Discovery of Molecular Hydrogen in White Dwarf Atmospheres
aa r X i v : . [ a s t r o - ph . S R ] F e b Discovery of Molecular Hydrogen in White Dwarf Atmospheres
S. Xu( 许 偲 艺 ) a , M. Jura a , D. Koester b , B. Klein a , B. Zuckerman a ABSTRACT
With the Cosmic Origins Spectrograph onboard the
Hubble Space Telescope ,we have detected molecular hydrogen in the atmospheres of three white dwarfswith effective temperatures below 14,000 K, G29-38, GD 133 and GD 31. Thisdiscovery provides new independent constraints on the stellar temperature andsurface gravity of white dwarfs.
Subject headings: white dwarfs, atmospheres
1. INTRODUCTION
An important discovery from the
International Ultraviolet Explorer (IUE) was two broadabsorption features at 1400 ˚A and 1600 ˚A in hydrogen-dominated (DA) white dwarfs coolerthan 20,000 K and 13,500 K, respectively (Greenstein & Oke 1979; Wegner 1982). Subse-quently, these features were explained by Koester et al. (1985) and Nelan & Wegner (1985)as Ly α satellite lines from collisions among H-H (1600 ˚A) and H-H + (1400 ˚A). These quasi-molecular absorption lines help to constrain stellar temperature and surface gravity for DAwhite dwarfs (Allard & Koester 1992). G29-38, GD 133 and GD 31, the targets described inthis letter, all show these quasi-molecular absorption features when observed with the IUE [Holm et al. (1985); Kepler & Nelan (1993),
IUE archive]. Here, we report first detectionsof true molecular hydrogen in these three white dwarfs, by employing the Cosmic OriginsSpectrograph (COS) on the
Hubble Space Telescope (HST) . These three stellar atmospheresare among the hottest stellar environments where photospheric molecular hydrogen has everbeen detected. a Department of Physics and Astronomy, University of California, Los Angeles CA 90095-1562;[email protected], [email protected], [email protected], [email protected] b Institut fur Theoretische Physik und Astrophysik, University of Kiel, 24098 Kiel, Germany;[email protected]
2. DATA
G29-38 and GD 133 were observed as part of program 12290 during
HST cycle 18,which focuses on using externally polluted white dwarfs to assess the volatile abundancesof accreted extrasolar planetesimals [see Jura (2003); Zuckerman et al. (2007); Klein et al.(2010); Dufour et al. (2012); Jura et al. (2012); G¨ansicke et al. (2012) and references within].G29-38 and GD 133 were chosen because they show a high degree of atmospheric pollu-tion from optical studies (Koester et al. 1997, 2005) and display excess infrared radiation(Zuckerman & Becklin 1987; Reach et al. 2005, 2009; Jura et al. 2007). As part of our pro-gram to study Hyades white dwarfs (Zuckerman et al. 2013), ultraviolet data for GD 31 wereretrieved from the
HST archive of the SNAPSHOT program 12169 (PI: B. G¨ansicke).For G29-38 and GD 133, the COS set-up was similar to that employed for GD 40and G241-6, as previously reported in Jura et al. (2012). The G130M grating was used withcentral wavelength 1300 ˚A and wavelength coverage of 1142 -1288 ˚A (strip B) and 1298 -1443˚A (strip A). The spectral resolution was ∼ ∼ ν ′′ = 2, 3, 4, 5 to ν ′ = 0 and Werner band transitions from ν ′′ =2, 3to ν ′ = 0 (Abgrall et al. 1993a,b). It turns out that we have accidentally found H in whitedwarf atmospheres! We present the strongest Lyman band H lines in Table 1; they are allappreciably stronger than the Werner bands in our observed wavelength interval.We serendipitously identified molecular hydrogen in a third star, GD 31, which may bea high-mass escaping member of the Hyades cluster (Zuckerman et al. 2013). The data arenoisier but clearly four absorption features are seen; these correspond to the strongest H Lyman band lines and their blends in G29-38, at the correct wavelengths, as presented inTable 1 and Figure 1.
