Discovery of a bright eclipsing cataclysmic variable
D. K. Sing, E. M. Green, S. B. Howell, J. B. Holberg, M. Lopez-Morales, J. S. Shaw, G. D. Schmidt
aa r X i v : . [ a s t r o - ph ] A ug Astronomy&Astrophysicsmanuscript no. Sing.J0644.arxiv c (cid:13)
ESO 2018November 19, 2018
Discovery of a bright eclipsing cataclysmic variable
D. K. Sing , , E. M. Green , S. B. Howell , J. B. Holberg , M. Lopez-Morales ⋆ , J. S. Shaw , G. D. Schmidt Institut d’Astrophysique de Paris, CNRS / UPMC, 98bis boulevard Arago, 75014 Paris, Francee-mail: [email protected] Lunar and Planetary Laboratory, Sonett Bld., University of Arizona, Tucson, AZ 85721, USAe-mail: [email protected] Steward Observatory, University of Arizona, 933 North Cherry Avenue, Tucson, AZ 85721, USAe-mail: [email protected],[email protected] WIYN Observatory & NOAO, P.O. Box 26732, 950 N. Cherry Ave., Tucson, AZ 85726-6732, USAe-mail: [email protected] Carnegie Institution of Washington, Dept. of Terrestrial Magnetism, 5241 Broad Branch Road NW, Washington, DC 20015, USAe-mail: [email protected] Department of Physics and Astronomy, University of Georgia, Athens GA, 30602, USAe-mail: [email protected]
Received; accepted
ABSTRACT
Aims.
We report on the discovery of J0644 + Methods.
Optical photometric and spectroscopic observations were obtained to determine the nature and characteristics of this CV.
Results.
Spectral signatures of both binary components and an accretion disk can be seen at optical wavelengths. The optical spectrumshows broad H I, He I, and He II accretion disk emission lines with deep narrow absorption components from H I, He I, Mg II and CaII. The absorption lines are seen throughout the orbital period, disappearing only during primary eclipse. These absorption lines areeither the the result of an optically-thick inner accretion disk or from the photosphere of the primary star. Radial velocity measurementsshow that the H I, He I, and Mg II absorption lines phase with the the primary star, while weak absorption features in the continuum,between H α and H β , phase with the secondary star. Radial velocity solutions give a 150 ± − semi-amplitude for the primarystar and 192.8 ± − for the secondary, resulting in a primary to secondary mass ratio of q = − ⊙ for the primary and 0.49 − ⊙ for the secondary, with the uncertainty largely due to the inclination. Conclusions.
The bright eclipsing nature of this binary has helped provide masses for both components with an accuracy rarelyachieved for cataclysmic variables. This binary most closely resembles a nova-like UX UMa or SW Sex type of cataclysmic variable.J0644 + + Key words. accretion, accretion disks – binaries: spectroscopic: eclipsing – novae, cataclysmic variables
1. Introduction
Cataclysmic Variables (CVs) are close binary systems consistingof a white dwarf (WD) primary and late type secondary whichoverflows its Roche lobe. In these systems, mass is transferedfrom the secondary onto the primary star, often forming an ac-cretion disk. CVs are thought to have been produced through thecommon-envelope (CE) process. In the CE process, one compo-nent of a binary star system has a rapidly expanding envelope,evolving on either the red giant branch (RGB) or the asymp-totic giant branch (AGB). The giant fills its Roche lobe startingrapid mass transfer, quickly filling its companion’s Roche lobeas well. The companion star spirals in helping to eject the enve-lope which has formed around both stars. The initial conditionsof the binary and the e ffi ciency of ejecting the envelope deter-mine the final separation of the binary. After the ejection of theenvelope, the core of the giant is left behind which has formedinto either a subdwarf or white dwarf, while the secondary’smass remains nearly unchanged. Those post-CE systems which Send o ff print requests to : D. K. Sing ⋆ Carnegie Fellow have not coalesced are thought to then go through a detachedphase, where further orbital angular momentum is lost throughgravitational radiation and magnetic braking. Post-common en-velope systems that have evolved to the point where the sec-ondary’s Roche-lobe is filled, either by the secondary evolvingo ff of the main sequence, or by loss of orbital angular momentumand favorable mass ratios, then become CVs. There are approxi-mately 28 known post-common envelope systems which are alsobelieved to be pre-CVs (Sing 2005).Although the accretion disk usually dominates the opticalcontinuum out-shining the primary, the white dwarf primary canbe observed if it is su ffi ciently hot or in cases of low mass trans-fer. The masses, radii, and spectral line profiles measured forthe primary star in CVs have thus far all been consistent withwhite dwarf characteristics. Certain nova-like (NL) CVs werepreviously thought to contain a hot subdwarf, such as UX UMa(Walker & Herbig 1954), because of the appearance of broad ab-sorption lines. These CVs, however, were shown to contain op-tically thick inner accretion disks which gives rise to the broadabsorption features. Sing et al.: Discovery of a bright eclipsing cataclysmic variable
Fig. 1.
