An Observational Study of Post-Asymptotic Giant Branch Stars

Transcription

An Observational Study of Post-Asymptotic Giant
Branch Stars
A thesis submitted for the degree of
Doctor of Philosophy
by
Timur ŞAHİN, M. Sc.
Armagh Observatory
Armagh, Northern Ireland
&
Faculty of Science and Agriculture
Department of Pure and Applied Physics
The Queen’s University of Belfast
Belfast, Northern Ireland
January 2008
To my family,
Abstract
In this thesis, we present an LTE model atmosphere analyses of a group of early B-type postasymptotic giant branch (pAGB) stars. With initial masses ≤ 9M⊙, post-AGB stars form an
important group of evolved stars and provide a unique opportunity to study stellar evolution
almost on a human time-scale. Post-AGB stars have spectral types ranging from K to B and
luminosities between 103 and 104 L⊙ . These objects ended their asymptotic giant branch (AGB)
evolution phase with a period of strong mass loss (10−7 − 10−4 M⊙ yr−1 ) and have been evolving
from cooler to hotter temperatures at almost constant luminosity on a timescale of ∼ 104 yr.
B-type pAGB stars span a wide range in effective temperature (10 000 − 30 000K). Their expected surface gravities (log g ) and effective temperatures ( Teff ) coincide with those of B stars
evolving from the main sequence. Therefore systematic observational analyses are required to
to distinguish these two groups. Furthermore, post-AGB stars may be divided into four distinct
groups based on their chemical composition. In this thesis, groups I and II represent post-AGB
stars which are very metal deficient with C/O≈1 and metal poor with C/O<1, when compared
with the Sun, respectively. The question is whether hot pAGB stars belong to either of these
four groups. Three further objectives included:
1. to discover whether post-AGB star have helium-normal or helium-rich photospheres.
2. the detection and measurement of s-process element abundances (e.g. Sr, Y, Ba, Hf).
3. to determine whether they show any anomaly in phosphorus abundance such as that seen
in the extreme helium stars (EHes).
High-resolution échelle spectra of several post-AGB stars were obtained at the AAT in 1999
and 2005 in order to study chemical composition, rotation velocities and other fundamental
properties. Echelle spectra present many difficulties for data reduction, including the problems
i
ii
Abstract
of order rectification and merging. To address these problems we developed an échelle spectrum
reduction package, known as TIGER. These spectra were analyzed using model atmospheres
and synthetic spectra computed with the Armagh LTE stellar atmospheres software. The semiautomated spectral fitting package SFIT was used to measure the stellar surface parameters and
composition.
The results show that Teff of the programme stars are in the range 15 000 – 25 000 K and
log g are in the range 2.5 –3.0. In addition to being metal-poor stars, they show mostly C/O<1.
Several of our programme stars, namely HD119608, LSS4331, LSS5112, and LB3116 confirm
this. The majority of hot post-AGB stars can be identified with the group II, metal-poor and
C-deficient post-AGB stars. The model atmosphere parameters, LTE element abundances and
estimated distance obtained here support the idea that programme stars are in true post-AGB
stars.
We detected helium enrichment in the post-AGB stars Hen3-1428 and LSS4331. We did not
detect any evidence of s-process elements, primarily because of the high Teff of our targets. Our
results do not show overabundance in phosphorus for any hot pAGB stars. Since we used the
same atomic data and methods, we conclude that the enhancement of phosphorus previously
found in some EHe stars is real.
We studied stellar wind signatures for the post-AGB star LSIV-12 111. Emission line equivalent
widths for Balmer lines show changes between two different epochs. Hen3-1428 and LSIV-12
111 show blue shifted absorption lines. A stellar wind is clearly present in both stars.
We compared variability of a group of post-AGB and a group EHe stars using archival photometry. We did not detect variability in EHe stars. We detected variability in five post-AGB
stars. Large variations in HR4049, HD213985, and HD52961 appear to be related to the binary
period.
Acknowledgments
I would like to thank my supervisor Dr. C. Simon Jeffery, whose guidance made my thesis
work possible, for providing me with the opportunity to complete my PhD thesis at the Armagh
Observatory and for introducing me to the world of post-AGB stars. I am very grateful for his
patience.
My special thanks to Prof. Mark Bailey for his support, encouragement and as a director, with
the help of computer and administrative staff, for providing such a nice place to work these past
three years.
I wish to express my gratitude the Northern Ireland Department of Culture Arts and Leisure
(DCAL) for providing financial support as a research studentship by the Armagh Observatory,
through a grant.
I would like to convey my deep regards to Dr. Apostolis Christou and Dr. Igor Savanov for
proofreading the manuscript partially and providing me their valuable suggestions. I would
also like to thank Prof. C. Andrew Cameron for sharing his deep knowledge and performing
some tests on STARLINK package ECHOMOP.
I would like to thank my wife, Petek, for her endless care, love, constant support and encouragement in all my professional endeavors. Especially, the time Petek spared for me from our little
daughter Lara Begüm was just precious and thanks to my daughter Lara Begüm for bringing
happiness to our lives.
...and my dear mom, dad: the following paragraphs are for you and of course will be in Turkish.
iii
iv
Acknowledgments
Sevgili Annem ve Babam,
Bunca yıl vermiş olduğunuz karşılıksız destek, sevgi ve emeğin karşılığını elbetteki ödemem
mümkün değil. Fakat en azından yüzünüzde ufak bir tebessüm oluşturacak şu cümleleri yürekten
yazdığımı bilmenizi istiyorum.
‘Sizin dolaylı olarak sarfetmiş olduğunuz emeğinizin bir ürünü olan bu tez sizlere adanmıştır
ve bilmenizi isterim ki sizin yardımınız ve desteğiniz olmaksızın bunu başarmam mümkün olamazdı. İyi ki varsınız.’
Table of Contents
Abstract
i
Acknowledgements
iii
Table of Contents
v
List of Tables
vii
List of Figures
ix
1 Introduction
1.1 AGB Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1.1 Evolution from Main-Sequence to Asymptotic Giant Branch
1.1.2 Thermal pulses and third dredge-up . . . . . . . . . . . . .
1.1.3 Hot-bottom burning . . . . . . . . . . . . . . . . . . . . .
1.1.4 S-process nucleosynthesis . . . . . . . . . . . . . . . . . .
1.2 Post-AGB Evolution . . . . . . . . . . . . . . . . . . . . . . . . .
1.2.1 Detection and Spectral Classification . . . . . . . . . . . .
1.2.2 Galactic distribution . . . . . . . . . . . . . . . . . . . . .
1.2.3 Previous abundances measurements, s-process abundances .
1.3 Related objects . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.1 Luminous High Latitude (B) stars . . . . . . . . . . . . . .
1.3.2 RV Tau stars . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.3 RCrB stars . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.4 Extreme Helium Stars . . . . . . . . . . . . . . . . . . . .
1.4 Overview of the thesis . . . . . . . . . . . . . . . . . . . . . . . .
1.4.1 Statement of the problem . . . . . . . . . . . . . . . . . . .
1.4.2 Outline of the thesis . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1
1
1
5
6
7
8
9
10
10
13
13
14
15
15
16
16
17
2 Observations and Data Reduction
2.1 Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2 Data reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.1 Reduction package: ECHOMOP . . . . . . . . . . . . . . . . . . . . .
19
19
21
22
v
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
vi
TABLE OF CONTENTS
2.3
2.4
2.5
2.2.2 Reducing the data: Problems and solutions
Échelle Reduction Package in IDL: TIGER . . . .
2.3.1 TIGER: Spectral order fitting . . . . . . . .
2.3.2 TIGER: Interpolation of blaze profiles . . .
2.3.3 TIGER: Pseudo-continuum normalization .
Evaluation of the normalization . . . . . . . . . . .
Summary . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3 Method of Analysis
3.1 Assumptions . . . . . . . . . . . . . . . . . . . . . .
3.2 Model atmosphere codes: LTE codes . . . . . . . . . .
3.3 Atomic data for LTE model calculations – LTE LINES
3.4 Determination of model atmosphere parameters . . . .
3.5 Line Identification . . . . . . . . . . . . . . . . . . . .
3.6 Radial Velocity . . . . . . . . . . . . . . . . . . . . .
3.7 Equivalent Widths: . . . . . . . . . . . . . . . . . . .
3.8 Microturbulent Velocity . . . . . . . . . . . . . . . . .
3.9 Colour Temperatures . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
22
25
27
27
28
32
32
.
.
.
.
.
.
.
.
.
35
36
37
39
40
40
41
41
42
43
4 High-resolution optical spectroscopy of the B-type abundance standard HR1765
4.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2 Model atmospheres analysis and results for HR1765 . . . . . . . . . . . . . . .
4.2.1 Determining Teff and log g : Renormalization . . . . . . . . . . . . .
4.2.2 Radial velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.3 Metallicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.4 Microturbulent velocity . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.5 Rotational velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.6 Abundances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3 Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.1 Comparison of abundances . . . . . . . . . . . . . . . . . . . . . . . .
4.3.2 Testing Teff and log g : ionization equilibrium and
Balmer line fitting . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.3 Testing cosmic argon abundance . . . . . . . . . . . . . . . . . . . . .
4.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
49
49
51
51
51
51
55
57
57
62
62
5 High resolution spectroscopy of the rapid rotating hot post-AGB star: HD119608
5.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2 Model atmosphere analysis and results for HD119608 . . . . . . . . . . . . . .
5.2.1 Metallicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2.2 Microturbulent velocity . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2.3 Rotational velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . .
67
67
68
72
72
73
63
64
65
TABLE OF CONTENTS
5.3
vii
5.2.4 Radial velocity . . . . . . . . . . . . . . . . . . . . . .
5.2.5 Testing model atmosphere parameters . . . . . . . . . .
5.2.6 Testing Teff : ionization equilibrium . . . . . . . . . . .
5.2.7 Testing log g : Helium line fitting . . . . . . . . . . . .
5.2.8 Testing helium abundance . . . . . . . . . . . . . . . .
5.2.9 Notes on individual lines . . . . . . . . . . . . . . . . .
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3.1 Discrepancy between photometric and spectroscopic Teff
5.3.2 Evolutionary status of the object . . . . . . . . . . . . .
5.3.3 The radial velocity of the star: rapid rotation and binarity
.
.
.
.
.
.
.
.
.
.
6 High resolution spectroscopy of two helium-rich hot post-AGB stars
6.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.1.1 Hen 3-1428 . . . . . . . . . . . . . . . . . . . . . . . . .
6.1.2 LSS4331 . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2 Analysis of Hen3–1428 . . . . . . . . . . . . . . . . . . . . . . .
6.2.1 Testing model atmosphere parameters . . . . . . . . . . .
6.2.2 Notes on individual lines . . . . . . . . . . . . . . . . . .
6.3 Analysis of LSS 4331 . . . . . . . . . . . . . . . . . . . . . . . .
6.3.1 Radial Velocity . . . . . . . . . . . . . . . . . . . . . . .
6.3.2 Rotational Velocity . . . . . . . . . . . . . . . . . . . . .
6.3.3 Notes on individual lines . . . . . . . . . . . . . . . . . .
6.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.4.1 Abundances . . . . . . . . . . . . . . . . . . . . . . . . .
6.4.2 Teff derived from photometry . . . . . . . . . . . . . . .
6.4.3 Evolutionary status of the objects . . . . . . . . . . . . .
7 High resolution spectroscopy of the hot post-AGB stars:
LSIV-04 01, LSIV-12 111, LSS5112, and LB3116
7.1 Background . . . . . . . . . . . . . . . . . . . . . .
7.1.1 LSIV-04 01 . . . . . . . . . . . . . . . . . .
7.1.2 LSIV-12 111 . . . . . . . . . . . . . . . . .
7.1.3 LSS 5112 . . . . . . . . . . . . . . . . . . .
7.1.4 LB 3116 . . . . . . . . . . . . . . . . . . .
7.2 Analysis of LSIV-04 01 . . . . . . . . . . . . . . . .
7.2.1 Metallicity . . . . . . . . . . . . . . . . . .
7.2.2 Microturbulent velocity . . . . . . . . . . . .
7.2.3 Testing model atmosphere parameters . . . .
7.2.4 Chemical composition . . . . . . . . . . . .
7.3 Analysis of LSIV-12 111 . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
74
74
74
76
76
77
82
82
84
84
.
.
.
.
.
.
.
.
.
.
.
.
.
.
87
87
87
88
90
90
91
95
96
96
96
97
97
97
101
.
.
.
.
.
.
.
.
.
.
.
103
103
103
104
106
107
109
109
109
110
110
114
viii
TABLE OF CONTENTS
7.4
7.5
7.6
7.3.1 Notes on individual lines . . . . . . .
Analysis of LSS 5112 (IRAS 18379-1707) . .
7.4.1 Metallicity . . . . . . . . . . . . . .
Analysis of LB3116 . . . . . . . . . . . . . .
