New Halo White Dwarf Candidates in the Sloan Digital Sky Survey
Kyra Dame, A. Gianninas, Mukremin Kilic, Jeffrey A. Munn, Warren R. Brown, Kurtis A. Williams, Ted von Hippel, Hugh C. Harris
aa r X i v : . [ a s t r o - ph . S R ] A ug Mon. Not. R. Astron. Soc. , 000–000 (0000) Printed 31 August 2018 (MN L A TEX style file v2.2)
New Halo White Dwarf Candidates in the Sloan DigitalSky Survey ⋆ Kyra Dame , A. Gianninas , Mukremin Kilic , Jeffrey A. Munn , Warren R. Brown ,Kurtis A. Williams , Ted von Hippel , Hugh C. Harris Homer L. Dodge Department of Physics & Astronomy, University of Oklahoma, 440 W. Brooks St, Norman, OK 73019, USA US Naval Observatory, Flagstaff Station, 10391 W. Naval Observatory Road, Flagstaff, AZ 86005, USA Smithsonian Astrophysical Observatory, 60 Garden St., Cambridge, MA 02138, USA Department of Physics and Astronomy, Texas A&M University-Commerce, P.O. Box 3011, Commerce, TX 75429, USA Embry-Riddle Aeronautical University, Physical Sciences, 600 South Clyde Morris Boulevard Daytona Beach, FL 32114-3900, USA
31 August 2018
ABSTRACT
We present optical spectroscopy and near-infrared photometry of 57 faint ( g =19 −
22) high proper motion white dwarfs identified through repeat imaging of ≈ ugriz and JH photometry to perform a model atmosphere analysis, and identify tenultracool white dwarfs with T eff < T eff = 3550 ± >
11 Gyr. There are four white dwarfs in oursample with large tangential velocities ( v tan >
120 km s − ) and UVW velocities thatare more consistent with the halo than the Galactic disc. For typical 0 . M ⊙ whitedwarfs, the cooling ages for these halo candidates range from 2.3 to 8.5 Gyr. However,the total main-sequence + white dwarf cooling ages of these stars would be consistentwith the Galactic halo if they are slightly undermassive. Given the magnitude limitsof the current large scale surveys, many of the coolest and oldest white dwarfs remainundiscovered in the solar neighborhood, but upcoming surveys such as GAIA and theLarge Synoptic Survey Telescope (LSST) should find many of these elusive thick discand halo white dwarfs.
Key words: stars: atmospheres, stars: evolution, techniques: photometric, whitedwarfs
As the remnants of some of the oldest stars in the galaxy,cool white dwarfs offer an independent method for datingdifferent Galactic populations and constraining their starformation history (Winget et al. 1987; Liebert et al. 1988).The current best estimates for the ages of the Galacticthin and thick discs are 8 ± . >
10 Gyr (Gianninas et al. 2015), re-spectively. Extended
Hubble Space Telescope observing cam-paigns on 47 Tuc, M4, and NGC 6397 revealed the end ofthe white dwarf cooling sequence in these globular clus- ⋆ Based on observations obtained at the MMT Observatory, ajoint facility of the Smithsonian Institution and the University ofArizona. ters (Hansen et al. 2004, 2007; Kalirai et al. 2012), whichreveal an age spread of 11 to 13 Gyr for the Galactic halo(Campos et al. 2015).Field white dwarfs provide additional and superior in-formation on the age and age spread of the Galactic discand halo. Recent large scale surveys such as the SloanDigital Sky Survey (SDSS) have found many cool fieldwhite dwarfs (Gates et al. 2004; Harris et al. 2006, 2008;Kilic et al. 2006, 2010b; Vidrih et al. 2007; Tremblay et al.2014; Gianninas et al. 2015). These stars are far closer andbrighter than those found in globular clusters, allowing forrelatively easy optical and infrared observations in multiplebands from ground-based telescopes. Modeling the spectralenergy distributions (SEDs) of these white dwarfs providesexcellent constraints on their atmospheric composition andcooling ages and gives us an alternate method for calibrating c (cid:13) K. Dame et al. the white dwarf cooling sequences of globular clusters. How-ever, there are only a handful of nearby halo white dwarfscurrently known.Kalirai (2012) use four field white dwarfs with halokinematics to derive an age of 11 . ± . ≈