Digital Sky Survey (Poss II, blue) finding chart forJ0644 + I NS VS ∼ . + + = ± = ± = ± = ± = ± = ± µ a = -5 ± µ d = -3.3 ± > Table 1.
J0644 + Parameter Value ValueRA(2000) 06:44:34.637DEC(2000) + ± ± ± ± ± ± ± ± K P Velocity 150 ± − He II solution K Velocity 192.8 ± − γ system -7.1 ± − M P ( i ) 0.63 − ⊙ Mass range M ( i ) 0.49 − ⊙ Mass range a ( i ) 1.82-1.88 R ⊙ i > ◦ Inclination surements to allow estimation of the component masses. We de-scribe these observations in §
2, analyze the data in §
3, discussthe results in §
4, and present our conclusions in §
2. Observations
Time resolved di ff erential photometry of J0644 + × ′′ pixel − ′ × ′ field of view. A new facilityimager with a blue-sensitive dual-amplifier Fairchild 4096 × ′ × ′ . With both instruments, 3 × routines forbias subtraction, flat-fielding, and cosmic ray cleaning. Precisedi ff erential photometry was obtained using IRAF’s APPHOTpackage to measure the magnitude of J0644 + The Image Reduction and Analysis Facility, a general purposesoftware package for astronomical data, is written and supported bythe IRAF programming group of the National Optical AstronomyObservatories (NOAO) in Tucson, AZ.ing et al.: Discovery of a bright eclipsing cataclysmic variable 3
Table 2.
Log of Kuiper 1.55m Photometric observations.
UT date HJD Start HJD End Filters N Inst.02 Feb 05 3403.63100 3403.86903 R 300 4K04 Feb 05 3405.61762 3405.87962 R 319 2K01 Mar 05 3430.60707 3430.80255 R + B 203 2K02 Mar 05 3431.59528 3431.83146 R + B 256 2K03 Mar 05 3432.60205 3432.79780 R + B 209 2K04 Nov 05 3678.81874 3679.03925 R 338 4K05 Nov 05 3679.78777 3680.03462 B 343 4K06 Nov 05 3680.85072 3681.03434 B 255 4K05 Dec 05 3709.71712 3710.04482 R + B 429 4K06 Dec 05 3710.71717 3711.01314 R + B 305 4K07 Dec 05 3711.70199 3712.01251 R + B 406 4K08 Dec 05 3712.69835 3713.03623 R + B 445 4K06 Jan 06 3741.63524 3741.29503 R + B 412 4K07 Jan 06 3742.62904 3742.95755 U + I 410 4K10 Jan 06 3745.60197 3745.95607 V + R 452 4K13 Oct 06 4021.91698 4022.02386 B 109 4K ary of 5.0 times the stellar FWHM) were held constant for allstars in a given image, although the values varied from image toimage, according to the seeing. An upper limit to the apertureradius of 4.5 ′′ was imposed for the few occasions when the im-age FWHM’s were greater than 2.0 ′′ , to avoid any contaminationfrom a faint visual companion 6.5 ′′ SSE of J0644 + + ff ect on the relative pho-tometry was minimal, however, due to our rather long exposuresand the selection of fairly bright reference stars. Di ff erential lightcurves for one of the comparison stars relative to the mean of allthe others indicates that the typical 1 σ noise in the J0644 + + ffi cients for Mt. Bigelow were used to correct all of theJ0644 + Spectroscopic observations of J0644 covering multiple orbitalcycles were obtained on the Steward Observatory 2.3m Boktelescope, located on Kitt Peak, during 2005 January 16-17,2005 December 04-07, and 2006 January 09 (see Table 3).The optical spectra were obtained using the Boller and ChivensSpectrograph at the Ritchey-Chretien f / ×
