Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://star.arm.ac.uk/~ambn/csjpaper.ps Äàòà èçìåíåíèÿ: Mon Nov 11 19:22:12 1996 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 22:02:54 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: ultraviolet

A&A manuscript no.
(will be inserted by hand later)
Your thesaurus codes are:
06 (08.01.1; 08.03.2; 08.09.2 HD 144941;
ASTRONOMY
AND
ASTROPHYSICS
9.10.1996
The iron abundance in helium star HD144941 ?
C.S.Jeffery 1 and P.M.Harrison 2
1 Armagh Observatory, College Hill, Armagh BT61 9DG, Northern Ireland
2 Dept of Physics & Astronomy, University of St Andrews, Fife KY16 9SS, Scotland
Received . . . , 1996; accepted . . .
Abstract. The iron abundance in the helium star
HD144941 has been measured from ultraviolet Fe iii lines.
It is 1:9 \Sigma 0:2 dex below solar and confirms the low metal­
licity (Z = 0:0003) previously determined from measure­
ments of CNO and other light elements. The result is
important because the low CNO metallicity must be re­
flected in the iron­group elements in order to explain why
HD144941 does not show pulsations.
Key words: stars: helium -- stars: individual (HD 144941)
-- stars: abundances --
1. Introduction
It has been well­established that high­luminosity helium
stars (L=M ? ¸ 10 4 ) are unstable against radial or non­
radial pulsations at effective temperatures up to at least
30000 K (Saio & Jeffery 1988). An explanation for pulsa­
tions in the lower­luminosity helium star V652 Her (Hill et
al. 1981) was only obtained (Saio 1993) with the publica­
tion of new opacities. These opacities (Iglesias et al. 1992)
show a peak due to iron­group elements in the region of
10 6 K capable of driving pulsations in low­luminosity he­
lium stars with T eff ¸ 25 000K. The success of the theory
was vindicated by the prediction (Saio 1995) and discovery
(Kilkenny & Koen 1995) of pulsations in the helium star
LSS 3184. A problem was that the helium star HD144941
shows no evidence for pulsations (Jeffery & Hill 1996, Pa­
per I), although it lies in the same pulsational instability
strip (Harrison & Jeffery 1996, Paper II). A possible solu­
tion and an important test of the theory lay in the predic­
tion that the extent of the pulsational instability finger to
low luminosity is strongly metal­dependent (Saio 1995).
Measurements of light elements (C,N,O,Si,S) in V652Her
(Jeffery et al. 1986), LSS 3184 (Drilling et al. 1996) and
HD144941 (Paper II) appeared to confirm this.
Send offprint requests to: C.S.Jeffery
? Based on observations obtained with the IUE satellite
However pulsations are driven by iron­group elements
and the optical spectra of these stars contain only a few
weak lines of Fe iii. In particular, the iron abundance in
HD 144941 had been measured from only one very weak
line (Paper II) and indicated an iron underabundance (rel­
ative to solar) roughly 0.5 dex smaller than for lighter el­
ements. Thus it is vital to confirm the iron abundance of
HD 144941 in order to interpret the absence of pulsations
correctly.
2. Observations
Two well­exposed high­resolution spectra of HD 144941
were obtained with the IUE SWP camera (image numbers
SWP23961 and SWP23962) and have already been used
to investigate the stellar wind (Jeffery et al. 1987). The
region between 1830 š A and 1960 š A is known to contain a
number of strong Fe iii lines. Kurucz' (1992) list of lines
with accurate wavelengths reveals 1020 Fe iii transitions,
of which 280 have gf ? 0:1, many of which are clearly
present in the IUE spectrum of HD 144941.
Since T eff and log g have already been derived from a
high­resolution high­S/N optical spectrum (Paper II), and
an appropriate model atmosphere exists it was decided to
synthesize the Fe iii spectrum of HD 144941 in the ultra­
violet in order to determine the iron abundance.
3. Iron abundance
The LTE radiative transfer program SPECTRUM
(Dufton, Lennon, Conlon & Jeffery, unpublished) and the
Kurucz list of theoretical probabilities for transitions be­
tween measured energy levels were used to synthesize spec­
tra including contributions from all species in the wave­
length interval 1880 -- 1920 š A. As a starting approxima­
tion, all abundances were set 1.5 dex below solar. Compar­
ing the result with a spectrum containing only Fe iii lines
it was clear that very few lines of other species contribute
in this region.
The synthetic spectrum was convolved with an instru­
mental profile (Gaussian FWHM=0.2š A). The continuum

2 C.S.Jeffery et al.: The iron abundance in helium star HD144941
Fig. 1. The IUE high­resolution spectrum of HD 144941
(histogram) compared with three synthetic spectra, with
log nFe = \Gamma5:8 (smooth line) \Sigma0:4 (dotted lines) and v t = 5
km/s.
level of the observed spectrum was adjusted to match the
continuum of the synthetic spectrum, which can be identi­
fied with reasonable confidence despite the line crowding.
There are line­to­line variations in the quality of the fit
(Fig. 1). These are due to errors in the adopted gf­values,
observed lines not present in Kurucz' linelist and noise
in the observations. A comparison of thirteen predicted
transition probabilities (Ekberg 1993) revealed one signif­
icant discrepancy (–1901:10 š A) in favour of the observed
spectrum.
With both strong and weak Fe iii lines accessible, it
was possible to check the value of the microturbulent ve­
locity ( v t ). Jeffery et al. (1987) had determined v t = 10
km/s from fits to the C ii –1324 š A line profile, which is sen­
sitive to the adopted microturbulence. For the Fe iii lines,
both v t and nFe were determined by minimizing the inte­
grated square residual between the observed and synthetic
spectra in the interval 1880 ­ 1920 š A. The best fit was ob­
tained with v t = 5 \Sigma 5 km/s and log nFe = n He = \Gamma5:8 \Sigma 0:2
(Fig. 1). There is not sufficient S/N and resolution in the
observed spectrum, or confidence in the atomic data, to
rule out the value of v t = 10 km/s adopted in Paper II.
Increasing v t by 5 km/s reduces nFe by !! 0:1 dex. It
reduces the CNO abundances by ! 0:05 dex.
Normalizing to log \Sigma¯ i n i = 12:15 gives log nFe =
5:7 \Sigma 0:2, compared with a solar value of log nFe = 7:48
(Holweger et al. 1990), implying that iron is underabun­
dant in HD 144941 by 1:8 \Sigma 0:2 dex. This is in excel­
lent agreement with the well­determined CNO abundances
(paper II).
4. Pulsations in helium stars
The improved iron abundance indicates a fractional heavy
element abundance (by mass) of Z = 0:00024 and that
the iron­group abundances follow the CNO abundances.
It confirms that the absence of pulsations in HD 144941
(Paper I) is due to its very low metallicity.
References
Drilling J.S., Jeffery C.S., Heber U.: 1996. A&A to be submit­
ted
Ekberg J.O.: 1993. A&AS 101, 1
Harrison P.M., Jeffery C.S.: 1996, A&A in press (Paper II)
Hill P.W., Lynas­Gray A.E., Kilkenny D., Walker H.J.: 1981,
MNRAS 197, 81
Holweger H., Heise C., Kock, M.: 1990, A&A 232, 510
Jeffery C.S., Heber U., Hamann W.­R., 1987. New Insights in
Astrophysics, ESA SP­263, 369.
Jeffery C.S., Heber U., Hill P.W.: 1986, Hydrogen­deficient
stars and related objects, IAU Coll. 87, p. 101, Hunger K.,
Sch¨onberner D. & Rao N.K., Reidel, Dordrecht, Holland.
Jeffery C.S., Hill P.W.: 1996, Observatory 116, 156 (Paper I)
Kilkenny D., Koen C.: 1995, MNRAS 275, 326
Kurucz R.L.: 1992, Stellar atmospheres: beyond classical mod­
els, p. 441, Crivellari L., Hubeny I., & Hummer D.G.,
Kluwer, Dordrecht, The Netherlands
Iglesias C.A., Rogers F.J., Wilson B.G.: 1992, ApJ 397, 717
Saio H., Jeffery C.S.: 1988, ApJ 328, 714
Saio H.: 1993, MNRAS 260, 465
Saio H.: 1995, MNRAS 277, 1393
This article was processed by the author using Springer­Verlag
L A T E X A&A style file L­AA version 3.