11 Gyrin a sample of 634 DA white dwarfs. These have accurateradial velocities determined from Balmer absorption lines,allowing for the determination of accurate 3D space veloci-ties (Pauli et al. 2003, 2006; Richter et al. 2007). Similarly,Kilic et al. (2012) use optical and infrared photometry andparallax observations of two cool white dwarfs with halokinematics, WD 0346+246 and SDSS J110217.48+411315.4to derive an age of 11.0-11.5 Gyr for the local halo.Ongoing and future photometric and astrometric sur-veys like the Panoramic Survey Telescope & Rapid Re-sponse System (Tonry et al. 2012), Palomar Transient Fac-tory (Rau et al. 2009), the Large Synoptic Survey Telescopeand the
GAIA mission will significantly increase the numberof field white dwarfs known. Previously, Liebert et al. (2007)performed a targeted proper motion survey for identifyingthick disc and halo white dwarfs in the solar neighborhood.Munn et al. (2014) present the proper motion catalog fromthis survey, which includes ≈ T eff < Reduced proper motion is defined as H = m + 5 log µ + 5(where µ is the proper motion and m is the apparent magni-tude), which is equivalent to M + 5 log V tan − . g = 8 white dwarfs with V tan = 40 and 150 km s − . Themodel colors become redder until the white dwarfs become Figure 1.
The reduced proper motion diagram for a portion ofthe Munn et al. (2014) proper motion survey. White dwarf evo-lutionary tracks for tangential velocities of 40 and 150 km s − are shown as solid lines. Filled circles mark spectroscopicallyconfirmed white dwarfs with H g >
21 mag and triangles showour targets with SWIRC near-infrared photometry, but with nofollow-up spectroscopy. cool enough to show infrared absorption due to molecularhydrogen (Hansen 1998). We selected targets for follow-upspectroscopy and near-infrared photometry based on theirreduced proper motion and colors. To find the elusive halowhite dwarfs and other white dwarfs with high tangentialvelocities, we targeted objects with H g >
21 mag and belowthe V tan = 40 km s − line. We obtained follow-up optical spectroscopy of 32 whitedwarf candidates at the 6.5 m MMT telescope equippedwith the Blue Channel Spectrograph (Schmidt et al. 1989)on UT 2009 June 18-23 and 2009 November 19-20. We useda 1 . ′′
25 slit and the 500 line mm − grating in first order toobtain spectra over the range 3660-6800 ˚A and with a re-solving power of R = 1200. We obtained all spectra at theparallactic angle and acquired He-Ar-Ne comparison lampexposures for wavelength calibration. We use observationsof the cool white dwarf G24-9 for flux calibration.Out of the 32 candidates with spectra, only two (SDSSJ024416.07-090919.7 and J172431.61+261543.1) are metal-poor halo subdwarfs. The remaining objects are confirmedto be DA, DC, or DZ white dwarfs. This relatively small (2out of 32) contamination rate from subdwarfs demonstratesthat our white dwarf sample is relatively clean.Figure 2 shows the spectra for the five DA WDs in oursample. Two of the DAs, J1513+4743 and J1624+4156, arewarm enough ( T eff ≈ α and a few of thehigher order Balmer lines, while the remaining three DAs c (cid:13) , 000–000 ew Halo WD Candidates J1513+4743J1624+4156J2230+14151604+3923J2230+1415
Figure 2.
Optical spectra for the five DA WDs in our sample.The dotted line marks H α . only show H α , which implies effective temperatures near5000 K.Figure 3 shows the MMT spectra for the 24 DC whitedwarfs in our sample, including the three cool DCs presentedin Kilic et al. (2010a). All of these 24 targets have feature-less spectra that are rising toward the infrared, indicatingtemperatures below 5000 K. We obtained J- and H-band imaging observations of 40 ofour targets using the Smithsonian Widefield Infrared Cam-era (SWIRC, Brown et al. 2008) on the MMT on UT 2011March 23-24. SWIRC has a 5 . × .