800 15 µ m pixel CCD. For2005 January observations, we used a 1st order 600 line / mmgrating blazed at 3568 Å. With a 2.5 arcsec slit width, a typ-ical spectral resolution of 5.5 Å was achieved at a reciprocaldispersion of 1.87 Å / pixel on the CCD. Typical exposure timesof 300 seconds yielded characteristic S / N ratios of 60 −
70. The2005 December and 2006 January observations were taken withthe 832 line / mm grating and used in 2nd order at two di ff erent Table 4.
Log of Spectroscopic and SpectropolarimetricObservations with the 2.3m Bok Telescope
Date MHJD Start MHJD End Spec. Range N2005 Jan. 16 3387.2321 Blue 12005 Jan. 16 3388.2599 3388.3950 Blue 42005 Dec. 04 3709.2061 3709.5283 Blue 382005 Dec. 05 3710.2056 3710.5262 Red 372005 Dec. 06 3711.1840 3011.5262 Blue 462005 Dec. 07 3712.1995 3012.5306 Red 402005 Dec. 31 3735.2401 3735.3214 Spol 92006 Jan. 09 3745.1532 3745.4544 Red 28Blue: 3880 - 5040 Å; Red: 5500 - 6200 Å; Spol: spectropolametric
Fig. 2.
R and B photometry showing the eclipse of the primarystar. The points at the bottom of the plot are the R residuals forthe 2005 March 02 data which characterize the standard di ff er-ential photometry errors.grating tilts to cover the blue region ( ∼ − ∼ − ff ects of atmosphericdispersion.Standard IRAF routines were used to reduce the data. Thewavelength calibration was established with He-Ar arc-lampspectra, interpolated between exposures taken before and aftereach observation, to account for any small wavelength shifts thatmay occur while the telescope tracks an object. The spectra wereflux calibrated using the Massey et al. (1988) spectrophotometricstandards G191-B2B and BD + ◦ − . − .
75 (as determinedbelow) were observed. These observations, however, showed nosignificant signs of circular polarization with a polarimetric frac-tion of + ± Sing et al.: Discovery of a bright eclipsing cataclysmic variable
Table 3.
J0644 + Object RA DEC V U-B B-V V-R V-I n m(J2000) (J2000) obs. nightsJ0644 06:44:34.4 + ± ± ± ± ± + ± ± ± ± + ± ± ± ± + ± ± ± ± + ± ± ± ± ± + ± ± ± ± ± + ± ± ± ± ± + ± ± ± ± ± + ± ± ± ± ± + ± ± ± ± ± + ± ± ± ± ± + ± ± ± ± ± + ± ± ± ± + ± ± ± ± ± + ± ± ± ± ± + ± ± ± ± + ± ± ± ± ± + ± ± ± ± ± + ± ± ± ± ±
3. Analysis
Photometric coverage in the B and R bands show deep 1 to 1.2magnitude primary eclipses of the hot primary star (see Fig.2). The eclipse lasts around 1.3 hours and has a characteristic“V” shape. There also is considerable ’flickering’ in the lightcurve, indicative of mass transfer, but no evidence of a secondaryeclipse. With multiple night coverage of J0644, the fundamen-tal orbital period of the system was subsequently determined tobe P = ± T mid − eclipse = HJD (2453403 . + (0 . × E , (1)with associated errors of ± and ± × − daysin P .Although there is a strong He II λ Spectroscopically, the out-of-eclipse observations show a strongblue continuum, on which are superimposed broad emissionlines due to H I and He I which have a deep, narrow absorptioncomponent not easily seen at lower resolution, Fig. 3. At higher
Fig. 3.
