EXOSAT observations of the contact binary VW Cephei

EXOSAT OBSERVATIONS OF THE CONTACT BINARY VW CEPHEI Osmi Vilhu Observatory and Astrophys. Lab. University of Helsinki SF-00130 Helsinki 13 Finland John Heise Space Research Lab. Beneluxlaan 21 3527 HS Utrecht The Netherlands ABSTRACT. EXOSAT observations of the contact binary VW Cephei on 19th March 1984 are presented. The L1-telescope with CMA+thick Lexan filter was used. The observations cover one orbital revolution showing an asymmetrical X-ray light curve. This can be modelled by an active neck, connecting the two stars, and with enhanced coronal regions on the primary star. Nearly simultaneous IUE observations are also presented. The observations form a part of the program to observe contact binaries with EXOSAT. I. INTRODUCTION A dumb-bell shaped contact binary (of W UMa type) consists of two low mass dwarfs. The components are so close that a shallow, Roche lobe filling, common envelope is formed. These interesting celestial bodies have been the subject of intense research for the last fifteen years. Still we do not have a completely satisfactory structure and evolutionary theory (for reviews see Mochnacki, 1981; Vilhu, 1981; Rahunen and Vilhu, 1982; Smith, 1984). Some of the computed evolutionary models use ad hoc angular momentum loss, stressed by van't Veer (1979), to avoid the unpleasant oscillations around the break of contact. The loss needed is roughly of that size which one gets by extrapolating fromothe observed braking of slowly rotating single stars (Rahunen, 1981; Vilhu, 1982). This braking is most probably due to magnetic wind, and here we have the obvious link to the high chromospheric-coronal activity observed with the IUE and EINSTEIN satellites (Dupree, 1981; Rucinski and Vilhu, 1983; Eaton, 1983; Cruddace and Dupree, 1984). This activity is apparently magnetic in origin (see eg. Linsky, 1983), since contact binaries are rapid rotators (periods 0.25 - 0.65 days) and have convective outer envelopes. The Rossby-number clearly determines the overall activity level (Vilhu, 1984). In addition to the dynamo-generated hot coronal loops, one may expect to find hot plasma also close to the neck connecting the two stars, although we have no clear physical picture for that. This narrow Space Science Reviews 40 (1985) 55-62. 0038-6308/85.15. 9 1985 by D. Reidel Publishing Company. 56 OSMI VILHU AND JOHN HEISE neck is one of the contact binary specialities. Through it the primary pushes large amounts of energy into the secondary, amounting roughly to AL = Q/(I+Q) Lp where Lp is the primary's core (= nuclear) luminosity and Q is the mass ratio (Ms/M~) , typically much less than unity. For our target VW Cephei this luminosity transfer equals to 0.1 L| = 4 1032 erg/s (see Table I). The degree of contact for VW Cep is very shallow (the fill-out factor F in Table I close to unity; Rucinski, 1973) and the thickness of the neck can be estimated to about 10 -2 R e . At this depth p ~ 10 -3 g/cm 3, and using the side-ways convection formalism by Hazlehurst and Meyer-Hofmeister (1973), velocities of the order of I km/sec are sufficient to carry 0 . 