7.5.1 Metallicity . . . . . . . . . . . . . .
7.5.3 Testing model atmosphere parameters
Discussion . . . . . . . . . . . . . . . . . . .
7.6.1 Abundances . . . . . . . . . . . . . .
7.6.2 Teff derived from photometry . . . . .
7.6.3 Evolutionary status of the objects . .
.
.
.
.
.
.
.
.
.
.
.
.
8 Conclusions And Future Work
8.1 Teff – log g diagram . . . . . . . . . . . . . .
8.2 Helium Abundance . . . . . . . . . . . . . . .
8.3 Phosphorus Abundance . . . . . . . . . . . . .
8.4 Relative Abundances . . . . . . . . . . . . . .
8.5 Temperature Sequence for the programme stars
8.6 Stellar Wind Signatures . . . . . . . . . . . . .
8.7 Future Directions . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
115
119
119
119
123
123
124
124
124
124
128
128
.
.
.
.
.
.
.
131
132
134
134
136
137
137
144
Bibliography
146
A Log files for observations
157
B Photometry
B.1 Variability and Evolution in Various Classes of Post-AGB Stars
B.1.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . .
B.1.2 Introduction . . . . . . . . . . . . . . . . . . . . . . .
B.2 Hipparcos Light Curves . . . . . . . . . . . . . . . . . . . . .
B.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . .
161
161
161
161
162
167
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
C Broadband photometry in the literature for the programme stars
169
D SPECTRAL ATLAS of HR1765
173
E SPECTRAL ATLAS of HD119608
179
F SPECTRAL ATLAS of LSIV-12 111
185
G Some comments on capabilities of ECHOMOP
195
List of Tables
2.1
Basic data for programme stars-I. . . . . . . . . . . . . . . . . . . . . . . . . .
20
2.2
The change in the shape of blaze profiles for the post-AGB source HD119608
can be seen in a0 coefficients. . . . . . . . . . . . . . . . . . . . . . . . . . . .
27
3.1
Photometric temperatures for some of the program stars. The rms scatter in
Teff was given as 3 percent for [u-b] method of NAP93 and 9 percent for GUL89. 46
4.1
Photospheric chemical abundances of HR1765 given as logn, normalized to
logΣµn= 12.15. It is compared to solar abundances (Grevesse & Sauval 1998).
Fundamental parameters for HR1765 with B, V, J, H, K magnitudes were also
presented. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
50
4.2
NII and OII lines used for micro-turbulent velocity determination. . . . . . . .
58
4.3
Equivalent Widths and LTE abundances for HR 1765. The abundances are calculated for vt =5, 6, and 9 km/s. . . . . . . . . . . . . . . . . . . . . . . . . . .
59
60
61
Equivalent Widths of OII lines in the spectrum of HR 1765 are compared to
those measured by Cunha & Lambert (1992) (CL). . . . . . . . . . . . . . . .
62
Measured equivalent widths of HeI lines in the spectrum of HR 1765 are compared to those measured by Leone & Lanzafame (LL), 1998. Please note that
4471 and 4921 are not included in our atomic list used to calculate LTE abundances. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
62
Photospheric line abundances for HR 1765 is compared to γPeg (Munn et al.
2004). Number of the lines used in the analysis are included only for HR1765. .
63
Photospheric line abundances for HR1765 is compared to Hambley et al. (1997)
(HAMB97). They agreed within error limits. . . . . . . . . . . . . . . . . . .
63
4.10 Atmospheric parameters of HR1765. . . . . . . . . . . . . . . . . . . . . . . .
65
4.4
4.5
4.6
4.7
4.8
4.9
ix
x
LIST OF TABLES
5.1
Atmospheric parameters and photospheric chemical abundances of HD119608
given as logn, normalized to logΣµn= 12.15, and compared to solar abundances
(Grevesse& Sauval 1998). Fundamental parameters are also presented. . . . . .
72
5.2
Line identification for HD119608. . . . . . . . . . . . . . . . . . . . . . . . .
79
5.3
80
5.4
81
5.5
82
6.1
Atmospheric and fundamental parameters for IRAS17311-4924 in the literature. Photospheric chemical abundances of IRAS17311-4924 given as logn,
normalized to logΣµn= 12.15. It is compared to solar abundances (Grevesse &
Sauval 1998). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
89
Atmospheric and fundemental parameters for IRAS17381-1616. Johnson BJ
and VJ magnitudes were obtained from Tycho magnitudes of the star. Photospheric chemical abundances of IRAS17381-1616 given as logn, normalized to
logΣµn= 12.15. It is compared to solar abundances (Grevesse & Sauval 1998). .
90
6.2
7.1
Atmospheric and fundamental parameters of LSIV-04 01. Photospheric chemical abundances of the star are given as logn, normalized to logΣµn= 12.15.
These are compared to solar abundances (Grevesse & Sauval 1998). Broadband photometry in the literature is also included. . . . . . . . . . . . . . . . . 104
7.2
Atmospheric and fundamental parameters of LSIV-12 111. Photospheric chemical abundances of the star are given as logn, normalized to logΣµn= 12.15.
These are compared to solar abundances (Grevesse & Sauval 1998). Broadband photometry in the literature is also included. . . . . . . . . . . . . . . . . 106
7.3
Atmospheric and fundamental parameters of LSS5112. Photospheric chemical
abundances of the star are given as logn, normalized to logΣµn= 12.15. These
are compared to solar abundances (Grevesse & Sauval 1998). Broadband photometry in the literature is also included. . . . . . . . . . . . . . . . . . . . . . 108
7.4
Atmospheric and fundamental parameters of LB3116. Photospheric chemical
abundances of LB3116 given as logn, normalized to logΣµn = 12.15. These are
compared to solar abundances (Grevesse & Sauval 1998). Broadband photometry in the literature is also included. . . . . . . . . . . . . . . . . . . . . . . . 109
7.5
Line identification for LSIV-04 01. SiII line at 3853.66 Å is weak. HeI at
4026.36 and FeII at 4508.28 and 4522.63 Å are too weak. . . . . . . . . . . . . 115
8.1
Logarithm of abundances relative to the standard star HR1765. Abundances of
HR1765 (this work) and the Sun (Grevesse & Sauval 1998) are also shown in
their usual form. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
8.2
Comparison of CNO abundances and metallities for the programme stars and
other groups of post-AGB stars in the literature. . . . . . . . . . . . . . . . . . 138
LIST OF TABLES
xi
8.3
Measurement results for emission and absorption line profiles for both AAT and
ESO runs. ESO spectrum was kindly provided by Dufton at QUB. . . . . . . . 143
A.1
A.3
A.5
A.7
A.9
A.11
AAT log file for the 19990728 run.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
157
158
158
158
159
159
B.1 Statistics of Hipparcos photometry. Sources for spectral type are given. . . . . 163
C.1
C.2
C.3
C.4
C.5
C.6
C.7
Broadband photometry for HD119608 in th literature. . . . .
Broadband photometry for LSIV-04 01 in th literature. . . .
Broadband photometry for IRAS17381-1616 in th literature.
Broadband photometry for IRAS17311-4924 in th literature.
Broadband photometry for LSIV-12 111 in th literature. . . .
Broadband photometry for LSS 5112 in th literature. . . . .
Broadband photometry for LB3116 in th literature. . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
170
170
170
171
171
172
172
List of Figures
1.1
L - Teff diagram for a 2M⊙ evolution track from main-sequence to white dwarf
phase. Blue track shows a born-again evolution caused by a very late thermal
pulse of the same mass while the numbers besides to each evolutionary phase
show logarithm of the approximate duration in units of 10 3 years (Herwig 2005).
3
Evolutionary tracks for theoretical model stars of low (1M⊙ ), intermediate (5M⊙ ),
and high mass (25M⊙ ) from Iben (1985). Bold sections define the locations
where major core nuclear burning phases occur. . . . . . . . . . . . . . . . . .
5
Two consecutive thermal pulses and illustration of third dredge-up process by
Falk Herwig. Convective regions are shown in green (Herwig 2005). . . . . . .
7
2.1
Flat-field image obtained with UCLES in 2005 AAT run. . . . . . . . . . . . .
25
2.2
Balance frame produced from flat-field image (see text). . . . . . . . . . . . .
25
2.3
Diagram depicting the parameters used in calculations of pseudo-continuum.
The x-axis shows wavelength (λ0,i < λ1,i ) while y-axis presents raw counts. . . .
29
The example plot is for a post-AGB star, HD119608. Interpolation result for
order 23 is given in red. Order 22 and 24 are used for interpolation. . . . . . . .
30
Pseudo-continuum-I: selection window. Overlapping sections with adjacent orders are shown in red. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
30
Pseudo-continuum-II: order 23 with overlapping regions of two consecutive orders 22 and 24. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
31
Pseudo-continuum-III: continuum selection and normalised result of order 23
for HD119608. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
31
Testing normalization. ESO/CASPEC (in black) and AAT/UCLES(in red) spectra are compared. H γ and Hδ show emission in the core. The forbidden line of
[SII] at 4068 Å , OII lines around 4350 Å and CII lines (multiplet no.28) at
4319 and 4322 Å are also shown. . . . . . . . . . . . . . . . . . . . . . . . .
33
1.2
1.3
2.4
2.5
2.6
2.7
2.8
xiii
xiv
LIST OF FIGURES
3.1
Diagram is to illlustrate steps in the analysis of broad-band spectrometry and
high resolution optical spectrum. Given either a high-resolution optical spectrum or low-resolution spectrophotometry covering ultraviolet and visual wavelengths, those programs allow us to drive various physical quantities. The
outputs are either Teff , log g, vsini, vt and chemical composition from a highresolution spectrum, or Teff , θ, E B−V from spectrophotometry. Programs, inputs, and outputs are presented as boxes, ellipses and oval boxes respectively
(Jeffery et al. 2001b). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
38
HR1765 (in black) and best model fit (in red). Both model and observed spectra
are moved to 0 velocity reference frame - part I . . . . . . . . . . . . . . . . .
52
are moved to 0 velocity reference frame - part II . . . . . . . . . . . . . . . . .
53
are moved to 0 velocity reference frame - part III . . . . . . . . . . . . . . . .
54
Determination of microturbulent velocity (v t ) via NII lines for HR 1765. For
vt =6 km/s the slope of the regression line vanishes as expected if the choice of
vt is correct. Solid line is the least-squares fit to the data points. . . . . . . . . .
55
Determination of microturbulent velocity (v t ) via OII lines for HR 1765. For
vt =6 km/s the slope of the regression line vanishes as expected if the choice of
vt is correct. Solid line is the least-squares fit to the data points. . . . . . . . . .
56
Determination of microturbulent velocity ( vt ) with a different method. vt ≈ 9
km/s is min. of parabola. The effect of [Fe/H] on calculated vt is also represented for [Fe/H]=-1 and [Fe/H]=1 respectively. . . . . . . . . . . . . . . . . .
57
Determination of micro-turbulent velocity (v t ) with a different method. vt ≈6
km/s is min. of parabola. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
58
Comparison of equivalent widths from Cunha & Lambert (CL) (1992) and Table
4.8 of this section. There seems to be no systematic differences between two
data sets; a formal least-square solution gives a gradient of 0.93 and a zero point
difference of 0.001 A. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
63
Comparison of equivalent widths from Leone & Lanzafame (LL) (1998) and
Table 4.9 of this section. There seems to be no systematic differences between
two data sets; a formal least-square solution gives a gradient of 1.04 and a zero
point difference of 0.002 A. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
63
4.10 Model atmosphere parameters for HR1765. Error bars for this work have been
also plotted. Filled circles show published model atmosphere parameters for
the star. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
64
4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
LIST OF FIGURES
5.1
5.2
5.3
xv
Two model atmosphere fits of Teff =23 000 K and Teff =26 000 K are presented. The fit in red presents the model atmosphere with Teff =26 000 K while
Teff=23 000 K model fit is shown in green. Both model and observed spectra
are moved to zero velocity reference frame - part I . . . . . . . . . . . . . . . .
69
are moved to zero velocity reference frame - part II . . . . . . . . . . . . . . .
70
are moved to zero velocity reference frame - part III . . . . . . . . . . . . . . .
71
5.4
Determination of microturbulent velocity ( vt ) via SiIII multiplet no.2 for HD119608.
For vt = 5 km/s the slope of the regression line vanishes as expected if the choice
of vt is correct. The dashed line is the least-squares fit to the data points. The
equations obtained from the least-squares fit are also given. . . . . . . . . . . . 73
5.5
χ2 method for checking model atmosphere parameters of HD119608. The red
curve represents the result for χ2 method when whole spectrum used while the
results for each half when the spectrum divided into two halves are shown in
black. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
75
5.6
Ionization equilibrium for SiIII multiplet no. 2. . . . . . . . . . . . . . . . . .