Optical spectra of J0644 + λ λ M ,longer period CVs. About 1 / λ The He II λ ing et al.: Discovery of a bright eclipsing cataclysmic variable 5 Fig. 4.
Optical spectra of J0644 + Fig. 5. λ / or in the hot-spot.Furthermore, the FWHM of the He II line is seen to decreaseby ∼
70% during eclipse as would be expected when emissionfrom an inner accretion disk with higher Keplerian velocity iseclipsed, decreasing the emission line width.A flux curve of the H I emission line flux for H γ is seenin Fig. 6 produced by averaging the Dec 04 and 06 2005 data.This line, while out of eclipse, is comprised of a broad emissionline component associated with the accretion disk and a narrowabsorption component which can reach well below the contin-uum (see Fig. 4 & Fig. 7). The line flux is summed over boththe absorption and emission line components, making it possi-ble to track the full emission component’s relative contributionto the flux during primary eclipse. The flux also provides infor-mation on the component’s line formation location. When theeclipse begins at phase − γ fluxis seen to abruptly decrease, implying the emission line forms Fig. 6.
Average line flux from the H γ line using the 2005 Dec 04& 06 data.in the outer accretion disk. The flux continues to decrease upuntil phase − − γ flux. Thespectra change from a blue continuum dominated by deep ab-sorption cores, to a fainter flat continuum containing only emis-sion lines at phase − § ν ( H α/ H β ) = I ν ( H α ) / I ν ( H β ). When determining theBalmer decrement from spectra, where fluxes are measured, thedirect comparison of flux-ratios to intensity-ratios assumes thatthe measured emission lines have the same profiles (see Masonet al. 2000 and references therein). The Balmer emission linesof J0644 are complex due to the absorption line components,changing the line profile in a non-uniform manner. The emissionlines in eclipse, however, have no absorption line cores and havesimilarly broadened profiles, making a measurement of D ν morereliable. Measuring the flux-ratio centered on primary eclipsegives a value of D ν ( H γ/ H β ) = .
91. Williams (1991) modeledthe Balmer decrement for H I emission lines from accretion disksusing grids of temperature, density and inclination. AlthoughD ν alone does not provide a unique solution for the accretiondisk temperature or density, it can provide valuable ranges forthose parameters. Seen in Fig. 8, our eclipse measured Balmerdecrement corresponds to a number density range between logN o = . − . o = . − . Sing et al.: Discovery of a bright eclipsing cataclysmic variable
Fig. 7. + Fig. 8.
Balmer decrement (H γ / H β ) vs. log density (g cm − )for four di ff erent temperatures calculated by Williams (1991).The values are for an inclination of 80 ◦ , except for 6,000 Kwhich is calculated at 52 ◦ . The value measured for J0644 + γ / H β = Radial velocity measurements were performed on the H I andHe I absorption lines as well as the H I, He I, and He II emis-sion lines using the 2005 December and 2006 January 2.3m Boktelescope observations. The spectra were velocity shifted to a he-liocentric rest frame where the HJD of the midpoint of each ex-posure was calculated. The radial velocities were measured us-ing a center-of-wavelength, λ ′ , for each line profile to calculatea Doppler velocity. For each line profile of interest, λ ′ was esti- mated following the formula, λ ′ = P λ f λ P f λ . We assumed a circularorbit and then fit the radial velocity curves with a sine functionof the form, V ( t ) = γ + K sin " π ( t − T ) P − πφ , (2)using a gradient-expansion algorithm to compute a non-linearleast squares fit. The fitted parameters were; the system velocity γ , the stellar Keplerian velocity K , and the phase o ff set φ . Thephotometric ephemeris was used, where φ = . K velocities between 150 −
180 km s − for the absorption lines. Between phases − § ff ects,non-uniform disk emission, or e ff ects from an accretion stream.The emission lines of H I, He I, and He II are also seen tophase with the primary star. The radial velocity curve derivedfrom the He II emission line is seen in Fig 11. This line is single-peaked and is sinusoidal throughout the orbital period, possiblyproviding a much more precise measure of the K velocity thanthe H I and He I emission lines. The He II velocity curve givesa K HeII = ± − and γ = -96 ± − . The He II emis-sion line appears, therefore, to be the only velocity feature thatis observed to behave sinusoidally throughout the entire orbitalphase. As we have seen, the bulk of the He II emission arisesin the inner-accretion disk which should accurately follow the ing et al.: Discovery of a bright eclipsing cataclysmic variable 7 Table 5.