12 arcmin field of viewat a resolution of 0 . ′′
15 per pixel. We observed each target ona dozen or more dither positions, and obtained dark framesand sky flats each evening. We used the SWIRC data reduc-tion pipeline to perform dark correction, flat-fielding, andsky subtraction, and to produce a combined image for eachfield in each filter. We use the 2MASS stars in the SWIRCfield of view for photometric and astrometric calibration.In addition, near-infrared photometry for two more targets,J0040+1458 and J1649+2932, are available from the UKIRTInfrared Deep Sky Survey (UKIDSS, Lawrence et al. 2007)Large Area Survey. Table 1 presents the ugriz and JH pho-tometry for our sample of 57 targets with follow-up spec-troscopy and/or near-infrared photometry. J0040+1458J0047-0852J0108-0954J0147-0935J0734+3728J0811+3842J0820+3904J0822+3903J0910+3744J1534+5624J1555+4940J1601+4120J1643+4438J1647+3946J1649+29321657+2638J1711+2940J1715+2600J2127+1036J2137+1050J2145+1106NJ2145+1106SJ2316-1044J2320-0845
Figure 3.
Optical spectra for 24 DC WDs in our sample.c (cid:13) , 000–000
K. Dame et al.
Figure 4.
Color-color diagrams for our sample of 57 white dwarfcandidates. Solid and dashed lines show the predicted colors forpure H ( T eff > T eff > g = 8 (Bergeron et al. 1995), respectively. Figure 4 presents optical and infrared color-color dia-grams for the same stars, along with the predicted colorsof pure H and pure He atmosphere white dwarfs. The dif-ferences between these models are relatively minor in theoptical color-color diagrams, except for the ultracool whitedwarfs with T eff < r − i color, whereas it occurs at 4500 K for the r − H color.The colors for our sample of 57 stars, including the targetswith and without follow-up spectroscopy, are consistent withthe white dwarf model colors within the errors. The major-ity of the targets with g − r > . T eff r − H colors than the pure He model sequence,indicating that the coolest white dwarfs in our sample haveH-rich atmospheres.We use the SWIRC astrometry to verify the proper mo-tion measurements from our optical imaging survey. Giventhe relatively small field of view of the SWIRC camera andthe limited number of 2MASS stars available in each field,the astrometric precision is significantly worse in the SWIRCimages compared to the Bok 90 inch and USNO 1.3m op-tical data. We find that the proper motion measurementsfrom the SWIRC data are on average 54 ±
44 mas yr − higher. Nevertheless, all but one of our targets, J1513+4743,have SWIRC-SDSS proper motion measurements consistentwith the proper motion measurements from our optical datawithin 3 σ . J1513+4743 is spectroscopically confirmed to bea DA white dwarf. Hence, the contamination rate of our sam-ple of 57 stars by objects with incorrectly measured propermotions should be relatively small. Our model atmospheres come from the LTE model atmo-sphere code described in Bergeron et al. (1995) and refer-ences within, along with the recent improvements in thecalculations for the Stark broadening of hydrogen linesdiscussed in Tremblay & Bergeron (2009). We follow themethod of Holberg & Bergeron (2006) and convert the ob-served magnitudes into fluxes, and use a nonlinear least-squares method to fit the resulting SEDs to predictions frommodel atmospheres. Given that all our targets appear to bewithin 150 pc, we do not correct for extinction. We con-sider only the temperature and the solid angle π ( R/D ) ,where R is the radius of the white dwarf and D is its dis-tance from Earth, as free parameters. Convection is modeledby the ML/ α = 0.7 prescription of mixing length theory.For a more detailed discussion of our fitting technique, seeBergeron et al. (2001); for details of our helium-atmospheremodels, see Bergeron et al. (2011). Since we do not haveparallax measurements for our objects, we assume a surfacegravity of log g = 8. This is appropriate, as the white dwarfmass distribution in the Solar neighborhood peaks at about0.6 M ⊙ (Tremblay et al. 2013). We discuss the effects of thischoice in Section 4.Below