A ccepted for publication in T he A stronom ical Journal P reprint typeset using LA TEX style em ulateap jv. 08/22/09

A B SO LU T E PR O PERT IES O F T H E SPO T T ED EC LIPSIN G B IN A RY STA R C V B O O T IS G uillerm o T orres1,Luiz Paulo R .Vaz2,and C laud H .Sandberg Lacy3

arXiv:0809.2287v1 [astro-ph] 12 Sep 2008

D raft version A pril23, 2013

A B ST R A C T W e present new V -band di erentialbrightness m easurem ents as wellas new radial-velocity m easurem ents ofthe detached,circular,0.84-day period,double-lined eclipsing binary system C V B oo. T hesedata along w ith otherobservationsfrom theliteraturearecom bined to deriveim proved absolute dim ensions of the stars for the purpose of testing various aspects of theoreticalm odeling. D espite com plicationsfrom intrinsic variability we detect in the system ,and despite the rapid rotation ofthe com ponents, we are able to determ ine the absolute m asses and radiito better than 1.3% and 2% , respectively. W e obtain M A = 1:032 0:013M and R B = 1:262 0:023R for the hotter,larger, and m ore m assive prim ary (star A ),and M B = 0:968 0:012M and R B = 1:173 0:023R for the secondary.T he estim ated e ective tem peraturesare5760 150K and 5670 150K .T he intrinsic variability w ith a period 1% shorterthan the orbitalperiod isinterpreted asbeing due to m odulation by spots on one or both com ponents. T his im plies that the spotted star(s) m ust be rotating faster than the synchronousrate,w hich disagreesw ith predictionsfrom currenttidalevolution m odels according to w hich both stars should be synchronized. W e also nd that the radius ofthe secondary is larger than expected from stellar evolution calculations by 10% ,a discrepancy also seen in other (m ostly lower-m ass and active) eclipsing binaries. W e estim ate the age of the system to be approxim ately 9 G yr. B oth com ponents are near the end oftheir m ain-sequence phase,and the prim ary m ay have started the shellhydrogen-burning stage. Subject headings: binaries: eclipsing | stars: evolution | stars: fundam entalparam eters | stars: individual(C V B oo) | stars: spots 1. IN T R O D U C T IO N

C V B oo (= B D + 37 2641 = G SC 2570 0843; = s 00 15h 26m 19: 54, = + 36 58053: 4, J2000.0; V 10:8, SpT = G 3V ) was discovered as a possible eclipsing binary star by Peniche et al. (1985). B usch (1985) conrm ed itasan eclipsing binary oftype EA and found its period to be 0.8469935 days.In hislastpublished paper, a study of4 lowerm ain sequence binaries,Popper(2000) determ ined a spectroscopic orbit for C V B oo. Popper was pessim istic about the prospects for determ ining accurate absolute properties ofthem because \It appears unlikely that de nitive photom etry w illbe obtained for these stars,partly because ofintrinsic variability." R ecently,a lightcurve and radialvelocity study ofthe system were done by N elson (2004b),resulting in the rst estim ates ofits absolute properties. T he param eters of C V B oo m ake it potentially interesting as the m ost evolved system am ong the wellstudied double-lined eclipsing binaries w ith com ponents near1 M (seeFigure1),a regim ew heresom ediscrepancies w ith theoreticalm odels have been pointed out. W e describe in the follow ing our extensive new photom etric and spectroscopic observationsofthe objectintended to im prove our know ledge ofthe system . T he presence of starspotsdoesin factlim itsom ew hatourability to determ ine highly accurate absolute properties for this binary star,but the results are stillaccurate enough for m ean1 H arvard-Sm ithsonian C enter for A strophysics, 60 G arden Street,C am bridge,M A 02138,e-m ail: gtorres@ cfa.harvard.edu 2 D epto. de F sica,IC E x-U FM G ,C .P.702,30.123-970 B elo H orizonte,M G ,B razil,e-m ail: lpv@ sica.ufm g.br 3 D epartm ent of P hysics, U niversity of A rkansas, Fayetteville, A R 72701,e-m ail: clacy@ uark.edu

ingful tests of current stellar m odels. A s we describe here,C V B oo contributessigni cantly to the body ofevidence concerning the di erences w ith theory m entioned above. 2. O B SE RVA T IO N S A N D R E D U C T IO N S 2.1. D i erentialand absolute photom etry

N ew di erentialbrightness m easurem ents of C V B oo were obtained w ith the facilities available atthe K im pel O bservatory (ursa.uark.edu). T hey consist of a M eade 10-inch f/6.3 LX -200 telescope w ith a Santa B arbara Instrum ents G roup ST 8 C C D cam era (binned 2 2 to produce 765 510 pixelim agesw ith 2.3 arcsecsquare pixels) inside a Technical Innovations R obo-D om e, controlled autom atically by an A pple M acintosh G 4 com puter.T he observatory is located on top of K im pel H all on the Fayetteville cam pus ofthe U niversity ofA rkansas,w ith the controlroom directly beneath the observatory inside thebuilding.Sixty-second exposuresthrough a B essellV lter(2.0m m ofG G 495 and 3.0m m ofB G 39)wereread out and dow nloaded by Im ageG rabber (cam era control software w ritten by J.Sabby) to the controlcom puter over a 30-second interval, and then the next exposure was begun. T he observing cadence was therefore about 90s per observation. T he variable star would frequently be m onitored continuously for 4{8 hours. C V B oo was observed on 89 nightsduring partsoftwo observing seasons from 2001 D ecem ber 1 to 2003 June 9. T he im ages were analyzed by a virtualm easuring engine application w ritten by Lacy that at- elded the im ages,autom atically located the variable and com parison stars in the im age, m easured their brightnesses, subtracted the corresponding sky brightness,and corrected

2

Torres et al.

F ig.1.| M ain-sequence eclipsing binaries in m ass range of C V B oo w ith accurate determ inations of their absolute properties (m asses and radiigood to better than 2% ). D ata are taken from A ndersen (1991) and updates from the literature. P rim ary and secondary com ponents are connected w ith solid lines. C V B oo is represented w ith open circles. T he dashed line show s the solarm etallicity zero-age m ain sequence from the m odels by Y iet al. (2001),for reference.

for the di erences in airm ass between the stars. Extinction coe cients were determ ined nightly from the com parison star m easurem ents. T hey averaged 0.20 m ag/airm ass.C V B oo is also know n as G SC 2570 0843. T he com parison stars were G SC 2570 0511 (\com p", V = 10:26, as listed in the Tycho C atalogue), and G SC 2570 0423 (\ck").B oth com parison starsarew ithin 8 arcm in of the variable star (\var"). T he com parison starm agnitude di erenceshcom p ckiwere constant at the level of 0.013 m ag (standard deviation w ithin a night), and 0.007 m ag for the standard deviation of the nightly m ean m agnitude di erence. T he di erentialm agnitude hvar com pi ofthe variable star was referenced only to the m agnitude of the com parison star, com p. T he resulting 6500 V -band m agnitude di erences hvar com piare listed in Table 1 and plotted in Figure 2. T hetypicalprecision ofthevariablestardi erentialm agnitudes is about 0.013 m ag per m easurem ent. In addition to our ow n, di erential photom etry of C V B oo was obtained by N elson (2004b) in V and C ousins I between 2003 M arch and June (253 and 265 m easurem ents,respectively). T he com parison star was G SC 2570 0869,and the check star wasG SC 2570 0511, w hich isthe sam e starwe used asthe com parison.T hese observations are incorporated into our analysis below . A bsolute photom etry of C V B oo is available in the literature from several sources, and color indices can be used to estim ate a m ean e ective tem perature for the com bined light of the system (assum ing no interstellar reddening). T he results are collected in Table 2. W e used the color/tem perature calibrations of R am rez & M elendez(2005) for dwarf stars for all but the Sloan g r index; for that color we used the calibration ofG irardiet al.(2004). In allcases we assum ed

F ig.2.| T he top panelshow s our V -band light curve ofC V B oo (+ sym bols), consisting of 6500 points, together w ith the V ( ) and I ( ) light curves from N elson (2004b), m arked \ V n" and \ In". N elson’s light curves are shifted as indicated, for clarity. O ur theoretical solution w ithout spots is overplotted (continuous grey lines;x3.1). T he low er panels show the O C residuals from these ts,and in the upper left corner,the standard deviation for a single m easurem ent.

solar m etallicity. T he value ofJohnson V is that listed in the Tycho C atalogue w ith no uncertainty given there. W e have assum ed a conservative error of 0.10 m ag for V . T he tem perature estim ate from the Johnson B V index uses the value ofthat index as listed in the Tycho C atalogue,w ith its listed error.T he tem perature values estim ated in these ways agree quite well,except for the estim atesfrom B V and B T VT ,w hich happen to have the largest form alerrors. T he weighted average of the 7 estim ates is 5706 60K ,w here the uncertainty does not account for possible system atic errorsin the various calibrations.T hiscolor-index-based tem peratureisquite consistentw ith spectroscopic estim ates discussed below , and thissuggeststhatthe interstellarreddening value,if any,is very sm all. C V B oo is identi ed as a strong X -ray source in the R O SAT catalog (Voges et al.2000). T his is presum ably due to an active chrom osphere/corona associated w ith its spot activity (see below ). 2.2. Ephem eris

Photoelectric or C C D tim es of m inim um light for C V B oo carried out over the past decade have been reported by a num berofauthors(A gerer & H ubscher2002, 2003; B akis et al. 2003; D iethelm 2001; D ogru et al. 2006; H ubscher 2005; H ubscher et al. 2005, 2006; K im et al.2006;Lacy 2002,2003;M aciejew ski& K arska 2004; N elson 2000, 2002, 2004a). A dditional tim es of eclipse including oldervisualand photographic m easurem ents reaching back to 1957 were kindly provided by J. M .K reiner(seeK reiner,K im & N ha 2000)ortaken from

C V B oo the literature (Locher2005;M olik 2007). Separate leastsquares tsto the 98 available prim ary and 50 secondary m inim a yielded ephem erides w ith virtually the sam e period w ithin the errors. A sim ultaneous t to allm inim a was then perform ed assum ing a circular orbit. U ncertainties were initially adopted as published,or assigned by iterationsand by type ofobservation so asto achieve a reduced 2 ofunity. T he resulting linear ephem eris is M in I(H JD )= 2;452;321:845322(50)+ 0:846993420(69)E : (1) w here the gures in parentheses represent the uncertainty in units ofthe last decim alplace. N o signi cant trends indicative ofperiod changesare seen in the O C residuals. A test solving for separate prim ary and secondary epochsw ith a com m on period yielded a phasedifference between the eclipses of = 0:49991 0:00013. T his is consistent w ith 0.5, supporting our earlier assum ption ofa circular orbit. 2.3. Spectroscopy

