Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://star.arm.ac.uk/~ambn/kam.ps Äàòà èçìåíåíèÿ: Tue Nov 19 17:52:15 1996 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 20:16:01 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: ngc 253

A&A manuscript no.
(will be inserted by hand later)
Your thesaurus codes are:
04.19.1, 08.01.2, 08.02.1, 08.03.5
ASTRONOMY
AND
ASTROPHYSICS
11.11.1996
EUV Emission from RS CVn binaries
C.K. Mitrou 1;2 , M. Mathioudakis 1;3 , J.G. Doyle 1 , and E. Antonopoulou 2
1 Armagh Observatory, College Hill, Armagh BT61 9DG, N. Ireland
email: kam@star.arm.ac.uk & mm@star.arm.ac.uk & jgd@star.arm.ac.uk
2 Sect. of Astrophysics, Astronomy and Mechanics, Dept. of Physics, Univ. of Athens, Athens 15783, Greece
email: kam@rigel.da.uoa.gr & eantonop@rigel.da.uoa.gr
3 current: Dept. of Pure & Applied Phys, Queens University Belfast, Belfast BT7 1NN, N. Ireland
email: M.Mathioudakis@Queens­Belfast.AC.UK
received date, accepted date
Abstract. We performed a study of 104 RS CVn systems
in the extreme ultraviolet (EUV) using the all--sky sur­
vey data obtained by the Extreme Ultraviolet Explorer
(EUVE). The present sample includes several new RS
CVn detections; 11 more than in the published EUVE
catalogs, and 8 more than in the ROSAT Wide Field Cam­
era catalog. The ratio of detections to non--detections re­
mained constant throughout the sky, implying that our de­
tections are not limited by the exposure time but are most
likely limited by absorption from the interstellar medium.
A general trend of increasing Lex/B (50--180 š A) flux with
decreasing rotational period is clear. The dwarf systems
exhibit a leveling­off for the faster rotators. In contrast,
the evolved systems exhibit no such effect. For the RS
CVn systems the losses in the EUV represent a smaller
fraction of the coronal radiative losses, as compared to
active late--type dwarfs.
Key words: RS CVn binaries -- extreme ultraviolet
(EUV) -- coronal activity -- interstellar hydrogen
1. Introduction
The RS CanumVenaticorum binaries typically consist of
a late--type (G or K) chromospherically active and usually
evolved star, forced into faster rotation than its age and
evolutionary status implies by the presence of a companion
which may (or may not) be active as well. The phenomena
associated with these stars originate from high magnetic
activity due to very active dynamos operating as a result
of deep convective zones and fast rotation. The study of
this activity in the UV, EUV and X--ray part of the spec­
trum has become possible with the launch of the IUE,
Send offprint requests to: C.K. Mitrou
and subsequently EXOSAT, Einstein, HST, ROSAT, and
EUVE satellites. The RS CVn systems occupy a promi­
nent position among the late--type stars insofar as they
exhibit the highest luminosities in the X--ray part of the
spectrum (Dempsey et al. 1993b), indicating that they
possess very active coronae. As a result of their detectabil­
ity, there have been many papers dealing specifically with
their behaviour in the X--ray and more recently, the EUV.
Among these studies, the most relevant to this present
work are the ROSAT--based papers by Pye et al. (1995)
and Dempsey et al. (1993b).
The EUV region is an important part of the electro­
magnetic spectrum providing numerous hot coronal lines,
some of which are sensitive to electron density variations.
However, its overall contribution to the radiative losses
from the stellar atmosphere is largely unknown. In the
present study, we use broad­band data collected with the
EUVE satellite (Bowyer et al. 1994) during an all--sky
survey lasting six months up to July 1993. The survey
was conducted in four bands, namely Lex/B (50--180 š A),
Al/Ti/C (160--240 š A), Ti/Sb/Al (345--605 š A) and Sn/SiO
(500--740 š A), with peak sensitivities at approximately 100,
180, 400 and 550 š A respectively. The primary objectives
of our work are to: (i) identify new RS CVn systems with
detectable EUV emission, (ii) determine the EUV activity
levels of all RS CVn binaries and (iii) investigate the valid­
ity of correlations between the EUV radiative losses and
other parameters. Our findings are compared to previous
studies based on ROSAT data.
2. Observations and data reduction
2.1. Count rates
The sample of stars investigated consists of the 104 sys­
tems classified as RS CVn­type binaries in the Strassmeier
et al. (1988) Catalog of Chromospherically Active Binary
Stars (CCABS). Their distribution in the sky is shown

2 C.K. Mitrou et al.: EUV Emission from RS CVn binaries
on an Aitoff projection in ecliptic coordinates in Fig. 1,
with the detections depicted as filled circles. Due to the
geometry of the EUVE survey, the ecliptic poles received
maximum exposure time with minimum exposure time at
the ecliptic equator. Note however, that the ratio of detec­
tions to non--detections remains constant throughout the
sky, implying that our detections are not limited by the
exposure time, but are more likely limited by absorption
from the interstellar medium.
The source count rates were determined by applying a
maximum likelihood method to a group of photon events
collected from a circular area with a radius of 24 arc min­
utes centered on the optical position of the source. Here
we examine the Lex/B and the Al/Ti/C bands. In both fil­
ters we consider significant all the detections with a max­
imum likelihood significance greater than 9 corresponding
to 3oe. Detector background dominates the Lex/B back­
ground, while the most important source of background
in Al/Ti/C is due to the HeII 304 š A geocoronal emission.
For weak detections the main source of uncertainty in the
count rates comes from counting statistics, whereas for
stronger detections the principal source of uncertainty is
introduced by vignetting effects.
To obtain our count rates we have used the full EUVE
sky survey dataset, the latest vignetting maps, and for
the derivation of fluxes as described in Sect. 3, the most
up--to--date effective areas of the EUVE photometers. The
results are shown in Table 1 for the valid (– 3oe) detec­
tions. For the non--detections, 3oe upper limits are given
in Table 2. We found 38 detections in the Lex/B, and 11
in the Al/Ti/C band. For the detected sources the coinci­
dence between the EUVE and optical positions is better
than 60 00
. If a discrepancy of more than 60 00
between the
two positions was found, then the source was rejected as
spurious and an upper limit was determined. At this point
we emphasize that, given the size of our sample and based
on a random sample test, (see Fruscione (1996) for a de­
tailed description), we would expect 2 spurious detections
in each band at the 3oe level.
The initial EUVE Bright Source List (BSL) by Ma­
lina et al. (1994) lists 27 RS CVn detections in the Lex/B
band and 4 in the Al/Ti/C. The subsequent EUVE Bright
Source Catalog (BSC) by Bowyer et al. (1994) includes
21 Lex/B and 7 Al/Ti/C detections in the main table.
The remaining six BSL detections did not meet the BSC's
criteria for inclusion in the main table and were listed
separately. The Al/Ti/C counts quoted in the BSC for
096 = AR Lac and 104 = II Peg were obtained by Pat­
terer et al. (1993), when using the two systems as EUVE
in--flight calibration targets. Our detections (Table 1) in­
clude all of BSC and BSL's detections in both bands, ex­
cept for 104 = II Peg and 096 = AR Lac in the Al/Ti/C
band where we only quote upper limits, that are 2.4
and 7.4 times higher respectively than the Patterer et al.
(1993) values. The RS CVn systems 027 = TZ Pic and
094 = RT Lac were detected only in Al/Ti/C. We shall
return to these objects in Sect. 5.
Fig. 1. The RS CVn systems in Ecliptic coordinates (circles).
Filled symbols represent the sources with a 3oe detection.
Fig. 2. Our Lex/B count rates against the ROSAT WFC
(S1+S2) values; the symbols are as follows. Circles: systems
comprising of giant stars. Squares: subgiants. Triangles: dwarfs.
The source index numbers are explained in Table 1.
2.2. Comparison to the ROSAT results
During 1990--1991 the Wide Field Camera (WFC) on
board the ROSAT satellite conducted an all--sky survey
at EUV wavelengths, using two broadband filters, namely
S1 and S2. Filter S1 covered the range 60­140 š A, and S2 the

C.K. Mitrou et al.: EUV Emission from RS CVn binaries 3
Table 1. The 104 RS CVn systems in the Extreme Ultraviolet. This table contains the significant (i.e. – 3oe) detections in
Lex/B and/or Al/Ti/C. For the non­detections (3oe) upper limits (UL) are given. Columns 1 & 2 contain an index number
for each system and its most common name. The systems are numbered in increasing RA order. Column 3 gives the distance
in parsec followed by the corresponding hydrogen column density to the source in column 4. In column 5 we give the CIV
observed flux (our own reduction of IUE spectra) in units of 10 \Gamma13 erg cm \Gamma2 s \Gamma1 . Columns 6, 7, and 8 contain the data for the
Lex/B passband in the following order: the count rate and the 1oe statistical error in counts ks \Gamma1 , its significance (oe), and the
broadband flux in units of 10 \Gamma12 erg cm \Gamma2 s \Gamma1 assuming a coronal temperature model of log T = 6.8. In columns 9, 10 and 11
the corresponding values for the Al/Ti/C filter are given in the same units.
