1: \documentclass[12pt,preprint]{aastex}
2:
3: \newcommand{\che} {\log\ ({\rm C/He})}
4: \newcommand{\mv} {M_V}
5: \newcommand{\msun} {$M_{\odot}$}
6: \newcommand{\halpha} {H$\alpha$}
7: \newcommand{\hbeta} {H$\beta$}
8: \newcommand{\Te} {T_{\rm eff}}
9: \newcommand{\logg} {\log g}
10: \newcommand{\nh} {\log\ ({\rm H/He})}
11:
12: \begin{document}
13:
14: \title{Rare White dwarf stars with carbon atmospheres}
15:
16: \author{P. Dufour\altaffilmark{1},
17: James Liebert\altaffilmark{1},
18: G. Fontaine\altaffilmark{2}, and
19: N. Behara\altaffilmark{3}}
20:
21:
22: \altaffiltext{1}{Steward Observatory, University of Arizona, 933 North Cherry Avenue, Tucson, AZ 85721; dufourpa@as.arizona.edu, liebert@as.arizona.edu}
23: \altaffiltext{2}{D\'{e}partement de Physique, Universit\'{e}
24: de Montr\'{e}al, C.P. 6128, Succ. Centre-Ville, Montr\'{e}al, Qu\'{e}bec,
25: Canada H3C 3J7; fontaine@astro.umontreal.ca}
26: \altaffiltext{3}{CIFIST, GEPI, Observatoire Paris-Meudon, 92195, France; natalie.behara@obspm.fr}
27:
28:
29: \begin{abstract}
30:
31: White dwarfs represent the endpoint of stellar evolution for stars
32: with initial masses between approximately 0.07 \msun~and 8-10 \msun, where
33: \msun~is the mass of the Sun (more massive stars end their life as
34: either black holes or neutron stars). The theory of stellar evolution
35: predicts that the majority of white dwarfs have a core made of carbon
36: and oxygen, which itself is surrounded by a helium layer and, for
37: $\sim$80 per cent of known white dwarfs, by an additional hydrogen
38: layer$^{1-3}$. All white dwarfs therefore have been traditionally
39: found to belong to one of two categories: those with a hydrogen-rich
40: atmosphere (the DA spectral type) and those with a helium-rich
41: atmosphere (the non-DAs). Here we report the discovery of several
42: white dwarfs with atmospheres primarily composed of carbon, with
43: little or no trace of hydrogen or helium. Our analysis shows that the
44: atmospheric parameters found for these stars do not fit satisfactorily
45: in any of the currently known theories of post-asymptotic giant branch
46: evolution, although these objects might be the cooler counterpart of
47: the unique and extensively studied PG1159 star H1504+65 (refs
48: 4-7). These stars, together with H1504+65, might
49: accordingly form a new evolutionary sequence that follow the
50: asymptotic giant branch.
51: \end{abstract}
52:
53:
54:
55:
56: Traces of carbon are typically observed as either neutral carbon lines
57: or molecular C$_2$ Swan bands (defining the DQ spectral type) in cool
58: helium-rich white dwarfs with effective temperatures ($\Te$) below
59: $\sim$ 13,000~K. The presence of carbon in the atmospheres of these
60: objects has been explained successfully by a model in which carbon is
61: dredged-up from the underlying carbon/oxygen core by the deep helium
62: convection zone$^8$. This model predicts a maximum contamination of
63: carbon at an effective temperature of $\sim$10,000~K (corresponding
64: approximately to the temperature at which the surface convection zone
65: is maximal) before gradually decreasing with lower temperature, in
66: agreement with atmospheric analysis determinations$^9$. We note that
67: although some of these objects show a very high level of carbon
68: pollution, helium always remains the dominant constituent of the
69: atmosphere (the highest carbon abundances by number (N) found are
70: around logN(C)/N(He) $\approx -3$).