C V B oo wasobserved spectroscopically w ith an echelle instrum enton the 1.5m T illinghastre ectoratthe F.L. W hipple O bservatory (M t.H opkins,A rizona).A totalof 66 spectra were gathered from 1991 June to 2005 A pril, each of w hich covers a single echelle order (45 A ) centered at 5188.5 A and was recorded using an intensi ed photon-counting R eticon detector. T he strongest lines in this w indow are those ofthe M g I b triplet. T he resolving power of this setup is = 35;000,and the observations have signal-to-noise ratios ranging from 13 to 36 per resolution elem ent of8.5 km s 1 . R adial velocities were obtained using the twodim ensional cross-correlation algorithm T O D C O R (Zucker & M azeh 1994). Tem plates for the cross correlations were selected from an extensive library of calculated spectra based on m odelatm ospheresby R .L. K urucz4 (see also N ordstrom et al.1994;Latham et al. 2002). T hese calculated spectra cover a w ide range of e ective tem peratures (T e ),rotationalvelocities(vsin i w hen seen in projection), surface gravities (logg), and m etallicities. Experience has show n that radial velocities are largely insensitive to the surface gravity and m etallicity adopted forthe tem plates,aslong asthetem perature is chosen properly. C onsequently,the optim um tem plate for each star was determ ined from extensive grids of cross-correlations varying the tem perature and the rotationalvelocity,seeking to m axim ize the average correlation weighted by the strength of each exposure. Solarm etallicity wasassum ed. T he results,interpolated to surface gravities of logg = 4:25 for both stars (see x3), are Te = 5800 K and vsin i = 73 km s 1 for the prim ary star,and Te = 5650 K and vsin i= 67 km s 1 for the secondary. Estim ated uncertainties are 200 K and 10 km s 1 for the tem peratures and projected rotationalvelocities,respectively. Tem plate param eters near these values were selected for deriving the radial velocities. Typical uncertainties for the velocities are 5.6 km s 1 for the prim ary and 5.9 km s 1 for the secondary,w hich are considerably worse than usualw ith this instrum ent because of the signi cant rotational broadening ofboth stars. 4

A vailable at http://cfaku5.cfa.harvard.edu.

3

T he stability ofthe zero-point ofour velocity system was m onitored by m eans of exposures of the dusk and daw n sky,and sm allrun-to-run correctionswere applied in the m anner described by Latham (1992). A dditional correctionsforsystem atics were applied to the velocities as described by Latham et al. (1996) and Torres et al. (1997) to account for residualblending e ects and the lim ited wavelength coverage ofour spectra. T hese correctionsare based on sim ulationsw ith arti cialcom posite spectra processed w ith T O D C O R in the sam e way as the realspectra. T he nalheliocentric velocities are listed in Table 3. T he lightratio between the com ponentswasestim ated directly from the spectra follow ing Zucker & M azeh (1994). A fter corrections for system atics analogous to those described above, we obtain ‘B =‘A = 0:71 0:04 at the m ean wavelength ofour observations (5188.5 A ), w herewereferto starA asthem orem assiveone(theprim ary)and to the otherasstarB .T hisvalue isin reasonable agreem ent w ith estim ates by Popper (2000) based on the relative strength ofthe N a I D lines in C V B oo. G iven thatthe starshave slightly di erenttem peratures (see below ), a sm all correction to the visual band was determ ined from synthetic spectra integrated overthe V passband and the spectralw indow of our observations. T he corrected value is (‘B =‘A )V = 0:73 0:04. R adialvelocities for C V B oo have been reported previously also by Popper (2000),w ho observed the star as part ofhis program focusing on binary system s ofspectraltype F to K .H is45 m easurem entsfrom 1988 February to 1997 June w ith the H am ilton spectrom eteratthe Lick O bservatory partially overlap in tim ew ith ours,and are ofexcellentquality. H owever,they require a num ber ofadjustm ents before they can be com bined w ith ours. O ne ofthese adjustm ents has to do w ith corrections he applied to his raw velocities. T he raw velocities (w hich he referred to as\O bserved")were reported forC V B oo alongside \O rbital" velocities w hich di er from the raw ones by the application oftwo corrections. T he rst is analogousto the correctionsforsystem atic e ectswe applied to our ow n velocities,and was derived in a sim ilar m anner using synthetic binary spectra (for details see Popper & Jeong 1994). T he second correction accounts for distortions and m utualirradiation in the close orbit, and was com puted by Popper (2000) using the form alism developed by W ilson (1990) as im plem ented in the W ilson-D evinney (W D )program thatwe also use below , and added to the velocities. G iven our plan to use the W D program to com bine the lightcurvesw ith the velocity m easurem ents in a sim ultaneous solution,the latter corrections in the data by Popper (2000) need to be rem oved prior to use or they would be applied tw ice. W e estim ated these corrections from a prelim inary solution w ith W D (presum ably em ulating Popper’s procedure), and applied them w ith the oppositesign to the \O rbital" velocities.T hese correctionsare no largerthan 1 km s 1 , w hich is sm aller than the form aluncertainties in the velocities (2.7 km s 1 for the prim ary and 2.1 km s 1 for the secondary,from the residuals ofprelim inary ts). A second adjustm ent we found necessary to apply to Popper’s m easurem ents is to correct for an o set of 1.60 km s 1 between his prim ary and secondary velocities,as indicated by prelim inary sine-curve ts (see Ta-

4

Torres et al.

ble6 by Popper2000,w heretheo sethe ndsissim ilar). E ectively thetwo starsyield di erentcenter-of-m assvelocities. W e found no such o set in our ow n m easurem ents,butexperience indicatesitcan som etim esappear w hen thereisa signi cantm ism atch between theadopted tem plates and the realstars,and ifnot corrected it can biasthe sem i-am plitudes w hen enforcing a com m on centerofm assin the t. W e have thus added 1:60 km s 1 to Popper’ssecondary velocities. Finally,a third adjustm entisto bring Popper’soverallvelocity zero pointinto agreem ent w ith ours. From trial orbital ts we found this required a shift of + 0.37 km s 1 to his velocities.5 Popper’s corrected velocities are listed in Table 4. Separate ts to his data and ours give sim ilar values for the sem i-am plitudes,and yield m assesthatdi er by less than tw ice their com bined uncertainties. W e therefore proceed to m erge the two data sets below . A third setofradialvelocitiesforC V B oo wasreported by N elson (2004b),but they were obtained at lower resolution,they are few in num ber (12),and show a m uch largerscatterthan the two otherdata sets( 15 km s 1 ) so they are oflittle use for our purposes. 3. M O D E LIN G O F T H E P H O T O M E T R IC O B SE RVA T IO N S

T he overallshape ofC V B oo’s light curves (Figure 2) show sratherm oderate proxim ity e ects despite the system ’s relatively short period of slightly m ore than 20 hours, w ith the curvature between the m inim a being m ostly due to the deform ation of the com ponents and, to a sm aller degree,to the m utualillum ination. A num ber of sm all-scale features are obvious to the eye that are possibly due to spots, other intrinsic variability,or even instrum entale ects,especially in the sm aller data sets ofN elson (2004b). O ther features described below are revealed through a m ore detailed exam ination,and introduce som e com plications into the analysis. 3.1. Initialsolutions withoutspots To begin w ith, we chose to m odel all the observations together in order to obtain a baseline solution against w hich to com pare m ore com plex solutions that attem pt to account for the features m entioned above.W e used a version ofthe W ilson-D evinney (W D ) m odeling program (W ilson & D evinney 1971; W ilson 1979, 1993) w ith extensive m odi cations as described in Vaz,A ndersen & C laret (2007) and references given therein. T he m odi cations pertinent to C V B oo include the capability to use m odelatm ospheres(now also available in the distributed versionsofW D ),consistency checksbetween variousparam eters,and theability to use the dow nhill sim plex algorithm (N elder & M ead 1965) instead of di erential corrections. W e com bined our ow n V -band light curve w ith the sparser V and I light curves of N elson (2004b) in order to im prove the constraint on the e ective tem perature ratio,and w ith our radial velocities from x2.3 as well as those of Popper (2000).T huswesolved sim ultaneously 3 lightcurvesand 4 radial-velocity curves. T he param eters adjusted were 5 W hile in principle the latter tw o adjustm ents (o sets) could be accounted for in our com bined photom etric and radialvelocity solution described below by sim ply adding free param eters to the t,lim itations in the current version ofthe W D code do not allow this,so w e have applied the o sets externally.