about 5000 K, H α is not visible. However, thepresence of hydrogen can still be seen in the blue from thered wing of Ly α absorption (Kowalski & Saumon 2006), andin the infrared from CIA due to molecular hydrogen. Coolwhite dwarfs with pure helium atmospheres are not subjectto these opacities, so their SEDs should appear similar toa blackbody. Because of this, atmospheric composition canstill be determined from ultraviolet and near-infrared data.Table 2 presents the best-fit atmospheric compositions, tem-peratures, distances, and cooling ages for our targets, as wellas their proper motions and tangential velocities. Below, wediscuss the pure H, pure He, and mixed H/He atmospheretargets separately, and highlight the most interesting objectsin the sample. Of our 57 targets, only 45 have the near-infrared data thatare needed to observe the CIA that allows us to detect thepresence of hydrogen. Of these 45 objects, twelve have SEDsbest fit by pure hydrogen models. Figure 5 shows the SEDsand our model fits for four of these objects (full sample isavailable online). We show the photometric data as errorbars and the best-fit model fluxes for pure H and pure Hecomposition as filled and open circles, respectively.J1513+4743 is the only DA white dwarf in our sam-ple with near-infrared photometry available (the four otherDAs are discussed in Section 3.5), and the pure H modelis a better fit to the SED than the pure He model. Forthe remaining objects, we chose the composition based onthe solution that best fits the SED. Our sample includesthree previously published H-atmosphere DC white dwarfs:J2137+1050, J2145+1106N, and J2145+1106S (Kilic et al.2010a). Our temperature estimates of 3670 ± ± ±
100 K, respectively agree with the previouslypublished values of 3780, 3730 K, and 4110 K (Kilic et al.2010a) within the errors. c (cid:13) , 000–000 ew Halo WD Candidates Table 1.
Optical and Near-Infrared Photometry of White Dwarf Candidates
SDSS u g r i z J H
J213730.86+105041.5 23.30 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± − ± ± ± ± ± − ± ± ± ± ± − ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± − ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± − ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± − ± ± ± ± ± − ± ± ± ± ± With the exception of the DA WD J1513+4743, allof the remaining 11 objects that are best explained bypure H atmosphere models have T eff H − band than ex-pected from the blackbody-like SEDs of pure He atmospherewhite dwarfs, indicating that they have H-rich atmospheres.In addition to the previously published J2137+1050 andJ2145+1106N (Kilic et al. 2010a), we identify three newwhite dwarfs with T eff T eff = 3550 ±
100 K) with an SED thatis matched relatively well by a pure H atmosphere model.The implied cooling age for such a cool white dwarf is 10.1Gyr assuming an average mass, log g = 8, white dwarf. For our remaining 33 objects with infrared data, 29 showno evidence of CIA and are best fit by pure He atmospheremodels. Figure 6 shows the SEDs for a sample of these tar-gets. All 29 of these objects have T eff in the range 4240 − K − band photometry would be useful to confirm theatmospheric composition for these stars. However, Bergeron(2001) and Kilic et al. (2010b) also find an overabundanceof pure He atmosphere white dwarfs in the temperaturerange 4500-5000 K. Kilic et al. (2010b) discuss a few poten- c (cid:13) , 000–000 K. Dame et al.
Figure 5.
Fits to the SEDs for four of our WDs with pure H atmospheres (full sample available online). Filled circles are pure H models,and open circles are pure He models (included for comparison).
Figure 6.
Fits to the SEDs for four of our WDs with pure He atmospheres (full sample available online). Filled circles are pure H models(included for comparison), and open circles are pure He models. c (cid:13) , 000–000 ew Halo WD Candidates Table 2.
Physical Parameters of our White Dwarf Sample. Source of the optical spectroscopic observations: (1) This paper, (2) Kilic et al.(2010b), and (3) Kilic et al. (2010a).