71:
72: Atmospheric abundance determination for DQ white dwarfs has been most
73: successful for stars on the cool side of the maximum contamination and
74: very few analyses have been done for hotter objects located on the
75: ascending side of the curve$^{10-12}$. Recently$^{13}$, several new hot
76: DQ white dwarfs have been discovered in the Sloan Digital Sky Survey
77: (SDSS), providing a unique opportunity to test the dredge-up theory on
78: the ''hot side''. However, appropriate atmospheric models were not
79: available at that time. Because the atmospheric compositions of these
80: white dwarfs were expected to be helium-dominated, effective
81: temperature estimates (T$_{He}$) were obtained by comparing the SDSS
82: colours with those of pure helium composition models. Although it was
83: expected that such an approximation would overestimate the effective
84: temperature by several thousand kelvin, some of these stars were so
85: hot (T$_{He} \ge$ 25,000~K) that an overlap with most of the helium-rich
86: star (the DB spectral type) temperature range seemed unavoidable. It
87: was hypothesized that these stars had thinner outer helium envelopes
88: so that dredge-up occurred earlier in the cooling sequence. These
89: highly carbon-polluted white dwarfs are expected to be massive, so it
90: was argued that they might represent the missing high-mass tail of the
91: DB mass distribution$^{14}$.
92:
93: Thus, it is with this scientific rationale in mind that we proceeded
94: with the calculation of the appropriate atmospheric models for these
95: objects. Because the continuum opacity of heavy elements might not be
96: negligible for these objects, these new models have been updated with
97: the latest C and O photoionization cross-sections from the Opacity
98: Project$^{15}$. Although the analysis of the coolest objects ($\Te \le
99: $15,000~K) is straightforward (results will be presented elsewhere;
100: manuscript in preparation), we quickly realized that no combination of
101: carbon and helium could successfully reproduce the observed features
102: (mostly C\textsc{ii} lines) in the optical spectrum of the hottest ones by
103: assuming a helium-dominated atmosphere. Indeed, such models predict
104: the presence of a strong He\textsc{i} $\lambda$=4,471 line that is not observed
105: spectroscopically in our sample of hot DQ stars. We thus concluded
106: that a good fit to both the spectra and the energy distribution was
107: possible only by considering atmospheres made primarily of carbon,
108: with little or no trace of hydrogen and helium.
109:
110: Figure 1 shows our fits to the optical spectrum and photometric energy
111: distribution of one such star using a grid of pure carbon atmosphere
112: models, as well as models containing traces of hydrogen and
113: helium. The fitting method is similar to that used for cooler DQ and
114: needs not to be repeated here$^9$. This particular case shows a small
115: trace of hydrogen which we believe is more likely to be the result of
116: accretion from the interstellar medium than of primordial origin. We
117: found eight more similar objects in the SDSS white dwarf
118: catalogue$^{16}$, all of which were found to have a carbon-dominated
119: atmosphere and a temperature between 18,000 and 23,000~K, although for
120: some faint stars, detailed analysis is more uncertain because of the
121: poor signal-to-noise ratio of the observations. Only crude upper
122: limits for hydrogen and helium (of the order of logN(C)/N(H) $\ge$ 1.5 and
123: logN(C)/N(He) $\ge$ 1.5) can be obtained from the other spectra. Our
124: detailed model atmosphere fits provide fair estimates of the effective
125: temperature and chemical composition, but higher-quality data (which
126: we hope to obtain in December 2007, weather permitting) are essential
127: for a precise measure of the surface gravity (and thus mass) of these
128: objects.
129:
130: These new stars are too hot to be explained by the standard convective
131: dredge-up scenario and so we sought another explanation. Natural
132: progenitor candidates to consider are the hot PG1159 stars. The
133: latter are probably the result of a late He-shell flash$^{17}$ (the
134: so-called ''born-again'' scenario) at the end of the
135: post-asymptotic-giant-branch phase, which almost completely eliminates
136: the remaining hydrogen and mixes material from the interior with the
137: helium envelope. As a result, these objects re-enter the white dwarf
138: cooling track, but this time they have a surface composition that is a
139: mixture of helium, carbon, oxygen, and little or no hydrogen ($\Te$
140: between $\sim$75,000 and 200,000 K; typical abundances$^{18}$, in mass
141: fraction, of He, C, O and Ne are 33$\%$, 50$\%$, 15$\%$ and 2$\%$). However, as
142: these stars cool down, gravitational diffusion rapidly separates the
143: heavier elements from the helium that tends to float to the surface
144: and by the time they have reached 25,000 K, helium completely
145: dominates the surface composition.