SOURCE NAME Dist. log NH fCIV Lex/B Al/Ti/C
(Index) Common or HD (pc) (cm \Gamma2 ) (10 \Gamma13 ) ct ks \Gamma1 oe f 6:8 ct ks \Gamma1 oe f 6:8
003 HD 3196 21 18.79 2.8 36\Sigma11 5 1.2 59\Sigma20 4 6.3
004 i And 31 19.81 18.1 37 \Sigma 5 10 7.2 16 \Sigma 5 3 11.8
005 CF Tuc 54 19.84 4.2 67 \Sigma 12 8 14.4 30 UL ­ 22.9
006 UV Psc + 125 19.09 ­ 19 \Sigma 7 3 0.8 56 UL ­ 21.7
008 HD 8358 60 18.73 1.4 33 \Sigma 9 5 1.1 58 UL ­ 5.5
009 AR Psc 17 18.73 6.4 157 \Sigma 16 17 5.1 40 \Sigma 13 4 3.8
011 6 Tri + 75 19.65 8.5 24 \Sigma 9 4 3.1 38 UL ­ 23.3
013 VY Ari 21 19.35 9.2 126 \Sigma 16 13 6.4 80 UL ­ 31.5
015 UX Ari 50 19.09 15.5 668 \Sigma 32 43 28.4 99 \Sigma 20 7 20.9
016 HR 1099 36 17.96 38.1 632 \Sigma 31 39 17.0 84 UL ­ 3.8
017 HD 22403 55 19.09 6.5 104 \Sigma 16 10 4.4 40 UL ­ 8.7
018 EI Eri 75 19.46 7.9 114 \Sigma 16 12 8.5 74 UL ­ 34.5
021 Capella 13 18.73 390.0 441 \Sigma 26 33 14.4 144 \Sigma 19 12 13.7
027 TZ Pic 60 17.96 1.1 12 UL ­ ­ 15 \Sigma 5 3 0.7
032 oe Gem 59 18.85 35.2 384 \Sigma 29 25 13.4 52 \Sigma 21 3 6.4
033 AE Lyn + 38 18.28 2.6 18 \Sigma 6 4 0.5 24 UL ­ 1.3
037 TY Pyx 55 19.06 4.5 59 \Sigma 11 9 2.4 70 UL ­ 14.0
040 IL Hya 138 18.99 4.0 32 \Sigma 10 4 1.2 44 UL ­ 7.5
042 DM UMa 130 19.49 1.0 15 \Sigma 7 3 1.2 44 UL ­ 22.4
043 ¸ UMa 7.9 18.00 8.3 140 \Sigma 18 13 3.8 88 UL ­ 4.0
047 DQ Leo 36 19.39 ­ 42 \Sigma 12 5 2.7 76 UL ­ 32.3
049 DK Dra + 130 19.60 4.6 29 \Sigma 8 5 3.0 36 UL ­ 20.8
051 IL Com 86 19.28 2.1 24 \Sigma 9 4 1.3 44 UL ­ 15.0
056 BH CVn 53 18.58 17.2 157 \Sigma 15 18 4.8 44 \Sigma 15 4 3.3
059 HD 127535 63 18.68 1.2 20 \Sigma 10 3 0.7 96 UL ­ 8.6
067 TZ CrB 21 18.56 18.1 472 \Sigma 21 44 14.3 90 \Sigma 15 8 6.6
069 ffl UMi + 71 19.44 8.58 15 \Sigma 4 5 1.1 12 UL ­ 5.5
071 V824 Ara 39 18.68 7.1 96 \Sigma 15 10 3.1 54 UL ­ 4.8
074 DR Dra 88 19.46 5.1 33 \Sigma 4 12 2.4 24 UL ­ 11.1
077 V772 Her 42 19.18 3.2 75 \Sigma 11 11 3.5 42 UL ­ 11.3
078 V815 Her 31 19.17 4.5 75 \Sigma 10 12 3.5 46 UL ­ 12.1
082 V478 Lyr 26 19.12 2.7 92 \Sigma 11 14 4.0 48 UL ­ 11.1
083 HR 7275 + 90 19.68 6.3 17 \Sigma 5 5 2.2 28 UL ­ 17.9
087 HD 190540 + 110 19.65 ­ 26 \Sigma 11 3 3.1 56 UL ­ 34.4
090 ER Vul 46 19.32 4.0 55 \Sigma 11 7 3.2 78 UL ­ 28.9
093 42 Cap + 34 18.60 4.3 41 \Sigma 11 6 1.3 50 UL ­ 3.9
094 RT Lac \Lambda 205 20.52 1.1 18 UL ­ ­ 43 \Sigma 16 3 92.5
096 AR Lac 47 18.73 13.9 107 \Sigma 11 16 3.5 52 UL ­ 5.1
102 – And 24 18.73 30.2 315 \Sigma 19 31 10.2 36 \Sigma 11 4 3.5
104 II Peg 29 18.73 9.1 165 \Sigma 18 17 5.4 72 UL ­ 7.0
+ Not detected by ROSAT's WFC
\Lambda Not detected by ROSAT's WFC, but our Al/Ti/C detection may be spurious

4 C.K. Mitrou et al.: EUV Emission from RS CVn binaries