Object Spectral Source Composition T eff d Cooling Age µ RA µ Dec V tan (SDSS) Type (log He/H) (K) (pc) (Gyr) (mas yr − ) (mas yr − ) (km s − )J2137+1050 DC 2 H 3670 ±
160 75 9.8 − − ±
110 68 9.7 191.9 − ±
100 65 9.1 185.9 − ±
90 94 6.3 128.1 18.5 57.6J0047 − ±
160 68 11.0 211 − ±
120 79 8.5 79.2J0108 − ±
520 77 9.9 − − ±
150 100 7.5 93.6J0147 − ±
320 108 8.2 211.5 − ±
190 126 6.9 127.3J0734+3728 DC 1 H 3700 ±
140 94 9.8 − −
114 50.5J0749+2947 . . . . . . He 4690 ±
60 90 6.7 216.1 − ±
120 94 8.7 68.1 − ±
80 118 6.4 − − ±
110 131 7.0 99.3 − ±
50 67 7.1 231.4 − ±
150 104 8.9 − −
129 106.0J0822+3903 DC 1 H 4190 ±
150 109 8.5 274.2 − ±
60 65 6.4 − −
161 49.9J0848+4204 . . . . . . He 4820 ±
120 146 6.4 − − ±
80 119 6.5 − − − ±
190 63 9.5 − − ±
70 94 6.5 − − ±
90 42 9.9 216.5 − ±
100 104 7.7 − − ±
120 128 6.8 − − ±
70 104 7.0 − − ±
80 97 6.6 73.9 − ±
120 148 6.5 − − ±
70 102 7.0 − − ±
170 116 8.4 − − ±
60 95 7.0 − − − ±
120 92 8.8 − ±
100 111 7.5 − −
15 82.3J1434+5345 . . . . . . He 4600 ±
70 100 7.0 −
143 136.6 93.5J1444+6025 . . . . . . He 4730 ±
100 132 6.6 − − ±
140 104 7.5 − ±
110 114 7.6 13.3 − ±
180 86 8.7 − − ±
120 124 2.3 − − ±
120 87 9.8 − − ±
100 107 7.8 − − ±
120 84 6.2 − +120 −
93 5.7 81.0J1552+4638 DA 1 H 5100 +120 −
88 5.2 − − ±
180 94 8.5 27.1 − ±
120 36 8.8 74.8 − ±
60 44 6.9 50.5J1601+4204 . . . . . . He 4540 ±
110 119 7.1 − ±
140 99 5.6 15.3 − ±
150 133 2.4 27.7 − ±
50 57 6.8 − − +60 −
70 6.2 42.8 − ±
70 91 7.5 − − ±
80 99 6.2 121.3 16.1 57.6J1657+2638 DC 1 H 3550 ±
100 67 10.1 − − ±
70 97 7.1 57.7 − ±
80 88 7.6 − − ±
370 93 9.1 − − ±
160 113 7.2 69.7J2230+1415 DA 1 H 5210 ±
140 107 4.6 − − − ±
790 93 9.8 237.7 −
72 109.7He 4310 ±
200 115 7.6 136.0J2320 − ±
190 67 10.8 166.7 20.3 53.3He 4010 ±
130 80 8.3 63.7 tial problems that could lead to misclassification of spectraltypes for these stars, including problems with the CIA cal-culations, or small shifts in the ugriz or JH photometriccalibration. The last four targets with infrared data (J0910+3744,J0927+4852, J1513+4743, and J1555+4940) have SEDsthat are inconsistent with either a pure H or pure He at-mosphere solution. We fit the SEDs of these stars with amixed H/He atmosphere model. The mixed models allow for c (cid:13) , 000–000 K. Dame et al. significant H -He CIA at higher temperatures than seen for H - H , as CIA becomes an effective opacity source at highertemperatures in cool He-rich white dwarfs due to lower opac-ities and higher atmospheric pressures (Bergeron & Leggett2002).Figure 7 shows our mixed H/He atmosphere modelfits for these four objects. The models yield log ( He/H ) of − . , . , − .
2, and − . µ m thatare never observed in cool white dwarfs. Hence, the tem-perature and composition estimates for such infrared-faintstars is problematic (see the discussion in Kilic et al. 2010b;Gianninas et al. 2015).The coolest object among these four stars, J0927+4852appears to be similar to WD0346+246. Oppenheimer et al.(2001) originally found a T eff = 3750 K and log ( He/H )= 6.4 for WD0346+246, for an assumed surface gravity oflog g = 8. However, Bergeron (2001) showed that such aHe-rich atmosphere would require accretion rates from theinterstellar medium too low to be realistic. With the ad-dition of parallax observations to constrain the distance,they estimated a more realistic solution with T eff = 3780K, log ( He/H ) = 1 .