theorbitalinclination i,thesecondary Te ( ux-weighted m ean surface tem perature), the bandpass-speci c prim ary lum inosity (see W ilson 1993),both stellar surface gravitationalpseudo-potentials (related to the stellar radii),the center-of-m ass radialvelocity ,the m ass ratio q M B =M A ,and an arbitrary phase shift. T he sim ilar depths of the two m inim a in both V and I (Figure 2)im ply thatthe com ponentsm usthave rathersim ilartem peratures,consistentw ith indicationsfrom spectroscopy. T he prim ary tem perature was held xed at a value determ ined from our results based on photom etry and spectroscopy, as follow s. A photom etric estim ate ofthe prim ary tem perature was derived from the m ean system tem perature (x2.1) using approxim ate values for the radius ratio and tem perature ratio from prelim inary light-curve solutions. T he result,5755 60 K ,was then com bined w ith the spectroscopic value of the prim ary tem perature (x2.3),giving a weighted averageforStarA of5760 150 K ,w hich we adopt. T he re ection albedos for both com ponents were held xed at the value 0.5,appropriate for stars w ith convective envelopes,and the gravity-brightening exponents were com puted internally in W D using the local value of Te for each point on the stellar surface taking into account m utual illum ination, follow ing A lencar & Vaz (1997) and A lencar et al. (1999). T he ux from each of the com ponents is represented by N extG en atm osphere m odels based on the PH O EN IX stellar atm osphere code (A llard & H auschildt 1995; A llard et al. 1997;H auschildt et al.1997a,b). T he lum inosity ofthe secondary is calculated by the program from its size and Te . T he lim b-darkening coe cients for both com ponents,xA and xB ,were taken from the tables by C laret (2000),and interpolated using a bi-linearschem e for the currentvaluesoflogg and Te ateach iteration.W ehave only considered the linearlaw here in view ofthe distortions in the light curve, w hich w ill tend to overw helm the rather subtle e ect oflim b-darkening. A lthough we have no evidence ofanother star in the system ,the possibility of third light (‘3 ) was explored carefully for its potentialin uence on the geom etric param eters,particularly in the solutions described below in x3.2, w hich include spots. A chieving convergence w hen solving for third light was found to be very di cult due to the intrinsic variability and the large num ber offree param eters,even w hen considering m ultiple param eter subsets (W ilson & B ierm ann 1976). W e found that the solution was not im proved, and ‘3 was not considered further. W e estim ate a conservative upper lim it to ‘3 of 1% , w hich does not produce signi cant changes in the geom etric param eters. A circular orbit has been assum ed in the follow ing,based on ourinvestigation ofthe eclipse tim ingsin x2.2,and thelack ofany indication in thelight curvesofa displacem entofthe secondary m inim um from phase0.5.T hisisconsistentw ith expectationsfrom tidal theory foran orbitw ith such a shortperiod (see x6). W e havealso assum ed heretidalforceshavesynchronized the com ponents’rotation w ith theorbitalm otion ofC V B oo. In these initialcalculationswe applied both least-squares di erential corrections and/or the sim plex m ethod between successive iterations. T he lim b-darkening coe cients,norm alization m agnitudes,surface gravities,and individualvelocity am plitudes were allupdated between consecutive runs to correspond to the solution from the

C V B oo

F ig.3.| P rim ary ( ) and secondary (+ ) radial velocity m easurem ents collected at C fA (Table 3) along w ith the theoretical curves obtained w ith W D and no spots (top panel). V elocities are shifted so thatthe center-of-m assvelocity isatzero (dashed line). T he large deviation from K eplerian m otion in the predicted velocity around both conjunctions is due to the R ossiter-M cLaughlin e ect (see Schlesinger 1909a,b; R ossiter 1924; M cLaughlin 1924), caused by partial eclipses of the rotating stellar surfaces (an effect built into the W D m odel). T he O C residuals are show n at the bottom . T he standard deviation of the unw eighted residuals is rv = 6:09 km s 1 for both com ponents and is show n in the upper left corner. T he reduced 2 values w ere 1.000 and 0.998, respectively.

previous iteration. T he solutionsare show n in Figures2,3 and 4,w ith the corresponding residuals. T he residuals from the radial velocity ts m atch the quality ofthe observations. T he photom etry isreasonably wellrepresented on averageby the theoreticalcurves,but the residuals for our V -band observations show an rm s scatter of0.0196 m ag that is m uch largerthan the m ean internalerror( 0.013 m ag). T his im m ediately suggests there m ay be unm odeled effects. T he intrinsic errors ofN elson’s observations were not reported in the originalpublication. 3.1.1. Study ofthe light-curve residuals

Partofthe extra scatterisno doubtdue to featuresin thelightcurvealluded to earlierthatareseen in Figure2, such as changes in the light levelfrom night to night at the sam e orbital phase (e.g., near phase 0.1), or other short-term deviations (e.g.,near phase 0.8),both in our data and in N elson’s. W e investigated the residuals of ourm orenum erousV -band photom etry furtherto search forperiodic signalsthatm ightadditionally contribute to the excessscatter.O urinitialexploration ofpossible signals near the orbitalperiod using the La er & K inm an (1965)m ethod revealed severalsim ilarperiodicities that appear signi cant. W e then extended the search to a m uch w iderrangeoffrequenciesby com puting theLom bScargle periodogram , and found other signals. T his is show n in the top panel of Figure 5. W e refer to this as the \dirty" power spectrum , since it is a ected by

5

F ig.4.| P rim ary ( ) and secondary (+ ) radial velocity data from Popper (2000),w ith the sam e lim its on the verticalaxes as in Figure 3. T he standard deviations ofthe unw eighted residuals are 1 and 1 for the prim ary and rv;A = 3:14 km s rv;B = 2:14 km s secondary velocities, respectively (show n in the upper left corner ofthe low er panels). T he reduced 2 values w ere 1.005 and 1.002.

the particular tim e sam pling of the observations (w indow function). T he highest peak corresponds to a period of 0.837 days, w hich is shorter than the orbital period of0.846993420 days.T he two nexthighestpeaks (indicated w ith arrow s)turn out to be 1-day aliases.To illustrate this, we have applied the C LEA N algorithm as im plem ented by R oberts et al.(1987) to rem ove the e ects ofthe w indow function. T he second panelofFigure 5 show s that only the 0.837-day peak survives this process,suggesting it is a realsignal. In the third panel an enlargem entofthe dirty powerspectrum in the vicinity ofthe m ain peak reveals ne structure that was also seen w ith the La er & K inm an (1965) m ethod. H ow ever,none ofthese peaksagree w ith the frequency corresponding to the orbitalperiod,w hich is represented for reference w ith a dotted line. T he two m ain sidelobes indicated w ith arrow s are 1-year aliases ofthe m ain peak. O nce again they disappear after application ofC LEA N , as seen in the bottom panel, supporting the reality of the rem aining signal. T he statisticalsigni cance ofthis signalwasestim ated by num ericalsim ulation.W e generated 100;000 arti cialdata setsusing the actualtim esof observation and the variance ofthe originalresidualsassum ing a G aussian distribution oferrors,and com puted theLom b-Scarglepowerspectrum forthesedata setsover the sam e frequency intervalconsidered above. W e then selecting the highest peak in each case. N one of them cam e close to the height ofthe peak we see in the real data, indicating a false alarm probability sm aller than 10 5 . T he precise frequency of this signalwas m easured in the C LEA N ed spectrum , and its uncertainty was estim ated from the half w idth at half m axim um of the peak. T he corresponding period is 0:83748 0:00052

6

Torres et al.

F ig. 6.| R esiduals ofour V -band K im pelO bservatory observations ofC V B oo show n as a function ofphase,using the period of 0.83748 days inferred from the pow er spectrum analysis (see Figure 5). T he tim e origin has been set arbitrarily to H JD 2;450;000.

F ig.5.| (a) Lom b-Scargle pow er spectrum of the residuals of our V -band K im pelO bservatory observations ofC V B oo from the no-spot solution described in the text. T he arrow s indicate 1day aliases of the central peak; (b) C LE A N ed pow er spectrum of the sam e m easurem ents using the algorithm of R oberts et al. (1987) to rem ove the e ects ofthe tim e sam pling;(c) E nlargem ent of panel (a), w ith the 1-year aliases of the m ain peak indicated w ith arrow s. T he dotted line represents the orbitalfrequency;(d) E nlargem ent ofpanel(b). T he period corresponding to the signal is 0:83748 0:00052 days.

days, w hich is di erent from the orbital period at the 18 level. A plot of the photom etric residuals folded w ith this period is show n in Figure 6, and indicates a peak-to-peak am plitude ofabout 0.04{0.05 m ag. T heK im pelO bservatory data covertwo observing seasons. Separate Lom b-Scargle power spectra show that the sam e signal is present in both seasons, along w ith the 1-day aliases, indicating the phenom enon is persistent from one year to the next. It is not seen as clearly, however,in the residuals from our baseline t ofthe observations of N elson (2004b), w hich are m uch sparser than ours (and span only 71 days instead of549 days). H isV -band data show a hintofthe m ain peak and its1day aliases,butnotthe I-band data,w hich have a larger scatter. W e carried out a sim ilar power spectrum analysis of the hcom p ckidi erentialm agnitudesfrom K im pelO bservatory,to explore the possibility that either the com parison or the check star m ight be the source of this variation. N o signi cant periodicity was seen. T hus, the phenom enon is intrinsic to C V B oo. Possible explanations for this variation include stellar pulsation, and star spots on one or both com ponents. N either of the stars appear to be in an evolutionary state that favors pulsationalinstability.Forexam ple,the absolute dim ensions derived below place both com ponents welloutside

the C epheid or Sct instability strips in the H -R diagram indicated by K j rgaard et al. (1983).O n the other hand,C V B oo is a know n strong X -ray source detected by R O SAT (Voges et al.2000),w ith an X -ray lum inosity oflogL X = 30:656 thatissom e 4000 tim esstrongerthan the Sun. In term s ofits bolom etric lum inosity C V B oo has logL X =L bol = 3:39, w hich is near the high end for active binaries. T hus it seem s likely that the underlying reason for the 0.837-day periodicity is related to spots on the surface ofone or both com ponents,and we proceed under this assum ption. It is interesting to note that these features on C V B oo seem to have lasted for an unusually long tim e (1.5 years in our case),at least com pared to sunspots,although even m ore extrem e exam ples have been docum ented in the literature. O ne is thewell-know n activebinary H R 1099 (Vogt et al.1999), w ith surface features persisting for at least 11 years. A n im portant im plication of this spot hypothesis is that the com ponent having spots would appear to be rotating slightly m ore rapidly than synchronously w ith the m otion in the circularorbit,w hich is unexpected for such a short-period binary.W ediscussthisin m oredetail below . 3.2. Solutions with spots

A n accurate m easurem entofthe vsin ivaluesforboth com ponents would allow for a direct test ofour hypothesis ofnon-synchronous rotation,and could even distinguish w hich of the stars is the culprit (or if both are). U nfortunately,however,the quality ofour spectroscopic m aterial is insu cient for that purpose. In principle, m odern light-curve m odels such as W D enable the user to solve for various param eters that describe the spots. 6 L 1 X isin units oferg s ,and w as determ ined from the R O SA T count rates and hardness ratios,the distance estim ate in x4,and the energy conversion factor ofFlem ing et al.(1995).