146:
147: Models exploring the evolution of stars 9-11 \msun~have produced white
148: dwarfs with O-Ne-Mg cores and CO envelopes and possibly little He and
149: H; such models may explain our peculiar stars$^{19-21}$. However, it
150: is unclear what quantity of He and H should remain for such
151: stars. Furthermore, the resulting white dwarfs have high surface
152: gravities ($\logg \approx 9$, in centimetre-gram-second units) and
153: would thus show extremely broad carbon lines that are incompatible
154: with the observed widths of the lines (Fig. 1 shows that $\logg
155: \approx 8$ is more likely, although analysis of a better spectrum is
156: needed for more precision). The most likely scenario to explain the
157: very existence of hot DQ white dwarfs with carbon-dominated
158: atmospheres is that these stars are the progenies of objects such as
159: H1504+65. The latter is a unique object among the known PG1159 stars
160: so far. It is the hottest specimen of its class at $\Te \approx$
161: 200,000 K, and its atmospheric composition very unusual, with a mass
162: fraction of $\sim 50\%$ C and $\sim 50\%$ O plus small traces of heavier
163: elements, but no detectable helium or hydrogen. It is currently
164: believed that H1504+65 is essentially a bare stellar nucleus produced
165: by a particularly violent post-asymptotic-giant-branch very late
166: thermal pulse that has destroyed in large part the remaining stellar
167: envelope containing helium and hydrogen. The best available
168: simulations of such an event, those of ref. 17, suggest indeed that,
169: following the last He flash, the full envelope becomes convective and
170: extends to deep enough layers for H and He to be completely consumed,
171: except that very small traces of He may survive.
172:
173: The uniformity of the chemical composition of the envelope of H1504+65
174: is believed to be maintained for a while by a residual wind but, with
175: time, diffusion becomes dominant and He, C and O must separate under
176: the influence of gravitational settling. The idea is that the wind
177: slowly dies out with cooling (decreasing luminosity) and that the
178: braking effect it provides on element separation becomes less and less
179: efficient. Hence, the small residual trace of helium believed to exist
180: in the envelope of H1504+65 eventually diffuses upward to form a thin
181: layer above a C-enriched and O-depleted mantle. The total mass of an
182: atmosphere is tiny ($\sim 10^{-14}$ to $10^{-15}$ M$_*$, where M$_*$
183: is the mass of the star), so there is ultimately enough accumulated He
184: to form a full atmosphere and the descendant of H1504+65 would now be
185: ''disguised'' as a He-atmosphere white dwarf after the PG1159 evolutionary
186: phase. With further cooling, a convection zone develops in the
187: C-enriched mantle owing to the recombination of that element while the
188: overlaying He layer remains radiative. At some $\Te$, whose exact
189: value is not yet known due to a current lack of proper models, it is
190: conjectured that the subphotospheric C convection zone becomes active
191: enough to be able to dilute from below the overlaying He layer. At
192: that point the star would undergo a dramatic spectral change,
193: transforming itself from a He-dominated atmosphere white dwarf to a
194: star with a C-dominated atmosphere, because the mass in the C
195: convection zone is orders of magnitude larger than the mass of the He
196: layer.
