3 On the nature of GSC2U J131147.2+292348

The flux-calibrated spectrum of GSC2U J131147.2+292348 is shown in Fig. 2. The signal-to-noise is around 10 for the whole spectrum, increasing slightly to the red. This noise level is clearly visible in the spectrum, and limits our ability to detect weak features.

The crosses in Fig. 2 represent the fluxes at different effective wavelengths as derived from the $B_{\rm J}$ , V₁₂, $R_{\rm F}$, and $I_{\rm N}$ photographic magnitudes in Table 2. The ultraviolet flux was derived from the photographic U magnitude of Moreau et al. (1995). The agreement appears reasonably consistent with the 10% and 20% accuracy levels of the flux-calibrated spectroscopy and the photographic photometry, respectively.

The spectrum appears dominated by strong absorption bands due to C₂ molecules. The four Swan bands with bandheads at $\lambda= 4382$, 4737, 5165, and 5636 Å are clearly identified, along with the less common Swan band at 6191 Å. In addition, strong Deslandres and d'Azambuja (D-d'A) absorption bands are also present in the blue part of the spectrum at 3600, 3852, and 4102 Å. These bands have been observed in the spectra of WDs with carbon rich atmospheres (DQ WDs) and temperatures above 6500 K. Finally, the spectrum in Fig. 2 shows an evident depression of the continuum in the Swan band region between 4500 and 6200 Å.

The spectral energy distribution (SED) of DQ stars changes with $T_{\rm eff}$ and carbon abundance as shown by the model atmosphere spectra presented in Koester et al. (1982) and Wegner & Yackovich (1984). Figure 5 of Wegner & Yackovich gives an indication on what to expect for different combinations of $T_{\rm eff}$ and C:He abundance. Swan bands are generally present, while D-d'A bands start to become visible in models with C: $\rm He\ga 10^{-6}$ at $T_{\rm eff}\simeq 6600$ K and with C: $\rm He \ga 10^{-2}$at $T_{\rm eff} \simeq 10~000$ K.

A SED with C₂ bands similar in strength to those observed in our spectrum requires a much enhanced C:He ratio for the given $T_{\rm eff}$. This can be seen by comparing the models in Fig. 5 of Wegner & Yackovich with those in their Figs. 2 and 3. At temperatures between 6000 K and 7000 K, deep absorption bands are produced with C: $\rm He\approx 10^{-4}$. At $T_{\rm eff} = 8000$ K, carbon abundance has to increase to a rather extreme value, C: $\rm He=0.9$, for the simultaneous presence of strong D-d'A and Swan bands in the synthetic SED (bottom panel of Fig. 3 of Wegner & Yackovich). This model bears the most resemblance with the spectrum of our WD, however, it does not show any evidence of the continuum depression seen in the observed spectrum. Theoretical evidence that such depression of the continuum emission could occur is provided in Koester et al. (1982). Their Fig. 1 displays theoretical C₂ spectra at $T_{\rm eff} = 8000$ K and increasingly higher C:He ratios. The effect is to boost band strengths, thus depressing the continuum in the Swan-band region.

Although the models with $T_{\rm eff} = 8000$ K just examined seem consistent with the appearance of the C₂ band systems observed in the spectrum of our WD, the relative flux at blue wavelengths (below $\sim$4100 Å) is probably too high compared to the observed SED in Fig. 2. In this regard, an attempt to find a black body compatible with the observed spectrum at $\lambda > 7000$ Å, the NIR fluxes from our $JHK_{\rm s}$ magnitudes, and with the blue peaks in the D-d'A region, resulted in a black-body temperature of $\sim$6000 K. (Note that in this case the depressed continuum occurs in the region of maximum black-body emission.)

From the discussion above, it is evident that much is still to be learned about the properties of this new DQ star, and the reliable determination of its temperature and chemical composition must await more detailed atmosphere models. Also, improved spectral coverage in the UV, below 3500 Å, would probably be of help in better constraining model calculations.

Finally, an approximate photometric parallax for GSCU J131147.2+292348 was estimated from the absolute magnitudes of theoretical models of non-DA stars. From the values in Tables 2 and 4 of Bergeron et al. (1995) for pure helium atmosphere WDs and averaging the distance moduli computed for the $IJHK_{\rm s}$ bands (which are not affected by the strong C₂ absorption bands) we estimate the distances $d\approx 70$, 80, and 90 parsecs for $T_{\rm eff}=6000$ K, 7000 K, and 8000 K, respectively.

This distance interval corresponds to a range of tangential velocity $V_{\rm tan}=4.74 \cdot \mu ~ d \simeq 160{-}200$ km s^-1 and galactic components with respect to the LSR from $(U,V)\simeq (-148.1, +9.6)$ to $(U,V)\simeq (-193.3, +10.8)$ km s^-1, for d=70 pc and 90 pc, respectively. These relatively high values are not consistent (3$\sigma$) with the velocity distribution of the thin disk, while they are consistent with the kinematics of the halo or thick disk stellar population.