C V B oo H owever,w ith only photom etric data atour disposalfor C V B oo, and m ost of it in a single passband, it is essentially im possible to tellw hich star has the spots,or w hether both com ponents have them . T his is a wellknow n di culty in light-curve m odeling. O ther inversion techniques such as D oppler im aging are m uch better suited to m apping surface inhom ogeneities,although even they are not w ithout their lim itations. M oreover, even ifwe knew w hich starhasthe spots,the determ ination oftheir param eters from light curves alone is a notoriously ill-posed problem ,on w hich there is abundant literature discussing issues of indeterm inacy and nonuniqueness in the presence oflim ited data quality (see, e.g.,Eker 1996,1999,and num erous references therein). H aving photom etry in m ultiple passbands m ay aleviate the problem som ew hat,butitdoesn’tsolve itand strong degeneraciesarelikely to rem ain w ith othersubtle e ects in the light-curves. T herefore,w hile we cannot hope to obtain an accurate picture ofthe distribution ofany surface featureshere,the consequencesofspotson the light curve are fairly clear in C V B oo (to the extent that our hypothesis is true), and we m ake an e ort in the follow ing to at least rem ove som e ofthose distortions and study their in uence on the geom etric param etersofthe system ,w hich are ofm ore im m ediate interest. In order to perm it the num ericaltreatm ent ofsurface features in this case,we introduced m odi cations in the W D code to allow for a precise tracking ofthe spot position at a period di erent than the orbitalone. In this schem e,wespecify the spotpropertiesata certain Julian date and, through the speci ed intrinsic rotation rate, the code keeps track of the spot m otion, w ith its longitude follow ing the com ponent’s rotation, and its colatitude,size,and e ective tem perature rem aining otherw ise constant. In view of the am biguities m entioned above regarding the location of spots in C V B oo, and ourinability to tellifthere m ighteven be m ultiple spots on one or both stars,we have taken our light-curve t from x3.1 as our starting point and investigated the follow ing three sim ple cases separately: (a) a single spot on the prim ary;(b) a single spot on the secondary;and (c) one spot on each com ponent. M ore com plex congurations becom e increasingly di cult to study due to convergenceproblem sin the solutions,and itisnotclear they are justi ed w ith the data available. T he in uence ofspots on the radialvelocity curves is very sm all com pared to our errors, so that those data are notofvery helpfulfor studying surface features. W e use only our m ore extensive V -band photom etry in the study of these three cases,although the solutions were checked using N elson’sV and I lightcurves.A sthe photom etric coverage does not necessarily overlap w ith the radialvelocity coverage,we have adopted forthe spotted casesthe m assratio q obtained from a no-spot tsim ilar to that in x3.1 that assum es asynchronousrotation (see below ),and held it xed. For lack ofother physicalconstraints,and given that the stars are quite sim ilar in all their properties, we assum ed that both com ponents rotate slightly super-synchronously,at a rate given by the ratio between the orbitalperiod and the residualperiod found in the previous section,equalto 1.0114. C ases (a) and (b) converged quite rapidly to sim ilar con gurations,in w hich the co-latitude,size,and the effective tem peratures ofthe spots are com parable,w hile

7

F ig. 7.| D i erence betw een the light curves w ith spots and those w ithout spots (show n in Figure 2),for the rst cycle of our observations. T he thin horizontal line show s the norm alization levelofthe theoreticallightcurves(phase 0.25, rstorbiting cycle).

their longitudes are such that the spots present always the sam e position relative to the center of the spotted com ponentand theobserver.T hiscan beseen in Table5, w here the longitudesofthe spotsin cases(a)and (b)are separated by nearly 180 .A nothersim ilarity isthatboth spotscoverthe com ponents’polarregions.A lthough attem pted,no solution could be obtained for \hot" spots (i.e.,w ith tem perature factors larger than unity;see below ). Solution (c) w ith one spoton each com ponentdid not converge as easily. W hen the param eters ofboth spots were left free to be adjusted, one of the spots (usually the one on the prim ary) tended to becom e very sm all and cold,w ith the tem perature factor Tfactor (ratio between the spot tem perature and the photospheric tem perature)becom ing sm allerthan allowed by theN extG en atm osphere tableswe used. T he solution we presentwas achieved by rst adjusting som e ofthe spot param eters w hile holding others xed,and then alternating and iterating untilconvergence. T he m axim um am plitude ofthe in uence ofthe spots on the light curve is 0.08m ag,and occurs for the onespot solutions,as show n in Figure 7. T his gure corresponds to the rst orbitalcycle ofour observations and, since the spots follow the com ponents’non-synchronous rotation, the dips change place at each orbiting cycle. T he two-spot solution of case (c) gives a slightly sm allerpeak-to-peak am plitude ( 0.06 m ag)thatseem s m arginally larger than indicated in Figure 6,suggesting thatperhaps a m ore com plex spot con guration m ay be needed. W e report in Table 6 the m odel param eters we obtain for the solution w ith no spots and for cases (a), (b),and (c),together w ith the radiiofthe com ponents in term s of the orbitalseparation. For the reasons described above,the solution w ithoutspotswasperform ed by solving sim ultaneously three lightcurvesand 4 radial velocity curves,w hereas the spotted ts are based only on our V -band light curve. T he m ain di erence in the param etervaluesisseen in the inclination angle,w hich is approxim ately onedegreehigherforthe solution w ithout spots. O ther param eters such as the secondary e ective tem perature and the sizesofthe com ponentstend to differ less between the spotted and unspotted solutions. Figure 8 gives a representation ofthe spot con gura-

8

Torres et al.

tion resulting from case (c) w ith the com ponents’ size and separation rendered to scale,and seen from the observer’s view point at six di erent orbital phases. T he stars are well detached from the corresponding R oche lobes, w ith ll-out factors (M ochnacki 1984) that are 0.7693 and 0.7501 forthe prim ary and secondary,respectively. W e noted above thatour tsyield polarspots,as has often been found (also from D oppler im aging techniques)forotheractive binariessuch asthe R S C V n system s. T here is considerable theoreticalsupport for this preference for high-latitude surface features in rapidlyrotating active system s (see, e.g., Schussler & Solanki 1992; G ranzer et al. 2000; Isik et al.2007). A curious result from our ts is that the spots happen to be positioned so as to avoid eclipses, although the reality of this con guration is di cult to assess. It is nevertheless an indication that the phenom enon responsible for the periodic behavior ofthe residuals in the unspotted solution doesnotlead to strong discontinuities,such asthose resulting from the eclipses. A lthough the rm s residual is m arginally sm aller for the solution obtained in case (c),asindicated in Table 6, this t as wellas the other two spotted solutions are visually indistinguishable from the solution w ithout spots show n in Figure 2. T he residuals for case (c) are displayed in Figure 9 for our V -band light curve as well as for the V and I light curves ofN elson (2004b). T he patterns clearly visible in these O C diagram s are not very di erent from those in Figure 2, w hich m ay give the im pression that not m uch progress has been m ade.7 T hey certainly indicate thatthere arestillfeaturesofthe brightnessvariation that are notcom pletely or correctly m odeled,possibly due to a m ore com plex spot con guration than we have assum ed,oreven som e com bination ofspots and m ulti-m odalpulsations. Problem sofan instrum entalnature in the photom etry cannot entirely be ruled outeither.H owever,w hatisnotim m ediately obvious to the eye is thatno signi cantperiodicities that we can detect rem ain in these residuals. T his is illustrated in Figure 10, in w hich the top curve show s our La erK inm an period study oftheK im pelO bservatory V -band residuals from the no-spot solution,and the lower curve show s the sam e study for the residuals from case (c). N ote the com m on verticalaxisforboth setsofresiduals, indicating the im provem ent in the overallvariance. 4. A B SO LU T E D IM E N SIO N S A N D P H Y SIC A L P R O P E RT IE S

Exam ination ofTable 6 show s that key geom etric param eterssuch asthe relative radii(rA ;vol,rB ;vol)vary by as m uch as 3{4% between the three spotted solutions, w ith solution (c) generally giving interm ediate results. A sindicated earlier,thisisthe tthatprovidesform ally the sm allest rm s residual,although the di erence com pared to the other two spotted solutions is m arginal. In allfour solutions the m ean light ratio outside ofeclipse accounting for spots, (‘B =‘A )V , is quite sim ilar to the spectroscopically determ ined value of0:73 0:04 (x2.3). From the e ective tem peraturesofC V B oo A and B the convective turnover tim e for both stars is estim ated to be 25 days, follow ing H all(1994). T he R ossby num 7 N ote,how ever,thatthose patternsare m ostobviousin N elson’s data, w hich do not actually enter into the nal solution adopted in x4.

F ig.8.| R epresentation ofthe com ponents ofC V B oo at di erentorbitalphasesasindicated on the left,show n to scale w ith their high-latitude spots as m odeled here. T hese spots resulting from our solution (c) are positioned in such a w ay that they practically avoid being eclipsed. T his is an indication of a rather sinusoidal behavior ofthe disturbing phenom enon causing periodic variations in the residuals ofthe unspotted solution (see text).

ber (ratio between the rotation period and the convective turnover tim e) is then R 0 0:033, w hich places both com ponents in the regim e w here stars usually display signi cant light variations due to spots (see H all 1994,Figure 6). O n the basis ofthe above we adopt t (c) w ith one spot on each com ponent as the best com prom ise for C V B oo, but we reiterate that this m odel is stillprobably only a crude approxim ation to the true spot con guration in the system , assum ing that spots are the underlying reason for the periodic signalfound in the light-curve residuals.For calculating the absolute dim ensions ofthe two stars we have chosen to use m ore conservativeuncertaintiesthan theform alerrorslisted in Table 6,to accountforthe spread am ong the three spotted solutionsgiven the uncertaintiesin the m odeling:we havecom bined the internalerrorsquadratically w ith half of the m axim um range in each param eter. T he values adopted are i = 86:24 0:33, a = 4:748 0:019 R ,

C V B oo

9

quadrature we use the average ofrpoint and rback . Finally,for com puting the absolute visualm agnitude M V and distance we have relied on the apparentV m agnitude listed in the Tycho C atalog,and ignored extinction.C V B oo wasnotobserved by the H ipparcos m ission (Perrym an et al. 1997), so no direct parallax m easurem ent is available. 5. C O M PA R ISO N W IT H ST E LLA R E V O LU T IO N T H E O RY

F ig.9.| O C residuals from the solution w ith one spot on each com ponent, based on the t to our V -band light curve. T he m odelused to com pute the residuals for the V and I light curves of N elson (2004b) in the top panels is based on the sam e lightcurve param eters as our V -band t, except for the m agnitude at quadrature and the w avelength ofthe observations. T he standard deviation of a single observation for each residualcurve is show n in the upper left corner ofeach panel.

