VLTFORS spectroscopy of faint cataclysmic variable

arXiv:astro-ph/0609196v17Sep2006Mon.Not.R.Astron.Soc.000,000–000(0000)Printed5February2008(MNLATEXstylefilev2.2)VLT/FORSspectroscopyoffaintcataclysmicvariablesdiscoveredbytheSloanDigitalSkySurveyJohnSouthworth1⋆,B.T.G¨ansicke1,T.R.Marsh1,D.deMartino2,P.Hakala3,S.Littlefair4,P.Rodr´iguez-Gil5,P.Szkody61DepartmentofPhysics,UniversityofWarwick,Coventry,CV47AL,UK2INAF-OsservatoriodiCapodimonte,ViaMoiariello16,80131Napoli,Italy3Observatory,UniversityofHelsinki,POBox14,Helsinki,Finland4DepartmentofPhysicsandAstronomy,UniversityofSheffield,S37RH,UK5InstitutodeAstrof´isicadeCanarias,38200LaLaguna,Tenerife,Spain6AstronomyDepartment,UniversityofWashington,Seattle,WA98195,USA5February2008ABSTRACTWepresentmedium-resolutionVLT/FORS2spectroscopyofsixcataclysmicvari-ablesdiscoveredbytheSloanDigitalSkySurvey.WedetermineorbitalperiodsforSDSSJ023322.61+005059.5(96.08±0.09min),SDSSJ091127.36+084140.7(295.74±0.22min),SDSSJ103533.02+055158.3(82.10±0.09min),andSDSSJ121607.03+052013.9(mostlikely98.82±0.16min,buttheone-dayaliasesat92minand107minarealsopossible)usingradialvelocitiesmeasuredfromtheirHαandHβemissionlines.Threeofthefourorbitalperiodsmeasuredhereareclosetotheobserved75–80minminimumperiodforcataclysmicvariables,indicatingthatthepropertiesofthepopulationoftheseobjectsdiscoveredbytheSDSSaresubstantiallydifferenttothoseofthecataclysmicvariablesfoundbyothermeans.AdditionalphotometryofSDSSJ023322.61+005059.5revealsaperiodicityofapproximately60minwhichweinterpretasthespinperiodofthewhitedwarf,suggestingthatthissystemisanintermediatepolarwithalowaccretionrate.SDSSJ103533.02+055158.3hasaperiodrightattheobservedminimumvalue,aspectrumdominatedbythecoolwhitedwarfprimarystarandexhibitsdeepeclipses,soisanexcellentcandidateforanaccuratedeterminationoftheparametersofthesystem.ThespectroscopicorbitofSDSSJ121607.03+052013.9hasavelocityamplitudeofonly13.8±1.6kms−1,implyingthatthissystemhasanextrememassratio.Fromseveralphysicalcon-straintswefindthatthisobjectmustcontaineitherahigh-masswhitedwarforabrown-dwarf-masssecondarycomponentorboth.Keywords:stars:novae,cataclysmicvariables–stars:binaries:close–stars:bina-ries:eclipsing–stars:whitedwarfs–stars:individual:SDSSJ023322.61+005059.5–stars:individual:SDSSJ091127.36+084140.7–stars:individual:SDSSJ093238.21+010902.5–stars:individual:SDSSJ101037.05+024914.9–stars:indi-vidual:SDSSJ103533.02+055158.3–stars:individual:SDSSJ121607.03+052013.9–stars:individual:SDSSJ133941.11+484727.51INTRODUCTIONCataclysmicvariables(CVs)areinteractingbinarystarscontainingawhitedwarfprimarystarandalow-masssec-ondarystarinacloseorbit.Inthemajorityofsystemsthesecondarystarisunevolved,fillsitsRochelobeandloses⋆E-mail:[e-mail](JS),[e-mail](BTG),[e-mail](TRM)masstotheprimarystarviaanaccretiondisc.Comprehen-sivereviewsofthepropertiesofCVshavebeengivenbyWarner(1995)andHellier(2001).TheevolutionofCVsdependsprimarilyonangularmo-mentumloss(AML)fromtheirorbit.CVswithlongperiodscontainasecondarystarwhichisonlypartiallyconvective.ItisthoughtthatthepresenceofamagneticfieldinthisstarcausesAMLbymagneticbraking(Verbunt&Zwaan1981;Rappaport,Joss,&Webbink1982)untiltheorbitalperioddecreasestoaboutthreehours.Atthispointthesecondaryc0000RAS2Southworthetal.starbecomesfullyconvectiveandmagneticbrakingceases.ThischangeinthestateofthestarcausesittorelaxtoitsthermalequilibriumradiusandsoshrinkinsideitsRochelobe.MasstransferceasesandweakerAMLisnowdomi-natedbygravitationalradiation(Paczy´nski1967).Oncetheperiodreachesabouttwohours,theshrinkingRochelobere-establishescontactwiththesecondarystar,restartingthemasstransfer.Theperiodcontinuestodecreasetoanobservedminimumofapproximately80minutes,atwhichpointthesecondarystarbecomesadegeneratebrown-dwarflikeobject.Asthemassdonorisnowdegenerate,furthermasstransfercausestheperiodtoincrease(Patterson1998).Systemsbeyondthis”periodbounce”aretheoreti-callypredictedtorepresentupto70%ofthesteady-statepopulationofCVs;populationsynthesismodelsalsopredictthatthevastmajority(99%)ofCVshaveevolvedtoperiodsshorterthanthe2–3hourgap(deKool1992;deKool&Ritter1993;Kolb1993;Politano1996;Kolb&Baraffe1999).ThemodelsalsopredictthatthelowrateofchangeofperiodneartheminimumvaluecausesanincreaseinthepopulationofCVsthere,whichshouldman-ifestitselfasaspikeintheperioddistribution.However,whilstseveralcandidatesexist,notoneCVhasyetbeencon-firmedtohaveasecondarycomponentwithamassappro-priateforaperiodbounceCV(Littlefair,Dhillon,&Mart´ın2003;Patterson,Thorstensen,&Kemp2005b).Inaddition,theobservedperioddistributionofCVsdoesnothaveaspikeattheminimumperiodanditdoesnotcontainfarmoreshort-periodsystemsthanthosewithperiodsbeyondthewell-known2–3hourgap(e.g.Downesetal.2001;Ritter&Kolb2003).Furthermore,thetheoreti-calminimumperiodisgenerallyfoundtobearound65minutes,whichissubstantiallyshorterthantheob-servedvalue.ManydifferentAMLmechanismshavebeenputforwardaspossiblesolutionstotheseproblems(e.g.Andronov,Pinsonneault,&Sills2003;Taam&Spruit2001;Schenkeretal.2002),withoutcompletesuccess.TheAMLratesgivenbyd