197:
198: Hence, the former PG1159 star H1504+65, showing initially a mixed C
199: and O atmosphere, would now show a C-dominated atmosphere after an
200: intermediate phase in which it would have been observed as a
201: He-atmosphere white dwarf. We note, however, that helium must
202: ultimately reappear at the surface in enough quantity to form a
203: helium-rich atmosphere because no carbon-rich object has ever been
204: discovered at lower temperatures ($\Te \le$ 15,000 K). The remaining
205: helium probably floats again at the surface when the star cools down,
206: perhaps turning the carbon-rich objects into DQ stars belonging to the
207: second sequence of high carbon abundances$^{11,22}$. We are at present
208: making evolutionary calculations to test the above scenario.
209:
210: To conclude, we roughly estimate that the space density of carbon-rich
211: white dwarfs, assuming a 50$\%$ completeness for the targeting in
212: SDSS$^{16}$, is between $\sim 2.1\times10^{-6}$~pc$^{-3}$ and $\sim
213: 7.0\times10^{-8}$~pc$^{-3}$ depending on the preferred value of
214: $\logg$ (and thus radius) and $\Te$ for the stars. For comparison, the
215: local space density of white dwarfs$^{23}$ is estimated to be $\sim
216: 5.0\times10^{-3}$~pc$^{-3}$. White dwarfs with carbon dominated
217: atmospheres are thus intrinsically rare and a relatively large volume
218: of space (as was done by the Sloan Digital Sky Survey) had to be
219: surveyed before many of them could be found.
220:
221:
222: \noindent 1. Iben, I., Jr. On the frequency of planetary nebula nuclei powered by helium burning and on the frequency of white dwarfs with hydrogen-deficient atmospheres. Astrophys. J. {\bf277}, 333-354 (1984).\\
223: 2. Koester, D. $\&$ Schoenberner, D. Evolution of white dwarfs. Astron. Astrophys. {\bf154}, 125-134 (1986).\\
224: 3. D'Antona, F. $\&$ Mazzitelli, I. Evolutionary times of white dwarfs-Long or short? IAU Colloq. {\bf95}, 635-637 (1987).\\
225: 4. Nousek, J. A. et al. H1504+65-an extraordinarily hot compact star devoid of hydrogen and helium. Astrophys. J. {\bf309}, 230-240 (1986).\\
226: 5. Werner, K. NLTE analysis of the unique pre-white dwarf H 1504 + 65. Astron. Astrophys. {\bf251}, 147-160 (1991).\\
227: 6. Werner, K. $\&$ Wolff, B. The EUV spectrum of the unique bare stellar core H1504+65. Astron. Astrophys. {\bf347}, L9-L13 (1999).\\
228: 7. Werner, K., Rauch, T., Barstow, M. A. $\&$ Kruk, J. W. Chandra and FUSE spectroscopy of the hot bare stellar core H 1504+65. Astron. Astrophys. {\bf421}, 1169-1183 (2004).\\
229: 8. Pelletier, C., Fontaine, G., Wesemael, F., Michaud, G. $\&$ Wegner, G. Carbon pollution in helium-rich white dwarf atmospheres. Time-dependent calculations of the dredge-up process. Astrophys. J. {\bf307}, 242-252 (1986).\\
230: 9. Dufour, P., Bergeron, P. $\&$ Fontaine, G. Detailed spectroscopic and photometric analysis of DQ white dwarfs. Astrophys. J. {\bf627}, 404-417 (2005).\\
231: 10. Wegner, G. $\&$ Koester, D. Atmospheric analysis of the carbon white dwarf G227-5. Astrophys. J. {\bf288}, 746-750 (1985).\\
232: 11. Thejll, P., Shipman, H. L., MacDonald, J. $\&$ Macfarland, W. M. An atmospheric analysis of the carbon-rich white dwarf G35 - 26. Astrophys. J. {\bf361}, 197-206 (1990).\\
233: 12. Desharnais, S., Wesemael, F., Chayer, P. $\&$ Kruk, J. W. FUSE observation of cool DB white dwarfs. ASP Conf. Ser. {\bf372}, 265-268 (2007).\\