F ig.10.| T he variance versus trial period in days (follow ing La er & K inm an 1965) for the K im pelO bservatory V -band residuals of the no-spot solution (top curve, corresponding to the t show n in Figure 2), and for the solution w ith tw o spots (bottom curve;see text). T he dotted line m arks the m ost signi cant period found for the O C ofthe solution w ith no spots,w hile the dashed line indicates the orbitalperiod.

q = 0:9378 0:0070, rA ;vol = 0:2658 0:0047, and rB ;vol = 0:2470 0:0048,and are based only on the K im pelO bservatory m easurem ents.T he nalresultsarepresented in Table 7,w here the uncertaintieswere obtained by propagating allobservationalerrorsin the usualway. T he starsin C V B oo departsom ew hatfrom the sphericalshape due to tidaland rotationaldistortions. T he relative di erence between the polar radius and the radius toward the inner Lagrangian point is 5.5% for the prim ary and 4.8% for the secondary.T he system is nevertheless welldetached: the sizes ofthe stars represent fractionsof70% and 66% oftheirrespective m ean R oche lobe sizes. T he tem perature for the secondary from the light-curve solution is in excellent agreem ent w ith the spectroscopic value (x2.3). Included in Table 7 are the predicted projected rotationalvelocities(vasync sin i)com puted w ith the adopted rotation period for the stars (P rot = 0:83748 days = P orb =1:0114; see x3.1.1), as well as the synchronous values (vsync sin i), for reference. T hese m ay be com pared w ith the m easured vsin ivaluesfrom spectroscopy (x2.3). T he stellar radiiused for these calculations are those presented to the observeratquadrature (w hich are 2.7% and 2.4% largerthan thevolum eradii;seeTable6), since that is the phase atw hich the spectroscopic observations are concentrated. A s a proxy for the radius at

In this section we com pare the absolute dim ensions of C V B oo w ith current stellar evolution m odels from the Yonsei-Yale series by Y iet al.(2001),incorporating an updated prescription forconvective core overshooting as described by D em arque et al.(2004). T hese m odels adopt a m ixing length param eter of M L = 1:7432,calibrated against the Sun. In Figure 11 we show evolutionary tracks com puted for the exact m asses we derive for each star (see Table 7), for a heavy-elem ent abundance equalto that ofthe Sun (w hich is Z = 0:01812 in these m odels; dotted lines). T he uncertainty in the location of the tracks that com es from our m ass errors is indicated w ith the error bar in the lower left. T he tracks show excellent agreem ent w ith the observations, suggesting the com position is near solar. T he m easured tem perature di erence between the com ponents is quite close to w hat the m odels predict. A m arginally better m atch is achieved w ith a slightly higher abundance of Z = 0:01955 (corresponding to [Fe/H ]= + 0:04,assum ing no enhancem ent ofthe elem ents),show n as solid lines in the gure. T he m odels indicate the prim ary is beginning its shellhydrogen-burning phase,and the secondary is near the end ofits m ain-sequence phase. T he age that best ts both com ponents in this logg{T e diagram is 9:0 1:8 G yr,and the corresponding isochrone is show n as a dashed line. W e have also considered a second set ofm odels,from the series by C laret (2004). T he physics in these calculations is sim ilar though not exactly the sam e as the previousones.Forexam ple,the solarcom position in this case is taken to be Z = 0:020,and the m ixing length param eter that best reproduces the observed properties of the Sun is M L = 1:68. T he com parison w ith the observations for C V B oo is show n in Figure 12. A lthough the C laret m odels m atch the m easured properties very well, we nd as w ith the Yonsei-Yale m odels that a slightly higher m etallicity (Z = 0:0225,or [Fe/H ] = + 0:05) provides an even better t. T his is show n by the solid lines in Figure 12. T he age ofthe system from these calculations is 9.8 G yr,consistent w ith the previous estim ate. Experim ents changing the m ixing length param eter show the sensitivity of the best- t com position to M L . In Figure 13 we com pare the observations w ith tracks com puted for a lower value of M L = 1:50, w hich has the e ect of yielding lower tem perature predictions. Solar-m etallicity m odelsare indicated w ith the dotted lines. In this case we nd thatthe best- tm etallicity (Z = 0:0185,or [Fe/H ]= 0:03) is slightly lower than solar (solid lines). T he preceding com parisons m ay give the im pression that the observations for C V B oo are very wellm atched by thepredictionsfrom theory,thatstellarphysicsiswell understood,and thatthereforethereisno reason forconcern.H owever,a m orecarefulexam ination indicatesthat thisisnotnecessarily true.O fthethreebasicparam eters

10

Torres et al.

F ig. 11.| A bsolute dim ensions for C V B oo com pared w ith evolutionary m odelsfrom the seriesby Y iet al.(2001). T he errorbars for log g are sm aller than the size of the sym bols. M ass tracks for the exact m asses w e m easure are indicated w ith solid curves for the best- tting m etallicity of Z = 0:01955 (w here Z = 0:01812 for these m odels). Solar m etallicity tracks are show n for reference (dotted curves). T he isochrone producing the best sim ultaneous m atch to both com ponents is show n w ith the short-dash line,and corresponds to an age of 9.0 G yr. T he long-dash lines represent sm allsectionsofthe tw o isochronescorresponding to the m axim um and m inim um age allow ed by the errors (9:0 1:8 G yr). T he uncertainty in the location of the m ass tracks is indicated w ith the error bar below the tracks for the prim ary, and is m uch sm aller than the tem perature uncertainty.

typically determ ined in eclipsing binaries (M , R , Te ), thetem peratureisusually theweakestsinceitoften relies on externalcalibrations.Figure14 displaysthe m easurem entsforC V B oo in a di erentdiagram ,them ass-radius plane,along w ith isochrones from the Yonsei-Yale series forthe sam e two m etallicitiesdiscussed in Figure 11. N o singlem odelm atchesboth com ponentsw ithin theerrors, and the secondary appearsnom inally olderthan the prim ary,the di erence in agebeing 25% .T hisisthe sam e phenom enon pointed out by Popper (1997) for several other system s including FL Lyr,RT A nd,U V Psc,and C en.A notherway ofinterpreting thisisthatthesecondariesin allthese binariesare too large fortheirm asses, com pared to theory or com pared to the prim aries. For C V B oo the o set in the secondary radius is 10% , w hich represents a very signi cant 5 deviation. Sim ilarradiusdiscrepancieshave been described recently by others (e.g.,C lausen et al.1999a;Torres & R ibas 2002; R ibas 2003;Lopez-M orales& R ibas 2005;Torres 2007), although early indications go as far back as the work of H oxie (1973) and Lacy (1977). T he prevailing explanation seem s to be that the enlarged radii of the secondaries,w hich are typically wellunder a solar m ass,are caused by strong m agnetic eldsand/orspotscom m only associated w ith chrom ospheric activity in these system s (see, e.g., M ullan & M acD onald 2001; C habrier et al. 2007,forthetheoreticalcontext).T hesignsofactivity in C V B oo arefairly obvious(spottedness,X -ray em ission), and are no doubt associated w ith the rapid rotation of

F ig.12.| A bsolute dim ensions for C V B oo com pared w ith evolutionary m odels from the series by C laret (2004) w ith a value of the m ixing-length param eter of M L = 1:68. M ass tracks for the exact m asses w e m easure and for solar com position (Z = 0:020, X = 0:70,in these m odels) are indicated w ith dotted curves. T he solid curvesgiving a som ew hatbetter tcorrespond to m odelsw ith a slightly higher m etallicity ofZ = 0:023. A n isochrone for log age = 9.99 is show n for reference (dashed line). T he uncertainty in the location ofthe m asstracksisindicated w ith the errorbarbelow the tracks for the prim ary,and is m uch sm aller than the tem perature uncertainty.

F ig.13.| Sam e as Figure 12,but for a m ixing-length param eter of M L = 1:50. M ass tracks for the m easured m asses and for solar com position (Z = 0:020,X = 0:70) are indicated w ith dotted curves. T he som ew hat better- tting solid curves correspond to a slightly low er m etallicity of Z = 0:0185 in this case, show ing the in uence of the M L param eter in the determ ination of the com position ofC V B oo.

the com ponents.

C V B oo

F ig.14.| M ass-radius diagram for C V B oo, show ing the m easurem ents against isochrones from the Y onsei-Y ale series for the sam e tw o m etallicities displayed in Figure 11. A ges are indicated along the top.

6. C O M PA R ISO N W IT H T ID A L T H E O RY

T he predictions of tidal theory were com pared w ith the observations by com puting the tim e of circularization and synchronization for C V B oo using the radiative dam ping form alism of Zahn (1977) and Zahn (1989), as well as the hydrodynam ical m echanism of Tassoul& Tassoul (1997), and references therein. T he procedure follow s closely that described by C laret et al. (1995) and C laret& C unha (1997). B oth theories predict that synchronization and circularization are achieved very quickly in thissystem by virtueoftheshort orbitalperiod,atan ageofm erely 157 M yr(logt= 8:197, orlessthan 2% oftheevolutionary age).T hefactthatwe m easure the orbit to be circular is therefore not surprising.O n theotherhand,theevidencefrom ourphotom etric observations (x3.1.1) suggesting the rotation m ay be slightly super-synchronousforatleastone ofthe com ponentsism oreinteresting,asitisnotpredicted by theory. G iven the nature of the system ,an activity-related explanation to this discrepancy is certainly possible. M ore precise m easurem ents ofthe projected rotationalvelocities vsin ifor the com ponents would be very helpful. 7. D ISC U SSIO N A N D C O N C LU SIO N S

D espite the system ’s intrinsic variability,the absolute dim ensionsforthecom ponentsofC V B oo havenow been established quite precisely. T he relative errors are better than 1.3% in the m asses and 2% in the radii. T he objectcan now be counted am ong the group ofeclipsing binaries w ith well-know n param eters. U nder di erent circum stances the large num ber and high quality ofthe photom etric observations we have collected m ight have perm itted a m ore detailed study of the lim b darkening law s and a com parison w ith theoretically predicted coe cients,but this possibility was thwarted here by the intrinsic variability. T his phenom enon is not itselfw ithout interest. If interpreted as due to the presence of spots,as we have done here,it im plies that at least one ofthe stars is rotating about 1% m ore rapidly than the