Table 2. The RS CVn systems in the Extreme Ultraviolet. This table contains the systems that are not acceptable detections
(! 3oe), in any of the two filters. Columns 1 & 2 contain an index number for each system and its most common name. The
systems are numbered in increasing RA order. Column 3 gives the distance in parsec followed by the hydrogen column density
to the source in column 4. Columns 5 and 6 contain the data for the Lex/B passband, in the following order: the 3oe Upper
Limit (UL) and the corresponding flux upper limit in units of 10 \Gamma12 erg cm \Gamma2 s \Gamma1 for a temperature model at log T = 6.8. In
columns 7 and 8 similarly we give the Al/Ti/C filter 3oe UL count rate, and the corresponding flux value for log T = 6:8 in units
of 10 \Gamma11 erg cm \Gamma2 s \Gamma1 .
SOURCE NAME Dist. log NH Lex/B UL Al/Ti/C UL
(Index) Common or HD (pc) (cm \Gamma2 ) ct ks \Gamma1 f 6:8 ct ks \Gamma1 f 6:8
001 33 Psc 111 18.90 32 1.2 66 0.9
002 BD Cet 71 18.90 30 1.1 82 1.1
007 HD 8435 100 19.62 20 2.2 56 3.4
010 HD 12545 310 20.68 16 71.2 42 12.4
012 HD 14643 77 20.00 10 3.8 32 3.1
014 LX Per 130 20.79 32 197.0 22 7.8
019 RZ Eri 143 19.84 30 6.4 38 2.9
020 12 Cam 134 20.60 36 127.0 40 9.8
022 TW Lep \Lambda 220 19.65 28 3.3 50 3.1
023 HD 37824 164 20.19 44 37.6 50 6.3
024 HD 39576 85 18.21 28 0.8 48 0.2
025 SZ Pic 30 17.96 22 0.6 20 0.1
026 CQ Aur 220 20.76 28 150.0 42 14.2
028 SV Cam 74 19.49 28 2.2 12 0.6
029 VV Mon 380 20.02 16 6.9 30 3.0
030 SS Cam 255 20.31 10 13.6 16 2.4
031 AR Mon 525 20.12 32 20.6 52 5.8
034 GK Hya 220 19.99 32 12.1 78 7.2
035 RU Cnc 300 20.05 34 16.4 50 5.0
036 RZ Cnc 395 20.05 24 11.6 52 5.3
038 WY Cnc 160 19.85 22 4.9 46 3.6
039 XY UMa 100 19.28 26 1.4 24 0.8
041 IN Vel 500 20.13 30 20.1 34 3.9
044 HD 101309 62 17.79 16 0.4 34 0.1
045 HD 101379 140 20.16 24 18.3 36 4.3
046 RW UMa 150 19.72 16 2.3 34 2.3
048 HU Vir 220 19.40 26 1.7 52 2.3
050 AS Dra 29 17.40 30 0.8 28 0.1
052 UX Com 350 19.91 38 10.6 34 2.8
053 HD 113816 165 20.00 18 7.1 38 3.6
054 RS CVn 180 19.80 42 7.9 42 3.0
055 BM CVn 250 19.89 40 10.3 34 2.8
057 HD 119285 \Lambda 80 19.63 58 6.5 148 9.0
058 BH Vir 166 20.02 16 6.9 72 7.0
060 RV Lib 270 20.53 40 115.3 34 7.4
061 SS Boo 220 19.99 20 7.6 32 3.0
062 UV CrB 230 19.92 34 9.7 42 3.5
063 GX Lib 219 20.37 30 49.6 68 11.4
064 LS TrA 54 18.68 26 0.8 14 0.1
065 RT CrB 360 19.93 18 5.4 56 4.8
066 RS UMi 350 20.03 6 2.8 24 2.3
068 WW Dra 180 19.98 12 4.4 12 1.1
070 V792 Her 310 19.98 14 4.9 24 2.3
072 HR 6469 69 19.22 28 1.4 36 1.1
073 HD 158393 400 20.85 30 213.2 52 20.7
075 Z Her 100 20.37 22 37.6 50 8.4
076 MM Her 190 19.57 14 1.3 64 3.6
079 PW Her 285 19.90 12 3.3 36 3.0
080 AW Her 315 20.31 12 16.9 50 7.5

C.K. Mitrou et al.: EUV Emission from RS CVn binaries 5
Table 2 (cont.)