75
5.7
Ionization equilibrium for SiIII multiplet no. 9. . . . . . . . . . . . . . . . . .
75
5.8
Ionization equilibrium for SiIII/IV. Both SiIII triplet belong to multiplet no. 2
and 9. were included. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
76
Hen3–1428 (in black) and best model fit (in red). Model and observed spectra
were corrected for radial velocity. . . . . . . . . . . . . . . . . . . . . . . . .
92
93
94
6.1
6.2
6.3
6.4
2
χ method for checking model atmosphere parameters of IRAS17311-4924.
Teff in x-axis presented in 1000 K units. The red represents the result when
the whole spectrum is used while the results when the spectrum divided into
two halves are shown in black. . . . . . . . . . . . . . . . . . . . . . . . . . .
95
6.5
Ionization equilibrium for Si II/III, Si II/IV, and Si III/IV. . . . . . . . . . . . .
96
6.6
LSS 4331 (in black) and best model fit (in red). Model and observed spectra
were corrected for radial velocity. Observed spectrum was binned to a pixel
size of 0.15 Å . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
98
LSS 4331 and best model fit. . . . . . . . . . . . . . . . . . . . . . . . . . . .
99
6.7
xvi
LIST OF FIGURES
6.8
LSS 4331 and best model fit. . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
7.1
LSIV-04 01 (in black) and best model fit (in red). . . . . . . . . . . . . . . . . 111
7.2
continued - part II. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
7.3
continued - part III. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
7.4
χ2 method for checking atmospher parameters of LSIV-04 01. . . . . . . . . . 114
7.5
LSIV-12 111 (in black) and best model fit (in red). Both model and observed
spectra are moved to zero velocity reference frame - part I. . . . . . . . . . . . 116
7.6
7.7
continued - part III. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
7.8
LS5112 and best model fit. Both model and observed spectra are moved to zero
velocity reference frame. Observed spectrum was binned to a pixel size of 0.15
Å- part I. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
7.9
7.10 continued - part III. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
7.11 χ2 method for checking model atmosphere parameters of LB3116. . . . . . . . 123
7.12 LB3116 (in black) and best model fit (in red). Both model and observed spectra
are moved to 0 velocity reference frame - part I. . . . . . . . . . . . . . . . . . 125
7.13 continuing part-II. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
7.14 continuing part-III. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
8.1
The log g – Teff diagram for post-AGB stars showing the positions of the
programme stars with R CrB (Asplund 1997; Lambert et al. 1997) and EHe
(Drilling et al. 1998; Pandey et al. 2001) stars. Hydrogen main sequence, horizontal branch, AGB, and the classical Eddington limit for radiative stability
are included. Evolutionary tracks of post-AGB (Schönberner 1983 for postAGB masses of 0.546 and 0.565M⊙ ; Blöcker 1995 for 0.605, 0.625, 0.696,
0.836, and 0.940M⊙; Gingold (1976) for a 0.519M⊙ star) are also shown. 1.
LSS5112; 2.IRAS17311-4924; 3.HD119608; 4.IRAS17381-1616; 5.LSIV-12
111; 6.LB3116; 7.LSIV-04 01. . . . . . . . . . . . . . . . . . . . . . . . . . . 133
8.2
Temperature-sequence for the programme stars. The Teff increases from top
to bottom. Each spectrum is vertically offset by 0.5 continuum units. The observed spectra of LSS5112, IRAS17311-4924, IRAS17381-1616, LB3116, and
LSIV-04 01 were binned to pixel sizes of 0.25, 0.15, 0.25, 0.15, and 0.15 Å
respectively - I. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
8.3
respectively - II. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
LIST OF FIGURES
8.4
8.5
xvii
respectively - III. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
respectively - IV. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
B.1 Hipparcos periodograms for post-AGB variables and HD119608. Abscissa is
log10 frequency (day−1 ) and ordinate is amplitude. . . . . . . . . . . . . . . .
B.2 Hipparcos window functions for post-AGB variables and HD119608. . . . . .
B.3 Phased light curve for post-AGB star HR4049 at a period of 472 d. The crosses
show the phased data binned with a width of 0.1 in phase. The average magnitude and scatter in magnitude were calculated for each phase bin. . . . . . . . .
B.4 Phased light curve for post-AGB star HD213985 at a period of 270 d . The
crosses show the phased data binned with a width of 0.1 in phase. The average magnitude and scatter in magnitude were calculated for each phase bin. . .
B.5 Phased light curve for post-AGB star HD52961 at a period of 69 d . The crosses
show the phased data binned with a width of 0.1 in phase. The average magnitude and scatter in magnitude were calculated for each phase bin. . . . . . . . .
.
.
.
.
.
165
165
166
166
166
D.1
D.2
D.3
D.4
D.5
SPECTRAL ATLAS for HR1765-I. .
SPECTRAL ATLAS for HR1765-II. .
SPECTRAL ATLAS for HR1765-III.
SPECTRAL ATLAS for HR1765-IV. .
SPECTRAL ATLAS for HR1765-V. .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
174
175
176
177
178
E.1
E.2
E.3
E.4
SPECTRAL ATLAS for HD119608-I. . .
SPECTRAL ATLAS for HD119608-II. .
SPECTRAL ATLAS for HD119608-IIII. .
SPECTRAL ATLAS for HD119608-IV. .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
180
181
182
183
F.1
F.2
F.3
F.4
F.5
F.6
F.7
SPECTRAL ATLAS for LSIV-12 111-I. .
SPECTRAL ATLAS for LSIV-12 111-II.
SPECTRAL ATLAS for LSIV-12 111-III.
SPECTRAL ATLAS for LSIV-12 111-IV.
SPECTRAL ATLAS for LSIV-12 111-V. .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
186
187
188
189
190
191
192
xviii
F.8
LIST OF FIGURES
SPECTRAL ATLAS for LSIV-12 111-V. . . . . . . . . . . . . . . . . . . . . . 193
Chapter 1
Introduction
1.1 AGB Evolution
1.1.1 Evolution from Main-Sequence to Asymptotic Giant Branch
During the main sequence (MS) stage, stars burn hydrogen in their cores via nuclear fusion.
They spend most of their lifetimes on the main-sequence. Their initial masses affect the time
spent on this stage directly. Hydrogen exhaustion in the core is followed by formation of a helium core where the star can not produce enough energy to stop contraction. As the central part
of the star contracts, hydrogen-burning starts in a shell surrounding the core via the CNO process. Energy production is concentrated in the regions of highest hydrogen content and highest
temperature (ǫ ∝ XH XCNO T16 ; T ∼ 2×107 K). This high temperature causes high pressure just
outside of the core causing the hydrogen envelope to expand hence the star leaves the MS and
evolves onto the sub-giant branch. This expansion is accompanied by a drop in the effective
temperature of the star. Energy production in the hydrogen burning shell causes the star to
expand to larger radii (> 105 R⊙ ) and evolve along the red-giant branch (RGB), where it is characterised by a convective envelope penetrating into the deep layers. This penetration may be
so deep that the convective envelope reaches material already processed by hydrogen-burning,
enriched in helium(4 He), nitrogen (14 N) and carbon (13 C) while depleted in carbon (12 C) and
1
2
Chapter 1. Introduction
oxygen. This brought to the surface in a process known as “dredge-up” or first “dredge-up” if
it occurs on the RGB. Most of the mass is in a very small fraction of the total radius and the
outer layers can be easily lost via a stellar wind. The star now has a hydrogen-burning shell
surrounding a helium core which heats as it contracts. This contraction lasts until the core becomes degenerate. At the tip of the RGB, electrons in the core are completely degenerate. That
means gravitational contraction will be balanced by degeneracy pressure which is independent
of temperature and pressure. When the central temperature has reached nearly 108 K, helium
ignites and heats up the core. As energy production and temperatures increase, degeneracy is
lifted, temperature increases, pressure increases and the core expands to find a new equilibrium
configuration. At this stage, the hydrogen-burning shell has also expanded and has lower temperature and density causing the shell to produce less energy than before. The star becomes a
Horizontal-Branch (HB) star.
When helium in the core is used up, energy production shifts to helium burning in a shell. The
outer layers of the star expand. The star evolves towards the early-asymptotic giant branch
(AGB) and becomes a red-giant for the second time. The expansion of the outer layer extinguishes the hydrogen-burning shell in the intermediate-mass stars, which causes the convective
envelope to move inward for a second time. This second dredge-up brings the hydrogen burning products, mainly helium and nitrogen, to the surface but it does not occur in low mass stars
since the hydrogen-burning shell is not extinguished.
Following the helium exhaustion in the shell, the hydrogen-burning shell takes over until enough
helium has been produced. When the helium shell is sufficiently massive, it will re-ignite
and new helium-burning will control the evolution. The helium-burning shell is known to be
thermally unstable and it undergoes thermal pulses recurrently. This is the thermally-pulsing
AGB (TP-AGB) phase (Figure 1.1), characterised by two nuclear burning shells surrounding
an electron-degenerate carbon-oxygen core and a deep convective envelope. According to
Schwarzschild & Härm (1967) and Weigert (1966) thermal pulses may occur in both low- and
intermediate mass stars and huge amounts of energy are released. This allows envelope convection to penetrate to the inter-shell region and mix carbon and oxygen enriched material to
the surface (third dredge-up). It is important to note that mixing events occurring during these
1.1. AGB Evolution
3
Figure 1.1: L - Teff diagram for a 2M⊙ evolution track from main-sequence to white dwarf
phase. Blue track shows a born-again evolution caused by a very late thermal pulse of the same
mass while the numbers besides to each evolutionary phase show logarithm of the approximate
duration in units of 103 years (Herwig 2005).
thermal pulses are responsible for the transport of carbon and s-process elements to the surface
(Iben & Renzini, 1983). It is also equally important to note that observed abundances of postAGB stars are mostly determined during this stage of mixing episodes driven by consecutive
thermal pulses.
The AGB represents the final phase of nuclear-driven evolution for low (M ≤ 2.5 M⊙ ) and
intermediate-mass (2.5 M⊙ ≤ M ≤ 8 M⊙ ) stars. Almost 99% of the stars in the galaxy are in this
mass range and will, most probably, become AGB stars (Aller 1994; Iben 1995). During this
stage, stars experience severe mass-loss which is responsible for removal of most of the stellar
envelope. This stage is also known as the beginning of proto-planetary nebulae (PPN) stage and
Figure 1.2 shows possible evolutionary scenarios for model stars of low, intermediate and high
mass.
4
The AGB is relatively well understood in general terms. However detailed comparisons between
theory and observation show that there is much to learn about thermal pulses, third dredge-up,
s-process nucleosynthesis and the formation of carbon and oxygen rich AGB stars. Challenges
include the observed hydrogen-deficiency and/or overabundance of helium burning products
in some groups of post-AGB stars, as well as explaining the observed s-process element enhancements/deficiencies in these stars. The discrepancies in observed abundances between hot
post-AGB stars and their cooler counterparts is also one of those problems waiting to be solved.
These all must be explained on the basis of AGB stellar evolution models.
Furthermore, it is useful to draw a picture of some drawbacks regarding observations of AGB
stars in general. It is difficult to study AGB stars for several reasons:
• They show rather complex molecular spectra.
• Atomic and molecular data for AGB stars are incomplete.
• Non-LTE and spherical geometry are important processes in the stellar atmosphere.
A similar list can be compiled for PNe as well. PNe also do not provide a reliable method in
determination of element abundances for several reasons:
• Most of the ejected material containing refractory elements such as Ca, Si, Fe etc. is
locked up in grains, hence they are not detectable via emission line spectroscopy.
• Some elements can exist in multiple ionization stages and hence an unknown fraction
may be left unobserved.
In this sense, post-AGB stars provide a unique opportunity as testbeds of AGB nucleosynthesis.
The determination of the model atmosphere parameters effective temperature Teff and surface
gravity g allows one to estimate the mass of the post-AGB star by comparison with theoretical
stellar evolutionary tracks.
This thesis is concerned with early B-type hot post-AGB stars and their surface abundances.
The problems to be answered in the framework of the thesis presented in detail in Section 1.4
are summarised as follows:
1.1. AGB Evolution
5
Figure 1.2: Evolutionary tracks for theoretical model stars of low (1M⊙ ), intermediate (5M⊙ ),
and high mass (25M⊙ ) from Iben (1985). Bold sections define the locations where major core
nuclear burning phases occur.
• Do hot post-AGB stars have hydrogen-normal and/or helium-rich surfaces?
• Do they show s-process elements in their spectra? If so, what are these s-process abundances?
• What is the phosphorus abundance in those hot post-AGBs?
1.1.2 Thermal pulses and third dredge-up
During AGB evolution, the helium-burning shell is thermally unstable and causes energy bursts
called thermal pulses or flashes repeating apparently every 105 years leading to the development
of a convective zone in the shell. The phase where hydrogen-shell burning dominates is known
as the “inter-pulse phase” which is interrupted by the helium-shell instabilities. The thermal
pulse drives a convection zone between the helium and hydrogen burning shells. The fusion
6
products of the helium-burning shell including
12
C and some
16
O are mixed throughout this
region.