11

synchronous rate, a result that was unexpected for a close but well detached system such as this. W e conclude that our current understanding oftidalevolution is still incom plete, or that other processes are at play in this system that theory does not account for. O ne interesting possibility is di erentialrotation. T he interpretation of m easurem ents of the rotation period of a star m ade by photom etric m eans,as we have im plicitly done here,usually relieson the assum ption ofsolid-body rotation.M oreoften than not,spotsare located atinterm ediate latitudes ratherthan on the equator,or at high latitudes in m ore active stars, and di erential rotation is such that the stellar surface revolves m ore slow ly at higher latitudes,at least in the Sun. T his w illtend to bias photom etric rotation m easurem ents towards longer periods,ifdi erentialrotation is signi cant enough. In C V B oo we see the opposite: the period is shorter than theequatorialrate,assum ing thatsynchronization holds. T hus di erentialrotation can only explain the signalwe have detected ifit is \anti-solar",w ith the polar regions rotating m ore rapidly.A handfulofstarsdo indeed show evidence of weak anti-solar di erential rotation (e.g., IL H ya,H D 31933, G em ,U Z Lib;W eber et al.2003; Strassm eier et al. 2003; K ~ ovariet al. 2007; V ida et al. 2007). T hey allhappen to be very active (som e ofthem w ith high-latitude spots,as in C V B oo),although they tend to be giants or subgiants rather than dwarfs. It is thought that this phenom enon m ay result from fast m eridional ow s(see,e.g.,K itchatinov & R udiger2004). Further progress in understanding the rotation of the C V B oo com ponents could be m ade w ith additionaldifferentialphotom etric observations in severalpassbands, along w ith sim ultaneous high-resolution,high signal-tonoise ratio spectroscopy over a fullorbitalcycle. A nother signi cant discrepancy we nd w ith theory is in the radius of the secondary, w hich appears to be som e10% too largecom pared w ith predictionsfrom stellar evolution m odels. T his di erence is in the sam e direction as seen for a num ber of other low -m ass eclipsing binaries,as m entioned in x5. In those cases one of the explanations m ost often proposed is that the strong m agnetic elds associated w ith activity (w hich is com m on in rapidly-rotating K and M dwarfs in close binaries) tend to inhibit convective m otions,and the structure ofthe staradjustsby increasing itssize to allow the surface to radiate the sam e am ount of energy. A t the sam e tim e, the e ective tem perature tends to decrease in order to preserve the total lum inosity. Spot coverage can produce sim ilar e ects. T heoreticaland observationalevidence for the conservation ofthe lum inosity in these system s has been presented by D elfosse et al. (2000), M ullan & M acD onald (2001), Torres & R ibas (2002),R ibas(2006),Torres et al.(2006),C habrier et al. (2007),and others (see also M orales et al.2008). W e do not see any obvious discrepancy in the tem perature of C V B oo B com pared to m odels,although ouruncertainties are large enough that the e ect m ay be m asked. If we restrict ourselves to well studied double-lined eclipsing binaries in w hich the m ass and radius determ inations are the m ost reliable,deviations from theory such as those described above have usually been seen in stars that are considerably less m assive than the Sun, w hich have deep convective envelopes. H owever,the recent study by Torres et al.(2006) pointed out that the

12

Torres et al.

problem is not con ned to the lower m ass stars, but extends to active objects approaching 1 M , such as V 1061 C yg A b, w ith M = 0:93 M . C V B oo B has an even largerm assof0.968 M ,and also appearsto be oversized. Sim ilarly w ith the virtually identical active star FL Lyr B (M = 0:960 M ). T he convective envelopesofthese objects are considerably thinner than in K and M dwarfs and representonly a few percentofthe totalm ass,yet they appear su cient for m agnetic elds to take hold and alter the globalproperties ofthe star, ifthat is the cause ofthe discrepancies.T hese exam ples show once again thatourunderstanding ofstellarevolution theory isincom plete,even forstarsnearthe m assof the Sun. T he spectroscopic observationsofC V B oo used in this paperwereobtained w ith thegeneroushelp ofP.B erlind, M .C alkins,R .J.D avis,E.H orine,D .W .Latham ,J.Peters,and R .P.Stefanik. R .J.D avis is also thanked for

m aintaining the C fA echelledatabase.W earegratefulas wellto J.M .K reiner for providing unpublished tim es of eclipse for C V B oo,to A .C laret,for calculating speci c m odels for the stars studied here,and to the referee for helpfulcom m ents. G T acknow ledges partialsupport for this work from N SF grant A ST -0708229. LPRV gratefully acknow ledges partial support from the B razilian agencies C N Pq,FA PEM IG and C A PES.Sum m er 2004 A rkansasR EU studentS.L.W altersisthanked by C H SL forherprelim inary analysisofthe absolute propertiesof this binary star (W alters & Lacy 2004). T his research has m ade use of the SIM B A D database, operated at C D S,Strasbourg,France,ofN A SA ’s A strophysics D ata System A bstractService,and ofdata products from the T wo M icron A llSky Survey,w hich is a joint project of the U niversity of M assachusetts and the Infrared Processing and A nalysisC enter/C alifornia Institute ofTechnology,funded by N A SA and the N SF.

R E FE R E N C E S A gerer,F.,& H ubscher,J.2002,IB V S N o.5296 A gerer,F.,& H ubscher,J.2003,IB V S N o.5484 A lencar,S.H .P.,& V az,L.P.R .1997,A & A ,326,257 A lencar,S.H .P.,V az,L.P.R .,& N ordlund, A .1999,A & A ,346, 556 A llard,F.,& H auschildt,P.H .1995,A pJ,445,433 A llard,F.,H auschildt,P.H .,A lexander,D .R .,& Starr eld,S. 1997,A R A & A ,35,137 A ndersen,J.1991,A & A R ev.,3,91 B akis,V .,B akis,H .,E rdem ,A .,C icek,C .,& D em ircan,O .,& B udding,E .2003,IB V S N o.5464 B usch,H .1985,IB V S,N o.2788 C habrier,G .,G allardo,J.,& B ara e,I.2007,A & A ,472,L17 C laret,A .2000,A & A ,363,1081 C laret,A .2004,A & A ,424,919 C laret,A .,G im enez,A .,& C unha,N .C .S.1995,A & A ,299,724 C laret,A .,C unha,N .C .S.1997,A & A ,318,187 C lausen,J.V .,B ara e,I.,C laret,A .,& V andenB erg,D .A . 1999a,in T heory and Tests ofC onvection in Stellar Structure, eds.A .G im enez,E .F.G uinan,& B .M ontesinos,A SP C onf. Ser.173 (San Francisco:A SP ),265 C utri,R .M .et al.2003,\2M A SS A llSky C atalog ofPoint Sources",N A SA /IPA C Infrared Science A rchive D elfosse,X .,Forveille,T .,Segransan,D .,B euzit,J.-L.,U dry,S., Perrier,C .,& M ayor,M .2000,A & A ,364,217 D em arque,P.,W oo,J.-H .,K im ,Y .-C .,& Y i,S.K .2004,A pJS, 155,667 D iethelm ,R .2001,IB V S N o.5027 D ogru,S.S.,D ogru,D .,E rdem ,A .,C icek,C .,& D em ircan,O . 2006,IB V S N o.5707 E ker,Z.1996,A pJ,473,388 E ker,Z.1999,A pJ,512,386 Flem ing,T .A .,M olendi,S.,M accacaro,T .,& W olter,A .1995, A pJS,99,701 Flow er,P.J.1996,A pJ,469,355 G irardi,L.,G rebel,E .K .,O denkirchen,M .,& C hiosi,C .2004, A & A ,422,205 G ranzer,T h.,Schussler,M .,C aligari,P.,& Strassm eier,K .G . 2000,A & A ,355,1087 H all,D .S.1994,M em .Soc.A str.Italiana,65,73 H auschildt,P.H .,B aron,E .,& A llard,F.1997a,A pJ,483,390 H auschildt,P.H .,A llard,F.,A lexander,D .R .,& B aron,E . 1997b,A pJ,488,428 H g,E .,Fabricus,C .,M akarov,V .V .,U rban,S.,C orbin,T ., W yco ,G .,B astian,U .,Schw ekendick,P.,& W icenec,A .2000, A & A ,355,27 H oxie,D .T .1973,A & A ,26,437 H ubscher,J.2005,IB V S N o.5643 H ubscher,J.,Paschke,A .,& W alter,F.2005,IB V S N o.5657 H ubscher,J.,Paschke,A .,& W alter,F.2006,IB V S N o.5731

Isik,E .,Schussler,M .,& Solanki,S.K .2007,A & A ,464,1049 K im ,C .-H .,Lee,C .-U .,Y oon,Y .-N .,Park,S.-S.,K in,D .H .,C ha, S.-M .,& W on,J.-H .2006,IB V S N o.5694 K itchatinov,L.L.,& R udiger,G .2004,A N ,325,496 K j rgaard A ndreasen,G .,H ejlesen,P.M .,Petersen,J.O .1983, A & A ,121,241 K~ ovari,Zs.,B artus,J.,Strassm eier,K .G .,V ida,K ., Sanda,M ., & O lah,K .2007,A & A ,474,165 K reiner,J.M .,K im ,C .H .,& N ha,I.S.2000,A n A tlas ofO C diagram s ofeclipsing binary stars,W ydaw nctw o N aukow e A p, K rakow Lacy,C .H .1977,A pJS,34,479 Lacy,C .H .S.2002,IB V S N o.5357 Lacy,C .H .S.2003,IB V S N o.5487 La er,J.,& K inm an T .D .1965,A pJS,11,216 Latham ,D .W .1992,in IA U C oll.135,C om plem entary A pproaches to D ouble and M ultiple Star R esearch,A SP C onf. Ser.32,eds.H .A .M cA lister & W .I.H artkopf(San Francisco: A SP ),110 Latham ,D .W .,Stefanik,R .P.,Torres,G .,D avis,R .J.,M azeh, T .,C arney,B .W .,Laird,J.B .,& M orse,J.A .2002,A J,124, 1144 Latham ,D .W .,N ordstrom ,B .,A ndersen,J.,Torres,G ., Stefanik,R .P.,T haller,M .,& B ester,M .1996,A & A ,314,864 Locher,K .2005,O pen E uropean Journalon V ariable Stars,3,1 Lopez-M orales,M .& R ibas,I.2005,A pJ,631,1120 M aciejew ski,G .,& K arska,A .2004,IB V S N o.5494 M cLaughlin,D .B .1924,A pJ,60,22 M ochnacki,S.W .1984,A pJS,55,551 M olik,P.2007,O pen E uropean Journalon V ariable Stars,60,1 M orales,J.C .,R ibas,I.,& Jordi,C .2008,A & A ,478,507 M ullan,D .J.,& M acD onald,J.2001,A pJ,559,353 N elder,J.A .,& M ead,R .1965,C om puter Journal,V ol.7,308 N elson,R .H .2000,IB V S N o.4840 N elson,R .H .2002,IB V S N o.5224 N elson,R .H .2004a,IB V S N o.5493 N elson,R .H .2004b,IB V S N o.5535 N ordstrom ,B .,Latham ,D .W .,M orse,J.A .,M ilone,A .A .E ., K urucz,R .L.,A ndersen,J.,& Stefanik,R .P.1994,A & A ,287, 338 Peniche,R .,G onzalez,S.F.,& Pena,J.H .1985,IB V S,N o.2690 Perrym an,M .A .C .,et al.1997,T he H ipparcos and Tycho C atalogues (E SA SP -1200;N oordw jik:E SA ) Popper,D .M .1980,A R A & A ,18,115 Popper,D .M .1997,A J,114,1195 Popper,D .M .2000,A J,119,2391 Popper,D .M .,& Jeong,Y .-C .1994,PA SP ,106,189 R am rez,I.,& M elendez,J.2005,A pJ,626,465 R ibas,I.2003,A & A ,398,239 R ibas,I.2006,A p& SS,304,89