SOURCE NAME Dist. log NH Lex/B UL Al/Ti/C UL
(Index) Common or HD (pc) (cm \Gamma2 ) ct ks \Gamma1 f 6:8 ct ks \Gamma1 f 6:8
081 o Dra 67 19.21 12 0.6 20 0.6
084 HD 181809 210 20.59 46 161.0 60 14.8
085 HR 7428 302 20.80 20 118.7 20 7.6
086 HD 185151 390 20.50 20 52.9 42 8.6
088 AT Cap 99 20.20 66 58.8 38 4.8
089 CG Cyg 63 19.59 22 2.2 62 3.6
091 AS Cap 93 20.19 24 20.3 80 10.0
092 AD Cap 250 20.50 22 56.3 40 8.2
095 HK Lac 150 20.48 20 50.0 32 6.4
097 V350 Lac 69 19.44 22 1.6 30 1.4
098 IM Peg 50 19.40 56 3.7 54 2.3
099 RT And 95 19.97 12 4.4 20 1.8
100 SZ Psc \Lambda 125 19.92 54 15.7 34 2.9
101 EZ Peg 83 19.75 38 5.7 84 5.8
103 KT Peg 25 18.73 36 1.2 94 0.9
\Lambda Detected by ROSAT's WFC S1 band
range 112­200 š A. The range covered by S1+S2 is similar to
the EUVE Lex/B band (50--180 š A); note however the over­
lap of S1 and S2 between 112 and 140 š A. The WFC results
can be found in Pounds et al. (1993), Bright Source Cat­
alogue, and more recently, after an improved reprocessing
of the database, in the 2RE Source Catalogue (2RESC)
by Pye et al. (1995).
Out of our sample, 2RESC identifies 33 signifi­
cant detections. The systems not detected by the WFC
are flagged in Table 1. With the exception of the
three systems: 022 = TW Lep, 057 = HD 119285 and
100 = SZ Psc (flagged in Table 2) that were detected only
in the S1 filter of the WFC, all the remaining objects
were also detected by EUVE in the Lex/B band. The
count rates we found for our sample are in agreement
with the WFC count rates, ranging between 50 and 100%
of the S1+S2 value. Given the slightly different spectral
range, the differences in effective areas, the overlap be­
tween S1 and S2 and the possible variability of the sources,
we consider this as reasonable agreement. Strong emis­
sion lines of Fe ions are present in the spectra of RS CVn
stars in the overlapping wavelengths, e.g. Fe XXIII/Fe XX
132.85 š A, Fe XXII 135.78 š A, Fe XXII 117.17 š A and Fe XXI
128.73 š A (see Fig. 4 later). The counts for 021 = Capella
and 104 = II Peg, two strong detections, are found at the
lower end of this range (ú 50%). In Fig. 2 we show our
count rates plotted against the rates quoted in the 2RESC.
Notable differences from the ROSAT WFC val­
ues are derived for 015 = UX Ari, 056 = BH CVn,
074 = DR Dra, and 096 = AR Lac. Systems UX Ari and
BH CVn have EUVE count rates higher by factors of 2
and 3 times respectively. An initial timing analysis of the
EUVE data however, showed that the EUVE counts were
affected by flaring. DR Dra has 20% higher EUVE counts.