Once a thermal pulse has occurred, a huge amount of energy is released. This episode may lead
to some structural re-adjustments in the AGB star causing a series of mixing events and nuclearburning episodes in the star. Transportation of nuclear-burning products to the stellar surface
via third-stage dredge-up occurs as a two-stage process (Lattanzio & Wood 2004). Conditions
in the helium-burning shell force it to expand pushing all the material above it, including the
hydrogen-burning shell, outward. Therefore, the hydrogen shell expands and cools so much
that it is almost extinguished. The flash-driven convection creates a carbon pocket, below the
hydrogen burning shell, containing about 25% carbon, 2% percent oxygen, and the rest mainly
helium (Iben 1976, 1982; Lattanzio 1987; Boothroyd and Sackmann 1988c). Helium in the
shell burns via 13 C(α, n)16 O and generates neutrons. S-process elements start to be produced by
slow-neutron capturing onto iron group elements until the next thermal pulse occurs. The next
mixing episode may bring these s-process elements to the surface (3rd dredge-up). This process
may repeat itself after each thermal pulse depending on the initial mass and composition of the
star. During a thermal pulse, one should only see a drop in luminosity; the energy generated by
helium burning is used up in driving the expansion (Iben 1975).
1.1.3 Hot-bottom burning
For massive AGB stars (Mstar >4 M⊙ ), the convectively unstable envelope may penetrate into the
top of the hydrogen shell where the temperature can reach as high as 108 K . Penetration of the
envelope into the hydrogen shell results in nuclear burning at the bottom of convective envelope.
At this stage, the CN-cycle, processing 12 C into 13 C and 14 N, is activated. This process is known
as hot-bottom burning (HBB) and may destroy the newly-formed carbon and prevent the star
from becoming a carbon star. During HBB, the star may also experience mass loss which is
of major importance to the dredge-up process. Both hydrogen burning at the bottom of the
envelope and mass loss from the stellar surface will reduce the mass of the envelope. AGB
evolution ends when the envelope mass becomes too small to support nuclear burning and the
star evolves to the white dwarf stage.
1.1. AGB Evolution
7
Figure 1.3: Two consecutive thermal pulses and illustration of third dredge-up process by Falk
Herwig. Convective regions are shown in green (Herwig 2005).
One important consequence of HBB is production of 7 Li (Cameron-Fowler 1971) and
26
Al by
the Mg-Al cycle (Mowlavi & Meynet 2000).
1.1.4 S-process nucleosynthesis
For production of elements beyond iron, a neutron capturing process is required. The typical
time scales for the s(slow) and r(rapid) neutron capture processes are thought to be 10 4 year
and 104 seconds respectively.
The helium rich intershell region in AGB stars provides a suitable environment for the production of s-process elements.
8
Neutrons are produced via mainly two reactions (Equation 1.1 and 1.2).
14
N + α→γ + 18 F
18
F + + β→18 O
(1.1)
18
22
O + α→γ + Ne
22
25
Ne + α→ Mg + n
12
C + p→13 N + γ
13
N + + β→13C + ν
13
C + α→16 O + n
Similarly, additional reactions involving 16 O, 17 O, 18 O, 21 Ne,
(1.2)
25
Mg or 26 Mg have occasionally
been considered as possible neutron producing reactions.
16
18
O + 16 O→31 S i + n
O + α→21 Ne + n
17
21
O + α→20 Ne + n
Ne + α→24 Mg + n
(1.3)
18
21
O + α→ Ne + n
25
28
Mg + α→ S i + n
26
Mg + α→29 S i + n
It has been recently suggested that the dominant neutron source in AGB stars is
13
C(α, n)16 O
(Lattanzio & Lugaro 2005).
1.2 Post-AGB Evolution
The material expelled away (Iben & Renzini 1983) from the system during the AGB stage
forms a slowly expanding circumstellar shell. When the mass of the hydrogen-rich envelope
drops to ≈10−3 M⊙ , it starts to contract at a constant luminosity. This contraction phase causes
an increase in effective temperature and at the same time leads to a radiatively driven wind
which can compress the circumstellar shell and may result in ionization of the circumstellar
material. In other words, the star may become hot enough to ionize its circumstellar material,
1.2. Post-AGB Evolution
9
observable as a planetary nebula (PN), in a few hundred years (Oudmaijer 1996). The star at
this stage is known as a post-AGB star. These have luminosity classes ranging from I (supergiant) to III (giant) with spectral types from B to K. Typical post-AGB stars are expected to
have luminosities around 103 –104 L⊙ (Blöcker 1995). Masses of these objects are in the range
0.6M⊙ – 1 M⊙ . The star leaves the AGB with Teff < 5000 K . When it reaches Teff > 30 000, it
may ionize the remnant nebula.
1.2.1 Detection and Spectral Classification
Although some post-AGB stars were known before the launch of IRAS, their number started
to grow after the IRAS mission in 1983 (IRAS Explanatory Supplement, 1985). Most postAGB candidates were detected and identified on the basis of their far-infrared (IRAS) colors
(Parthasarathy & Pottasch 1986; Pottash & Parthasarathy 1988; Parthasarathy 1993a; Kwok
1993) or discovered in the IRAS Point Source Catalog (IRPSC), generally via cross-correlation
with sources in optical catalogs.
Systematic searches for detection of stars with infra-red excess based on these kinds of crosscorrelation methods revealed many other new post-AGB stars. While IRAS based selections
in the literature like e.g. Volk & Kwok (1989), van der Veen et al. (1989) and Slijkhuis
(1992), seem to indicate a large fraction of G-type supergiants, another cross-correlation between Smithsonian Astronomical Observatory (SAO) Catalog and IRPSC (1985) by Oudmaijer
et al. (1992) yielded F-type supergiants. Several IRAS sources classified on the basis of their
low resolution spectra by Hrivnak et al. (1989) were F and G type and a few were hotter. So
there appears to be a clear trend that IRAS based selections yield cool post-AGB stars. These
observations are in contrast with the evolutionary scenarios/models in the literature for a star
evolving through the AGB – PN sequence. According to these models, post-AGB stars as transition region objects spend 70% of their lifetime as a B-type star. This can be understood from
the 5 M⊙ evolutionary track of Schönberner (1981, 1983) calculated during this transition. The
result is be qualitatively the same for other core masses (Oudmaijer et al. 1992).
Two-peaked structure of the spectral energy distribution (SED) is a characteristic feature of a
10
post-AGB star. This double-peaked SED indicates cessation of mass loss as well as a detached
circumstellar envelope around the star. Another method to identify these objects is focused on
optically bright objects with an IR excess because of surrounding circumstellar gas and dust
(Hrivnak et al. 1989; Trams et al. 1991; Oudmaijer et al. 1992).
1.2.2 Galactic distribution
More recent studies on these objects show that spectral types B and F dominate. Many of the
known B type post-AGB stars were discovered in population studies of B-stars in the galactic
halo. Moehler and Heber (1998), in a survey of hot sub-dwarf candidates from the PalomarGreen Survey (Green et al. 1986), found that some of the stars very far from the galactic center
could be post-AGB stars. In another survey of B-type stars in the halo of our Galaxy (Keenan
1992), stars previously classified as young population-I objects were identified as hot post-AGB
stars by Mooney et al. (2002) on the basis of their surface gravities, effective temperatures, and
chemical compositions (Conlon et al. 1991, 1993; Conlon, Dufton & Keenan 1994; McCausland et al. 1992).
1.2.3 Previous abundances measurements, s-process abundances
First of all, it should be noted that it is difficult to identify that a star is clearly in post-AGB evolutionary phase. Interpretation of their observed abundances is mostly difficult as they present
an extremely diverse chemical composition pattern. There are currently four distinct subgroups
of post-AGB stars identified in the literature:
1. Group I post-AGB stars typically show C/O≈1 with very low iron abundance. Even
though, post-AGB stars show effective temperatures and surface gravities which are typical for normal B-type population-I super-giants, the observed abundances for many of
them differ from those of normal population-I supergiants 1 but are similar to those observed in the interstellar medium (Venn & Lambert 1990). This similarity could be understood in the following way: most of these stars are known to have been surrounded
1
Such Galactic halo post-AGB examples in the literature are PG 1704+222, HD341617 and LSIV-0401
(Mooney et al. 2002).
1.2. Post-AGB Evolution
11
by a circumstellar dust shell and as they evolve through this transition phase, their circumstellar envelope may be dissipated. It is probable that this dissipation may also be
accompanied by a gas and dust fractionation or, more specifically, by dust grain formation. So the idea is that this dust grain formation may have caused observed deficiencies
and extraordinary abundance peculiarities by selective removal of the elements condensed
onto these dust grains. Examples with C/O≈1 include HR4049 ([Fe/H]=-5.0, [Ca/H]=6, [C/H]=-0.2, [O/H]=-0.5; Takeda et al. 2002), HD52961 ([Fe/H]=-4.8, [C/H]=-0.4,
[O/H]=-0.6; Van Winckel et al. 1992; Waelkens et al. 1992), BD+39 4926 ([Fe/H]=2.1-3.3,[C/H]=-0.3, [O/H]=-0.1; Kodaira 1973; Lambert et al. 1988; Waelkens et al.
1991a) and HD44179 ([Fe/H]=-3.3, [C/H]=0.0, [O/H]=-0.4; Waelkens et al. 1992). All
are extremely deficient in Fe and other refractory elements. Such observed over/underabundances may also be due to the fact that some of these stars left the AGB before
experiencing a 3rd dredge-up. Observationally, there is a clear indication for carbon deficiency in hot B-type post-AGB stars implying that these stars might have evolved off
the AGB before undergoing the 3rd dredge-up. Examples are HD 341617 (Mooney et al.
2002), Barnard 29 (Conlon et al. 1994), ZNG-1 (Mooney et al. 2004) and ROA 5701
(Thompson 2006).
2. Group II post-AGB stars are O-rich (C/O<1) and slightly metal deficient. They do not
show s-process elements. However, they display a double-peaked SED similar to the
first group as well as a metallicity range similar to those C-rich post-AGB stars. 89 Her,
SAO239853, HD133656, and HD161796 are such examples in this group.
3. In addition, a third group have C/O>1 and s-process elements. HD 56126 (Klochkova
1995), HD187885 (Van Winckel et al. 1995) and IRAS06530-0213 (Reddy et al. 2002)
are the only examples showing s-process enhancement with the high C/O ratios expected
for a post-AGB star which has experienced the 3rd dredge-up. These objects (C/O>1)
show two peaks in their SED and their metallicities ([Fe/H]) range between -0.3 and -1.0.
4. A fourth group is hot post-AGB stars. These stars are also the subject of this thesis. These
metal-poor stars show remarkably high deficiency in C and overabundance in O (Conlon
12
et al. 1991; McCausland et al. 1992; Moehler & Heber 1998; Mooney et al. 2002).
They also show less depletion in N, O, Mg and Si. LSIV-04 01 (see Chapter 7 for details)
is a typical example of this group evolving through a low mass post-AGB evolutionary
track and showing over-abundance of He. Since the HGL(B)S were believed to be young
massive Pop I stars previously, the depletion in C, N, O, Mg, and Si and as well as He
being over-abundant are incompatible with their being young massive Pop I stars. LSIV12 1112 and LSII+34 263 are examples of this group. They both have infrared excess
similar to post-AGB stars.
Many of the group members do not show IR excess.
The question is which of the groups of cool post-AGB stars do these stars belong to ?
The disagreement between the chemical compositions found for cool and hot post-AGB stars
and the structure of their circumstellar envelopes are some of the challenges awaiting to be
elucidated by means of spectroscopy. Mooney et al. (2001) report that although the post-AGB
sample included in their work were hotter analogues of cooler post-AGB objects, they were
showing notably large carbon deficiencies. Today, the nature of the carbon rich envelope and its
changing structure from one proto-planetary star to another still remains a mystery.
Both dredge-up and s-process nucleosynthesis triggered by the thermal pulses cause major
structural re-adjustment on the AGB in terms of both chemical composition and abundance
enhancement. However, in general, [C/Fe], [N/Fe], [O/Fe] and [S/Fe] abundance ratios and
particularly depletion of some elements such as Al, Ca, Mg, Si, Ti and Fe (refractory elements)
relative to the non-refractories or lack of depletion in CNO and/or S are not well understood
on the basis of these processes. It may be that this is only due to inappropriate LTE stellar
atmosphere models used in spectroscopic analysis of these objects (Pandey et al. 2004).
Generally, most post-AGB spectra contain weak and blended lines. In addition, the number of
lines that can be measured for analysis is relatively small. As a result abundance measurements
are not always of sufficiently high accuracy. It should also be noted that there are post-AGB
2
3
The star has recently started to ionze its circumstellar material and become PNe.