J0644 + Feature K Phase O ff set γ χ ν km s − φ − φ phot km s − He I λ ± ± ± λ ± ± ± λ ± ± ± λ ± ± ± λ ±
20 0.033 ± ±
16 2.27He I λ ± ± ± λ ±
17 0.015 ± ±
13 4.18He II λ ±
14 0.039 ± ±
10 1.06
Fig. 9.
Radial velocity curve of the He I λ K HeII , asthe best fit for the K velocity of the primary star, K P .The best fit γ velocities, measured from both the emissionand absorption lines, give widely di ff erent values ranging from ∼ -100 to 25 km s − . These di ff erences are similar to those typ-ically seen in CVs such as the dwarf novae WZ Sge (Masonet al. 2000) and VY Aqr (Augusteijn 1994). The velocity dif-ferences would suggest that the accretion disk emission is non-symmetric, which leads to the inconsistencies observed. The 2005 December 07 and 2006 January 09 observations, takenwith a first order 1200 / mm grating blazed at 5346Å, were opti-mized to detect the weak spectral signature of the late type sec-ondary and measure its radial velocities using a cross-correlationmethod. J0644’s 2MASS H-K color suggested that the sec-ondary might be a K star. K stars have numerous strong absorp-tion lines in our chosen wavelength range, 5035–6195Å, in par-ticular the Mg triplet near 5100Å, whereas J0644 exhibited analmost pure continuum spectrum, except for one He emissionline at 5876Å. Given that the secondary star features are week,we note that the observed spectral region used here is well to theblue of where the CCD starts to show fringing near 6800Å, so Fig. 10.
Radial velocity measurements of J0644 + Fig. 11.
Radial velocity measurements of J0644 + ff ected by incompletely removedfringes, as might occur in the red.Along with the J0644 + / N, in-cluding several radial velocity standards and MK spectrum stan-dards ranging from F7V to M0V. For the January run, the grat-ing was slightly tilted in order to shift the spectra by 5Å ( ∼ Sing et al.: Discovery of a bright eclipsing cataclysmic variable onto the same logarithmic wavelength scale. We fit the contin-uum for each spectrum, divided by the fit, and then subtracted1.0 from the result, in order to get a continuum value of zero. Allof the cross-correlations used the double-precision version of theIRAF FXCOR task, with a ramp filter function, and a gaussian fitto the cross-correlation peak. The filter cuton and cuto ff parame-ters were optimized for highest sensitivity to narrow lines at theobserved 2.3Å resolution. The high S / N main sequence spectrawere cross-correlated against the radial velocity standards, thenall were shifted to the rest velocity. The error of the mean forthe velocity zero point is 4 km s − . Next, the J0644 spectra werecross-correlated against each of the available template spectra inthe wavelength range 5040-5800Å + σ error bars can be seenin Fig. 12. The solid curve is a sinusoidal fit to the datapointsbetween phases − + − − K velocity of K = ± − and a system gamma ve-locity of γ = -7.1 ± − . As stated above, no obvious trendswith changes in the secondary’s spectral type with orbital phasewere observed. The binary star parameters can be estimated using the eclipse asa constraint on the orbital inclination. With the orbital period and K velocities determined in § M P , becomes a function of the inclination, i , given by, M P ( i ) = PK ( K P + K ) π G sin i , (3)where G is the gravitational constant and the other parametersare listed in Table 1. The minimum mass of the primary corre- FXCOR calculates velocity errors that are relatively correct, butinclude an unknown multiplicative factor. We tried several Monte Carlosimulations to estimate the size of the true internal errors, and found thatthey were always small compared to the scatter of the points around thefitted curve. Slit-centering errors are most likely responsible for a largepart of the external error. When compared to the final sinusoidal fit,however, the fxcor errors were observed to closely match an expected1 σ error distribution and were thus adopted as our velocity errors. Fig. 12.