It comprises of a K0­2 giant and a hot white dwarf com­
panion (of T = 30; 000K according to Fekel et al. 1993),
so the WD's contribution in the EUV might explain this
discrepancy. The AR Lac counts though, are at 35% of
the ROSAT value. Patterer at al. (1993), report similar
findings when using AR Lac as an EUVE in--flight cali­
bration target, and attributed this deviation to long--term
EUV variability of the system. On the other hand, Mc­
Gale et al. (1995), who performed a timing analysis of the
ROSAT all--sky survey data, found temporal variability
for this system, and suggested the possibility of a flare.
Apart from the WFC, one more instrument on board
ROSAT, the Position Sensitive Proportional Counter
(PSPC) performed an all--sky survey. The PSPC covers
energies between 0.1 and 4 keV (or equivalently, the spec­
tral range 5--124 š A). A thorough study of a large sample
of RS CVn stars (112 detections out of 136 observed) is
presented in Dempsey et al. (1993b). All of our detec­
tions (as well as all the WFC detections) were detected by
the PSPC. In Fig. 3 we have plotted the Lex/B against
the PSPC count rates. The strikingly deviating systems
are the flaring 015 = UX Ari (see previous paragraph)
and 087 = HD190540, which appears overluminous in the
EUV (see also Sect. 4).
3. Derivation of fluxes
In order to derive fluxes from the count rates, the neutral
hydrogen column density along the line of sight (NH ) has
to be determined, and a specific coronal model adopted. In
the present study, only the Lex/B (50--180 š A) and Al/Ti/C
(160--240 š A) passbands are considered.

6 C.K. Mitrou et al.: EUV Emission from RS CVn binaries
Fig. 3. Our Lex/B count rates against the ROSAT PSPC val­
ues; the symbols are as in Fig. 2
3.1. Interstellar medium attenuation
We determined the values for the hydrogen column densi­
ties by using a model of the interstellar medium developed
by Jelinsky & Fruscione (1996). In order to estimate the
amount of hydrogen in any given direction and distance,
this method performs a three--dimensional interpolation
on a large database of HI column densities (Fruscione et
al. 1994). The ISM attenuation was calculated using the
hydrogen and helium photoionization cross sections com­
piled by Rumph et al. (1994). A ratio of HeI/HI = 0.1 was
used. However, since the distribution of neutral hydrogen
in the interstellar medium (ISM) is extremely irregular
and given the fact that the software employs an interpo­
lation, the accuracy of the method relies on the number
of stars with directly measured NH used in the fit.
Dempsey et al. (1993a), in their paper on ROSAT data,
quote hydrogen column densities for some of the EUVE
detections. Comparing their column densities to the ones
obtained with the method described above, we find that
most are in agreement in the sense that they yield fluxes
within 50%. For system 015 = UX Ari, the Jelinsky & Fr­
uscione (1996) method provided us with an unrealistically
high value for NH (log NH = 20:33). For this object we
adopted a log NH = 19:09 from Dempsey et al. (1993a).
The distances to our objects come into the picture
twice. Initially, they affect the value for NH and thus the
conversion of count rates into observed fluxes, and sub­
sequently, when converting the observed fluxes to surface
fluxes or luminosities. Direct parallax measurements exist
for only few of our stars and unfortunately, for many of our
targets, we have encountered in the literature substantial
differences concerning this parameter. We have used, as a
rule, the distances and radii quoted in the second CCABS
Table 3. Effects of assumed temperature model and hydrogen
column density on obtaining fluxes from count rates. The fluxes
for the one--temperature models correspond to a count rate of
1 ct s \Gamma1 and are given in units of 10 \Gamma11 erg cm \Gamma2 s \Gamma1 .
Log Lex/B filter Al/Ti/C filter
NH f 6:5 f 6:8 f 7:0 f 6:5 f 6:8 f 7:0
17.50 3.81 2.62 2.85 4.50 3.90 3.82
18.00 3.97 2.70 2.95 5.50 4.50 4.34
18.30 4.20 2.83 3.15 6.98 5.50 5.17
18.60 4.65 3.07 3.56 10.60 7.70 7.00
18.70 4.91 3.20 3.78 12.80 9.02 8.27
18.90 5.59 3.60 4.47 20.74 13.34 12.22
19.10 6.67 4.30 5.78 39.00 21.70 20.12
19.25 7.90 5.17 7.59 66.36 31.65 29.97
19.35 9.00 6.04 9.54 92.79 39.42 38.26
19.50 11.27 8.08 14.51 136.60 50.80 51.80
19.65 14.75 11.76 24.31 172.70 61.40 65.91
19.88 24.72 24.83 60.24 224.20 80.70 90.77
20.00 34.40 39.44 98.01 261.60 94.50 107.04
20.20 65.91 88.89 214.30 357.00 126.70 142.91
20.35 114.38 157.93 371.60 466.70 160.80 181.73
20.50 199.12 262.29 596.35 625.20 208.00 234.80
20.65 330.56 409.69 896.30 850.50 274.50 311.30
(Strassmeier et al. 1993), except for a few systems men­
tioned in Sect. 4.