LSII+34 26 was shown to have a detached circumstellar shell by Parthasarathy (1993).
1.3. Related objects
13
stars which do not show s-process enhancement in their spectra at all (i.e. HR6144) for the
probable reasons explained above.
1.3 Related objects
Although the post-AGB stage is relatively well established, there are still many details to be
resolved before it can be fully understood from both an observational and a theoretical point
of view. While several other groups of stars occupy the same parts of the HR diagram, how
or whether these are linked or fit into the theoretical evolution of post-AGB stars is one of the
major fields of post-AGB research requiring attention.
1.3.1 Luminous High Latitude (B) stars
Bidelman (1951) first drew attention to the existence of peculiar A and F-type super-giants (e.g.
89 Her, HR6144, and HD161796) at high Galactic latitudes. After the launch of IRAS, the
circumstellar characteristics of these objects were revealed. IRAS data also pointed out the fact
that these high galactic latitude supergiants (HGLS) had a considerable infrared excess, possibly
due to circumstellar dust ( Parthasarathy & Pottash 1986; Trams 1991). This led them to the
idea that these objects might be in a kind of transition region, believed to last several thousand
years, from the AGB to the PNe phase.
From the evolutionary point of view, there are many possible scenarios to explain HGLS stars.
1. HGLS are luminous young massive stars which formed in situ in the galactic halo.
2. They reached their large distance as runaway massive stars within their lifetime (Conlon
et al. 1990).
3. They are less luminous low-mass members of the old stellar population such as in late
stages of stellar evolution or evolved off AGB (Trams et al. 1991).
Parthasarathy & Pottasch (1986) concluded that the HGLS were in fact low-mass post-AGB
stars. Bond & Pottasch (1987) showed that there are several features distinguishing the old
14
population post-AGB supergiant stars from young massive stars. First of all, young massive
super-giants are not expected to have high galactic latitudes. If a star is a member of the old
stellar population then it should show a metal deficiency. Furthermore, if a star shows CNO
and s-process element enhancement, this is probably due to the fact that it has undergone a
dredge-up process during its AGB stage.
Some of the HGLS have far-infrared colours similar to PNe suggestive of them being post-AGB
candidates (Manchado et al. 1989; Garcia-Lario et al. 1990).
Although the HGLS may show near-population-I abundances (Hambly et al. 1996, and references therein) they do not show normal abundance patterns for all elements, in contrast to young
B-type stars. Hence a detailed abundance analysis of high resolution spectra of these objects
(post-AGBs) will differentiate between the two types of stars (Hambly et al. 1996). There are
such examples of high galactic latitude stars in the literature classified as post-AGB objects on
the basis of their spectral energy distributions and spectroscopic analyses. These objects are
IRAS18095+2704 (Hrivnak et al. 1988), HR 4912, 7671, 6144, 161796 (Luck et al. 1990),
HR4049, HD52961 (Waelkens et al. 1991), HD56126, HD101584, HD161796 (Parthasarathy
& Pottasch 1986; Parthasarathy et al. 1992), and HD105262 (Reddy et al. 1996).
1.3.2 RV Tau stars
RV Tau stars occupy the high-luminosity end of the Pop II Cepheids. These high latitude and
metal deficient stars are luminous pulsating stars of F, G, and K spectral types. They show
deep minima in their light curves. Their cycle-to-cycle variability and amplitude variations are
among those properties of RV Tau stars. Although their evolutionary status is not precisely
known, there is a general agreement that they are post-AGB objects. Since they show strong
infra-red excess (Jura 1986) which can probably be attributed to circumstellar dust shell, Jura
(1986) placed these stars in the post-AGB evolutionary stage. However, Giridhar et al. (2000)
and Maas (2003) showed that they do not share chemical characteristics of post-AGB stars and
do not show an enhancement in C or s-process elements. Instead, a depletion pattern was seen in
refractory elements. It is probable that this depletion is caused by dust-gas separation occurring
1.3. Related objects
15
at the end of AGB where mass-loss has ceased recently. On the other hand, SX Cen, BZ Sct,
RU Cen, UY CMa, DY Ori, AR Pup, HP Lyr, AC Her, EP Lyr, and AD Aql show abundance
patterns very similar to post-AGBs.
1.3.3 RCrB stars
RCrB stars are hydrogen-deficient supergiants and about 30 are known in the Galaxy. One peculiar feature of these stars is that their atmospheres are hydrogen depleted by factors of typically
104 (Bidelman 1953; Warner 1967). They also show sudden deep drops in luminosity by up to
eight magnitudes, caused by dust obscuration in the line of sight (e.g. Clayton 1996). The origin
of this extreme hydrogen deficiency is still unclear. According to Schönberner (1977), RCrB
stars are AGB stars which lost their hydrogen envelopes at an earlier stage and are evolving
into hot extreme helium stars. Renzini (1979, 1981) suggested that RCrB stars are born-again
post-AGB stars. They experience a final helium-shell flash which results in ingestion of residual hydrogen envelope. However, the time-scales associated with this loop mechanism are very
rapid. For example, required lifetimes for a 1.2 and 0.6 M ⊙ star are given as 20 and 2000 years
respectively by Renzini (1979). Another mechanism suggested to explain observed hydrogendeficiency in RCrB stars involves the merger of two white dwarfs in a binary system (Webbink
1984). Saio & Jeffery (2002) concluded that this was the most likely model for the origin of
RCrB are extreme helium stars.
1.3.4 Extreme Helium Stars
EHes are carbon-rich B and A spectral type pulsating super-giant stars with extremely low
surface abundances of hydrogen (Jeffery, Drilling, & Heber 1987; Jeffery 1996). Their temperatures lie in the range 8 000 ≤ Teff (K)≤ 32 000 K and they are known to have unusual chemical
compositions (large helium abundances ≥ 99 percent and weak or non-existent hydrogen lines;
see Woolf & Jeffery 2002). It was proposed that these objects are post-AGB stars on a white
dwarf cooling track and they undergo a ’late thermal pulse (LTP)’, (Iben et al. 1983; Iben &
Tutukov 1984; Webbink 1984). The late thermal pulse model has been intensively studied in
the literature (Iben & MacDonald 1995; Blöcker & Schönberner 1997; Herwig et al. 1999).
16
It seems that the LTP scenario might explain the s-process element production in some RCrB
stars (Bond, Luck & Newman 1979; Lambert & Rao 1994) but it fails for EHes for two reasons:
(i) a degenerate CO core with a helium-burning shell cannot account for the low-luminosity
EHes (Saio & Jeffery 2000) and (ii) surface abundances of carbon and oxygen are an order of
magnitude higher than observed in EHes according to current LTP models (Herwig et al. 1999).
Pandey et al. (2001) have emphatically rejected the LTP model for the origin of EHes on the
basis of this argument. They also added that there is a fundamental problem in classifying EHes
and RCB stars as post-AGB stars in terms of LTP as observed and predicted compositions of
EHes and RCB stars are different. Firstly, a helium shell for an AGB star is rich in oxygen and
carbon and secondly, the final He-shell flash does not have a great effect on this mixture in the
shell because the mass of hydrogen rich envelope is very small. Thirdly, post-AGB stars should
also be rich in 13 C, but 13 C abundance is relatively low for RCB stars and absence of s-process
enrichment also rules out an association with thermally pulsing AGB stars4 .
1.4 Overview of the thesis
1.4.1 Statement of the problem
We have performed high resolution spectroscopy of hot early B-type post-AGB stars with effective temperatures in the range ≈ 10 000 – 30 000 K. Their optical spectra are rich in absorption
lines, suitable for abundance measurement of many elements including He, C, N, O, Mg, Al, Si,
S, Fe. These data have been analyzed using LTE, line-blanketed model atmospheres. We compared observed abundances with the expected abundances of first time normal post-AGB stars of
the same spectral type and luminosity class. Extreme-helium stars (EHes) showing abundances
anomalies were also used in this comparison, as we aimed to answer the question “whether hot
post-AGB stars show similar anomalies to EHes regarding overabundance of phosphorus and/or
neon ?”
The evolutionary tracks of post-AGB stars cross the hydrogen-burning main sequence, so that
4
V1920 Cyg and HD124448 are such an exceptions that most of the neutron-capture elements (Sr, Y, Zr, Ba
and lanthanides) are undetectable in their optical spectra.
1.4. Overview of the thesis
17
main-sequence and (hot) post-AGB stars cannot be differentiated on the basis of their stellar atmosphere parameters alone. These parameters must be corroborated by quantitative abundance
analysis in order for a ”typical post-AGB abundances pattern” to be established. These results
will also help related theoretical evolutionary tracks to be examined.
For some of the examined program stars, extended lists of line identifications and chemical
element abundances are presented for the first-time. Some previous abundance measurements
are questioned by the spectroscopic analysis results presented in the following chapters.
We will mainly try to answer whether there is a connection between genuine post-AGB stars
and related objects. Phosphorus is observed as overabundant in EHes and its abundance will
be investigated for the hot post-AGB stars in our sample. Hopefully, the work presented in
this thesis will lead to a clearer picture of chemical characteristics for post-AGB stars with
the following questions in mind: If phosphorus is found to be depleted, what happened to the
missing phosphorus ? If it is not the case, then why is phosphorus overabundant ? Is there any
connection between overabundance of phosphorus and observed hydrogen deficiency in some
post-AGB sources ?
1.4.2 Outline of the thesis
In this chapter, we have drawn a general picture of post-AGB stellar evolution. We have identified four distinct groups of post-AGB stars. We have introduced a number of important questions to be addressed in this thesis.
• Chapter 2: Here, high resolution observations of the program stars with some basic
information are presented. The difficulty in normalization of echelle spectra and proposed
solution with an IDL package namely TIGER is also presented.
• Chapter 3: In this Chapter, properties of the Armagh Stellar Atmosphere codes with
the assumptions made are briefly summarized. In a separate study, the literature was
searched for photometric data for the programme stars. Results of this photometric analysis which can be used in initializing model atmosphere parameters ( Teff and/or log g )
18
are also presented for some of the post-AGB stars for the first time. The discrepancies
between spectroscopic and photometric temperatures are discussed.
• Chapter 4: In Chapter 4, an LTE model atmosphere analysis of the standard B star
HR1765 is presented with an extended line list. A spectral atlas of the star is included
in Appendix-D. Since the star will be included in a differential abundance analysis of the
programme stars, we have presented the abundances derived from both spectrum synthesis and fine analysis of the spectrum.
• Chapter 5: In Chapter 5, a model atmosphere analysis of our prototype rapid rotating
post-AGB star HD119608 is presented. A study to explain the discrepancy between photometric and spectroscopic temperatures of the star is presented. The model atmosphere
parameters are interpreted with the help of available scenarios for observed abundances
in the spectrum. A spectral atlas of the star is included in Appendix-E.
• Chapter 6: In Chapter 6, we present high resolution spectroscopy of two helium-rich
post-AGB stars, IRAS17311-4921 and IRAS17381-1616.
• Chapter 7: In Chapter 7, LTE model analyses of post-AGB star LSIV-04 01, LSIV-12
111 being a PN, LSS5112 and LB3116 are presented. A spectral stlas of LSIV-12 111 is
included in Appendix-F.
• Chapter 8: In Chapter 8, we summarise the work performed in this thesis and present
results for the current analysis. The main emphasis is on phosphorus and helium abundance. The CNO abundances for the programme stars are compared to those of other
group of post-AGB stars in the literature e.g. EHe.
We also performed a photometric study which is presented in Appendix-B. In this self-contained
study, we searched SIMBAD database for B-type post-AGB and EHes. For those stars with
Hipparcos photometry, light curves were analyzed for possible variability. We aimed to trace
characteristics of variability for different groups of post-AGB stars.
Broadband photometric data for each programme stars were collected and compiled in table
format. This is presented in Appendix-C.
Chapter 2
Observations and Data Reduction
2.1 Observations
Journal of observations and sources: High resolution optical spectra were obtained in two
different runs on the nights 1999 July 28, and 29 and 2005 August 26, 27, 28, and 29 with the
3.9-m Anglo-Australian Telescope (AAT). The University College London échelle spectrograph
(UCLES; Walker & Diego 1985) was employed with the 31.6-grooves-mm−1 grating. For the
1999 and 2005 runs, a thinned Tektronix CCD (1024X1024; 24 µm pixels) and an EEV2 CCD
(2048X4096; 13.5 µm pixels) were used respectively. For the 1999 run, two separate exposures
were made for each target, with central wavelengths of 4307 Å, giving a spectral coverage from
3847 to 5010 Å, and 6043 Å(4915 to 8150 Å). For the 2005 run, one exposure was made for
each target, with a central wavelength of 4340 Å, giving a spectral coverage from 3848 to 5200
Å. Stellar exposures were bracketed with thorium-argon (Th-Ar) exposures in both runs for
wavelength calibration. For both runs, the mean wavelength residuals for the Th-Ar exposures
were less than 0.02 Å and the spectral resolution was 0.05 Å pixel−1 for the 2005 run. Flat
fields were made using quartz continuum lamps. A list of the stars observed together with some
basic data is given in Table 2.1.