Radial velocity of the secondary star derived from cross-correlation of the continuum between H α and H β . A sine-curvefitted between the phases of -0.2 and 0.4, where the velocity vari-ation appears sinusoidal, is overplotted.sponds to i = ◦ giving M P (90 ◦ ) = . M ⊙ . The mass ra-tio, measured from the radial velocity curves and independent ofinclination assuming circular orbits, is found to be q = M P along with q , give a minimum secondary mass of M (90 ◦ ) = . M ⊙ . The orbital separation, a , then comes fromKepler’s third law given by, a ( i ) = G [ M P + M ] (cid:20) P π (cid:21) , (4)resulting in a minimum separation of a min = R ⊙ .The inclination can be constrained by the eclipse geometrywith estimates for the e ff ective size of the primary (star + disk),the cooler secondary star radius, and the semi-major axis of thebinary. A minimum inclination angle corresponds to one whichstill produces primary eclipses. By e ff ective size of the primary,we mean the aggregate size of the primary star and the inneraccretion disk which dominates the blue optical continuum andis observed to eclipse. The primary’s e ff ective size is determinedin § ∼ R ⊙ while the secondary radius can be estimatedby the volume radius of its Roche lobe to be 0.78 R ⊙ , using q and a min . These radii estimates constrain the inclination to i > ◦ .The resulting masses are therefore determined to fall in the rangeM P = . − .
692 M ⊙ for the primary and M = . − . ⊙ for the secondary. The uncertainty in the absolute massesis dominated by the inclination uncertainty. Although detailedeclipse modeling may ultimately provide a better estimation ofthe inclination, the masses of each component are determined inthis study to ∼
4. Discussion
In cataclysmic variables, the primary star spectrum can some-times be obtained by subtracting a spectrum taken just beforeprimary eclipse with one taken at mid-eclipse when the primaryis fully eclipsed. This technique has previously been used to ex-tract a white dwarf spectrum for eclipsing CVs such as Z Cha(Marsh et al. 1987). On our 2005 December 04 run, two spectrawere obtained such that the exposures corresponded to phasesjust before and after the completion of primary eclipse (see Fig. ing et al.: Discovery of a bright eclipsing cataclysmic variable 9
Fig. 13.