The analysis of spectroscopic observations of RS CVn
systems to date (Dupree et al. 1993, Mitrou et al. 1996)
have shown that, in the spectral range covered by the
Lex/B filter, a significant fraction of the total counts cor­
respond to emission lines of highly ionised Fe which are
formed between log T = 6.5 and 7.2. The hydrogen col­
umn densities used in the present study (see Tables 1 &
2) extend over two orders of magnitude. So, we examined
the dependence of the conversion factor (from counts per
second to flux units) on the assumed temperature model,
and on the adopted NH value. This is presented in Table
3 for the Lex/B and Al/Ti/C filters.
Columns 2, 3 and 4 give the Lex/B fluxes resulting
from the use of single--temperature models at log T = 6.5,
6.8, and 7.0 in units of 10 \Gamma11 erg cm \Gamma2 s \Gamma1 per ct s \Gamma1 , for
various values of log NH . The values in Tables 1, 2 & 3
were determined using the Monsignori--Fossi & Landini
(1994) line emissivities.
3.2. Single--temperature coronal model
We note that for log NH between 17.50 and 19.10 cm \Gamma2 ,
the choice between the 6.5 and the 6.8 models affects the
resultant Lex/B flux within 50%, while the difference be­
tween the 6.8 & 7.0 models is ú 25%. At higher column
densities though (e.g. log NH ú 20 cm \Gamma2 ), the reverse is

C.K. Mitrou et al.: EUV Emission from RS CVn binaries 7
Fig. 4. Plasma models for log T = 6.5, 6.8 & 7.0 folded through the EUVE Lex/B and Al/Ti/C effective areas. The models
were determined using the Monsignori­Fossi & Landini (1994) line emissivities with log NH = 18.70 cm \Gamma2 at a distance of 30pc.
true: little difference exists between the two lower temper­
ature models, but a factor of two difference in the resul­
tant Lex/B flux occurs between the higher temperature
models. For the Al/Ti/C filter fluxes (columns 5, 6 and
7) there is negligible difference between the 6.8 and 7.0
models. If we assume the model at 6.5 though, we obtain
significantly higher fluxes, indicating a strong dependence
on lower temperatures.
Considering each temperature model separately, we
note that, for log NH below 18.60 cm \Gamma2 , the conversion
factor is relatively insensitive to the column density value
adopted, and remains within 100% for both filters. For
higher column densities though, the picture changes dra­
matically as the resultant fluxes increase very sharply with
increasing log NH . In Fig. 4, we present sample spectra
for the coronal models considered. These spectra have
been folded through the EUVE Lex/B and Al/Ti/C effec­

8 C.K. Mitrou et al.: EUV Emission from RS CVn binaries
Fig. 5. Conversion factors from counts per second into flux
(in units of 10 \Gamma11 erg cm \Gamma2 s \Gamma1 ) for various values of log NH .
Linestyles are as follows: Dotted : Lex/B & Al/Ti/C for
log T = 6:5. Solid : Lex/B & Al/Ti/C for log T = 6:8. Dashed
: Lex/B & Al/Ti/C for log T = 7:0.
tive areas. Figure 5 shows graphically a summary of the
conversion from count rates (ct s \Gamma1 ) to observed fluxes
(10 \Gamma11 erg cm \Gamma2 s \Gamma1 ), plotted against hydrogen column
density for the three temperature models (i.e. log T = 6.5,
6.8 and 7.0).
The Lex/B filter spectral range roughly coincides with
the range (80--160 š A) covered by the short wavelength
spectrometer on board EUVE. To test the quality/validity
of the single--temperature assumption, we have analyzed
the available EUV spectra for seven of our detections
(i.e. HR 1099, Capella, oe Gem, ¸ UMa, AR Lac, – And
and II Peg), and compared the observed integrated flux
from these spectra to the flux predicted by the single--
temperature model at log T = 6:8. The observed flux of
102 = – And is equal to 62% of the predicted value, while
for the remaining six objects the agreement is better than
25%. Taking into consideration the errors from counting
statistics, (for these objects, between 5 and 10%), possible
flaring incidents and the general variability of the systems,
the single--temperature approximation gives fluxes good to
ú 25%. Hence, in our tables and figures throughout the
paper, we used as flux density the value resulting from the
adoption of a log T = 6:8 coronal model.