3, and log g = 8 .
34. A re-analysis byKilic et al. (2012) that include the red wing of the Ly α opacity indicate a similar solution with T eff = 3650 K,log ( He/H ) = − .
4, and log g = 8 .
3. Adopting a similarlog g value for J0927+4852 would yield a T eff of 3730 K andlog ( He/H ) of 0.3.This exercise shows the problems with constraining theatmospheric composition of ultracool white dwarfs, and theneed for parallax observations to derive accurate parame-ters for such white dwarfs. Regardless of these issues, allfour mixed atmosphere white dwarfs appear to be ultracool( T eff < There are twelve spectroscopically confirmed white dwarfsin our sample that lack infrared photometry. Figure 8 dis-plays the SEDs along with the pure H and pure He modelfits for a subsample of these objects. The spectra of four ofthese objects; J1552+4638, J1604+3923, J1624+4156, andJ2230+1415 confirm that they are DA white dwarfs, and thepure H models reproduce the SEDs and spectra reasonablywell. This brings our final number of pure H solutions tosixteen stars.The remaining eight objects without infrared data areconfirmed to be DC white dwarfs. For the most part, thepure H and pure He models are nearly indistinguishable inthe optical for these objects and we cannot determine theircomposition. Table 2 shows the results for both pure H andpure He solutions for these objects.All of these targets have SEDs rising toward 1 µ m, hencethe lack of infrared data limits the precision of these tem-perature measurements. However, given the lack of He at-mosphere white dwarfs below 4240 K, we do not expectJ0047 − − T eff of 3140 ± ±
190 K respectively. J0108 − − − − The estimated temperatures for our targets yield whitedwarf cooling ages between 5 and 10 Gyr, with the only no-table exceptions being J1513+4743 and J1624+4156, whichhave cooling ages of 2.3 and 2.4 Gyr respectively. Eight ob-jects have cooling ages longer than 9 Gyr, with the oldest be-ing J1657+2638 at 10.1 Gyr. However, in order to associatea white dwarf with the thick disc or halo, it is important todetermine the total stellar age (Bergeron et al. 2005). Themain-sequence lifetime of the ≈ M ⊙ progenitor of a 0 . M ⊙ white dwarf is 1.0-1.3 Gyr; therefore, the total ages of ourobjects on average range from 6 to 11 Gyr, with J1513+4743and J1624+4156 having total ages between 3.3 and 3.7 Gyr.Figure 9 shows U versus V (bottom) and W versusV (top) velocities of our objects (assuming a radial ve-locity of 0 km s − and calculated using the prescriptionof Johnson & Soderblom (1987)), as well as the 3 σ ellip-soids of the halo, thick disc, and thin disc populations(Chiba & Beers 2000). The filled, open, and red circles rep-resent the objects best fit by pure H, pure He, and mixedH/He atmosphere models, respectively. For the eight objectswith undetermined compositions, velocities were calculatedassuming the pure H solution for simplicity. The choice ofthe pure H or pure He solution has a negligible effect on thefinal UVW velocities (see Table 2).J2137+1050 shows velocities inconsistent with thickdisc objects in U, consistent with the analysis in Kilic et al.(2010a), while the results for the J2145+1106 common-proper motion binary are consistent to 2 σ , but not 3 σ . In ad-dition, three other targets in our sample show velocities in-consistent with thick disc objects: J0822+3903, J1024+4920,and J1513+4743, with cooling ages of 8.5, 6.8, and 2.3 Gyrrespectively. The Toomre diagram for our targets is shownin Figure 10, with thin disc and thick disc boundaries fromFuhrmann (2004); the differentiation between our halo can-didates and the rest of our sample is clearer here than inFigure 9.The total main-sequence + white dwarf cooling agesof these objects are relatively young for halo objects, butwithout parallax measurements, we cannot constrain theirmasses, velocities, and cooling ages precisely. For example,if these objects have M ≈ . M ⊙ (Bergeron 2001), theprogenitor mass would be closer to 1 M ⊙ and their main-sequence lifetimes would be on the order of 10 Gyr, mak-ing them excellent candidates for membership in the halo.A lower surface gravity would also imply a larger and moredistant white dwarf, and UVW velocities that are even moreinconsistent with the thick disc population. Conversely, forlog g = 8 . c (cid:13) , 000–000 ew Halo WD Candidates Figure 7.