C V B oo R oberts,D .H .,Lehar,J.,& D reher,J.W .1987,A J,93,968 R ossiter,R .A .1924,A pJ,60,15 Schlegel,D .J.,Finkbeiner,D .P.,& D avis,M .1998,A pJ,500, 525 Schlesinger,F.1909a,P ubl.A llegheny O bs.,1,134 Schlesinger,F.1909b,P ubl.A llegheny O bs.,3,28 Schussler,M .,& Solanki,S.K .1992,A & A ,264,L13 Strassm eier,K .G .,K ratzw ald,L.,& W eber,M .2003,A & A ,408, 1103 Tassoul,M .,& Tassoul,J.-L.1997,A pJ,481,363 Torres,G .,& R ibas,I.2002,A pJ,567,1140 Torres,G .,Stefanik,R .P.,A ndersen,J.,N ordstrom ,B .,Latham , D .W .,& C lausen,J.V .1997,A J,114,2764 Torres,G .,Lacy,C .H .S.,M arschall,L.A .,Sheets,H .A .,& M ader,J.A .2006,A pJ,640,1018 Torres,G .2007,A pJ,671,L65 V az,L.P.R .,A ndersen,J.,& C laret,A .2007,A & A ,469,285 V ida,K .,K ~ ovari,Zs., Svanda,M .,O lah,K .,Strassm eier,K .G ., & B artus,J.2007,A N ,328,1078 V oges,W .et al.2000 \R O SA T A ll-Sky Survey Faint Source C atalog",IA U C .7432R .1V

13

V ogt,S.S.,H atzes,A .P.,& M isch,A .A .1999,A pJS,121,547 W alters,S.L.,& Lacy,C .H .S.2004,B A A S.,36,1370 W eber,M .,Strassm eier,K .G .,& W ashuettl,A .2003,in C ool Stars,Stellar System s,and the Sun,eds.A .B row n,G .M . H arper,and T .R .A yres,(U niversity ofC olorado),p.922 W ilson,R .E .1979,A pJ,234,1054 W ilson,R .E .1990,A pJ,356,613 W ilson,R .E .1993,in:N ew Frontiers in B inary Star R esearch, eds.K .C .Leung abd L.-S.N ha,A P S C onf.Series,38,91 W ilson,R .E .,& B ierm ann,P.1976,A & A ,48,349 W ilson,R .E .,& D evinney,E .J.1971,A pJ,166,605 Y i,S.K .,D em arque,P.,K im ,Y .-C .,Lee,Y .-W .,R ee,C .H ., Lejeune,T .,& B arnes,S.2001,A pJS,136,417 Zahn,J.-P.1977,A & A ,57,383 Zahn,J.-P.1989,A & A ,220,112 Zucker,S.,& M azeh,T .1994,A pJ,420,806

14

Torres et al. TA B LE 1 D ifferential V -band m easurem ents of C V B oo. H JD

2;400;000

52250.99816 52250.99907 52251.00000 52251.00091 52251.00182

P hase

V

0.35467 0.35575 0.35685 0.35792 0.35900

+ 0.452 + 0.416 + 0.445 + 0.508 + 0.433

N ote. | Table 1 is available in its entirety in the electronic edition ofthe A stronom ical Journal. A portion is show n here for guidance regarding its form and contents.

TA B LE 2 P hotom etric indices and inferred m ean effective tem perature of C V B oo. P hotom etric Index Johnson V ................. Johnson B V ............. T ycho-2 B T VT .......... Johnson/2M A SS V J ..... Johnson/2M A SS V H ..... Johnson/2M A SS V K s .... T ycho-2/2M A SS VT K s .. Sloan g r .................

V alue 10.75 0.73 0.82 1.18 1.47 1.55 1.629 0.473

0.10 0.11 0.13 0.10 0.10 0.10 0.081 0.002

Te (K )

R ef. 1

5417 5448 5693 5666 5692 5679 5760

350 329 103 155 157 129 100

1 1 1,2 1,2 1,2 1,2 3

N ote. | R eferences: (1) H g et al.(2000);(2) C utriet al.(2003); (3) Sloan D igitalSky Survey data.

C V B oo

15

TA B LE 3 N ew radial velocity m easurem ents of C V B oo. H JD 2 440 000

P hase

Star A

(O C ) A

Star B

km =s

48408.8881 48428.7637 48435.7679 52336.9667 52362.8251 52391.8515 52395.7780 52419.8759 52420.8446 52424.9263 52481.7239 52537.6015 52657.0283 52681.9996 52687.0564 52688.0019 52690.9568 52712.0278 52718.9482 52720.9903 52721.8866 52743.0017 52743.8681 52745.9296 52748.8870 52751.9548 52752.8104 52769.7356 52771.8062 52773.8521 52800.6851 52802.6741 52804.7941

0.1792 0.6453 0.9148 0.8530 0.3826 0.6526 0.2884 0.7395 0.8832 0.7022 0.7601 0.7319 0.7327 0.2150 0.1853 0.3016 0.7903 0.6677 0.8383 0.2493 0.3075 0.2370 0.2599 0.6938 0.1854 0.8074 0.8176 0.8003 0.2449 0.6604 0.3407 0.6890 0.1920

127.53 + 113.28 + 70.53 + 101.70 96.29 + 115.23 126.96 + 133.04 + 80.62 + 119.97 + 131.60 + 145.12 + 136.61 139.26 126.70 136.00 + 137.79 + 117.71 + 103.39 142.32 131.97 138.96 138.48 + 127.96 124.72 + 122.44 + 118.39 + 123.65 139.56 + 103.84 116.42 + 120.03 127.64

2.12 + 6.33 + 1.73 6.29 2.48 + 4.57 + 7.58 1.97 9.96 9.29 3.44 + 10.69 + 2.10 3.99 + 0.86 4.61 + 6.78 + 0.12 11.69 3.84 2.26 0.91 0.26 + 1.02 + 2.87 4.20 4.95 5.00 1.14 10.53 + 0.57 5.45 + 2.03

+ 144.05 121.89 80.45 121.62 + 97.00 126.58 + 131.35 156.34 102.38 137.99 147.89 153.73 143.34 + 149.89 + 125.61 + 134.21 152.45 132.03 126.31 + 135.44 + 135.59 + 141.72 + 136.36 152.37 + 133.76 141.35 137.92 141.11 + 137.19 124.10 + 128.81 138.94 + 145.25

N ote. | T he O

km =s

(O C ) B

+ 13.68 4.64 4.19 3.39 + 0.31 5.38 8.81 9.07 2.82 + 3.09 0.59 7.10 + 3.38 + 8.92 7.06 2.57 9.46 3.42 0.46 8.97 + 0.61 2.22 7.76 13.76 + 1.05 3.07 3.18 0.66 7.15 + 1.06 + 7.42 1.90 + 10.30

H JD 2 440 000

52805.7304 52807.6959 52808.6829 52828.6620 52830.7704 52894.6202 53011.0543 53017.0656 53036.0482 53045.0203 53047.9883 53072.0214 53102.9332 53124.9836 53125.8812 53131.8024 53133.8280 53134.8008 53155.9103 53156.8078 53157.6883 53158.8651 53159.7571 53182.6826 53183.7214 53183.8170 53184.7284 53189.7763 53190.6836 53192.6964 53217.6764 53452.8782 53485.8782

P hase

Star A

0.2974 0.6180 0.7833 0.3715 0.8608 0.2449 0.7124 0.8097 0.2214 0.8143 0.3184 0.6931 0.1890 0.2227 0.2825 0.2733 0.6648 0.8134 0.7362 0.7959 0.8354 0.2248 0.2779 0.3449 0.5713 0.6842 0.7602 0.7200 0.7912 0.1676 0.6602 0.3505 0.3118

133.61 + 79.11 + 141.29 116.79 + 104.81 148.22 + 138.93 + 130.12 130.15 + 110.06 133.81 + 120.32 128.08 148.22 135.40 148.17 + 113.70 + 123.52 + 135.62 + 124.95 + 118.15 132.66 137.36 107.94 + 55.26 + 118.57 + 139.79 + 134.61 + 129.10 123.51 + 113.85 114.77 121.62

km =s

(O C ) A

1.12 12.00 + 8.91 16.14 + 0.97 9.80 + 7.38 + 4.16 + 6.19 14.40 7.71 6.41 + 0.67 11.69 + 0.26 11.15 2.64 1.25 + 0.82 4.81 + 1.81 + 4.16 0.97 + 7.08 2.71 5.32 + 4.76 + 1.70 1.72 2.64 0.42 2.51 + 6.74

Star B km =s

(O C ) B

+ 134.43 96.25 158.46 + 112.11 115.59 + 145.71 147.38 144.86 + 149.26 137.98 + 138.58 138.36 + 140.61 + 142.32 + 136.82 + 142.70 125.78 134.88 159.54 145.87 128.87 + 136.23 + 141.77 + 117.19 66.52 136.77 147.31 150.90 146.12 + 140.40 123.95 + 115.01 + 125.58

3.53 + 4.12 14.00 + 8.13 1.81 + 1.37 3.84 7.31 + 7.14 2.03 + 7.46 + 0.02 + 6.65 0.01 4.54 0.12 + 1.49 + 1.40 12.50 4.23 1.66 6.41 0.38 2.10 + 2.40 1.44 0.01 5.90 3.34 + 14.92 + 1.11 1.34 7.96

C residuals correspond to the solution described in x 3.1.