19
20
Chapter 2. Observations and Data Reduction
Table 2.1: Basic data for programme stars-I.
Star
α(2000)
1999 July 28-29 Post-AGBs
CD-53 5736
03 30 50.70
HD119608
13 44 31.31
LSIV-04 01
16 56 27.7
LSS5112
18 40 48.5
LSIV-12 111
20 01 49.84
PHL1580
21 30 25.25
2005 August 26-29 Post-AGBs
IRAS 17311-4924
17 35 02.50
IRAS 17381-1616
17 41 00.03
IRAS 19157-0247
19 18 22.71
LB 3116
19 18 49.19
2005 August 26-29 Comparisons
HR 1765
05 21 45.75
HR 18862
05 35 01.01
HR 7773
20 20 39.82
δ(2000)
Spectral Type
-52 47 55.2
-17 56 13.2
-04 47 20
-17 04 36
-12 41 17.6
-19 22 34.5
V MAG
B MAG
F6 10.15 10.47
B1Ib2 7.512a
7.43
B7Ib3 12.10
11.2
B1IIIpe 11.934
11.8
5
B1Ibe 11.33 11.3B
12.33
12.2
-49 26 26.4
-16 18 12.5
-02 42 10.8
-64 35 31.1
B1IIe
B1Ibe
B9Ib7
O
10.746
12.55
11.14
12.2
11.5
12.49
-00 22 56.9
-06 00 33.4
-12 45 32.7
B2V2
B1V
B9IV
4.722b
5.702c
4.75
4.54
5.44
4.71
2 : (Houk+, 1975); 2a :Fernie (1983), 2b :Crawford et al. (1971) ;2c :Walker (1969)
respectively; 3 : (Hardorp+ 1959-1965);4 : Drilling (1991); 5 : Drilling (1975); 6 : Klare
& Neckel (1977);7 : Reed & Niemczak (2000).
Échelle spectrum format: One of the goals of using an échelle spectrograph is to obtain a
high-resolution spectrum using an entrance slit and the higher orders of an diffraction grating.
Échelle gratings are a special kind of diffraction grating which are used in very high orders (e.g.
m≈ 200). They have generally lower ruling (groove) densities (e.g. 79 lines mm−1 ) compared
to those gratings used in low orders (e.g. 1200 lines mm −1 ). To achieve high efficiency on these
high orders, the gratings are blazed at a high angle. Due to the higher order numbers, light
coming from different orders overlaps in the beam dispersed by the échelle grating according to
the grating equation 2.1.
In order to separate these orders, a cross-disperser is used to separate the light perpendicular to
the dispersion direction of the échelle. In other words, a long spectrum is sliced up and stacked
into segments which is well suited to the CCDs’ format. These segments each have their own
blaze profiles and they overlap. Thus, a given wavelength at the long wavelength end of one
order is repeated in the short wavelength end of an adjacent order.
2.2. Data reduction
21
d(sinα + sinβ) = mλ,
(2.1)
where α is angle of incidence, β is angle of diffraction (blaze angle), d is distance between
adjacent grooves, m is order number, and λ is wavelength of incident light beam.
Another advantage of using an échelle spectrograph is its wide wavelength coverage. For a
given wavelength range, there are several échelle orders. As the spectrum of each order has a
blaze profile similar to a sinc function, the wavelength region which can be efficiently measured
is limited and it can be significantly different from one order to another. The wavelength range
in one spectrum which does not overlap the similar range in the neighboring order spectra is
called the free spectral range, and can be represented by means of wavelength and order number.
2.2 Data reduction
A CCD image contains additional signal (noise) caused by the electronics of the device. It
is mainly due to the metal-oxide semi-conductor (MOS) structure of the CCDs which include
many of these light sensitive MOS cells. The number of electrons collected in each MOS cell
as a result of the photo-electric affect is determined by the gain of the CCD. In order to transfer
these collected electrons to an external medium, they are converted to digital counts via an
analog to digital converter. This conversion contributes to those counts by a certain amount
called bias which should be removed from science and calibration frames.
As a first step, master-bias, master-dark and master-flat images were obtained by combining
bias, dark and flat frames taken at the beginning of each night. Subtraction of master-bias
frame from science, dark, flat and arc images was followed by the removal of thermal electrons,
which are present due to the temperature sensitive structure of MOS capacitors in the chip, from
science and flat images by help of the master-dark frame. For the 2005 run, the dark counts were
not higher than the bias counts so they were not used.
The next step was to correct CCD images for pixel-to-pixel variations, cosmetics, fringing, and
22
vignetting as degraded sensitivity near the edges of the CCD. Due to the manufacturing, all
MOS cells in a CCD do not have the same sensitivity over a certain wavelength range. So
that they need to be corrected for this pixel-to-pixel variations. This stage was performed by
division of science frames by a master-flat. This was followed by converting two-dimensional
images to one-dimensional wavelength calibrated ones via Th-Ar exposures. This last process
was performed in STARLINK package ECHOMOP.
Fringing on the CCD can be seen as diffraction patterns (for the details see section 2.2.2).
2.2.1 Reduction package: ECHOMOP
Reduction of the raw two-dimensional CCD images to one dimensional spectra was carried out
using tasks in the ECHOMOP environment (version 3.3-5, Mills & Webb 1994) which was
primarily written for the reduction of UCLES data. There are over 30 tasks, menu options and
help files available in the program. Almost 160 different parameters can be modified via a user
interface and these choices can be crucial if one is reducing faint sources in ECHOMOP (see
Appendix-G for comments on the menu options).
2.2.2 Reducing the data: Problems and solutions
Pre-processing of CCD images has been done in the usual way including trimming of the data
sections, bias subtraction, flat-fielding and cosmic-ray removal. The seeing was measured to
be in the range 1.0 - 2.0 arc-sec for both 1999 and 2005 data sets. Both 1999 and 2005 AAT
UCLES frames had to be rotated and reversed in dispersion direction to be processed with
ECHOMOP.
The individual orders were located and traced by fitting polynomials to their center of the gravities in ECHOMOP. The scatter was less than 0.5 pixel. After slit determination, a combined
and bias subtracted master flat-field was constructed. Dekker size was chosen as wide as possible in order to cover each individual order and its sky in the flat-field image. Sky and object
channels were also modeled by making use of the master flat image. In a typical data frame,
for most of the orders sky subtraction was done in an automated manner. For some it was not
2.2. Data reduction
23
possible because of cosmic ray contamination in the sky. This problem was solved by performing cosmic-ray removal prior to ECHOMOP. Otherwise sky determination was performed
manually1 in a way that these regions were not included in sky subtraction. It was followed by
extraction of spectra.
Locating arc lines in the arc spectra was the next step and for 2005 run it was done in an
automated manner via ECHOMOP ECH IDWAVE function by setting AUTO ID=YES. The
routine for automated wavelength calibration uses an arc line list, generally provided by an arclamp exposure and expects the candidate features to be identified as position/intensity pairs. It
then attempts to match using a reference database and generates polynomials to fit the observed
position of features. The wavelengths and positions of the Th-Ar comparison-lines in the arc
spectra were fitted with polynomials of degree 3 giving wavelength scales with root mean square
residuals of less then 0.02 Å.
For 1999 data, the automated wavelength calibration in ECHOMOP got stuck many times. Our
approach was to change the orientation of the frames since ECHOMOP needs échelle orders to
be running horizontally. If it is suspected that wavelength scale is decreasing from left to right
then TUNE REVCHK parameters should be set to YES to allow the algorithm to check for a
reversed arc. Otherwise FIGARO IREVX was used to flip the data frames in order to reverse
the wavelength scale.
The last step was the extraction of the orders. One can make extraction of spectra in two
different schemes. One is the optimal extraction method. By this method, it is aimed to obtain
best S/N ratio. The second is a simple extraction. In this method, extraction is performed in
an unweighted manner. All program stars have been extracted optimally (Horne 1986). Before
that, we did some experimental work with the 1999 data (see below). We did not perform blaze
removal and merging process in ECHOMOP due to unrevealed bugs in the code (A.C. Cameron,
private communication). Instead we had to develop our own IDL package to perform the rest of
the steps.
1
This process is pretty much the same in IRAF
24
Experimental work on blaze removal – Testing ECHOMOP: We performed some experimental work on blaze removal and spectral extraction algorithms2. All orders were extracted
simply and compared with optimally extracted spectra to check the reliability of the extraction
algorithm and to see which method should be preferred. For most of our program stars having
noisy spectra, the results of the optimal weighting scheme were satisfactory.
We also checked whether the extraction algorithm used for object spectra was being applied to
flat-field images successfully since we could use these images for correcting the blaze function.
For this, we first reduced a master flat-field image optimally, and then divided it by an optimally
reduced object spectrum. Blaze removal was not successful. We found some trend which could
be represented by first order polynomial fits. One common problem with échelle spectra was
ripples, saw-tooth like features mostly caused by incorrect dekker size determination.
The next step was to make the extraction non-optimally. We found that there was no distinguishable difference between optimally and simply reduced flat-field images so that we attempted to
divide optimally extracted object spectra with the simply reduced flat-field image. Neither of
these two methods worked well for blaze removal.
In the reduction of the 2005 AAT data, a balance frame (‘balance factor’ frame) was used3
instead of a raw flat image. The flat-field balance factors are the per-pixel values which are
multiplied into the raw data to perform photometric correction required to correct for differing
pixel-to-pixel sensitivity of the CCD. ECHOMOP fits low-degree polynomials; along traces,
and along the image columns. Each flat-field pixel in an order is then used to calculate a ‘balance
factor’. This is a number close to 1 which represents the factor by which a given pixel exceeds
its expected value (predicted by polynomial). It should be noted that this technique requires that
the flat-field orders run smoothly (no fringing or cosmetics) both along and across each order.
A master flat-field was then median filtered using a box size of 1 pixel in the spatial direction
and 50 pixels in the dispersion. The raw master flat image was divided by this smoothed flat
image. This gave us a balance frame. In this case the balance factors are simply copied from
the frame supplied and this gave better result. The same balance frame model was also applied
2
3
We were assisted by Prof. A.C. Cameron, University of Saint Andrews.
T  TUNE PREBAL    YES
2.3. Échelle Reduction Package in IDL: TIGER
25
Figure 2.1: Flat-field image obtained with Figure 2.2: Balance frame produced from flatfield image (see text).
UCLES in 2005 AAT run.
to science frames. This balance frame also helped to check intrinsic structures on the surface of
CCD such as fringing and cosmetics. It was also used for pixel to pixel correction on the CCD.
A comparison between a flat-field image and a balance frame can be seen in Fig.2.1 and 2.2. In
Fig.2.2, green circles in lower right and upper right show defected pixels on the CCD surface
and can not be seen in flat-field image. The rectangular box in the bottom right presents fringe
patterns which are traced by yellow parabolas to help eye. The red box in the upper center show
cosmetics and again can not be seen in the flat-field image.
Experimental work on blaze removal – Testing DIPSO: In an another experiment with 1999
AAT data, merging and normalization of the spectra were performed in DIPSO (Howarth et al.
1995). Normalization was done by dividing the merged spectra by a second degree polynomial
with DIPSO routines. The result was again unsatisfactory due to the changing structure of the
instrumental profile through échelle orders (see also Table.2.2).
2.3 Échelle Reduction Package in IDL: TIGER
It is difficult to combine many spectral orders of an échelle spectrum to produce a single continuous spectrum. Generally speaking, this difficulty arises because the blaze function varies
rapidly along each order.
It has not been possible to normalize the spectrum by processing the spectrum as a whole.
26
Although there have been studies of large scale blaze function correction (Katz et al. 1998)
it was not possible for our échelle spectra (see Section 2.2.2). The general approach to this
problem is to use flat images or a spectrum of a similar temperature and similar reddening
metal-poor star with as few lines as possible (e.g. a white dwarf or subdwarf). In our case,
removal of blaze function was performed on an order by order basis since the standard stars
observed by us were not convenient for this aim.
In some cases, a flat-field calibration lamp may define the blaze function but, in practice, it is not
easy. There are some optical components in the flat-field beam which may create some residuals
when applied to astrophysical observations so that using spectra of astrophysical sources instead
of flat-field images may provide a solution to this problem.
We chose to remove the instrumental response from each order individually. For this purpose,
we developed an IDL routine, which we called, TIGER. Third or fourth degree polynomials
gave the best result but one also has to take data in overlapping regions into consideration
since this information also could be important in defining a pseudo-continuum for some échelle
spectra.
. The capabilities of TIGER are as follows:
• Spectral order fitting.