Fits to the SEDs of the four white dwarfs best fit by mixed atmosphere models. published white dwarfs (J2137+1050 and J2145+1106 bi-nary), none of our objects with cooling ages above 9 Gyrhave UVW velocities inconsistent with the thick disc, nordo they show the high tangential velocities expected for haloobjects. In fact, the highest tangential velocity for these ob-jects is 72 km s − . Assuming these objects really do belongto the thick disc gives a thick disc age of ≈
11 Gyr.Our assumption of zero radial velocity has a negligi-ble effect on our results (see the discussion in Kilic et al.2010a). The UVW velocities of our halo white dwarf candi-dates remain inconsistent with the 3 σ distribution for thethick disc for positive and negative radial velocities up to100 km s − (though J0822+3903 only remains inconsistentin both U and W for radial velocities between -90 and 30km s − ). We present follow-up optical spectroscopy and/or near-infrared photometry of 57 cool white dwarf candidates iden-tified from a ≈ ugriz and JH photometry. The best-fit modelshave 29 pure He atmosphere white dwarfs with T eff = 4240 − T eff = 3550 − T eff = 3210 − T eff = 3550 ±
100 K and an SED that isreproduced fairly well by a pure H atmosphere. For an aver-age mass of 0.6 M ⊙ , J1657+2638 would be an ≈
11 Gyr old(main-sequence + cooling age) white dwarf at a distance of67 pc. The implied tangential velocity of 40 km s − demon-strates that J1657+2638 belongs to the Galactic thick disc.Our sample contains three new halo white dwarf can-didates. All three have high tangential velocities and UVWvelocities inconsistent with the Galactic thick disc. The old-est halo white dwarf candidate is J0822+3903 with a cool-ing age of 8.5 Gyr. However, without trigonometric parallaxobservations, we cannot accurately constrain the distances,masses, and ages of our white dwarfs.Our current sample of cool field halo white dwarfs islimited by a lack of deep proper motion surveys. Ongoing c (cid:13) , 000–000 K. Dame et al.
Figure 8.
Fits to the SEDs for a sample of our WD lacking IR data (full sample available online). Filled circles are pure H models, andopen circles are pure He models. As can be seen, no model is clearly better.
Figure 10.
Toomre diagram for our 57 targets. Symbols have the same meaning as Figure 9. Thin disc (dotted) and thick disc (dashed)boundaries taken from Fuhrmann (2004). c (cid:13) , 000–000 ew Halo WD Candidates Figure 9.
Plots of W vs. V (top) and U vs. V (bottom) velocitydistributions for our sample of H-rich (black dots), He-rich (whitedots), and mixed (red) WDs. Also plotted are the 3 σ ellipsoids forthe Galactic thin disc (dotted), thick disc (dashed), and stellarhalo populations (solid). and future large scale surveys such as GAIA and
LSST willfind a significant number of cool white dwarfs, including halowhite dwarfs, in the solar neighborhood. With g − band mag-nitudes of 20 −
22, we expect parallax errors from
GAIA torange from about 400 − µ as , corresponding to uncer-tainties of ≈
20 per cent in both mass and cooling age forthe majority of our targets. In addition,
GAIA will revealthe brighter population of halo white dwarfs near the Sun. ACKNOWLEDGEMENTS
We gratefully acknowledge the support of the NSF andNASA under grants AST-1312678 and NNX14AF65G, re-spectively.