TA B LE 4 R adial velocities for C V B oo from P opper (2000). H JD 2 440 000

P hase

Star A

(O C ) A

Star B

km =s

47198.0795 47254.9417 47397.6592 47695.7285 47696.7035 48080.7574 48081.6981 48312.0038 48344.9980 48345.8895 48345.9531 48819.6895 48819.7122 48819.7362 48820.6789 48820.7049 48822.7202 48822.8020 49116.9777 49117.8084 49117.8249 49117.8549 49117.8748

0.6418 0.7760 0.2749 0.1895 0.3407 0.7727 0.8833 0.7931 0.7476 0.8001 0.8752 0.1906 0.2174 0.2457 0.3587 0.3894 0.7688 0.8654 0.1830 0.1638 0.1832 0.2187 0.2421

+ 100.59 + 134.29 137.69 129.31 112.58 + 134.39 + 87.15 + 128.21 + 131.13 + 128.12 + 96.80 137.81 140.25 143.18 111.27 92.10 + 135.49 + 97.14 131.43 121.42 133.33 137.56 140.97

4.51 + 0.77 0.88 0.39 + 4.42 + 0.45 3.33 2.20 4.16 0.57 + 1.25 8.56 4.55 4.74 3.34 2.69 + 1.12 4.17 4.67 2.21 6.49 1.65 2.64

112.85 144.43 + 142.48 + 136.69 + 121.75 144.60 98.55 139.74 144.15 138.79 105.62 + 137.28 + 140.79 + 141.52 + 107.70 + 92.66 149.18 112.48 + 132.23 + 122.85 + 135.33 + 142.79 + 145.24

km =s

(O C ) B

+ 2.43 + 1.25 0.12 + 2.54 + 0.35 + 1.53 + 0.90 + 2.60 + 3.42 + 1.70 0.74 + 2.78 0.64 2.85 4.03 + 0.66 2.59 1.42 + 0.41 0.85 + 3.42 + 1.14 + 0.99

H JD 2 440 000

49117.9386 49202.6852 49204.7213 49204.7416 49204.7623 49496.9018 49496.9163 49583.6596 49583.6807 49907.6918 49907.7182 49907.7666 49907.7881 50176.9827 50177.0097 50177.9325 50177.9511 50177.9691 50177.9891 50178.0072 50178.0259 50616.7131

P hase

Star A

(O C ) A

Star B

km =s

km =s

(O C ) B

0.3175 0.3733 0.7772 0.8012 0.8256 0.7392 0.7563 0.1695 0.1944 0.7371 0.7683 0.8254 0.8508 0.6746 0.7065 0.7960 0.8179 0.8392 0.8628 0.8842 0.9063 0.8409

126.53 102.13 + 133.70 + 126.62 + 117.42 + 135.56 + 134.71 126.79 128.99 + 136.77 + 139.29 + 124.72 + 108.25 + 122.91 + 133.62 + 129.82 + 124.78 + 112.79 + 103.32 + 88.65 + 77.68 + 113.60

0.09 2.52 + 0.35 1.79 2.96 + 0.57 0.49 5.13 + 1.40 + 1.91 + 4.87 + 4.28 0.84 + 2.53 + 3.34 + 0.09 + 1.57 1.87 + 0.58 1.30 + 0.27 0.27

+ 133.81 + 101.37 145.93 140.09 130.86 147.61 146.30 + 126.42 + 132.75 148.30 148.07 135.86 121.30 131.47 140.19 138.96 131.51 120.04 113.06 99.35 84.92 121.75

+ 2.32 1.49 0.43 + 0.10 + 0.69 0.36 + 1.18 + 0.08 2.98 1.19 1.43 4.24 1.88 + 0.11 + 1.99 + 2.65 + 3.09 + 5.37 0.46 0.46 2.06 + 2.80

N ote. | Sm allcorrections to the published values have been applied as described in x 2.3. T he O C residuals correspond to the solution described in x 3.1.

16

Torres et al. TA B LE 5 Spot param eters for C V B oo. C o-latitude (deg)

Longitude (deg)

R adius (deg)

Tfactor

pri

8:062 59

5:47 90

44:50 50

0.82 5

b

sec

7:926 52

170:212 10

47:76 69

0.775 19

c

pri

4:589 63

13:058 43

35:835 46

0.5940 55

c

sec

4:665 43

151:43 38

55 1

0.881 11

C ase

C om ponent

a

N ote. | T he spot co-latitude is m easured from the pole visible to the observer, and the longitude is m easured from the line joining the com ponents’ centers and increasing in the direction of orbital m otion. T he radius is m easured as seen from the center of each com ponent,and the tem perature factor is relative to the unspotted photosphere. T he uncertainties listed are in units of the last decim alplace and correspond to the internalerrors from the least-squares m ethod.

TA B LE 6 Light-curve solutions for C V B oo based on our V -band photom etry. Param eter i ( )........... A

............

B

............

Te

;B

(K ).....

a (R )........ (km s 1 ).... q

M

B

=M

A

...

(‘B =‘A )V ...... (‘B =‘A )V ;0:25 .. V (m ag)...... xb olo;A ........ xb olo;B ........ xV ;A .......... xV ;B ..........

N o spots

C ase (a)

C ase (b)

C ase (c)

Param eter

87.651 42 4.6752 76 4.9014 78 5632.8 1:6 4.757 12

86.891 34 4.6591 48 4.9614 50 5628.1 1:4 (4.748) ( xed)

86.650 33 4.7844 49 4.8186 47 5656.1 2:4 (4.748) ( xed)

86.237 32 4.7495 46 4.8707 43 5672.6 4:3 (4.748) ( xed)

rA ;p ole ..........

15:877 31 0.9376 24 0.741 10 0.734 0.0196 0.428 0.437 0.715 0.724

( 15:889) ( xed) (0.9378) ( xed) 0.746 4 0.699 0.0148 0.429 0.437 0.715 0.724

( 15:889) ( xed) (0.9378) ( xed) 0.769 21 0.829 0.0147 0.428 0.435 0.715 0.722

( 15:889) ( xed) (0.9378) ( xed) 0.778 14 0.802 0.0146 0.428 0.434 0.715 0.721

rB ;p ole ..........

rA ;p oint ......... rA ;side .......... rA ;back ......... rA ;vol ...........

rB ;p oint ......... rB ;side .......... rB ;back ......... rB ;vol ........... A B

.............. ..............

N o spots

C ase (a)

C ase (b)

C ase (c)

0.26533 67 0.28105 90 0.27032 73 0.27726 82 0.27189 74

0.26646 49 0.28268 67 0.27167 53 0.27877 62 0.27252 55

0.25798 47 0.27188 61 0.26254 51 0.26870 58 0.26327 51

0.26028 46 0.27478 62 0.26501 50 0.27142 57 0.26577 51

0.24054 98 0.25177 117 0.24410 104 0.24935 111 0.24483 105 0.378 0.390

0.23699 86 0.24757 101 0.24042 91 0.24535 94 0.24107 91 0.378 0.390

0.24573 84 0.25825 101 0.24972 90 0.25548 96 0.25049 90 0.378 0.388

0.24247 81 0.25423 97 0.24624 86 0.25168 93 0.24697 87 0.378 0.386

N ote. | In allspotted solutions both com ponents w ere assum ed to rotate at a rate 1.0114 tim es faster than the orbitalm otion (see x 3.1.1). For the no-spot solution the rotation is assum ed to be synchronous. T he linear lim b-darkening coe cients (x) as w ellas the gravity-brightening exponents ( ) are m arked w ith an asterisk to indicate that they w ere changed dynam ically during the iterations as T eff and log g changed. T he gravity-brightening exponents varied over the m utually illum inated stellar surfaces follow ing A lencar & V az (1997) and A lencar et al. (1999), and the values presented here are for the non-illum inated hem ispheres. T eff;A w as held xed at 5760 K .T he quantity (‘B =‘A )V ;0:25 corresponds to the V -band light ratio at the rst quadrature w ithout considering the e ect of spots,and (‘ B =‘A )V is the m ean light ratio outside of eclipse accounting for the spots and proxim ity e ects. T he uncertainties given on the left-hand side of the table (in units of the last decim alplace) are the form al internal errors of the m inim ization procedure, w hile the ones on the right for the com ponent radii account for the uncertainties of the gravitationalpseudo-potentials as w ell as the m ass ratio. T he quantities rA ;vol and rB ;vol represent the \volum e radius" for each star, i.e., the radius of a sphere w ith the sam e volum e as the distorted stars.

C V B oo

17

TA B LE 7 P hysical param eters of C V B oo. Param eter A bsolute dim ensions M ass (M )................. R adius (R )............... log g (cgs).................. M easured v sin i (km s 1 )... vasync sin i (km s 1 )........ vsync sin i (km s 1 )......... R adiative and other properties Te (K ).................... log L =L ................... M b ol (m ag) ................ M V (m ag).................. L B =L A ..................... D istance (pc)...............

P rim ary

Secondary

1.032 1.262 4.249 73 78.5 77.6

0.013 0.023 0.016 10 1.1 1.1

0.968 1.173 4.285 67 72.7 71.9

0.012 0.023 0.017 10 1.1 1.1

5760 0.197 4.24 4.32

150 0.048 0.12 0.12 0.81 259

5670 0.107 4.46 4.57 0.13 16

150 0.049 0.12 0.13

N ote. | M V and M b ol w ere com puted using bolom etric corrections from F low er (1996) along w ith M b ol = 4:732. T he predicted asynchronous projected rotational velocities vasy n c sin i correspond to the values assum ing the rotationalperiod is P orb =1:0114 for both stars (see x 3.2),w hile the vsy n c sin i values give the result ifP rot = P orb . In both cases w e use the radius of the stars at quadrature.