1 L s through the neck. This corresponds to mass transfer 0.5 10 -0 M| a value very similar to Webbinks (1977) estimate in his energy transfer model. It should be noted that this mass exchange forms a closed circulation pattern between the components. When hot gas moves from the big star to the small one, cooler currents carry the same mass back (through the same neck): p ~ s 0.5 10 -3 M| Otherwise evolutionary time-scales would be much too short. We may speculate that the hot gas moving in the neck could be seen in X-rays, or else the moving currents may collide producing X-rays in shocks. In the present paper we report the first EXOSAT observations of contact binaries. We observed VW Cep during its whole 6h 40m orbital cycle. IUE observations were also performed two orbital revolutions earlier. Table I. Parameters of VW Cep HD/SAO B-V 197433/9828 0.86 Ref.: Binnendijk, 2. Period (d) 0.278 1970; Rucinski, Ms/M p Mp/M~ i F 0.37 0.80 67 ~ 0.9 AL 0.~ L| 1973 EXOSAT AND IUE OBSERVATIONS Exosat LI/CMA was used in a continuous run on 19th March 1984 starting UT 19h 20m. IUE observations with the LWP camera in the low dispersion mode were performed on the same day from UT 3h 43m, hence 2.3 orbital revolutions earlier. The observations covered one orbit. Fig. I shows a typical IUE spectrum with the prominent MGII 2800 emission, while Fig. 2 gives the Exosat run (the length is exactly one orbit) binned in 10 minutes intervals. The mean 4 Lx - count rate was 0.05 counts/sec, corresponding to 0.24 10 -11 erg/cm2/sec in the energy interval 0.3 - 4 keV for a thin 107 K plasma (Westergaard et al., 1984), EXOSAT OBSERVATIONS OF THE CONTACT BINARY VW CEPHE1 LWP 2 9 9 2 VWCEP L~D 57 I c~ r,,,~ Ld ~X -,4- LL o . . . . I 2700 . . . . I 2 8 OO . . . . I 2900 . . . . . . . ] 000 LAMBDA Fig. I. IUE low resolution LWP-spectrum of VW Cep at phase = 0.26, t(exp) = 5 min, units of erg/cm2/%. or to 0.72 10 -11 erg/cm2/sec in the interval 0.1 - 4 keV (NH = 0). This is roughly the same as the mean observed Einstein IPC-flux 0.96 10 -11 erg/cm2/sec (0.1 - 4 keV, Cruddace and Dupree, 1984). This le~0to Lx/Lbo I = 2500 10 -7 and can be compared with LMgll/Lbo I = 10 -7 and LCIV/Lbo I = 200 10 -7 9 It should be remarked that the UV-contamination of CMA for VW Cep is negligible (<< 10 -4 cts/sec), as can be estimated from the observed IUE's short wavelength SWPfluxes. To see better the overall trends, the observations were smoothed with a Gaussian profile with FMHW = 500 sec. This is shown in Fig. 3. In the same plot we give the IUE fluxes for the MGII emission and UV-continuum around 2800 %. The visual light curve, as given by the IUE's fine error sensor (FES), is also shown together with a synthetic light curve at 5500 A. To a first approximation, the changes of the 58 OSM[ V[LHU AND JOHN HEISE VW CEP EXOSAT L l / F l t ' : 3 S,s LIGHTCURVE fPOm 1.9-MAR-84 ( 79] J9: 20: 6 BI.NSIZE (SEC) = 6 0 0 . 0 0 0 0 , BACKGROUND SUBTRACTED vw cep T F 0.08 , , l - I 0.08 0.07 0.07 CO LLJ 0 . 0 6 O3 t 0.05 to "~ z o.o4 0 O4 o 0.03 co 0.02 -- 0 . 0 3 0.0~ 0.00 5 I I lO 15 i - JJ I__ 20 BIN 0.003 [ o.oo~ -~1~2 o.oo~ I " 5 L I L 25 30 36 # - i _ ] 0.5~,~ I i BACKG~OUNO(COUNTS/SEC) o.ooo o.o 0.06 0.05 J I 10 r ~ L 1 ~5 ~q-- 20 ~ 25 I _ I 30 --7 L [ _1 30 - < 1 35 - 002 -- 0.01 -- 0 . 0 0 4O - - O. 003 0.002 0 001 I0 000 40 LIVETI~E [Ss r 5 10 I I ~5 20 _ 25 I 35 o.o 40 Fig. 2. Exosat CMA/4Lx-light curve of VW Cep over one orbit on 19th March 1984 averaged in 10 min bins. The lower panels give the background and the satellite live time, respectively. visual and ultraviolet fluxes are caused by the apparent change of the visible area (surface brightness over the cormnon envelope is roughly constant). However, especially in the ultraviolet, the O'Connell effect (Linnell, 1982) is clearly seen, but now in the