• Interpolation of blaze profiles.
• Normalization of spectral orders via exclusion of unwanted regions from spectral orders.
• Pseudo-continuum normalization.
• Cosmic-ray removal.
• Quick-look fitting and overlapping inspection.
• Cutting unwanted regions from rectified orders interactively prior to merging.
• Merging of rectified spectral orders.
• Quick-look spectrum inspection.
27
• Production of merged and rectified spectrum in spectrum-2(SP2) file format.
2.3.1 TIGER: Spectral order fitting
Order fitting was performed via polynomials interactively. The user is able to exclude unwanted
regions such as absorption / emission lines, spikes etc from spectral orders following the overlapping inspection window (Fig.2.5 and 2.6). The normalized result is shown just after the
selection for inspection purposes (Fig.2.7). This process can be repeated for the best fit to the
spectrum.
In Table 2.2, we present 4th order polynomial fitting coefficients for 8 adjacent orders. It is given
for post-AGB star HD119608 and shows the example fitting database created in IDL Échelle
reduction package TIGER. The p and c abbreviations in the first column stand for normalisation with interactive (c)ontinuum-selection and (p)seudo-continuum normalisation (see below)
respectively. TIGER allows this information to be updated when needed.
T ype
c
p
p
p
p
c
c
p
...
OrderNo
1
2
3
4
5
6
7
8
...
a0
18530.9
19060.7
21698.4
23100.1
23886.0
25879.5
27300.0
28643.6
...
a1
-264.927
-200.206
-298.210
-323.383
-268.612
-302.241
314.245
-304.665
...
a2
-31.3081
-20.0993
-36.9143
-37.2063
-38.8436
-41.8036
43.7560
-43.6033
...
a3
0.280221
0.074270
0.330848
0.379793
0.167061
0.259233
0.255451
0.197593
...
a4
0.0173431
0.0124021
0.0253749
0.0195951
0.0288330
0.0270330
0.0284272
0.0255709
...
Table 2.2: The change in the shape of blaze profiles for the post-AGB source HD119608 can be
seen in a0 coefficients.
2.3.2 TIGER: Interpolation of blaze profiles
For some échelle orders, continuum fitting may not give the best result. This can be due to an
inappropriate dekker determination, or to diffuse light in the spectrograph, a common difficulty
with échelle spectra, or to wide absorption features in the spectrum. Interpolation of the blaze
functions from adjacent orders can be a solution, and TIGER makes this possible (Fig.2.3).
28
When this option is run, users are asked to select orders which will be used in the interpolation.
Let us presume that the orders selected are m and n, that polynomial coefficients for these
orders are given as vectors P m and Pn and the order we want to interpolate is given as i. Then
the coefficients of order i are given as Pi .
1 × (n − i) × Pm + (i − m) × Pn
Pi =
n−m
(2.2)
and the interpolated blaze function for order i as:
bi (λ) =
3
X
k=0
Pk,i × (λ − λ0 )k
(2.3)
In this example, it is calculated as a 3rd degree polynomial and depicted in blue colour in Figure
2.1.
2.3.3 TIGER: Pseudo-continuum normalization
If spectral order fitting via interactive mouse selection or interpolation of successive orders is
not a solution for the spectra, users are advised to use pseudo − continuum normalization option
in TIGER.
When this routine is run, overlapping regions in adjacent ȩchelle orders are determined, and
displayed on the screen. An example plot is given in Fig.2.5. The functions fi (λ), ci (λ), bi (λ)
represent observed flux, identified continuum and estimated blaze function for each order. In
our approach, we require in the overlapping regions that:
ci−1
1,
bi−1
ci
1,
bi
ci+1
1.
bi+1
(2.4)
We also have the additional constraint, within regions of overlap.
fi
fi+1
fi−1
= =
bi−1 bi bi+1
(2.5)
29
Figure 2.3: Diagram depicting the parameters used in calculations of pseudo-continuum. The
x-axis shows wavelength (λ0,i < λ1,i ) while y-axis presents raw counts.
Rewriting for orders i − 1 and i:
ci ′ = fi ×
bi−1
fi−1
(2.6)
defines additional new continuum points ci ′ by making use of data in the overlapping region
between order i − 1 and i. This process can be repeated for consecutive orders in the same way.
It should be noted that the function defined by ci ′ is also a function of blaze profile of the order
i − 1. The next step is to add these new pseudo-continuum points (ci ′ ) to the already available
continuum points- ci -of the order i and to fit a new blaze function. In both techniques namely
pseudo-continuum and interpolation, the wings of broad lines are always an issue since they
carry less weight in the fit and they are also under-illuminated (low S/N ratio). One approach
to this problem was to discard the outer part of wings. This was done via a procedure which is
included in TIGER. Cutting unwanted parts in individual orders can be performed either prior to
applying TIGER for both blue and red part of the wings or within the TIGER as an option which
is successful in only red part of the wings in pre-merging stage and just after the normalization.
30
Figure 2.4: The example plot is for a post-AGB star, HD119608. Interpolation result for order
23 is given in red. Order 22 and 24 are used for interpolation.
Figure 2.5: Pseudo-continuum-I: selection window. Overlapping sections with adjacent orders
are shown in red.
31
Figure 2.6: Pseudo-continuum-II: order 23 with overlapping regions of two consecutive orders
22 and 24.
Figure 2.7: Pseudo-continuum-III: continuum selection and normalised result of order 23 for
HD119608.
32
2.4 Evaluation of the normalization
In order to test the normalization for LSIV-12 111, we compared AAT/UCLES 1999 échelle
spectrum of the star to ESO/CASPEC 1994 data obtained by the QUB Belfast group (Fig.2.8).
This comparison spectrum was analyzed by Ryans et al. (2003) and provided to us by Philip L.
Dufton. The ESO spectra were not merged. We performed the merging in DIPSO.
This was followed by moving the spectra in wavelength to an arbitrary rest frame. First, the ESO
observations were moved to the rest frame of the AAT spectrum. We determined a velocity shift
of 14 km/s between two consecutive ESO spectra. Both spectra seem to match perfectly. The
normalization in all parts of the spectrum agrees very well.
2.5 Summary
The UCLES spectra of all our programme stars were reduced with ECHOMOP, and rectified
with TIGER to provide rectified spectra with a wavelength coverage 3847 – 5010 Å at a mean
resolution R≃48 000. For details of the observations and some examples of the rectified spectra
see Appendix A, C, D, E.
2.5. Summary
33
Figure 2.8: Testing normalization. ESO/CASPEC (in black) and AAT/UCLES(in red) spectra
are compared. Hγ and Hδ show emission in the core. The forbidden line of [SII] at 4068 Å , OII
lines around 4350 Å and CII lines (multiplet no.28) at 4319 and 4322 Å are also shown.
Chapter 3
Method of Analysis
A stellar atmosphere represents the medium between a stellar interior and outer space. The spectroscopic analysis of a star involves determining the structure of the stellar atmosphere and the
transfer of radiation from the stellar interior through the atmosphere. In other words, analysis of
a stellar atmosphere requires the best possible determination of the emergent spectrum. Finding the best match between observed and emergent stellar radiation is required for a satisfactory
analysis.
In order to perform model atmosphere analyses of hot early B-type post-AGB stars, we used
Armagh Stellar Atmosphere Codes which are a compilation of Fortran90 programs specifically
designed for the calculation of B-type spectra with extreme compositions of helium, carbon and
nitrogen. In this chapter, we present assumptions made in calculating model atmospheres and
give a basic description of the programs STERNE, SPECTRUM and SFIT2. This will be followed by a discussion of the methods used for determination of model atmosphere parameters
and chemical abundances for the programme stars.
35
36
Chapter 3. Method of Analysis
3.1 Assumptions
The general assumptions made here for stellar atmospheres for deriving chemical compositions
are plane-parallel geometry, hydrostatic, radiative and local thermodynamic equilibrium.
• Plane-parallel geometry may be assumed when the thickness of the atmosphere is small
compared to the radius of the star.
• Hydrostatic equilibrium may be assumed when the local gravity is exactly balanced by
the pressure(P). This assumption may not hold for post-AGB stars which show strong
stellar winds.
dP(r)
− ρ(r)g = 0
dr
(3.1)
where ρ is density.
• Radiative equilibrium assumes that the total radiation (F) is constant with depth (τ) in the
atmosphere and the only energy source lies below the atmosphere. This also means that
there is no flux of energy coming onto the atmosphere from above.
F=
Z
F λ dλ = σ Teff 4 ;
dF
=0
dτ
(3.2)
where σ is the Stephan-Boltzmann constant
• Local Thermodynamic Equilibrium (LTE) assumes that collisional processes dominates
the radiative ones in the stellar atmosphere, and hence all level populations are given by
the Boltzman equation1.
1
Since asymmetry in helium lines and emission in the Balmer line cores observed in those indicate a moderate
or strong stellar wind, this assumption may not hold for Hen3-1428, LSS4331 and LSIV-12 111. However, Kudritzki (1979) showed that non-LTE effects may be safely ignored for stars of Teff ≤ 35 000 K . It should be noted
that all of the programme stars included in this thesis have effective temperatures of Teff ≤ 35 000 K . Moreover,
Gies & Lambert (1992) and Killian (1994) also found that, for B-type stars, LTE abundances agree to 0.2 dex
with corresponding non-LTE values. On the other hand, in most cases, strictly LTE model atmospheres do not
take scattering into account but it is treated in Armagh Stellar Atmosphere code so the LTE codes used for the
model atmosphere analyses of the programme stars presented in the following chapters should not be considered
as strictly LTE but partially non-LTE. By all means, those results should be confirmed and/or tested by non-LTE
model atmospheres.
3.2. Model atmosphere codes: LTE codes
37
Since geometry has been defined as plane-parallel, physical parameters defining the characteristics of the stellar atmosphere only depend on optical depth which accompanied by only one
physical temperature.
3.2 Model atmosphere codes: LTE codes
We have carried out a line-blanketed LTE analysis for all our program stars to derive photospheric chemical compositions using a grid of pre- and new computed model atmospheres. The
model grids were computed by Armagh Stellar Model Atmosphere Code STERNE 2.6d.
STERNE: The code STERNE calculates the structure of stellar atmosphere for hot stars with
effective temperatures between 10 000< Teff <35 000 K and is optimised to calculate model atmospheres of extreme helium (H-deficient), carbon, and nitrogen compositions. It was originally written by Wolf and Schönberner in Berlin (1974) and again substantially extended by
Heber (Kiel and Bamberg) and Jeffery (St. Andrews and Armagh). Atomic bound-free opacities are from ATLAS and supplemented by additional tables for carbon and nitrogen (Peach
1970). For line opacities, odf’s are used as in ATLAS62 . Line blanketing caused by the metal
lines blocking the flux results in backwarming of the deeper layers of the atmosphere. This
backwarming modifies the atmosphere and is accounted for in STERNE. The radiative transfer equation is solved by Feautrier’s method (Feautrier 1964) with temperature corrected using
the Unsöld-Lucy procedure (Lucy 1964). The result of STERNE is a model atmosphere structure describing the temperature, pressure and emergent flux as a function of optical depth. The
emergent flux distribution is calculated by integrating the source function (S) over optical depth.
Fλ =
Z
∞
S (λ, τ)e−τ dτ.
(3.3)
0
SPECTRUM: High resolution theoretical synthetic spectra were computed using SPECTRUM
(Jeffery et al. 2001) from plane-paralleled line-blanketed STERNE 2.6d produced model atmo2
A newer version using opacity sampling (Behara 2007) was not completely operational at the time this work
was carried out.
38
composition
UV + visual
spectrophotometry
High resolution
spectrograms
STERNE
Model flux grid
Model structure grid
FFIT
SPECTRUM
Teff, θ, Eb−v
High resolution
model grid
v_t
SFIT
Atomic data
Teff, log g, v sin i
LTE_LINES
SFIT_SYNTH
v_t, composition
Figure 3.1: Diagram is to illlustrate steps in the analysis of broad-band spectrometry and high
resolution optical spectrum. Given either a high-resolution optical spectrum or low-resolution
spectrophotometry covering ultraviolet and visual wavelengths, those programs allow us to
drive various physical quantities. The outputs are either Teff , log g, vsini, vt and chemical composition from a high-resolution spectrum, or Teff , θ, E B−V from spectrophotometry. Programs,
inputs, and outputs are presented as boxes, ellipses and oval boxes respectively (Jeffery et al.
