Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://star.arm.ac.uk/~ambn/pjaact.ps Äàòà èçìåíåíèÿ: Thu Aug 1 19:32:49 1996 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 21:12:17 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: ultraviolet

A&A manuscript no.
(will be inserted by hand later)
Your thesaurus codes are:
08(08.01.2; 08.06.3; 08.12.1)
ASTRONOMY
AND
ASTROPHYSICS
30.7.1996
Effect of chromospheric activity on the mean colours of
late­type stars
P.J. Amado and P.B. Byrne
Armagh Observatory, College Hill, Armagh BT61 9DG, N. Ireland
Received; accepted
Abstract. In this paper we show, for the first time, the
effect of activity on the (U --B) 0 vs. (B--V ) 0 colour­colour
diagram for a large sample of late type stars. We clearly see
that for single­lined spectroscopic RSCVn systems there
exists an ultraviolet excess over that recorded for quies­
cent stars of the same spectral type. The same effect is seen
in main­sequence stars but to a lesser extent. From simple
arguments we demonstrate that the cause is unlikely to be
unresolved flaring, the combined effect of strong chromo­
spheric emission lines, or X­ray/UV back­heating of the
photosphere. White­light faculae are a reasonable candi­
date for the effect however. In our investigation we have
not taken into account the effect that metallicity might
have on the colour indices.
Key words: stars: activity -- stars: fundamental parame­
ters -- stars: late­type
1. Introduction
The solar chromosphere and corona are usually viewed
as regions of the solar atmosphere whose physical condi­
tions are controlled by magnetic heating processes orig­
inating below the photosphere. The energy content and
gas density of these outer regions are orders of magni­
tude lower than those prevailing in the underlying photo­
sphere. Thus photospheric processes determine the mag­
netic configurations which control both the heating of the
corona and its morphology. To a very close approximation
the chromosphere and corona have no effect on the so­
lar photosphere. Although the same processes which heat
the outer layers may give rise to heating effects visible at
near­photospheric levels (e.g. white­light faculae).
Chromospherically active late­type stars exhibit all of
the panoply of solar activity but on scales several or­
ders of magnitude larger than solar. The solar paradigm
Send offprint requests to: P.J. Amado
is widely (and successfully) adopted in interpreting such
phenomena. It is implicitly assumed that the stellar pho­
tosphere underlying these outer atmospheric effects may
be regarded as similar to that of a quiescent star of similar
spectral type, at least in a global sense.
Excesses in the mean global blue colours of some young
open cluster stars have been detected for a some time
now. Turner (1979) measured an UV excess in the mem­
bers of the young (ú70 Myr) Pleiades open cluster with
(B--V ) colour between 0:5 Ÿ B--V Ÿ 0:8, of 0.03 magni­
tudes with respect to the Hyades members (ú800 Myr).
He explained this as being due to an effect of metallicity,
with the Pleiades members being closer in metallicity to
average nearby stars and lower than that of the Hyades
members.
For later Pleiades objects, Burton (1972) explained the
position of members with 1:1 Ÿ B--V Ÿ 1:4 below the
ZAMS as due to gray extinction from circumstellar dust
shells. Later, Stauffer (1980) ruled out the dust shell hy­
pothesis and, in his paper, where he examined the ``turn­
on'' point of pre­main sequence objects in the Pleiades, ex­
plained this ultraviolet excess (detected as well by Landolt
(1979)) as due to flaring. He reasoned that, since many
of these stars are flare stars, this might affect the blue
colours, making them look bluer and, thus, producing the
excess.
This UV excess has been observed in some other types
of stars. For instance, the excess in the blue colours of the
RS CVn single­lined spectroscopic binary HU Virgo would
be fitted if a hot companion (early F) were responsible of
affecting the colours (Cutispoto 1996). However, Fekel et
al. (1986) did not find any evidence of a hot secondary
component in their IUE spectra of HU Vir. Strassmeier
(1994) confirmed the K0IV spectral classification for the
primary and did not detect any sign of a secondary.
So, up to now, there have been a few attempts at ex­
plaining this excess, which we call here UV excess, seen in
the blue colours of some stars. In this paper we examine
several classes of chromospherically active late­type stars

2 P.J. Amado et al.: Effect of chromospheric activity on the mean colours of late­type stars
for evidence of an effect of activity on the mean global
photospheric colours.
2. The sample
The sample of active and inactive stars was taken from
several sources in the literature. We studied for all of them
the necessity of correction for interstellar extinction. The
following criteria were applied. All those stars with dis­
tances d Ÿ100 pc were not corrected. It is generally agreed
that, for stars nearer than 100 pc, there is no need of any
correction for interstellar extinction, since this correction
would be very small (EB\GammaV ! 0:04 in the direction of
the galactic center at d = 100 pc). Furthermore, if any
small correction were needed, the direction of the correc­
tion would be almost parallel to the zero­reddening curve
in the colour­colour diagram. Stars with galactic latitude
jbj ? 20 ffi and d ? 100 pc were corrected using the ex­
pression ¯
EB\GammaV = 0:06 cosec(jbj) \Gamma 0:06 derived by Woltjer
(1975) for globular clusters outside the galactic absorbing
layer, putting upper limits to the correction. Stars with
jbj ! 20 ffi and d ? 100 pc were corrected using the ex­
pression A v = 0:14 (1 \Gamma exp (\Gamma10d sin(jbj)))=sin(jbj) (Van
Herck 1965) where d is in kpc. We take A v =EB\GammaV =3.2
and EU \GammaB =EB\GammaV =0.72 (Crawford & Mandwewala 1976).
We also took special care with the corrections for those
stars lying in directions of specially high extinction
(Deutschman et al. 1976).
2.1. Active stars
The sample of active stars was taken from the catalogue
of ``Chromospherically Active Binary Stars'' (CABS) by
Strassmeier et al. (1993) and from the papers by Leggett
(1992), Houdebine et al. (1996), Doyle (1996) and Math­
ioudakis et al. (1995).
The stars selected from the CABS catalogue are all
single­lined spectroscopic binaries. We imposed this con­
dition to prevent the UBV colours of the observed compo­
nent being affected by a companion of different spectral
type. CABS also provides mean colours in the standard
UBV system, which were then corrected for interstellar
extinction as explained above. This subsample contains
stars of three different luminosity classes, namely, giant,
subgiant and dwarf stars.
The subsample of stars from Leggett (1992) and
Houdebine et al. (1996) are all main sequence stars with
spectral types later than M0. We selected as active those
stars which Leggett classified as dMe or flare stars or
which showed variability, and those from Houdebine et al.
(1996) with Hff in emission. Most of these objects are
from the Gliese catalogue (Gliese and Jahreiss 1991) which
means that they have d ! 25 pc and therefore not needing
any correction for interstellar absorption.
Mathioudakis et al.'s (1995) stars are all main se­
quence stars of spectral type mid­F to mid­M. They
present in their paper the observed fluxes in the EUVE
(Bowyer et al. 1994) Lex/B and Al/Ti/C bands, and de­
tection in both was the criterion adopted for including
them in the sample of active stars.
Doyle (1996) uses in his paper the flux of the C iv
–1550 š A doublet observed by the IUE satellite (Boggess
et al. 1978) to derive the total radiative output from the
chromospheric/coronal plasma for dwarfs between early F
and middle M. We used the flux of the C iv line in the
IUE LORES spectra of these stars to measure the level
of activity. Stars with a clear detection of the line were
taken as active. None of them needed any correction for
interstellar absorption since their distances were all much
less than 100 pc.
2.2. Inactive stars
The sample of inactive giants is the same as that used
by Amado and Byrne (1996) for the derivation of their
colour­T eff and colour­surface brightness relationships for
late­type stars.
Inactive subgiants with d !100 pc were taken from the
Bright Star Catalog (Hoffleit 1982) avoiding those that
were referred to as binary stars or variable stars.
The sample of ``inactive'' main­sequence stars was se­
lected as follows: the less active dM stars from Leggett
(1992), those from Houdebine et al. (1996) with Hff in ab­
sorption, the upper limits in the IUE C iv line flux from
Doyle (1996) and the 3oe levels in the EUVE Al/Ti/C band
in the paper by Mathioudakis et al. (1995). We note that,
since these criteria are detector based some residual active
stars may be included in the sample.
3. Results
In Fig. 1, we plot a two­colour diagram (U \Gamma B) 0 against
(B \Gamma V ) 0 for the sample of active and inactive stars, to­
gether with the zero­reddening curves of Schmidt­Kaler
(1982), for dwarfs, subgiants and giants.
Giant stars extend over a range in spectral type from
G0 ((B--V ) 0 = 0:65) to M6 ((B--V ) 0 = 1:52), although the
active giants do not extend over the full range. The ac­
tive and inactive giants overlap in the range between K0
((B--V ) 0 = 1:00) and K4 ((B--V ) 0 = 1:39). Subgiant stars
range in (B--V ) 0 from 0.5 to 1.2 with again a slightly more
restricted range for the active stars. The main­sequence
objects are distributed over all the values of (B--V ) 0 in
the plot.
Figure 2 shows the difference between the unreddened
(U --B) 0 colour and the intrinsic (U --B) 0 colour linearly
interpolated from the tables of Schmidt­Kaler (1982) by
using (B--V ) 0 for each star. This plot shows that the ma­
jority of the active giants (89:50%) and all the subgiants
are above the mean zero­reddening curves, thus showing a
positive UV excess ffi (U --B). The non­active giants, on the

P.J. Amado et al.: Effect of chromospheric activity on the mean colours of late­type stars 3
Fig. 1. (U --B)0 vs. (B--V )0 colour­colour diagram for active
(solid symbols) and inactive stars (open symbols). Giants are
represented by circles, subgiants by squares and dwarfs by tri­
angles.
other hand, lie out on the line with a very small scatter
(!0.1).
This UV excess is not so evident in the case of main­
sequence objects, although a larger number (66:67%) lie
above the mean line than below (31:53%). For the inactive
dwarfs, the percentages are 47:24% above the curve and
49:61% below it. In Table 1, we give some statistics on the
excess.
4. Discussion
The question that immediately arises from this investiga­
tion is, what produces this UV excess?. We will consider
next the following possibilities:
ffl Flaring
ffl Chromospheric emission
ffl X­ray back­heating
ffl Faculae
4.1. Flaring
Flares on active late­type stars show many similar charac­
teristics to those on the Sun but are orders of magnitude
more energetic and frequent. They show flux enhance­
ment over the quiescent state at almost all wavelengths.
M dwarf flare (UV Cet) stars are characterized by ``white­
light'' emission, i.e. a continuum intensity increasing very
strongly towards the blue. The possibility then arises that
Fig. 2. Ultraviolet excess ffi(U --B) vs. (B--V )0 diagrams for a)
class III stars, b) class IV stars, c) active class V stars and d)
inactive class V stars. Symbols are as in Fig. 1.
the mean light of individually unresolved flares of this type
might produce the observed blue excess in active stars.
Lacy et al. (1976) analyzed statistically 386 flares of
eight UV Cet flare stars observed by Moffett (1974), ob­
taining mean values for the flare energy through the U
passband (E), the number of flares per interval of time
(N ) and the mean, time averaged, rate of energy loss due

4 P.J. Amado et al.: Effect of chromospheric activity on the mean colours of late­type stars
Table 1. Mean ultraviolet excess ¯ ffi (U\GammaB) , with their standard
deviations, oe, for the samples of active and inactive giant, sub­
giant and dwarf stars. In the last two columns we give the
maximum value of the deviation and the number of stars in­
cluded in the subsamples.
Lum. ¯
ffi (U\GammaB) oe ffi max n
III 0.1578 0.1193 +0.4679 38
active IV 0.1344 0.0597 +0.1942 7
V 0.0421 0.0872 +0.4921 111
III 0.0055 0.0405 +0.0706 24
inactive IV 0.0070 0.0463 +0.0816 42
V 0.0065 0.0872 \Gamma0:3100 127
to flaring (L \Lambda ). Three of these UV Ceti stars, namely, YZ
CMi, AD Leo and EV Lac were included in our sample
of dMe stars. So we chose them in order to compare their
extra­luminosity in the U passband necessary to match
their observed UV excesses with those ones, L \Lambda , produced
by observed flares in the work of Lacy et al. (1976). The
result is that the UV excess luminosities, that are between
10 28 --10 29 erg s \Gamma1 , are two order of magnitudes larger than
the flare luminosities, and closer to the X­ray luminosities
measured in these three stars.
Lacy et al. (1976), however, realized that their ob­
served flares were only the energetic end of the spectrum
of flare energies from these stars. They found that the dis­
tribution of flare energies could be well represented by a
power law relationship between flare energy and frequency,
log š = ff+ fi log EU . Integrating this expression we obtain
for the time­averaged energy per second emitted by flares
in the U photometric passband
L 0 = 10 ff fi
fi + 1 (E fi+1
min \Gamma E fi+1
max )
If we take E min = 0:0 and Emax equal to the energy
of the detection threshold for each of the three stars men­
tioned above, we will obtain an upper limit for the energy
of flares below the detection limit. In all three cases we
obtain luminosities between 10 26 --10 27 erg s \Gamma1 , still two
order of magnitudes below those necessary to produce the
observed excesses. Thus, we conclude that small­scale flar­
ing is unlikely to account for the observed effect.
4.2. Chromospheric emission
Ultraviolet emission lines originating in the chromospheres
and transition regions of active stars are very prominent in
their spectra. The presence of these strong emission lines
in the wavelength range covered by the Johnson U pass­
band would produce an increase in the flux in active stars
to inactive stars where no such emission would be present.
Such an effect would give rise to a shift towards the blue
of the (U --B) colour of the star, appearing consequently
as an UV excess.
The Johnson U passband extends from 3050--4200 š A
peaking near 3600 š A. Included in this spectral region are
the following chromospheric lines: Ca ii H&K –3934/69 š A
and many of the H i Balmer series lines from the series
limit (–3640 š A) to Hffi(–4101 š A). Pettersen and Hawley
(1989) give surface fluxes for many of these lines and these
results demonstrate that, for the late­K and M dwarf stars
the Ca ii lines are the strongest emitters in this part of
the spectrum. (Note that, at their spectral resolution Hffl
is blended with Ca ii H.) Their measured Ca ii line surface
fluxes are 10 5 --10 6 erg s \Gamma1 cm \Gamma2 for the most active dKe
and dMe stars. These fluxes translate to luminosities be­
tween 10 26 --10 27 , again two orders of magnitudes below
those necessary to produce the observed excesses. Thus
we conclude that the effect of the major chromospheric
emission lines is unlikely to account for the observed ef­
fect. We cannot rule out however, a contribution from a
distribution of a large number of metallic emission lines
spread throughout the spectral region.
4.3. X­ray back­heating
Since the effect we observe is clearly associated with chro­
mospheric and coronal activity, we have sought indications
of correlation between the observed excess and some mea­
sure of this activity in individual stars. This is done in
Fig. 3 and Fig. 4 where each star's (U --B) excess is plot­
ted against its X­ray and Mg ii luminosity, normalized to
bolometric luminosity (or surface area in the case of the
giants). It is immediately apparent that no such correla­
tion exists, the correlation coefficients being, for Fig. 3 a),
0.25786, and b), 0.24252 and for Fig. 4, 0.31912.
Furthermore, as noted above, the energy requirements
of the observed excess in the U band would demand heat­
ing at a rate at least equal to the total X­ray luminosity
of the star, implying 100% back­heating efficiency.
4.4. Faculae
In the Sun, faculae are patches of hot material (hotter by
about 300 K from the surrounding photosphere (Phillips
1992)) within active regions. They are composed of chains
or clusters of tiny bright points -- facular bright points --
very similar in brightness, size and lifetimes to the network
bright points in quiet areas. Plages are the extension of the
photospheric faculae into the chromosphere.
In active stars, where the filling factors of active re­
gions may approach unity, the higher brightness and the
hotter temperature of faculae and plages over those of the
photosphere would affect the colours of the star shifting

P.J. Amado et al.: Effect of chromospheric activity on the mean colours of late­type stars 5
Fig. 3. UV excess for active dwarfs vs. a) the logarithm of the X­ray luminosity over the bolometric luminosity of the star, b)
the logarithm of the Mg ii luminosity over the bolometric luminosity of the star. Arrows mean upper limits.
Table 2. Effective temperatures, fluxes and filling factors for three dMe stars
Name T ph (K) \DeltaF U F q FU = F q + \DeltaF U F fac f (%)
YZ CMi 3150 6:016 10 6 2:673 10 7 3:275 10 7 7:618 10 7 12.2
AD Leo 3450 1:040 10 8 7:618 10 7 1:802 10 8 1:847 10 8 95.8
EV Lac 3300 1:097 10 8 4:618 10 7 1:559 10 8 1:207 10 8 147.2
AD Leo 3:100 10 8 44.5
EV Lac 2:112 10 8 66.5
them towards the blue, thus producing an UV excess. This
effect, of course, will be more enhanced for stars with lower
effective temperature for which the contrast between the
faculae and the photosphere will be larger.
To quantify the effect of faculae on the average colours
of the stars, let us assume a temperature for the facula of
T fac = T ph + 300 K, where T ph is the effective tempera­
ture of the quiet photosphere and we assume black body
flux distributions for both active and quiet photospheres
(although these are likely to be far from the actual flux
distributions, they will suffice for an order of magnitude
estimate). We will take as examples for the dwarf stars the
three dMe active stars mentioned above, namely, YZ CMi,
AD Leo and EV Lac. In Table 2, we give all the param­
eters used in the calculations of the surface fluxes emit­
ted by the quiet photosphere (F q ) and the facula (F fac )
in these stars. The surface flux in the U band that we
would observe (FU ) would be the sum of the UV excess
we actually see for these stars (\DeltaF U ) plus F q . Making
use of the equation FU = fF fac + (1 \Gamma f)F q , we can de­
rive the filling factors for the active region necessary to
produce the observed UV excess. Those filling factors are
also given in Table 2, where in the last two rows, we also
set T fac = T ph + 500 K for the two more active stars. We
note that the required filling factors derived are substan­
tial fractions of the entire stellar surface.
Saar & Linsky (1985) found from Zeeman splitting in
the infrared spectrum of the dMe star AD Leo that ac­
tive regions covered 73%\Sigma6% of the surface of that star
with a mean field strength of B = 3800 \Sigma 260G. Since the
star did not show any evidence of flares at the time of
the observation they concluded that this value probably

6 P.J. Amado et al.: Effect of chromospheric activity on the mean colours of late­type stars
represented the quiescent magnetic flux level for AD Leo.
Utilizing the same technique for an optical spectrum of the
dMe star EV Lac, Johns­Krull & Valenti (1996) measured
magnetic fields of B = 3800 \Sigma 500 G covering 50%\Sigma13%
of its surface. We note that the implied magnetic filling
factors are similar to the figures we derive from the above
crude estimates based on black­body faculae.
Fig. 4. UV excess for active giants vs. the logarithm of the
X­ray luminosity over the squared radius of the star. Arrows
mean upper limits.
Moreover, if we take into account the effect of contrast
between the facula and the surrounding photosphere of
the star, we should be able to predict that the farther we
move towards later spectral types, the larger the excess
should become (if the temperature contrast is similar in
different spectral types). We can see exactly the expected
effect in Fig. 5, where we plot the excess in the (U --B)
colour against (B--V ) 0 for active giant stars. The solid
lines represent the effect of the contrast between the fac­
ular component and the photosphere assuming a constant
excess flux in the U passband. In other words, we took the
star labelled as CABS 16 (star number 16 in the CABS
catalog), which is a giant star with (B--V ) 0 = 1:15 or spec­
tral type K0, and calculated the extra­flux in the U band
neccesary to produce the excess we observe in (U --B). As­
suming this same constant extra­flux as the effect of activ­
ity on a normal star of spectral type G5, we produced an
excess in the (U --B) colour that positions the star at the
end of the solid line, i.e., with a ffi (U --B) = 0:21. We did
the same calculations for the star labelled as CABS 27,
where the ``parent'' star is of a spectral type G1 and the
Fig. 5. UV excess for active giants vs. (B--V )0 colour index.
The solid lines represent the effect of contrast between the
facula component and the photosphere.
``descendant'' star is a K2 giant star, and for another two
more stars. We can see that the tracks from CABS 16 and
CABS 27 envelope almost all the stars in the sample of ac­
tive giants, with the tracks from the other two stars falling
within the previous two. This indicates that a contrast ef­
fect could play an important role in the decrease of the
observed UV excess towards the earlier spectral types for
active giants. In the case of the dwarfs, we saw in Fig. 2
(the third panel from the top) the same trend, i.e., the
earlier the spectral type the smaller the excess, but it is
not so clear.
5. Conclusions
For the first time, we clearly show that activity affects
the overall properties of the photosphere of a star in its
intrinsic mean colours. We have demonstrated that the
UV excess can be due to a contribution to the overall
stellar colours of a facular component on the surface of
these stars.
In our discussion here we have not, however, taken
into account the effect that metallicity might have on the
colour indices. Estimating such effects is beyond the scope
of the present paper but would warrant a further investi­
gation, perhaps to be undertaken spectroscopically.
Acknowledgements. Research at Armagh Observatory is sup­
ported by a Grant­in­Aid from the Department of Education
for Northern Ireland. PJA acknowledges support of a Post­
graduate Research Assistantship from the Armagh Observa­

P.J. Amado et al.: Effect of chromospheric activity on the mean colours of late­type stars 7
tory. This research has made use of the SIMBAD database,
operated at CDS, Strasbourg, France.
References
Amado P.J., Byrne P.B., 1996, Ir. Astron. J. 23, 177
Boggess A., Carr F.A., Evans D.C., et al., 1978, Nature 275,
372
Bowyer S., Lieu R., Lampton M., et al., 1994, ApJS 93, 569
Burton F.J., 1972, ApJL 171, 57
Cutispoto G., 1996, A&AS in press
Crawford D. L., Mandwewala N., 1976, PASP 88, 917
Deutschman W.A., Davis R.J., Schild R.E., 1976, ApJS 30, 97
Doyle J.G., 1996, A&A 307, L45
Fekel F.C., Moffet T.J., Henry G.W., 1986, ApJS 60, 551
Gliese W., Jahreiss H. 1991, ``The Third Catalogue of Nearby
Stars''
Hoffleit, D. 1982, ``The Bright Star Catalogue''
Houdebine E.R., Mathioudakis M., Doyle J.G., Foing B.H.,
1996, A&A 305, 209
Johns­Krull C.M., Valenti J.A., 1996, ApJ 459, L95
Lacy C.H., Moffett T.J., Evans D.S., 1976, ApJS 30, 85
Landolt A.U., 1979, ApJ 231, 468
Leggett S.K., 1992, ApJS 82, 351
Mathioudakis M., Fruscione A., Drake J.J. et al., 1995, A&A
300, 775
Moffett T.J., 1974, ApJS 273, 29
Phillips K.J., 1992, Guide to the Sun. Cambridge Univ. Press,
Cambridge
Pettersen B.R., Hawley S.L., 1989, A&A 217, 187
Saar S.H., Linsky J.L., 1985, ApJ 299, L47
Schmidt­Kaler T., 1982. In: Schaifers K. and Voigt H.H.(eds.)
Landolt­B¨ornstein, Vol 2b. Springer, Heidelberg, p. 14
Stauffer J.R., 1980, AJ 85, 1341
Strassmeier K.G., Hall D.S., Fekel F.C., Scheck M., 1993,
A&AS 100, 173
Strassmeier K.G., 1994, A&A 281, 395
Turner D.G., 1979, PASP 91, 642
Van Herck G., 1965, Bull. Astron. Ins. Netherlands 18, 71
Woltjer L., 1975, A&A 42, 109
This article was processed by the author using Springer­Verlag
L A T E X A&A style file L­AA version 3.