3. DISCUSSION
According to the Saha equation, molecular hydrogen is most concentrated in low tem-perature, high density environments. Previously, ro-vibrational lines from H have beendetected in cool stellar atmospheres in the infrared (Spinrad 1964) and in the Sun’s ultravio-let emission spectrum (Jordan et al. 1978; Sandlin et al. 1986). Our ultraviolet detection ofphotospheric H introduces a new aspect to stellar physics. With stellar parameters of T ∗ =11,820 K, log g=8.40 for G29-38 (Xu et al. 2013) and T ∗ =12,121 K, log g = 8.005 for GD 133(Koester et al. 2009), we computed white dwarf model atmospheres (Koester 2010) and cal-culated the number density of molecular hydrogen relative to atomic hydrogen, n(H )/n(H),as shown in Figure 2. At maximum, molecular hydrogen is still 10 − . and 10 − . less thanthe amount of atomic hydrogen in G29-38 and GD 133, respectively. These computed ratiosare comparable to, but at the higher end of the concentration of trace elements in heavilypolluted white dwarfs (Jura et al. 2012). Because molecular hydrogen is distributed over alarge number of ro-vibrational levels, each individual line is relatively weak. As discussed inZuckerman et al. (2013), H is used to resolve a major temperature puzzle for GD 31. Due toits high gravity, the equivalent widths (EWs) of 4 detected H lines in GD 31 (T=13,700 K,log g=8.67) are comparable to those in GD 133; there is a substantial amount of molecularhydrogen in the atmosphere of GD 31.As shown in Figure 1, we computed the model spectra for G29-38 and GD 133 followingKoester (2010) with all the line data obtained from the Kurucz webpage (Kurucz 1995)and the H partition function from Irwin (1981). In a statistical sense, the model wellrepresents the data and reproduces most molecular hydrogen features. However, the fit to http://kurucz.harvard.edu/linelists.html lines.Another aspect of this discovery is that by constraining the abundance of the HDmolecule, we can place a limit on the D/H ratio. It is generally believed that all deu-terium is primordial from the Big Bang nucleosynthesis and destroyed in stellar interiors(Epstein et al. 1976). Any amount, if present, in the white dwarf atmosphere must comefrom some external source, likely relic planetesimals. The values are not very constraining forthe targets presented here. But if cooler stars can be observed with COS with a sufficientlyhigh signal-to-noise ratio, then more meaningful upper limits or actual detections of the HDmolecule in white dwarf atmospheres may be anticipated.In white dwarfs cooler than 12,000 K, H is present but usually there is not enoughultraviolet flux to make observations in a time-efficient manner. The environment in hotterstars is typically more hostile for molecular hydrogen but high pressure may still enabledetection of H .
4. CONCLUSIONS
With the
HST , molecular hydrogen was detected for the first time in white dwarf at-mospheres. The three stars, G29-38, GD 133 and GD 31, have temperatures between 11,800K and 13,700 K. H can be used as an independent constraint to white dwarf atmosphericconditions. This opens a door to many future explorations.Support for program REFERENCES
Abgrall, H., Roueff, E., Launay, F., Roncin, J. Y., & Subtil, J. L. 1993a, A&AS, 101, 273—. 1993b, A&AS, 101, 323Allard, N. F. & Koester, D. 1992, A&A, 258, 464Dufour, P., Kilic, M., Fontaine, G., Bergeron, P., Melis, C., & Bochanski, J. 2012, ApJ, 749,6Epstein, R. I., Lattimer, J. M., & Schramm, D. N. 1976, Nature, 263, 198Falcon, R. E., Rochau, G. A., Bailey, J. E., Ellis, J. L., Carlson, A. L., Gomez, T. A.,Montgomery, M. H., Winget, D. E., Chen, E. Y., Gomez, M. R., & Nash, T. J. 2012,astro-ph 1210.7197G¨ansicke, B. T., Koester, D., Farihi, J., Girven, J., Parsons, S. G., & Breedt, E. 2012,MNRAS, 424, 333Greenstein, J. L. & Oke, J. B. 1979, ApJ, 229, L141Holm, A. V., Panek, R. J., Schiffer, III, F. H., Bond, H. E., Kemper, E., & Grauer, A. D.1985, ApJ, 289, 774Irwin, A. W. 1981, ApJS, 45, 621Jordan, C., Brueckner, G. E., Bartoe, J.