opposite sense: the maximum following the secondary minimum (at phase 0.75) is brighter than the maximum following the primary minimum (at phase 0.25). The effect increases with decreasing wavelength. Clearly some redistribution of spots and active regions has taken place. The X-ray light curve is asymmetrical indicating a non-homogeneous corona. In the next chapter we present a simple geometrical model for the X-ray light curve. EXOSAT OBSERVATIONS OF THE CONTACT BINARY VW CEPHEI UW CEP , , , ~x x+ X J b_ I , , MRRCH , , ! , X~+x .~. % -.. x 9 ,~~ X J~ X CE 19TH , 59 ~ 1984 , , +~ , i x , , I , , , +x . 9 ". ~-u , I x'~ x .8.. .8 , 4" +X+x+ x :I. , ,. x+ X 0 .~+X o . , .' 1.0 o o "8' 0.3 0.8 0.7 Y/ 0.6 O.S i 0. gO i i i I i i l 0.7g Ii I i K 9O0 i J I 1 i J 1.2g J 1 i I. g0 M i | I 1.7S PHRSE +FES (IUE) visual x Sgnth SSO0 o UU-contlnuum MglI-emission EXOSRT C M R / 4 L x Model 1 w Model 2 - Fig. 3. Visual, ultraviolet-continuum (2600 % - 3 2 0 0 A), Mgll-emission and soft X-ray light curves of VW Cep drawn in the same scale. All are normalized with their maximum flux-value. The X-ray models refer to those in Fig. 4. The visual and ultraviolet curves are shifted vertically. 3. GEOMETRICAL MODELLING OF THE X-RAY PLASMA To get some idea of the distribution of the coronal plasma around the system, we divided the surface into 25 regions. For each region a separate light curve was computed, assuming surface corona with no limb-darkening (and no height above the photosphere). We used the synthetic light curve program kindly provided by Slavek Rucinski. These 25 active regions were further combined, to obtain the best fit 60 OSMI VILHU AND JOHN HEISE . . ' ; " ; ; ; . ...... 9 "" 9 0.25 Model 2 , . - : : : : : . . . 9 .-~ +§ .... ...... ...... .+++++.~ ....... .; ;. :. . . .. - 9 ~- +++++ -+~.t ~ , = . ~ ... "" - ......... - : ;.- .... ;; ...... : ...... ": :::::::::: .......... .......... ::::: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . r 3 , * * ~ : : ,++~ 9......... . ... . ... . .. . ... . .. . . . . . . . . + + *.++~s,~sl~ ; :;:::: ::!i!~ii!iii~ii~i:: -.;; 9. . . . . . . . 9......... ::::::: ........ . . . . . . . . . . . "" " " " ; ; : ' " 0.75 ..;;:1;;:;. = . . . . . . . § ........ ~ + + + + + + + + + + + + + ~ = ~ * ~ = + + + + . + + + + . + + ~ = ~ + + + + + + + + + + + + ~ + + + + + + + + + + + ~ * ~ * + + + + + + + + ~ + ~ . ~ + § ++++~-~-+~.~ + - ~ + + ~ : - .- : , 9 9- "++++++++ "++++++++++ ..... ~*~**** . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.25 .......... ..... iiiiiiiiiiiiiiiiili ++,~r .+. .......... ~++ ~+++ ~++++ ~+++++ ~++++++ ~+++~+++ ~++++++++ ~§ ~+++++++++ ~§247247 Mode I 1 - - - ..-, +..~,v,~,~,~,~,~ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . O 9 .. ,.**~,~ . .- ; - .-." ; - : ; ; : - ; . . . . . . . 9i 9 . ....... ........................... 9:iii!iiiii!iiii ii!i!ii: ............. . ~ - --.-- - ;-:;; "'" .- ; . - - 0.5 ........... .......... ........ ....... ..... +~-+ ++§ ~++++++§ ~+++§247 ~++++++++§ ~+++++++ ~+++++ "~+* + 0.75 Fig. 4. Modelling of the X-ray active regions to fit the Exosat CMA/4Lx-light curve on 19th March 1984. Model 2 assumes a constant background and fits only the variable component above the minimum level. The strength of the shading characterizes the local emission measure. The numbers refer to orbital phases. EXOSAT OBSERVATIONS OF THE CONTACT BINARY VW CEPHEI to the smoothed X-ray light curve. It is evident that the solution may not be unique, except