A New, Bright, Short-Period, Emission Line Binary in Ophiuchus1 M. A. Stark2 Department of Physics and Astronomy, University of Wyoming, Dept. 3905, 1000 E. University Ave., Laramie, WY 82071 8 [email protected] 0 0 2 Richard A. Wade n Department of Astronomy and Astrophysics, The Pennsylvania State University, 525 Davey Lab, a J University Park, PA 16802 8 [email protected] 1 ] John R. Thorstensen, Christopher S. Peters h p Department of Physics and Astronomy, 6127 Wilder Laboratory, Dartmouth College, Hanover, NH - 03755-3528 o r t s Horace A. Smith, Robert D. Miller a [ Department of Physics and Astronomy, Bio-Physical Sciences Bldg, Michigan State University, East Lansing, MI 48824-1116 1 v 3 E. M. Green 6 9 Steward Observatory, University of Arizona, Tucson, AZ 85721-0065 2 . 1 0 8 0 ABSTRACT : v The 11th magnitude star LSIV−08◦3 has been classified previously as an OB star in the Xi LuminousStarssurvey,oralternativelyasahotsubdwarf. Itisactuallyabinarystar. Wepresent spectroscopy,spectroscopicorbitalelements,andtimeseriesphotometry,fromobservationsmade r a attheKittPeakNationalObservatory2.1m,StewardObservatory2.3m,MDMObservatory1.3m and2.4m,Hobby-Eberly9.2m,andMichiganStateUniversity0.6mtelescopes. Thestarexhibits emissionof varyingstrengthinthe coresofH andHeI absorptionlines. Emissionis alsopresent at 4686˚A (He II) and near 4640/4650˚A (N III/C III). Time-series spectroscopy collected from 2005 July to 2007 June shows coherent, periodic radial velocity variations of the Hα line, which weinterpretasorbitalmotionwitha periodof0.1952894(10)days. High-resolutionspectrashow that there are two emission components, one broad and one narrow, moving in antiphase, as mightarisefrom anaccretiondisk andthe irradiatedface of the mass donorstar. Lesscoherent, low-amplitudephotometricvariabilityisalsopresentonatimescalesimilartothe orbitalperiod. Diffuse interstellar bands indicate considerable reddening, which however is consistent with a distance of ∼100–200 pc. The star is the likely counterpart of a weak ROSAT X-ray source, whosepropertiesareconsistentwithaccretioninacataclysmicvariable(CV)binarysystem. We ◦ classify LSIV−08 3 as a new member of the UX UMa subclass of CV stars. Subject headings: binaries: close — novae, cataclysmic variables — stars: emission-line — stars: indi- vidual(LSIV−08◦3) — stars: variables: other 1 1. Introduction approximately 4600–6750˚A and 6400–8900˚A at ∼1.3˚A pixel−1. The features observed in the ◦ ThestarLSIV−08 3hasV =11.475,(B−V)= GoldCam spectra include broad absorption lines +0.190 (Reed & Niemczak 2000), and was first with emission cores in Hα, Hβ, and He I 5875˚A; classified “OB” in the Luminous Stars in the emission from He I 6678˚A, He II 4685˚A, and NorthernMilkyWaycatalog(Nassau & Stephenson N III/C III ∼4640/4650˚A. The emission compo- 1963). Lateritwasreclassified“sdB”byKilkenny & Busse nents were observed to vary in both strength and (1992)basedonReticonspectrogramsandStro¨mgren profile in as little as ∼30 minutes (Fig. 1; see also photometry. Additionalphotometryofthisobject §2.2). is available in the literature (Reed et al. 1998; All GoldCam spectra were processed using Reed 1998; Reed & Niemczak 2000), but no ad- standard IRAF3 routines. Unfavorable observ- ditional spectroscopic observations have been re- ing conditions prevented reliable flux calibrations ported. Astrometricandphotometricdataforthis on mostnights, so we discuss only the normalized star, collected from the literature, are reported in spectra. The continua were fitted by hand using Table 1. splines. The equivalent widths (EWs) of features ◦ We obtained spectra of LSIV−08 3 as part of of interest were measured in each spectrum; the ourprogramoncompositespectrumhotsubdwarf wavelength interval encompassed by each feature stars (Stark & Wade 2003, 2005; Stark 2005; was determined by eye, noting where the line de- Stark & Wade 2006), owing to its unusually red viated from the continuum. For lines with broad V −Ks andJ−Ks colors as observedin the Two absorptionwingsandcoreemission,twoEWmea- Micron All-Sky Survey (2MASS; Skrutskie et al. surementsweremade. ThefirstisthetotalEWof ◦ 2006). We found that LSIV−08 3 exhibits broad thewholeline(absorptionandemissionincluded). absorptionlines with core emission, which is vari- ThesecondEWreferstojusttheemissioncompo- able in both strength and profile shape. We nent, measured above the point where the emis- obtained time-series photometry and additional sion emerges from the