REFERENCES
Bergeron, P., Saumon, D., & Wesemael, F. 1995, ApJ, 443,764Bergeron, P. 2001, ApJ, 558, 369Bergeron, P., Leggett, S. K., & Ruiz, M. T. 2001, ApJS,133, 413Bergeron, P., & Leggett, S. K. 2002, ApJ, 580, 1070Bergeron, P., Ruiz, M. T., Hamuy, M., et al. 2005, ApJ,625, 838Bergeron, P., Wesemael, F., Dufour, P., et al. 2011, ApJ,737, 28Brown, W. R., McLeod, B. A., Geary, J. C., & Bowsher,E. C. 2008, Proceedings of the SPIE, 7014, 70142Campos, F., Bergeron, P., Romero, A. D., et al. 2015, MN-RAS, in press, arXiv:1512.03114Chiba, M., & Beers, T. C. 2000, AJ, 119, 2843Fuhrmann, K. 2004, Astronomische Nachrichten, 325, 3Gates, E., Gyuk, G., Harris, H. C., et al. 2004, ApJ, 612,L129Hansen, B. M. S. 1998, Nature, 394, 860Hansen, B. M. S., Richer, H. B., Fahlman, G. G., et al.2004, ApJS, 155, 551Hansen, B. M. S., Anderson, J., Brewer, J., et al. 2007,ApJ, 671, 380Harris, H. C., Munn, J. A., Kilic, M., et al. 2006, AJ, 131,571Harris, H. C., Gates, E., Gyuk, G., et al. 2008, ApJ, 679,697Holberg, J. B., & Bergeron, P. 2006, AJ, 132, 1221Gianninas, A., Curd, B., Thorstensen, J. R., et al. 2015,MNRAS, 449, 3966Johnson, D. R. H., & Soderblom, D. R. 1987, AJ, 93, 864Kalirai, J. S. 2012, Nature, 486, 90Kalirai, J. S., Richer, H. B., Anderson, J., et al. 2012, AJ,143, 11Kilic, M., Munn, J. A., Harris, H. C., et al. 2006, AJ, 131,582Kilic, M., Munn, J. A., Williams, K. A., et al. 2010, ApJ,715, L21Kilic, M., Leggett, S. K., Tremblay, P.-E., et al. 2010, ApJS,190, 77Kilic, M., Thorstensen, J. R., Kowalski, P. M., & Andrews,J. 2012, MNRAS, 423, L132Kowalski, P. M., & Saumon, D. 2006, ApJ, 651, L137Lawrence, A., Warren, S. J., Almaini, O., et al. 2007, MN-RAS, 379, 1599Leggett, S. K., Ruiz, M. T., & Bergeron, P. 1998, ApJ, 497,294Liebert, J., Dahn, C. C., & Monet, D. G. 1988, ApJ, 332,891Liebert, J., Kilic, M., Williams, K. A., et al. 2007, 15thEuropean Workshop on White Dwarfs, 372, 129Munn, J. A., Harris, H. C., von Hippel, T., et al. 2014, AJ,148, 132 c (cid:13) , 000–000 K. Dame et al.
Oppenheimer, B. R., Saumon, D., Hodgkin, S. T., et al.2001, ApJ, 550, 448Pauli, E.-M., Napiwotzki, R., Altmann, M., et al. 2003,A&A, 400, 877Pauli, E.-M., Napiwotzki, R., Heber, U., Altmann, M., &Odenkirchen, M. 2006, A&A, 447, 173Rau, A., Kulkarni, S. R., Law, N. M., et al. 2009, PASP,121, 1334Richter, R., Heber, U., & Napiwotzki, R. 2007, 15th Euro-pean Workshop on White Dwarfs, 372, 107Schmidt, G. D., Weymann, R. J., & Foltz, C. B. 1989,PASP, 101, 713Tonry, J. L., Stubbs, C. W., Kilic, M., et al. 2012, ApJ,745, 42Tremblay, P.-E., & Bergeron, P. 2009, ApJ, 696, 1755Tremblay, P.-E., Ludwig, H.-G., Steffen, M., & Freytag, B.2013, A&A, 559, A104Tremblay, P.-E., Leggett, S. K., Lodieu, N., et al. 2014,ApJ, 788, 103Vidrih, S., Bramich, D. M., Hewett, P. C., et al. 2007, MN-RAS, 382, 515Winget, D. E., Hansen, C. J., Liebert, J., et al. 1987, ApJ,315, L77 c (cid:13)000