perhaps as regards some general features. This is a rather general problem with the light curves of eclipsing binaries. However, our solutions are at least possible and consistent with the present observations. We present two solutions (see Fig. 4): Model I assumes that the X-rays arise solely from the stellar surface, while Model 2 assumes a constant background, so that only the variable component is explained with active regions on the surface. The strength of the shading in Fig. 4 indicates the amount of hot plasma responsible for the emission. In both models an active neck-region, close to the secondary, had to be introduced. This is mainly responsible for the narrow asymmetrical minimum close to the phase 0. In Model I the broad secondary minimum (at phase 0.5) arises when the active back-side of the primary is not visible. In Model 2 the same is caused by the eclipse of the secondary's neck by the star itself. Model I indicates that the primary may be the more active star in X-rays (as compared with the secondary) and have a deeper convective zone, more suitable for the dynamo-processes to work. The primary may also have more star-spots (associated with the coronal regions) making the primary slightly cooler than the secondary, an observational fact which has been many times explained in this way. The solution for the neck (see Fig. 4) indicates that this most important region for the existence of a contact binary may be a source of intense X-radiation. However, the activity of the neck may just be transient, because the Einstein IPC-light curve (Dupree, 1981) shows only one broad maximum during the orbital revolution and can be modelled with active regions predominantly on one side of the primary. We note that if the contribution of the neck-region is removed from the solution of Model I, one obtains a qualitatively similar X-ray light curve to the Einstein one, but only phase-shifted. The complete discussion of the paper will appear elsewhere together with EXOSAT-observations of some other contact binaries (W UMa, XY Leo, 44 Boo). Acknowledgements We thank Dr. Slavek Rucinski for his synthetic light curve program and Dr. Karl Johan Donner for reading the manuscript. O. Vilhu is grateful to Prof. Cornelis de Jager and the Space Research Laboratory of Utrecht for hospitality during the visit in September 1984. References Binnendijk L 1970, Vistas in Astronomy 12, 2!7. Cruddace R G, Dupree A K 1984, Astrophys. J. 277, 263. Dupree A 1981, in R M Bonnet and A K Dupree "(eds.), Solar Phenomena in Stars and Stellar Systems, D. Reidel Publ. Co. p. 407. Eaton J A 1983, Astrophys. J. 268, 800. Hazlehurst J, Meyer-Hofmeister E 1973, Astron. Astrophys. 24, 379. 62 OSMI VILHU AND JOHN HEISE Linsky J L 1983, in J O Stenflo (ed.), Solar and Stellar Magnetic Fields: Origins and Coronal Effects, D. Reidel Publ. Co. p. 313. Mochnacki S W 1981, Astrophys. J. 245, 650. Linnell A P 1982, Astrophys. J. 50, 85. (Suppl.) Rahunen T 1981, Astron. Astrophys. 102, 81. Rahunen T, Vilhu 0 1982, in Z Kopal and J Rahe (eds.), Binary and Multiple Stars as Tracers of Stellar Evolution, D. Reidel Publ. Co. p. 289. Rucinski S M 1973, Acta Astronomica 23, 79. Rucinski S M, Vilhu 0 1983, Mon. Not. R. astr. Soc. 202, 1221. Smith R C 1984, (preprint) van't Veer F 1979, Astron. Astrophys. 80, 287. Vilhu 0 1981, Astrophys. Space Sci. 78, 401. Vilhu 0 1982, Astron. Astrophys. 109, 17. Vilhu 0 1984, Astron. Astrophys. 133, 117. Webbink R F 1977, Astrophys. J. 215, 851.