4. Results & Discussion
In Fig. 6a we plot the EUVE Lex/B fractional luminosity
against the X--ray data from Dempsey et al. (1993b). In
almost every case these authors used the distances quoted
in CCABS to obtain luminosities, so the distances are the
same as ours. The only exceptions were the distances to
systems 009 = AR Psc, 011 = 6 Tri, 040 = IL Hya, and
Fig. 6. (a) X--ray fractional luminosity versus the Lex/B,
the straight line corresponds to equation (1); (b) The CIV
fractional luminosity versus the Lex/B, the straight line corre­
sponds to equation (2); symbols as in Fig. 2.
083 = HR 7275. For systems AR Psc, 6 Tri and IL Hya
Dempsey et al. (1993b) use distances of 70, 14.7 and 138
pc respectively, while CCABS quotes 17, 75 and 263 pc.
For the first two objects, we modified the given X--
ray luminosity and flux to correspond to the CCABS dis­
tance, and for IL Hya we adopted their value. For object
HR 7275, Dempsey et al. (1993b) use a distance of 48 pc,
and the CCABS quotes 250 pc. CCABS mentions explic­
itly that, the distance of 48 pc is obviously wrong, but on
the other hand the 250 pc distance yielded unrealistically
high value for L Lex . Therefore, we used the value given
by Slee et al. (1987), and Drake et al. (1989), of 90 pc
(a distance of 250pc with the corresponding higher NH ,

C.K. Mitrou et al.: EUV Emission from RS CVn binaries 9
would produce a flux 200 times greater). Also for system
087 = HD190540 we adopted a distance of 110 pc from
Drake et al. (1989), instead of the 340 pc value quoted in
CCABS. This latter distance would yield very high NH
and consequently an even higher flux value. But even at
the lower distance and hence lower NH , the system is found
to be overluminous in the Lex/B band (Fig. 6a).
The apparent bolometric luminosities were obtained
from Basri et al. (1985); Drake et al (1989); Simon & Fekel
(1987); Gurzadyan & Rustambekyan (1994)---objects
005 = CF Tuc, 006 = UV Psc and 059 = HD127535; and
Katsova & Tsikoudi (1993)---object 013 = VY Ari. The
CIV 1550 š A data were the result of our own reduction us­
ing the IUEDR and DIPSO packages, which are available
on the UK STARLINK network. When possible, the aver­
age value of several spectra were used. IUE SWP images
do not exist for objects 006 = UV Psc, 087 = HD 190540,
and the only existing spectrum of 047 = DQ Leo is sat­
urated beyond 1450 š A. For systems where both compo­
nents have similar spectral type and luminosity class, in
the plots, but not in Table 2, we attributed part of the ap­
parent X--ray, EUVE and UV flux to both components in
proportion to their radii. In the case of Capella we credited
90% of the flux to the hotter, smaller and faster rotating
G1 III star (Ayres, 1984 and Simon & Fekel, 1987) and in
the case of HR 1099 we considered the less active's com­
ponent contribution as negligible (Dempsey et al. 1993b).
The systems deviating in Fig. 6a are 005 = CF Tuc,
015 = UX Ari, 042 = DM UMa and 087 = HD190540. A
timing analysis of EUVE Lex/B data however, showed
that the CF Tuc and UX Ari data are affected by flares.
We have no explanation for the deviation of the gi­
ants DM UMa and HD190540. A search of the SIMBAD
database for other sources near the expected positions of
these objects, yielded negative results. The solid line in
Fig. 6a was derived by using a least--squares linear fit to
the data and corresponds to
log(L X\Gammaray) =L bol ) = (1:00 \Sigma 0:07) log(L Lex =L bol )+0:81(1)
with a linear correlation coefficient of 0.93 if the obviously
deviating systems are omitted. The mean slope was found
by interchanging the independent variable.
This relation implies that the ROSAT PSPC X--ray lu­
minosities for the RS CVn systems are roughly one order
of magnitude greater than the Lex/B luminosities. Math­
ioudakis et al. (1995) examined the EUV activity of single
dwarfs (types F--M) and found the two quantities to be
of comparable magnitude (Mathioudakis