Extracted spectrum of the region containing the inneraccretion disk and primary star. The spectrum is consistent witha ∼ − − ∼ λ § ∼ ⊙ and the orbital period is 0.26937420 days. With theseparameters, the secondary star revolves around the primary, it itsrest frame, with a velocity of 348 km s − . Measured from theemission line profiles in § ∼ ⊙ .The narrow H I and He I absorption features in Fig 13 have acharacteristic Gaussian FWHM broadening of around 560 − − . These absorption line profiles clearly do not resemble aWD, being much to narrow. They are also narrower than the ex-pected widths of subdwarf or main sequence B stars. Given thestrong He II λ β , it seems likely that the Balmer line wings are partiallyfilled in with He II emission at λ R ⊙ , it is unclear as to the origin of the high veloc-ity components of the emission lines. These components requireKeplerian velocities of ∼ − and should originate closerto the primary star at distances of around 0.1 R ⊙ . The dilemmais highlighted by narrow absorption cores within wide emissionlines. In other optically thick disks, wide absorption lines areseen with narrow emission line cores. An accretion disk withboth an optically thick and thin region, however, could poten-tially explain the observed continuum eclipse light curve. Thelight curve shows a broad shallow “U”-shaped eclipse profile to-ward the beginning and end of the eclipse, and a much deeper“V”-shaped curve at the center of the eclipse. An outer opti-cally thin region could produce the “U”-shaped eclipses as seenin normal NL CVs, while an optically thick flared inner regioncould produce the “V”-shaped curve as seen during mid-eclipse(Knigge et al. 2000).An alternative explanation is to assume that the primary staris not hidden and the star’s photosphere can be seen with deep HI and He I absorption lines and a 25,000 K black body temper-ature. These characteristics (along with the eclipse light curve)suggests the star is not a WD, as the absorption lines are much tonarrow. If not a WD, the primary would most likely be a pre-WDor subdwarf (sdO or sdB), given its characteristic size of 0.31R ⊙ and mass 0.65 M ⊙ . These scenarios are complicated with thelack of any secondary eclipses thus far observed, the extremenarrowness of the absorption lines, and the high luminosity ofsubdwarf stars. With the stated characteristics mass and size, theprimary star would have a surface gravity around log g = thick-disk or UX UMa NL CV. AlthoughUX UMa type CVs exhibit a wide range of spectral features (seeWarner 1995), a main feature of these binaries are wide absorp-tion lines (typically in H I, He I, and Ca II) with narrow emissionline cores. These CVs have been shown to contain optically thickinner accretion disks, giving rise to the wide Doppler broadenedabsorption lines. An optically thin outer disk then provides thenarrow emission component. This spectral characteristic, how-ever, is the opposite of what is observed in J0644 which has anarrow absorption component in the center of wide emissionlines. Furthermore, for J0644 the inner accretion disk temper-atures would have to be ∼ these stars are narrow central absorption components in H I andHe I, which look a lot like the features seen in J0644. However,this central absorption component in SW Sex stars only appearsaround phase 0.5 and can appear and disappear quickly. Thesecharacteristics have been explained by an overflowing stream re-sulting in absorption which can only be viewed through a flareddisk during phases 0 . − .
6. In J0644, however, the absorp-tion components persist throughout the orbital cycle, only dis-appearing during primary eclipse. This would seem to rule outany overflows in J0644 as the absorption is constrained to theinner disk / primary star region and is nearly always visible. SWSex stars also show single peaked emission lines whereas othereclipsing CV binaries show double peaked lines expected froma Keplerian disk. This single peaked nature has been suggestedas at least partly due to a flared disk which produces a flat-topped emission line profile. Although J0644 does not show sin-gle peaked emission from H I and He I (except during eclipse),the He II emission line is observed to be single peaked, possi-bly indicating that it too has a flared disk. SW Sex stars alsoshow large radial velocity phase lags relative to the photometricephemeris and wind-formed P Cygni profiles, traits not observedin J0644. The emission lines of J0644 are significantly narrowerthan SW Sex. Dhillon et al. 1997 measured a FWZI (full widthat zero intensity) of 5700 km s − for He II in SW Sex whichcompares to ∼ − for J0644.
5. Conclusions
J0644 is a bright new example of an eclipsing binary with ongo-ing mass transfer. Spectroscopic and photometric observationshave revealed J0644 to be a young cataclysmic variable binarydominated by either an optically thick inner accretion disk ora white dwarf progenitor. J0644 is a unique binary that doesnot easily fit into any known cataclysmic variable subtypes, al-though it most closely resembles a UX Uma or SW Sex type NLCV. With no observed magnetic fields, P Cygni profiles, phaselags, or superhumps there is no evidence with which to evoke theusual physical interpretations. These include truncated accretiondisks, winds, and stream-disk hot spots. Numerical models andfurther observations in di ff erent wavelengths regimes will ulti-mately be needed to fully understand this unique system and itsproperties. Given its bright nature and multitude of features ob-served in this binary, J0644 will become an important binary formany future studies. Acknowledgements.
