Recent temperature determinations of hot H-rich (DA) and H-poor (DO, DB) white dwarfs and PG1159 stars are based on model Balmer line profile fits, supplemented by far/extreme UV continuum fits or ideally ionization equilibrium techniques, where observations permit.
Pure H-atmosphere models are generally used to analyse DA white dwarfs,
with temperatures obtained from optical and EUV analyses
showing good consistency (see Barstow et al. 1993), and show improved
temperature sensitive relative to earlier UV continuum/Ly
fits (e.g.
Holberg et al. 1986).
However, there is a well known `Balmer line problem' for some high
gravity H-rich central stars of Planetary
Nebulae (CSPNe), well illustrated by Napiwotzki
(1995b) who obtained discrepant temperatures using different Balmer series
members for the CSPNe of the Helix nebula NGC7293, ranging from
48kK (using H
) to 107kK (H
)! While unsolved,
the incorporation of
metal line blanketing improves the situation, and
temperatures obtained from higher Balmer lines are in
better agreement with alternative spectroscopic (
70kK
since HeI absent) and photoionization
(120kK from Clegg & Walsh 1989) results.
Napiwotzki (1995a) discussed the importance of non-LTE effects even for cool DO white dwarfs, despite their high surface gravities, while Dreizler & Werner (1996) have compared results based on blanketed non-LTE models with previous LTE analyses finding the greatest difference between 60--70kK. Werner et al. (1991) found that the non-LTE analyses are critical for the higher temperature PG1159 stars. In the absence of superior spectroscopic diagnostics, temperature determinations in DB white dwarfs rely on HeI profile fits using LTE models (e.g. Beauchamp et al. 1995).
Stellar temperatures of CSPNe (and OB stars in HII regions) can also be obtained by means of their nebulae, most frequently via Zanstra analyses (Zanstra 1931; Harman & Seaton 1966). However, results from HeII are frequently inconsistent with those from HI (the `Zanstra discrepancy') due to the use of inappropriate (often blackbody) flux distributions (Gabler et al. 1991), or an optically thin PNe. Use of realistic O-type model fluxes by Sellmaier et al. (1996) provides a possible solution to the [NeIII] problem in HII-regions, though model atmospheres are not yet fully capable of explaining both the stellar and nebular spectrum simultaneously.