absorption wings, giving a spectroscopy in order to correctly classify this lower limit for the “true” emission EW. Reported star. in Table 3 are the average EWs and their for- mal errors (not including any systematic errors 2. Observations or accounting for variability). Also reported are the number of measurements, their standard de- 2.1. Low-Dispersion Spectra viation, minimum and maximum values, and the 2.1.1. Kitt Peak National Observatory 2.1m difference between the maximum and minimum. Telescope (Note that EW> 0 indicates absorption, while EW<0 indicates emission.) ◦ LSIV−08 3 was observed from the Kitt Peak Significant reddening of this object is ev- NationalObservatory(KPNO)2.1mtelescopeus- idenced by strong interstellar Na I D lines ing the GoldCam spectrograph during 2003 June and several diffuse interstellar bands (DIBs). and September. A journal of observations is Schlegel, Finkbeiner, & Davis (1998) report total given in Table 2. All exposures were between galactic reddening of E(B−V)≈+0.48along the 5 and 20 minutes long. Two different spectro- ◦ line of sight to LSIV−08 3. The reddening and graph configurations were used, which covered distance are discussed in §3.2. 1Based in part on observations obtained with the 2.1.2. Steward Observatory 2.3m Telescope Hobby-EberlyTelescope,whichisajointprojectoftheUni- versityofTexasatAustin,thePennsylvaniaStateUniver- ◦ Aflux-calibratedspectrumofLSIV−08 3,cov- sity,StanfordUniversity,Ludwig-Maximillians-Universita¨t Mu¨nchen,andGeorg-August-Universita¨tGo¨ttingen. ering 3620–6895˚A at ∼3˚A pixel−1 dispersion 2VisitingAstronomer,KittPeakNationalObservatory, National Optical Astronomy Observatory, which is oper- 3IRAF is distributed by the National Optical Astronomy atedbytheAssociationofUniversitiesforResearchinAs- Observatories, which are operated by the Association of tronomy, Inc. (AURA) under cooperative agreement with Universities for Research in Astronomy, Inc., under coop- theNationalScienceFoundation. erativeagreement withtheNationalScienceFoundation. 2 (∼9˚Aresolution),wasobtainedonUT2002April telescope and modular spectrograph, equipped 3 at the Bok 2.3m telescope on Kitt Peak, Ari- with a SITe 20482 CCD; this gave 2˚A pixel−1 zona,usingtheB&Cspectrographwithathinned, dispersion from 4300 to 7500˚A, with severe vi- back-illuminated 1200 × 800 Loral CCD. This gnetting towardthe ends of this range. The spec- “pre-discovery”spectrum is shown in Figure 2. It tral resolution of this setup was slightly better extendstoshorterwavelengthsthentheGoldCam than that of the Mark III. Some of these expo- observations, including the Balmer lines up to sures were taken in bright twilight and bracketed the series limit. A second spectrum was obtained with comparison exposures; for those taken away with the same instrument setup on UT 2007 May from twilight, the night-sky features were used 8 and was flux-calibrated in a homogeneous way to find the zero point shift. Extensive internal with the first; the overall spectral energy distri- checks show that this procedure worked well with bution is very similar, with the flux level differing this setup. Aliasing related to daily cycle count by about 3%. The H emission lines were weaker uncertainties was suppressed by pushing to large in 2007, with the emission reversal in Hγ virtu- hour angles;the target’sbrightnessmade it possi- ally disappearing and the height of Hα emission ble to obtain good signal-to-noise spectra even at abovethe continuumdiminishing by a factorof2. large airmass. ◦ Kilkenny & Busse(1992)classifiedLSIV−08 3as Flux standard stars were observed to calibrate “sdB” based on uvby photometry and 2˚A resolu- the instrument response, when the twilight sky tion Reticon spectroscopy covering 3900–4600˚A, was reasonably clear. Experience suggests that so Hβ was not observed, and weak emission in the zero point of the MDM spectra should be ac- the absorption core of Hγ might easily have been curate to about 20%. Exposures of bright O and overlooked or even absent. The spectral energy B stars were also obtained and used to map and distribution (SED) is quite blue, even despite the correct for the telluric absorption features. reddening evidenced by the DIBs. The SED is An average flux-calibrated spectrum was con- further discussed in §3. structed from the 1.3m data taken in 2005 June and July. Using the IRAF sbands task and the 2.1.3. MDM Observatory 1.3m and 2.4m Tele- passband tabulated by Bessell (1990), we find a scopes synthetic V magnitude of 11.47. The MDM 1.3m Time-series spectra of LSIV−08◦3 were ob- average spectrum is very similar to the Steward tained at the MDM Observatory4 on Kitt Peak, 2.3mspectrum from 2002April in the wavelength Arizona. Table 4 gives a journal of observations. region common to both, except the overall flux Most of the MDM spectra were obtained using level is about 10% higher, and the emission lines the MarkIII spectrographanda SITe10242 CCD are more evident (the Hβ emission reversal pro- detector mounted on the McGraw-Hill 1.3m tele- trudes slightly above the continuum; the He I scope. A600linemm−1grismgave∼2.3˚Apixel−1 λ6678 line is in emission at about 15% of the in- dispersionoverthe range4660–6900˚A, with a full tensity of Hα). widthhalfmaximum(FWHM)resolutionof4.1˚A. 2.1.4. Radial Velocity Variations All stellar exposures were bracketed with spectra of Hg, Ne, and Xe comparison lamps, in order All ofthe MDM spectra showHα emission. To to track the spectrograph flexure. (In 2004 June, measuretheradialvelocity(RV),weconvolvedthe an attempt was made to use the λ5577 night-sky line profile with the derivative of a Gaussian and emission feature to set the wavelength zero point; took the line center to be the zero of this convo- unfortunately, this did not work well, and those lution (Schneider & Young 1980). The width of spectra are excluded from the present analysis.) the convolving function was optimized for a line A few MDM spectra are from the Hiltner 2.4m with a FWHM of 11.5˚A, which is approximately the observed line width. We searched the veloc- 4MDM Observatory is operated by a consortium consist- ity time series for periodicities using a “residual- ingofDartmouthCollege,ColumbiaUniversity,OhioState gram” technique (Thorstensen et al. 1996), which University, the University of Michigan, and Ohio Univer- isespeciallyeffectivewhenthemodulationisaccu- sity. ratelysinusoidal. Becausethedataspanhundreds 3 ◦ of days, we were careful to use a sufficiently fine shortened exposures obtained as LSIV−08 3 ap- gridoftrialfrequencies. Theperiodsearchyielded proachedtheendoftheHETtrack. Threeorfour a single, uniquely-defined frequency near 5.12 cy- separate exposures were obtained in quick suc- cle d−1; a least-squares fit to the velocities of the cession on each of four nights in 2004 and three form nights in 2007 (see Table 5 for a journal of obser- 2π(t−T ) 0 vations). The typicalsignal-to-noiseratiowas100 v(t)=γ+Ksin (cid:20) P (cid:21) at 5822˚A (range 75–120). Th-Ar lamps and flat then yielded a preliminary period of 0.195290 d, field exposures were taken either before or after with a velocity semiamplitude of 80 km s−1. the sequence of exposures each night. Twenty-three additional RVs were used to ex- These spectra were processed with standard tend the span of RV observations and to improve IRAF routines. At the dispersionof the HRS, the the precision of the ephemeris. These were mea- absorption components of the Balmer lines in the ◦ sured using the same 11.5˚A FWHM Gaussian- LSIV−08 3 spectrum are roughly as broad as a derivative technique, applied to high-resolution single spectral order. Thus, a pseudo-continuum spectra from the Hobby-Eberly Telescope (HET) was fitted to the broadH absorptionprofiles, and obtained in 2004 and 2007 (see §2.2 below). The each spectrum was normalized using this fit. The final heliocentric Hα RV ephemeris, based on 163 data were then rebinned in uniform heliocentric points, is given by: wavelength bins. Figure 4 shows the normal- ized Hα core emission profiles at several key or- T = HJD 2453628.6389±0.0016, bital phases. As the phase varies, the profile does 0 not simply shift back and forth in velocity, but P = 0.1952894±0.0000010d, rather changes shape. To the eye, a decomposi- K = 77±4 km s−1, and tion into a broad component and a narrow com- γ = −44±3 km s−1. ponent moving in anti-phase is suggested. This is made more evident in Figure 5, which gives a trailed-spectrogramrepresentationofallthe data. The rms residual between the data and the fit- The measured EW of the Hα emission (above ted curve is 17 km s−1. Figure 3 shows the RVs the pseudo-continuum) was roughly constant at folded on the period, with the best-fitting sinu- ∼3.0˚A in 2004, but was lower in 2007, averag- soidsuperposed. TheperiodicityintheHαRVsis ing 1.7, 2.1, and 1.8˚A on 2007 April 19, June 8, coherentoverthreeobservingseasons,andthe ve- and June 22 respectively. To make Figs. 4 and 5, locities admit no possibility of a cycle count error we have “stretched” each spectrum to achieve a — the data are consistent with a binary orbit. uniform emission EW of 3.0˚A, keeping the con- tinuum level at unity. In the trailed spectrogram 2.2. High-Resolution Spectra (Fig.5),eachspectrumhasthenbeenmadetofilla Follow-up observations were taken at the phase range of 0.05 cycles, centered on the phase HET using the High Resolution Spectrograph bin (bin width = 0.01 cycles) nearest to the or- (HRS; Tull 1998). Owing to the declination of bital phase at mid-observation. Note that data LSIV−08◦3, it can only be observed by the HET from 2004 and 2007 have been intermingled, and for about one hour per night as it is transit- different orbital phases were observed in different ing, from mid-April through mid-July. Queue years. The change in EW from one year to the observations of LSIV−08◦3 were made with the other may signal a change in the behavior of the HRS using 2′′ fibers (one object fiber and two profilevsorbitalphase,sowecautionthatquanti- sky fibers) and 2×3 on-chip binning, resulting in tativeconclusionsaboutthesedistinctnarrowand R ≡ λ/∆λ ≈ 30,000, covering the wavelength re- broadvelocitycomponents,basedonthesefigures, gion from Hβ to Hα. (Unfortunately, the λ5780 are tentative. DIB falls into the gap between the “red” and The phase of the broad component in the Hα “blue” CCDs with this setup.) Exposure dura- emission profile is given by the RV ephemeris de- tions were 10 minutes, with the exception of the rived in §2.1.4, with the narrow component mov- first observation of 20 min duration and a few ing in anti-phase. Thus, the Gaussian-derivative 4 methodusingFWHM=11.5˚Adescribedin§2.1.4 2.3. Photometry likely gives a “diluted” RV semiamplitude for the 2.3.1. Michigan State University 0.6m Telescope broad component. Given the possible change in the profile behavior between 2004and 2007noted DifferentialV bandphotometry of LSIV−08◦3 above, we did not attempt to decompose each was obtained using the Michigan State Univer- observed profile or otherwise make formal two- sity 0.6m telescope equipped with an Apogee component fits to the high-resolution data. A Ap47p CCD camera. Aperture photometry for rough estimate gives FWHM ∼8–10and ∼3˚A for LSIV−08◦3 was obtained relative to the star the broad and narrow components. Estimates of TYC 5642−00482−1, which has BT = 12.389, the RV semiamplitudes are Kbroad ∼ 120 and VT = 11.183. (This star is significantly redder Knarrow ∼ 90 km s−1, with errors of perhaps than the program star, but there is no blue star 20 km s−1. of comparable brightness to the target within the We tentatively identify the broad emission as CCD field of view.) Observations were made on arising from an accretion disk around a compact twelve nights: UT 2004 July 26; 2004 August 1, object and the narrow emission as arising from 7, 9, 16, 22, 31; 2004 September 18; 2005 July the mass donating star in a cataclysmic binary. If 15, 20, 23; and 2006 May 27. Exposure times the narrow emission arises on the hemisphere of were typically 30 s, varying somewhat depending the mass donor that faces the disk (see §3.1), one on observing conditions. Observing run lengths mightexpectthe EWofthenarrowcomponentto variedfromasshortasaboutanhourto6.3hron vary, such that the emission EW is largest when 2006 May 27, this being about the maximum run themassdonorisatsuperiorconjunction,whichis lengthpossibleatreasonableairmassfromthelat- orbital phase 0.5 in our convention. We searched itude of the observatory. Calibrations were done forthiseffectusingthelow-resolutionMDMspec- using standard procedures and sky flats. Since tra from the most extensive data set (2005 July), observations were only taken in the V filter, no but did not find any firm indication of an EW transformations to the standard system involving variation with orbital phase. The limit on the colortermscouldbe attempted. Thetypicalerror variation in the total EW is 15–20% full ampli- of a single observation is ±0.02 mag. tude. It is not possible to decompose these low- The photometry shows LSIV−08◦3 to be vari- resolution spectra into separate components. The ableonbothshort(flickering)andlongtimescales, high-resolutionHRSspectraareunfortunatelynot with no indication of eclipses. By itself, the pho- suitable for this test, since all data collected near tometry suggests no clear or consistent period of orbitalphase0.5comefromthe2007observations, variation — folding the data on numerous trial and all data colleced near orbital phase 0.0 come periods showedmanypossible aliasperiods. Fold- from 2004. While there is a tantalizing difference ing the data on the period derived from the RV in totalEW betweenthese two observingseasons, observations(§2.1.4)doesshowweakmodulations we cannot know whether it is related to orbital or waves, perhaps multi-periodic, in the bright- phase;amoreconcentratedobservingcampaignis ness of LSIV−08◦3 (Fig. 6). The rapid modula- needed. We note that there need not be a large tion in ∆V that is seen in the 2005 Jul 23 data, modulation of the EW of the narrow component, ofamplitude∼0.03mag andlasting severalcycles iftheorbitalinclinationofthebinaryismoderate. overthephaseinterval∼0.2–0.6,isnotseeninthe The Hβ emission line profile shows behavior photometriccheckstars;thismaybesimplyanin- similar to Hα, with the narrow component being stance of strong flickering. The median light level perhaps more pronounced (slightly higher inten- is about∼0.06mag brighterin2005–2006than in sity compared with the broad component). Sim- 2004. Additional time-series photometry, from a ilar behavior is also reflected in the He I λ5876 more southern site to permit longer nightly runs, emission profile, although it is weaker. wouldbedesirableincharacterizingthelightcurve and searching for persistent periodic signals. 2.3.2. Northern Sky Variability Survey ◦ LSIV−08 3 was detected as a variable star 5 in the Northern Sky Variability Survey (NSVS; subclass. These systems show persistent broad ◦ LSIV−08 3 ≡ NSVS 16408817; Wo´zniak et al. Balmer absorption-line spectra. Ironically, these 2004). The NSVS record shows 92 observations lines were once thought to indicate pressure- ofthis star. The typicalerrorofa singlemeasure- broadening in the atmosphere of a hot subdwarf mentis∼0.02mag. Thedataaresparse,collected orwhitedwarf—e.g.,Walker & Herbig(1954)— on45nightsduringtheinterval1999Aprilto2000 and only later were reinterpreted as arising from March, with no more than four observations per an optically thick accretion disk, where they are night. AlightcurvefoldedonourRVperiodshows (at least in part) kinematically broadened. no phase-dependent pattern. The median unfil- UX UMa stars also exhibit a wide range in tered optical magnitude is 11.855 with a scatter strengths of the emission line components. Emis- of 0.052 mag. The total range of observedmagni- sion from combinations of H I, He I, He II, N III, tudes is 11.698±0.026to 11.977±0.068. The data and C III have all been observed at a variety of exhibitvariationsinmedianbrightnessatthe0.05 strengths in various UX UMa stars, or variable mag level with characteristic time scales of 20– over time in a single star of the class. Emis- 40 days. Similar to the epoch-to-epoch variations sion is seen in all of these lines in LSIV−08◦3, seen in the MSU observations, this suggests that and the features are consistent with those ob- the system has been consistent in its photometric served in the UX UMa stars RW Sex and IX Vel behavior over the period of 1999-2006. (Beuermann et al. 1992; Beuermann & Thomas 1990, respectively). For UX UMa stars, the 3. Discussion Balmer decrement in the emission components is steeper than in the absorption lines. This can re- 3.1. Novalike CV Interpretation sultin asituationwhere Hα is purelyin emission, Wehavealreadysuggested(§2.1.2)thatLSIV−08◦3 while highermembersshowprogressivelystronger could have been misclassified as a hot subdwarf (less filled-in) absorption troughs. This effect is ◦ (sdB) star, given only photometry and spec- seen in LSIV−08 3 (Fig. 2). troscopyconfinedto wavelengthsbluewardofHβ. Narrow Balmer emission has been found in Indeed, the dwarf nova FO Per (aliases: RL 92 many UX UMa systems, superimposed on the and RWT 92) was classified by Chromey (1979) broaderemissionfromthedisk. Thisnarrowcom- asansdBstar,when,unbeknownsttohim,itmust ponent is attributed to irradiation of the atmo- have been near maximum light in its outburst cy- sphere of the secondary star, with reprocessing of cle. TwoStewardObservatory2.3mspectraofFO thelighttoBalmeremission. Asdiscussedin§2.2, Per, also near maximum light, obtained in 2004 the asymmetryandvariationsin the emissionline Dec and 2006 Dec with the identical setup as de- profile shape (Figs. 4 & 5) suggest the presence scribed in §2.1.2, show a spectrum that is almost of such a component, moving in anti-phase to the indistinguishable from a somewhat reddened sdB broad component, as required for this interpreta- star, except for emission cores in Hα, Hβ, and, tion. very faintly, in the next few higher order Balmer The orbital period of ∼ 0.195d (∼ 4.7 hours) lines, i.e., atalmostexactlythe same levelasseen ◦ determinedforLSIV−08 3iscomparabletoother ◦ in our 2007 May spectrum of LSIV−08 3. UX UMa systems. ◦ The SED ofLSIV−08 3 is considerablyredder thanthoseofknownsinglesdBstars. Dereddening 3.2. Distance and Reddening bytheequivalentofE(B−V)∼0.6ormorewould ◦ LSIV−08 3 lies at galactic coordinates (l,b)= be required to bring the Steward SED at optical ◦ ◦ (11.0,+20.8). The Schlegel et al. (1998) redden- wavelengths into agreement with sdBs that have ing maps show that the total line-of-sight extinc- T in the range 24,000 — 35,000 K. Likewise, eff tion varies on small angular scales in this direc- theE(B−V)neededtobringthe2MASSinfrared tion. We attempted to infer the E(B−V) color colorsintotherangeofsinglesdBstarsis0.6mag excess directly from the strength of the DIBs, us- or more. ing the linear relations, W = a E(B−V), estab- ◦ 0 On the other hand, LSIV−08 3 shares many lished by Herbig (1975). The λ5780 DIB has an properties with novalike variables ofthe UX UMa 6 EW of 0.32±0.03˚A, suggesting E(B−V) ≈ 0.50 Voges et al. 1999). The count rate in the po- magwithanuncertaintyofatleast10%,if weare sition sensitive proportional counter (PSPC) measuring the narrow component of this feature was 0.11±0.02 s−1, yielding an estimated 0.1– only, or E(B−V)≈0.24 mag if we are measuring 2.4keV flux of 1.5×10−12 erg cm−2 s−1. Taking ◦ the sum of the narrow and broad DIBs (the lat- LSIV−08 3 to be the optical counterpart, the X- ter is more likely). The λ4430 DIB has an EW of ray/optical flux ratio is given by logfX/fopt ≈ 0.26±0.04˚A, corresponding to E(B−V) ≈ 0.12 −2.0. with atleast15%uncertainty. Some ofthe uncer- This source was also observed serendipitously tainty arises, depending on whether “all-sky” or by ROSAT (1WGA J1656.4−0834; White et al. “Sco-Oph” values of a0 are adopted. Both DIBs 2000)duringapointedobservationofanothertar- imply reddening less than the full Schlegel et al. get. Inthis∼8000 sexposure,150netcountswere (1998) column: E(B−V)=+0.47. collected, and the count rate was 0.02±0.002 s−1 The heliocentric RV of the interstellar Na I D (5 times lower than in the RASS), correspond- lines is −14.5 km s−1 (average of D1 and D2). ing to fX ≈ 3.0 × 10−13 erg cm−2 s−1 and Most of this can be explained by the reflex of logfX/fopt ≈ −2.6. A best-fitting blackbody the Sun’s space motion with respect to the Local spectrum has kT ≈ 0.2 keV with an intervening Standardof Rest, so kinematic informationabout hydrogencolumndensityof1.3×1021cm−2,corre- thedistanceoftheabsorbingmaterialbetweenthe spondingtoE(B−V)≈0.2. Theunabsorbedflux Sun and LSIV−08◦3 is unavailable. is F0.5−2 keV ≈2.7×10−13 ergs−1cm−2. Fits us- IfLSIV−08◦3isassumedtobeaUXUMavari- ing a bremsstrahlung or thermal plasma emission ablestar,howfarawayis it,andis this consistent model would result in a similar kT value, given with the measured reddening? We adopt an ab- the highly absorbed spectrum and the PSPC’s solute magnitude MV ≈ +5 and a representative limited energy range and resolution. The inferred apparent magnitude V = 11.50 for LSIV−08◦3. total X-ray flux is quite uncertain. We use the standard relation AV = 3.1E(B−V) X-ray emission is not associated with sdB to compute distance d as a function of assumed stars. On the other hand, CVs of all subclasses E(B−V). Values between 100 and 200 pc re- show X-ray emission. Verbunt et al. (1997) dis- sult for E(B−V) in the range 0.0 to 0.5 mag, cuss the CVs detected in the RASS. A typical smaller distances corresponding to larger redden- luminosity in the 0.5–2.5 keV band for a no- ings. Ak et al. (2007) have put forward a calibra- valike variable is LX ∼ 1031 erg s−1. The re- ◦ tionofMJ forCVs,usingasinputstheorbitalpe- quired distance for LSIV−08 3 to have this LX, riod and the extinction-corrected 2MASS J mag- given its modeled flux, is d ∼ 500 pc. Us- nitude and J −H, H −Ks colors. Again varying ing a 2 keV bremsstrahlung spectrum to model theassumedE(B−V),wefinddintherange140– the ROSAT counts, Verbunt et al. (1997) find 180 pc by this method (minimum error at fixed logFX/FUV+opt < −2.5 to be typical for nova- reddening is 40%). Comparing the 2MASS colors likes. Given the uncertainties in the input spec- ◦ of LSIV−08 3 and IX Vel directly, we find a rea- trum and absorption column, we conclude that ◦ sonable match if E(B−V) ≈ 0.10 mag. We con- LSIV−08 3 has X-ray properties consistent with cludethat,ifE(B−V)intherange0.10–0.20mag membership in the UX UMa subclass of CVs. canbeaccommodatedwithindistancesof100–200 pc, the colors and magnitudes of LSIV−08◦3 are 4. Summary consistentwiththoseofaUXUMavariable. Such Giventhe manysimilaritiesbetweenour obser- reddening values within such distances are plausi- ◦ vations and other UX UMa novalike variables, we ble along the line of sight to LSIV−08 3, accord- ◦ proposethatLSIV−08 3shouldbeclassifiedwith ingtothemapsof,e.g.,Perry & Johnston(1982). theUXUMastars. Evidencesupportingthisclas- 3.3. ROSAT X-ray Observations sification includes the presence of emission rever- sals in the Balmer absorption series; other emis- A weak X-ray source detected in the ROSAT sion lines including He I, He II, and N III/C III; All-Sky Survey (RASS) appears to be coinci- the variability of the emission line strengths; the ◦ dent with LSIV−08 3 (1RXS J165630.2−083442; 7 intrinsically blue continuum; and the mild photo- by, e.g., the SloanDigital Sky Survey,the sky has metric variability. The key evidence in favor of not been thoroughly explored at V ∼11. a CV interpretation lies in the periodically mod- ulated Doppler shifts and profiles of the emission M.A.StarkthanksG.B.Berriman,forhishelp lines,withaperiodof4.7hr,consistentwithalow- interpretingthe2MASScatalog,andK.T.Lewis, mass quasi-mainsequence star that fills its Roche A. Narayanan, and K. Herrmann, for assisting lobe and transfers mass to a luminous accretion with the observations at KPNO. We are grate- disk around a white dwarf. The inner hemisphere ful to K. T. Lewis for carrying out the fit to the of the donor star may be irradiated by the ac- ROSAT data, and to L. Townsley, who provided cretion disk. Reddening and distance estimates helpful interpretation. We also thank the HET are consistent with the CV interpretation, from resident astronomers M. Shetrone, S. C. Ode- the standpoint of the expected luminosity of the wahn, and H. Edelmann and telescope operators system and the distribution of interstellar dust in F.Deglman,M.Soukup,M.Villarreal,andV.Ri- ◦ thedirectionofLSIV−08 3. Finally,weakX-rays ley. TheHobby-EberlyTelescopeisajointproject are apparently associated with the system, and of the University of Texas at Austin, the Penn- the X-ray luminosity and X-ray/optical flux ratio sylvania State University, Stanford University, are consistent with observations of other novalike Ludwig-Maximillians-Universit¨at Mu¨nchen, and CVs. Georg-August-Universit¨at G¨ottingen. The HET We refrain from deriving the dynamical mass is named in honor of its principal benefactors, ratioofthebinaryorindividualmassesofthestars William P. Hobby and RobertE. Eberly. This re- ◦ in LSIV−08 3 from any of the various RV semi- search has been supported by grants from NASA amplitudes that we have presented above, mainly (including NASA GSRP grant NGT5-50399), the because the HRS spectra on whichsuchestimates ZaccheusDanielFoundationforAstronomicalSci- would rely were obtained on widely separated ence, and the Pennsylvania Space Grant Consor- dates, and the EW of the emission lines varied tium. JRT andCSP gratefully acknowledgefund- significantly among the various epochs. It is thus ingfromtheNationalScienceFoundationthrough possiblethatoneorbothsitesoftheemissionvar- grantAST-0307413,andthankthe MDMstafffor ied as well. Work aimed at determining the mass expertandconscientiousobservingsupport. Holly ratio is best undertakenwith intensive time-series SheetstooksomeoftheMDMspectra. Asalways, spectroscopy that can provide full orbital phase we are grateful to the Tohono O’odham for leas- coverage in one or two nights, to minimize such ing us their mountain so that we may study the variations. Likewise, we cannot offer a definitive great universe around us. HAS thanks the Cen- interpretationofthephotometricvariability,given ter for Cosmic Evolution and the National Sci- ◦ our present limited data. LSIV−08 3 thus re- ence Foundation (grant AST-0440061) for their mains as an attractive and interesting target for support. EMG acknowledges support from NSF further study. grant AST-0098699. This publication makes use Despite the apparentline-of-sight extinction to of data products from the Two Micron All Sky the source, LSIV−08◦3 is one of the brighter UX Survey, which is a joint project of the Univer- UMastarssofardiscovered. Downes et al.(2001) sity of Massachusetts and the Infrared Processing list only three “UX Uma” systems (of two dozen and Analysis Center/California Institute of Tech- catalogued) that are brighter: IX Vel, V3885 Sgr, nology, funded by the National Aeronautics and and RW Sex. Only seven CVs of any “nova- Space Administration and the National Science like”sub-type(89systemstotal)arebrighterthan Foundation. LSIV−08◦3 at their maximum light. Bright, rel- Facilities: KPNO:2.1m(GoldCam),Bok(B&C atively nearby systems are not only appropriate spectrograph), Hiltner (modular spectrograph), for follow-up studies, but contribute dispropor- McGraw-Hill (Mark III spectrograph), HET tionately to studies of the space density andkine- (HRS) matics of the various subclasses of CVs. Even though hundreds of thousands of stars that are REFERENCES muchfainterhavebeenobservedspectroscopically Ak, T., Bilir, S., Ak, S., & Retter, A. 2007, New 8 Astronomy, 12, 446 Stark, M. A. & Wade, R. A. 2005, in ASP Conf. Ser. 334: 14th European Workshop on White Bessell, M. S. 1990,PASP, 102, 1181 Dwarfs, ed. D. Koester & S. Moehler, 439 Beuermann,K., Stasiewski,U., & Schwope,A. D. —. 2006, Baltic Astronomy, 15, 175 1992,A&A, 256, 433 Thorstensen, J. R., Patterson, J. O., Shambrook, Beuermann,K.&Thomas,H.-C.1990,A&A,230, A., & Thomas, G. 1996,PASP, 108, 73 326 Tull, R. G. 1998, in Proc. SPIE Vol. 3355: Chromey, F. R. 1979,AJ, 84, 534 Optical Astronomical Instrumentation, ed. S. D’Odorico, 387–398 Downes, R. A., Webbink, R. F., Shara, M. M., Ritter, H., Kolb, U., & Duerbeck, H. W. 2001, Verbunt, F., Bunk, W. H., Ritter, H., & Pfeffer- PASP, 113, 764 mann, E. 1997,A&A, 327, 602 Herbig, G. H. 1975, ApJ, 196, 129 Voges, W., Aschenbach, B., Boller, T., Høg,E.,Fabricius,C.,Makarov,V.V.,Urban,S., Bra¨uninger, H., Briel, U., Burkert, W., Corbin, T., Wycoff, G., Bastian, U., Schwek- Dennerl, K., Englhauser, J., Gruber, R., endiek,P.,&Wicenec,A.2000,A&A,355,L27 Haberl, F., Hartner, G., Hasinger, G., Ku¨rster, M., Pfeffermann, E., Pietsch, W., Predehl, P., Kilkenny, D. & Busse, J. 1992, MNRAS, 258, 57 Rosso, C., Schmitt, J. H. M. M., Tru¨mper, J., & Zimmermann, H. U. 1999, A&A, 349, 389 Nassau, J. J. & Stephenson, C. B. 1963, Ham- burger Sternw. Warner & Swasey Obs. Walker, M. F. & Herbig, G. H. 1954, ApJ, 120, 278 Perry, C. L. & Johnston, L. 1982,ApJS, 50, 451 White, N. E., Giommi, P., & Angelini, L. 2000, Reed, B. C. 1998, ApJS, 115, 271 VizieR Online Data Catalog, 9031 Reed, B. C., Carmody, P. M., & Landy, D. H. Wo´zniak, P. R., Vestrand, W. T., Akerlof, C. W., 1998,PASP, 110, 27 Balsano,R.,Bloch,J.,Casperson,D.,Fletcher, Reed, B. C. & Niemczak, P. J. 2000, PASP, 112, S., Gisler, G., Kehoe, R., Kinemuchi, K., Lee, 69 B. C., Marshall, S., McGowan, K. E., McKay, T. A., Rykoff, E. S., Smith, D. A., Szymanski, Schlegel, D. J., Finkbeiner, D. P., & Davis, M. J., & Wren, J. 2004, AJ, 127, 2436 1998,ApJ, 500, 525 Schneider, D. P. & Young, P. 1980,ApJ, 238,946 Skrutskie,M.F.,Cutri,R.M.,Stiening,R.,Wein- berg, M. D., Schneider, S., Carpenter, J. M., Beichman, C., Capps, R., Chester, T., Elias, J.,Huchra,J.,Liebert,J.,Lonsdale,C.,Monet, D. G., Price, S., Seitzer, P., Jarrett, T., Kirk- patrick, J. D., Gizis, J. E., Howard, E., Evans, T., Fowler, J., Fullmer, L., Hurt, R., Light, R., Kopan, E. L., Marsh, K. A., McCallon, H. L., Tam, R., Van Dyk, S., & Wheelock, S. 2006, AJ, 131, 1163 Stark, M. A. 2005, PhD thesis, The Pennsylvania State University This2-columnpreprintwaspreparedwiththeAASLATEX Stark, M. A. & Wade, R. A. 2003,AJ, 126, 1455 macrosv5.2. 9 Fig. 2.— Flux-calibrated spectrum of ◦ LSIV−08 3, obtained UT 2002 April 3 at the Steward Observatory 2.3m telescope, in- Fig. 1.— KPNO 2.1m Goldcam spectra of ◦ cluding higher-order Balmer lines. Absorption LSIV−08 3 showing the variable nature of the emission components at Hβ (top), He I 5875˚A bandheads at λ6277 and λ6867 are telluric; ab- partlyblendedwithinterstellarNaIDlines(mid- sorptionfeatures at λ4430(near the strongerHeI dle), and Hα/He I 6678˚A (bottom). Dates and λ4471 feature) and λ5780 are diffuse interstellar bands. times are UT 2003; times correspond to mid- exposureandareheliocentric. Phasescorrespond- ing to the ephemeris in §2.1.4 are also indicated. Datesthatareitalicizedinparenthesesinthebot- tom panel do not have corresponding data for Hβ andHeI5875˚A.Dottedlinesshowthelaboratory wavelengthsofthelabeledtransitions;thespectra themselves are plotted using observed (topocen- tric) wavelengths. Several gaps in the spectra re- sult from cosmic ray removal. 10