We wish to thank For, BiQing (University of Texas) inher observing help during four 61” nights and the referee Ed Guinan for hishelpfull comments. D.K.S. is supported by CNES. This publication makes useof the data from the Northern Sky Variability Survey created jointly by theLos Alamos National Laboratory and University of Michigan. The NSVS wasfunded by the US Department of Energy, the National Aeronautics and SpaceAdministration and the National Science Foundation. This publication makes useof data products from the Two Micron All Sky Survey, which is a joint projectof the University of Massachusetts and the Infrared Processing and AnalysisCenter / California Institute of Technology, funded by the National Aeronauticsand Space Administration and the National Science Foundation.
References
Augusteijn T. 1994, A&A, 292, 481Clarke, D. 2002, A&A, 386, 763Dhillon, V. S., Marsh, T. R.,& Jones, D. H. P. 1997, MNRAS, 291, 694Groot, P. J., Rutten, R. G. M., & Paradijs J. van 2004, A&A, 417, 283Hoard, D. W., Wachter, S., Clark, L. L., Bowers, T., P. 2002, ApJ, 565, 511Knigge, C., Long, K. S., Hoard, D. W., Szkody, P. & Dhillon V. S. 2000, ApJ,539, L49Landolt, A. U. 1992, AJ, 104,340 Marsh, T. R., Horne, K. & Shipman, H. L. 1987, MNRAS, 225, 551Mason, E., Skidmore, W., Howell, S. B., Ciardi, D. R., Littlefair, S. & Dhillon,V. S. 2000, MNRAS, 318, 440Mason, E. & Howell, S. B. 2005, A&A, 439,301Massey, P., Strobel, K., Barnes, J. V., Anderson, E. 1988, ApJ, 328, 315Schwarzenberg-Czerny, A. 1989, MNRAS, 241, 153Sing, D. K. 2005, PhD ThesisSzkody, P., Garnavich, P., Howell, S., & Kii, T. 1990, in Proceedings of the 11thNorth American Workshop on Cataclysmic Variables and Low Mass X-RayBinaries, Edited by Christopher W. Mauche, Cambridge University Press,p.251Schmidt, G. D., Stockman, H. S., & Smith P. S. 1992, ApJ, 398, L57Walker, M. F. & Herbig, G. H. 1954, ApJ, 120, 278Warner, B. 1995
Cataclysmic variable stars , Cambridge Astrophysics Series,Cambridge, New York: Cambridge University PressWilliams, G. A. 1991, AJ, 101, 1929Wozniak, P. R., Vestrand, W. T., Akerlof, et. al. 2004, AJ, 127, 3043Zacharias, N., Urban, S. E., Zacharias M. I., et al. 2004, AJ, 127, 3043
Appendix A: Photometric Solutions
We observed the field around J0644 on two clear nights, UT 06Dec 2005 and 05 Mar2006. UBVR observations were obtainedin December for 18 Landolt standard stars (when the I filter wasnot available), and UBVRI observations of 14 Landolt standardswere obtained in March. The photometric solutions for the twonights were:U = u − . + . ∗ ( u − b ) + . ∗ Xb − . ∗ Xu (0 . = b + . + . ∗ ( b − v ) + . ∗ Xv − . ∗ Xb (0 . = v + . − . ∗ ( b − v ) + . ∗ Xb − . ∗ Xv (0 . = r + . − . ∗ ( v − r ) + . ∗ Xv − . ∗ Xr (0 . = u − . + . ∗ ( u − b ) + . ∗ Xb − . ∗ Xu (0 . = b + . − . ∗ ( b − v ) − . ∗ Xv − . ∗ Xb (0 . = v + . − . ∗ ( b − v ) + . ∗ Xb − . ∗ Xv (0 . = r − . − . ∗ ( v − r ) + . ∗ Xv − . ∗ Xr (0 . = i − . + . ∗ ( v − i ) − . ∗ Xv − . ∗ Xi (0 ..