-D. F., Sandlin, G. D., & Vanhoosier, M. E. 1978,ApJ, 226, 687Jura, M. 2003, ApJ, 584, L91Jura, M., Farihi, J., & Zuckerman, B. 2007, ApJ, 663, 1285Jura, M., Xu, S., Klein, B., Koester, D., & Zuckerman, B. 2012, ApJ, 750, 69Kepler, S. O. & Nelan, E. P. 1993, AJ, 105, 608Klein, B., Jura, M., Koester, D., Zuckerman, B., & Melis, C. 2010, ApJ, 709, 950Koester, D. 2010, Mem. Soc. Astron. Italiana, 81, 921Koester, D., Provencal, J., & Shipman, H. L. 1997, A&A, 320, L57Koester, D., Rollenhagen, K., Napiwotzki, R., Voss, B., Christlieb, N., Homeier, D., &Reimers, D. 2005, A&A, 432, 1025 6 –Koester, D., Voss, B., Napiwotzki, R., Christlieb, N., Homeier, D., Lisker, T., Reimers, D.,& Heber, U. 2009, A&A, 505, 441Koester, D., Weidemann, V., Zeidler-K.T., E. M., & Vauclair, G. 1985, A&A, 142, L5Kurucz, R. L. 1995, in Astronomical Society of the Pacific Conference Series, Vol. 78, Astro-physical Applications of Powerful New Databases, ed. S. J. Adelman & W. L. Wiese,205Meyer, D. M., Lauroesch, J. T., Sofia, U. J., Draine, B. T., & Bertoldi, F. 2001, ApJ, 553,L59Nelan, E. P. & Wegner, G. 1985, ApJ, 289, L31Reach, W. T., Kuchner, M. J., von Hippel, T., Burrows, A., Mullally, F., Kilic, M., &Winget, D. E. 2005, ApJ, 635, L161Reach, W. T., Lisse, C., von Hippel, T., & Mullally, F. 2009, ApJ, 693, 697Sandlin, G. D., Bartoe, J.-D. F., Brueckner, G. E., Tousey, R., & Vanhoosier, M. E. 1986,ApJS, 61, 801Silvotti, R., Pavlov, M., Fontaine, G., Marsh, T., & Dhillon, V. 2006, Mem. Soc. Astron. Ital-iana, 77, 486Spinrad, H. 1964, ApJ, 140, 1639Thompson, S. E., van Kerkwijk, M. H., & Clemens, J. C. 2008, MNRAS, 389, 93Wegner, G. 1982, ApJ, 261, L87Xu, S., Jura, M., Klein, B., Koester, D., & Zuckerman, B. 2013, in preparationZuckerman, B. & Becklin, E. E. 1987, Nature, 330, 138Zuckerman, B., Klein, B., Xu, S., & Jura, M. 2013, ApJ submittedZuckerman, B., Koester, D., Melis, C., Hansen, B. M., & Jura, M. 2007, ApJ, 671, 872
This preprint was prepared with the AAS L A TEX macros v5.2. λ ( ˚A ) F λ (10 − erg s − cm − ˚A − ) G D G − G D + quasi-molecular feature. The spectra aresmoothed with a 5 point boxcar and shifted by 39 km s − , 58 km s − and 89 km s − forG29-38, GD 133 and GD 31 respectively, to be in the heliocentric reference frame (Xu et al.2013; Zuckerman et al. 2013). For clarity, the spectrum for GD 133 is offset by 0.5 × − erg s − cm − ˚A − and for GD 31 by 1.5 × − erg s − cm − ˚A − . The red dashed linesare our computed model spectra. Statistically, the models reproduce the data but the fitto individual lines is not ideal. For G29-38 and GD 133, the entire spectrum in strip A(1298-1443 ˚A) shows numerous ro-vibrational lines of H . The strongest absorption featuresaround 1345 ˚A, 1357 ˚A, 1366 ˚A and 1369 ˚A are also seen in GD 31. 9 –Fig 2. Ratio of number densities of molecular hydrogen compared to atomic hydrogenat different optical depths in G29-38 and GD 133. For G29-38, this ratio peaks at logn(H )/n(H) = -4.7, where τ = 0.04, T = 9900 K and ρ = 3.1 × − g cm − . For GD 133,the maximum of log n(H )/n(H) is -5.1, which occurs at τ = 0.05, T=10,200 K and ρ = 1.5 × − g cm − . 10 –Table 1: The Strongest Identified Lyman Band H Lines λ λ
Transition b G29-38 GD 133 GD 31(cm − ) a (˚A) EW (m˚A) EW (m˚A) EW (m˚A)74344.49 1345.1 0-4 R(5) 170 c c c c c c c c c c c c c c ...70889.48 1410.6 0-4 R(13) 112 c c ...70076.46 1427.0 0-4 P(13) 75 53 ... a From Abgrall et al. (1993a). b We follow the notational conventions in Meyer et al. (2001). c This line is blended and the reported number is the total EW for both lines together.
Note.
For G29-38 and GD 133, the EW uncertainty is dominated by the choice of continuuminterval because the whole region shows numerous H2