234: 13. Liebert, J. et al. SDSS white dwarfs with spectra showing atomic oxygen and/or carbon lines. Astron. J. {\bf126}, 2521-2528 (2003).\\
235: 14. Beauchamp, A. D\'etermination des Param\`etres Atmosph\'eriques des \'etoiles Naines Blanches de Type DB. PhD thesis, Montr\'eal (1995).\\
236: 15. Behara, N. $\&$ Jeffery, C. S. LTE model atmosphere with new opacities. 1. Methods and general properties. Astron. Astrophys. {\bf451}, 643-650 (2006).\\
237: 16. Eisenstein, D. J. et al. A catalog of spectroscopically confirmed white dwarfs from the Sloan Digital Sky Survey data release 4. Astrophys. J. {\bf167} (Suppl.), 40-58 (2006).\\
238: 17. Herwig, F., Bl\"ocker, T., Langer, N. $\&$ Driebe, T. On the formation of hydrogen-deficient post-AGB stars. Astron. Astrophys. {\bf349}, L5-L8 (1999).\\
239: 18. Werner, K. $\&$ Herwig, F. The elemental abundances in bare planetary nebula central stars and the shell burning in AGB stars. Publ. Astron. Soc. Pacif. {\bf118}, 183-204 (2006).\\
240: 19. Garcia-Berro, E. $\&$ Iben, I. On the formation and evolution of super-asymptotic giant branch stars with cores processed by carbon burning. 1: SPICA to Antares. Astrophys. J. {\bf434}, 306-318 (1994).\\
241: 20. Garcia-Berro, E., Ritossa, C. $\&$ Iben, I. On the evolution of stars that form electron-degenerate cores processed by carbon burning. III. The inward propagation of a carbon-burning flame and other properties of a 9 \msun~model star. Astrophys. J. {\bf485}, 765-784 (1997).\\
242: 21. Ritossa, C., Garcia-Berro, E. $\&$ Iben, I. On the evolution of stars that form electron-degenerate cores processed by carbon burning. V. Shell convection sustained by helium burning, transient neon burning, dredge-out, URCA cooling, and other properties of an 11 \msun~population I model star. Astrophys. J. {\bf515}, 381-397 (1999).\\
243: 22. Koester, D. $\&$ Knist, S. New DQ white dwarfs in the Sloan Digital Sky Survey DR4: confirmation of two sequences. Astron. Astrophys. {\bf454}, 951-956 (2006).\\
244: 23. Holberg, J. B., Oswalt, T. D. $\&$ Sion, E. M. A determination of the local density of white dwarf stars. Astrophys. J. {\bf571}, 512-518 (2002).\\
245:
246:
247: \acknowledgements This work has been partially supported by the NSF
248: for work on SDSS white dwarfs. This work was also supported in part by
249: the NSERC (Canada).
250:
251: {\bf Author Information} Reprints and permissions information is
252: available at www.nature.com/reprints. Correspondence and requests for
253: materials should be addressed to P.D. (dufourpa@as.arizona.edu).
254:
255: \clearpage
256:
257: \clearpage
258: \begin{figure}[p]
259: \figcaption[f1.eps] {{\bf Fit to the optical spectra and energy
260: distribution for a carbon-rich white dwarf. a}, The thick red line
261: represents our $\logg$ = 8 best fit (the parameters are indicated in
262: the panel) while the blue line represents a solution for $\logg$ =
263: 9. Hydrogen abundance is determined by fitting $H\beta$
264: ($\lambda$=4,861$\rm \AA$) while helium abundance is constrained by
265: the absence of the He\textsc{ii} $\lambda$ =4,471$\rm \AA$ line. {\bf
266: b}, Photometric measurements in the u, g, r, i and z bands are
267: represented by error bars, while the average model fluxes from the
268: same best $\logg$ = 8 model are shown by filled circles. Both {\bf a}
269: and {\bf b} have the same y axis.
270: \label{fg:f1}}
271: \plotone{f1}
272: \begin{flushright}
273: Figure \ref{fg:f1}
274: \end{flushright}
275: \end{figure}
276:
277:
278: \end{document}
279:
280: