the mean solar quiescent flux
in these lines. Figures are also given for continuum surface fluxes in IUE`s SW region
(1150--1950 š A).
Star II Peg 1 V711 Tau 2 IMPeg 3
– And 4 Quiet
Line logfq logff logfq logff logfq logff logfq logff Sun
CII 9 49 54 112 8 28 20 43 0.5
CIV 20 104 72 194 13 71 35 105 0.6
NV 2 15 8 40 9 18 12 22 0.1
cont ---­ 4460 ---­ ---­ ---­ ---­ 270 490 ---
1 Doyle et al. 1989; 2 Linsky et al. 1989; 3 Busazi et al. 1987; 4 Baliunas et al. 1984
3.2 RSCVn flaring emission
As with their quiescent emission, flaring emission in RSCVn`s is qualitatively
the same as in the dMe stars but RSCVn flares are considerably more energetic.
This greater energy arises both from a higher specific flux during the flare and
from a much longer flare duration. Indeed RSCVn flares may last up to sizeable
fractions of a day, arguing in favour of continual heating on these time scales.
As a result IUE has been able, in some cases, to observe the time evolution of
RSCVn flares in ultraviolet lines (cf. Fig. 2).
Use of LW HIRES spectra to study the behaviour of the MgII h&k resonance
lines during RSCVn flares has given important indications of long­lived mass
flows in material at chromospheric temperatures. In general, as in optical lines,
these velocities are seen as red­shifts, but blue­shifts are also recorded (Doyle
et al. 1988, 1989; Linsky et al. 1989). However, reports of representative velocities
derived from these observations should be treated with caution. The MgII h&k
lines are optically thick and, as a result, there is not a simple mapping of velocity
onto the line profile.
A few RS CVn`s have been monitored for flaring sufficiently often to allow
estimates to be made of the frequency and energy distribution of ultraviolet line
flares. For instance Mathioudakis et al. (1992), using earlier IUE data from a
variety of sources, estimated that the RSCVn system II Peg produced a flare of
integrated energy in transition region lines of ¸10 35 erg every 10 hours.
4 Physical parameters of flares
Under optically thin (coronal) conditions a simple relationship exists between the
observed line flux F line and a quantity known as the emission measure (EM ) of
the plasma, i.e.
F line = ¸(N 2
e dV ) = ¸ EM (1)
where the quantity ¸ contains a number of essentially atomic parameters such
as the abundance of the element concerned, the relative abundance of the ionic

6 P. Brendan Byrne
Fig. 3. An example IUE LW HIRES spectrum, in the region of the MgII h&k
chromospheric emission lines, of the RSCVn star, II Peg, in quiescence and while
flaring (P.B. Byrne, private communication). The solid line is the flare spectrum,
the dashed line is the quiescent spectrum and the dotted line is the difference.
Note the excess extended emission in the wings, especially the red wing.
species, etc. The validity of the coronal approximation has been tested by, for
instance, checking that the ratio of strong resonance doublets is in the ratio of
their statistical weights (see e.g. Byrne et al. 1987). Once the overall distribution
of EM has been determined from a number of well observed lines formed over
a range of temperature, it can be combined with a suitable radiative loss func­
tion to determine the total radiative losses, including unobserved lines, over the
temperature in question (Byrne et al. 1987, Mathioudakis et al. 1991).
It is also possible to determine local electron density, N e , in the transi­
tion region by observing density­sensitive inter­system line ratios, such as be­
tween CIII]–1176 š A, CIII]–1908 š A and SiIII]–1892 š A. In the dMe`s these lines
are extremely weak and the derived densities correspondingly uncertain (Math­
ioudakis et al. 1991). Furthermore, it is not known whether flare conditions are
constant during typical exposure times as explained above. However, values of
N e ¸10 10\Gamma10:5 cm \Gamma3 would appear to be indicated. Signals being much stronger
for the closest RSCVn`s, intersystem line fluxes can be measured with greater
confidence (Byrne et al. 1987). The long duration of RSCVn flares also gives us
greater confidence in the applicability of the resulting densities. Typical values
are N e ¸10 10:5\Gamma12 cm \Gamma3 .
Combining EM and N e yields the volume of the plasma at that temperature
via the relation
dV = F line =¸ N 2
e (2)

Stellar flares -- UV 7
which for dMe`s yields values for large flares of dV¸ 10 28\Gamma29 cm \Gamma3 and for
RSCVn flares of dV¸ 10 30\Gamma32 cm \Gamma3 . Assuming that such flares arise within mag­
netic loops, we can represent such loops as semi­circular in outline with radius
R loop and of constant circular cross­section of radius r ¸ 0:1R loop . N such loops
would have a volume, V loop , given by the following expression.
V loop ¸ 0:04ú 2 R 3
loop =N (3)
Equating this to derived volume leads to a typical loop dimension R loop (dMe) ¸
10 9\Gamma10 cm and dimension R loop (RS CV n) ¸ 10 10\Gamma11 cm Note that R loop is only
weakly dependent on N . Such dimensions, at least in the case of the dMe`s,
are remarkably solar­like. However, we should caution that such dimensions are
derived from the volume of material at transition region temperatures only.
It is possible, even likely, that the bulk of the flaring plasma may be at X­
ray temperatures. Whether the X­ray plasma occupies the same loops as the
transition region material is a matter of uncertainty, however, even in the solar
case.
5 HST flare observations
As pointed out above, HST has disadvantages (spectral coverage, Earth eclipses,
etc.) for flare observations, as well as obvious advantages (aperture, spectral
and temporal resolution, etc.). Several interesting results have been achieved
already. Maran et al. (1994) monitored the dMe flare star, AU Mic, spectroscop­
ically using GHRS in the wavelength region 1345--1375 š A without recording any
discrete flares. However, they found that the mean spectrum showed a strong
FeXXI–1354 š A coronal emission line, whose intensity is comparable to that ob­
served in solar flares. This result would support the conclusion from IUE data
that the most active dMe's flare almost continually (Butler et al. 1987, Byrne
et al. 1991).
Woodgate et al. (1992) also observed AU Mic with the GHRS on HST and
detected a UV line flare which resulted in a factor 7 increase in the flux of the
SiIII–1206 š A line on the blue wing of Lyff. During this flare they observed a red
asymmetry in the wing of Lyff which they attributed to recombining protons
from a proton beam.
Perhaps the most dramatic HST flare star result to date has been the obser­
vation of extended red wings on the CIV–1548/51 š A resonance lines during the
initial stages of a flare on the dMe flare star, AD Leo (Bookbinder et al. 1992).
Interpreted as velocities, these imply downflowing material at mid­transition
region temperatures of ¸2 000 km s \Gamma1 . Byrne (1993) discussed this result and
calculated that the downflowing material's kinetic energy represents ¸25 times
the energy radiated in the CIV lines. Furthermore, the observation of downflows
of this magnitude in the optically thin CIV lines lends credence to reports of
comparable velocities in optically thick chromospheric lines such as HI Balmer.
There is a clear implication that mass flows may be a common and important
energy sink in stellar flares.

8 P. Brendan Byrne
References
Baliunas, S.L., Guinan, E.F., Dupree, A.K. 1984, ApJ, 282, 733
Bookbinder, J., Walter, F., Brown, A. 1992, in ``Cool Stars, Stellar Systems and the
Sun'', eds. J. Bookbinder & M. Giampapa, ASP Conf. Ser., p. 27
Boggess, A. et al. 1978, Nature, 275, 372
Busazi, D.L., Ramsey, L.W., Huenemoerder, D.P. 1987, ApJ, 322, 353
Butler, C.J., Doyle, J.G., Andrews, A.D., Byrne, P.B., Linsky, J.L., Bornmann, P.L.,
Rodon`o, M., Pazzani, V., Simon, T. 1987, A&A, 174, 139
Byrne, P.B. 1993, in ``Physics of Solar and Stellar Coronae'', eds. J.L. Linsky & S.
Serio, Kluwer, p. 489
Byrne, P.B., McKay, D. 1990, A&A, 227, 490
Byrne, P.B., Doyle, J.G., Brown, A., Linsky, J.L., Rodon`o, M. 1987, A&A, 180, 172
Byrne, P.B., Agnew, D.J., Cutispoto, G., Kilkenny, D.W., Neff, J.E., Panagi, P.M.
1991, in ``Surface Inhomogeneities on Late­type Stars'', eds. P.B. Byrne & D.J.
Mullan, Springer, 255
Doyle, J.G., Butler, C.J., Byrne, P.B., van den Oord, G.H.J. 1988, A&A, 193, 229
Doyle, J.G., Byrne, P.B., van den Oord, G.H.J. 1989, A&A, 224, 153
Giacconi, R. et al. 1982, in ``The Space Telescope'', Space Telescope Institute, Baltimore
Linsky, J.L., Neff, J.E., Brown, A., Gross, B.D., Simon, T., Andrews, A.D., Rodon'o,
M., Feldman, P.A. 1989, A&A, 211, 173
Maran, S.P., Robinson, R.D., Shore, S.N., Brosius, J.W., Carpenter, K.G., Woodgate,
B.E., Linsky, J.L., Brown, A., Byrne, P.B., Kundu, M.R., White, S., Brandt, J.C.,
Shine, R.A., Walter, F.M. 1994, ApJ, 421, 800
Mathioudakis, M., Doyle, J.G., Rodon`o, M., Gibson, D.M., Byrne, P.B., Avgoloupis,
S., Linsky, J.L., Gary, D., Mavridis, L.N., Varvoglis, P. 1991, A&A, 244, 155
Mathioudakis, M., Doyle, J.G., Avgoloupis, S., Mavridis, L.N., Seiradakis, J.H. 1992,
MNRAS, 255, 48
Phillips, K.J.H., Bromage, G.E., Doyle, J.G. 1992, ApJ, 385, 731
Woodgate, B.E., Robinson, R.D., Carpenter, K.G., Maran, S.P., Shore, S.N. 1992, ApJ,
397, L95
This article was processed using the L a T E X macro package with LMAMULT style