2001b).
spheres. SPECTRUM is a spectrum synthesis code which will also compute line profiles and
specific intensities. This program was originally written by Dufton at Queen’s University of
Belfast. The code assuming plane-parallel geometry, hydrostatic, radiative and local thermodynamic equilibrium was extended by Lennon (QUB), Conlon (QUB) and Jeffery. It will read
ATLAS (Kurucz) or STERNE calculated model atmospheres with a line list to compute a normalized synthetic spectrum. The normalized spectrum (sλ ) can be calculated via dividing total
emergent flux ( fλ ) by the continuum flux ( fλc ) and will be calculated as a function of function
of Teff , log g, micro-turbulent velocity( vt ) and chemical composition of species of i(relative
3.3. Atomic data for LTE model calculations – LTE LINES
39
abundance by number; ni ) in the following form.
sλ (T eff , log g, vt , ni , i = 1, ...) =
fλ
fλc
(3.4)
In this computation, hydrogen, helium and metal line profiles were computed in the following
way.
The neutral hydrogen line profiles were computed from Vidal et al. (1973) and Lemke’s (1997)
broadening tables and helium line profiles for HeI 4026, 4388, 4471, and 4922 Å from Barnard
et al. (1969, 1974, 1975) and for HeI 4009 and 4144 Å from Gieske & Griem (1969) and
Dimitrijevic & Sahal-Brechot (1984) while the remaining helium lines were computed as Voigt
profiles. Ionized helium line profiles were computed using broadening tables of Schöning &
Butler (1989). The metal lines in the normalized spectrum were computed as Voigt profiles.
Radiative, collisional, and Doppler broadening as well as micro-turbulent broadening were also
included.
SFIT2: For each program star, a grid of model spectra was compiled with the stellar atmosphere code SFIT 2.7. SFIT is a combination of SFIT SOLVE and SFIT SYNTH routines and
originally designed for both single and binary stellar spectra. Levenburg-Marquardt, Amoeba
and Genetic Algorithms are the parameter optimization methods included in the code.
3.3 Atomic data for LTE model calculations – LTE LINES
In order for model spectra of the programme stars to be calculated, we used a database of atomic
data called LTE LINES3 which is a compilation of atomic data, for blue-visible (λ 4000 – 5000
Å ) absoption lines of the elements, suitable for LTE analysis of early-type stars (Jeffery 1991).
In this dynamic compilation, absorption lines are included with their wavelengths, oscillator
strengths, radiative, collisional and van der Waal’s damping constants, excitation energies of
the lower levels in the transitions, multiplet numbers (from Moore’s revised multiplet tables).
The references for damping coefficients and oscillator strengths are also given.
3
LTE LINES is reachable via http://www.arm.ac.uk/∼csj/lines/lte/html/lines.lte.html
40
3.4 Determination of model atmosphere parameters
The determination of atmospheric parameters using theoretical grids is an iterative process requiring consistent estimates for the effective temperature, gravity, micro-turbulent velocity and
metallicity. SFIT SOLVE was used to fit model spectra to observed spectra. The code uses
a non-linear least-square fitting technique for determination of the parameters. χ2 is used as
a measure of the match between theoretical and observational spectra; its minimization defines the most probable model atmosphere parameters for the star. Temperatures are obtained
simultaneously from relative strengths of both the hydrogen and helium lines in addition to
the excitation and ionization equilibrium of all elements, especially HeI/II and SiII/III/IV, in
the spectrum. This is done simultaneously with obtaining the helium abundance from relative
strengths of hydrogen and helium lines, while the gravity is determined from hydrogen and
helium profiles.
Basically, one has to compare an observed spectrum with several model spectra. If the results
are different from an initial assumed composition then the process is repeated until convergence
is established. It should be noted that all the model atmospheres used in the analysis of the
program stars were calculated using the LTE approximation since they are not hotter than 30 000
K , above which non-LTE effects become important (Napiwotzki 1997). It should also be noted
that these high-resolution normalized model spectra used in preparation of the theoretical grids
were calculated with a micro-turbulent velocity ( vt ) of 5 km/s unless stated otherwise. The
next stage is the measurement of abundances of minor elements with SFIT SYNTH. Once
abundances were estimated, the micro-turbulent velocity can be calculated (see section 3.8). A
final check was made for changing carbon abundance and metallicity.
3.5 Line Identification
Line identification has been done manually by making use of Moore multiplet tables (1945)
and the ILLSS4 catalogue (1993). The standard procedure is to look at a spectrum and determine whether any members of a given multiplet are present with appropriate relative strengths.
4
A revised version of the Identification List of Lines in Stellar Spectra (ILLSS) Catalogue (Coluzzi 1993-1999)
3.6. Radial Velocity
41
Then lines having similar excitation potentials and similar laboratory strengths are searched.
Attention has been paid to blended lines. For some of the stars analyzed in this work, we have
presented their line identifications for the first time and for the others with a better quality.
Spectral atlases for the standard star HR1765, the hot post-AGB star HD119608 and young PN
star LSIV-12 111 are presented in Appendices D, E, F respectively.
3.6 Radial Velocity
After having identified the lines in the spectrum of our program stars, their shifts relative to
laboratory wavelengths were measured. The spectra presented in the following chapters were
shifted to the laboratory reference frame. The wavelength shifts (δλ) correspond to a heliocentric velocity (V) for the programme stars are corrected to heliocentric frame via following
relation.
V =c×
δλ
− V⊙
λ
(3.5)
where V⊙ is the correction due to the motion of the Sun. The first term on right hand side of
the equation 3.5 is the measured velocity shift.
3.7 Equivalent Widths:
Equivalent widths (wλ ) of spectral lines of a certain element depend on the number of the absorbing particles or number density in the stellar atmosphere and therefore on elemental abundances as much as on the excitation potential of lower level of the lines and oscillator strengths.
The main advantage of using wλ for abundance determination is that it is essentially insensitive
to rotational and instrumental broadening.
log
5040
πe2 Nr HH
wλ
= log
+
logA
+
logg
f
λ
−
χ − logκν
λ
mc2 u(T )NE
T
(3.6)
42
where the wavelength of the line center is λ, the electron charge is e, the electron mass is m, the
temperature is T in K , the partition function as a function of temperature is u(T ), the number
of hydrogen particles per unit volume is Nr , the number of the element E per unit volume is
NE , the excitation potantial is χ, the number abundance for the element E is A, the statistical
weight is g, the transition probability is f , the continous absorption coefficient is κν and HH is a
function of λ, ∆λD , f, m, e, c ( H=H(λ,∆λD ,f,m,e,c)).
Wλ of unblended or slightly blended metal absorption lines were measured by normalizing
regions of continuum around identified lines with first order polynomials in DIPSO (Howarth
& Murray 1991). Wλ was measured using a trapezoidal integration of the area of the spectral
line between two user specified points in DIPSO. This allowed us to determine statistical errors
in wλ . There are also systematic errors in continuum placement.
3.8 Microturbulent Velocity
This additional velocity term in the definition of Doppler width or the Maxwellian velocity distribution was first indroduced into spectral line analysis by Struve & Elvey (1934) and attributed
to non-thermal motion in the stellar atmosphere. Since then it has been invoked in abundance
studies to reduce systematic errors in abundance analysis. The Doppler width including microturbulent velocity (vt ) is given by
∆νDoppler =
ν
c
×
2kT
mi
+
v2t
1/2
(3.7)
where ν is the frequency of the line centre, k is the Boltzman constant, T is the local temperature
and mi is the mass of species i. Weak and saturated lines act differently to a change in vt . An
increase in vt for a weak line causes broadening and keeps the wλ conserved but the line itself
becomes shallower. For a saturated line, the line depth remains the same but total amount of
absorption in the line is increased (desaturation of the strong lines).
Therefore the determination of v t is a crucial step in abundance analysis. Measured wλ values
3.9. Colour Temperatures
43
were plotted against derived abundances for selected ions for a range of vt . Ions chosen include
CII, NII, OII, SiIII, SII, III, FeII, III. Since the derived abundances must be independent of
individual line strengths, the slope of the fit to these abundances should be zero (zero-slope
method). The vt satisfying this condition is used for the final abundance analysis.
3.9 Colour Temperatures
Photometry can be used as a tool for determining stellar atmospheric parameters. The results
from a photometric system (broad-, intermediate- or narrow-band) can be compared to those
found via spectrum synthesis or fine analysis methods or can be used as input to stellar atmosphere codes for an abundance analysis. One has to first account for extinction, which is an
effect of dust obscuration. Since most astronomical objects are observed through some amount
of interstellar medium, it is required to remove the effects of this wavelength-dependent obscuration (dereddening) from those observations. Extinction can have both circumstellar and
interstellar components.
The correction for extinction should be done prior to determination of model atmosphere parameters. In order to do that an extinction law should be considered in the form of RV , which is
a ratio of total extinction (AV ) to selective extinction (E(B-V)). In this sense, using Strömgren
(Strömgren 1966) or Geneva (Cramer & Maeder 1979) systems might be an advantage since
both are built on the basis of reddening free parameters. Unfortunately, we could not find any
Geneva photometry data for any of our program stars except HD119608. These two systems
have relatively narrow band-passes. Temperatures calculated from Johnson photometry (Johnson & Morgan 1951) should be taken only as initial estimates prior to a model atmosphere
analysis.
Most of our program stars had photometric data in the literature. These data have been compiled for each program star in the following chapters. Whenever possible, Johnson, Strömgren,
Tycho-2 (Hog et al. 2000), 2MASS (Skrutskie et al. 1997), DENIS (Epchtein et al. 1994),
MSX (Egan et al. 2003), and IRAS point source catalogue (Beichmann 1985) photometry was
included in these tables. For some stars, Johnson photometry was not available. Instead Tycho-2
44
photometry was converted to Johnson via following formula:
B J = BT − 0.240(BT − VT )
(3.8)
V J = VT − 0.090(BT − VT )
So the data presented in those tables could be used to determine whether the program stars show
evidence of a circumstellar shell when plotted in logλ – λlog(flux).
In this section, we present our results for photometric temperatures using Strömgren and Johnson colours of some of the program stars. In certain cases, inappropriate usage of photometric
calibrations to determine stellar temperatures may lead us to wrong conclusions about evolutionary status of the objects as well as their chemical compositions5. Since these photometric
temperatures can be used to initialize model atmosphere parameters, we aimed to calculate these
colour temperatures for most of the our program stars for the first time and to compare them to
those obtained from our model atmosphere analysis. For those studied elsewhere, we aimed to
emphasize the effects of color excesses, E(B-V) and E(b-y) on calculated colour temperatures.
Because the gravities were obtained on the basis of hydrogen Balmer lines and helium lines by
fixing the temperature, in most cases at ionization equilibrium temperatures, we mainly focused
on determination of photometric temperatures and their reliability for further analysis.
Strömgren indices used in the related calibrations either in a reddening free form with brackets
(e.g. Equation 3.9) or dreddened form with zero subscript (e.g. (B-V)0 or m0 showing the
intrinsic value of Johnson colours (B-V) and Strömgren index m1). The source of photometric
colours was SIMBAD6 . Adopted temperatures are reported in Table.3.1.
5
For instance, in Table 3.1, for LSIV-12 111, three different values of E(B-V) are presented. Those three values
will indicate to three different colour temperatures changing from 19 660 to 56 978 K when (B − V)0NAP93 relation
is used.
6
http://simbad.u-strasbg.fr/simbad/sim-fid
45
[c1] = c1 − 0.2(b − y)
[m1] = m1 + 0.32(b − y)
[u − b] = [c1] + 2.0[m1]
(3.9)
(B − V) = (B − V)0 + E(B − V)
(b − y) = (b − y)0 + E(b − y)
Wherever possible, in addition to the uvbyHβ method (Moon 1985), Napiwotzki, Schönberner
& Wenske (1993) ([u-b]NAP93 ), ((b − y)0NAP93 ) and ((B − V)0NAP93) relations, as well as the calibrations of Gulati et al (1989) (GUL89) and Balona (1994) (BAL94) were used for the transformation of Johnson and Strömgren indices to effective temperatures respectively. We used
BAL94 to calculate gravities. The gravities with these calculated temperatures are presented in
Table.3.1.
The negative value of E(b-y) for PHL1580 calculated by us should be noted (Table 3.1). Negative extinction translates to a reddening rather than a de-reddening. We therefore used E(b-y)=0
for temperature estimation.
By looking at the Table 3.1, some of the program stars show discrepancies in calculated photometric temperatures. The findings are as follows:
1. For instance, for HD119608, almost half of its photometric temperatures are around
23 000 K , while Strömgren photometric indices indicate a higher temperature. Possible
explanation is as follows: when the 2MASS colours of the star were converted to fluxes
by using zero-magnitude fluxes from Zombeck (1990) and log(λFλ ) is plotted against
log(λ), it can be clearly seen that JHK magnitudes of the star are too high. In fact, the
star is even brighter in H. We conclude that HD119608 clearly show an infra-red excess
which may indicate a circumstellar material, or a companion star, or both. BAL94 gave
Teff = 15 081 and log g = 1.38. For the latter calculation, we used β given by Hauck &
Mermilliod (1998).
46
STAR
Teff (K)
[u − b]NAP93
Teff (K)
(b − y)0NAP93
HR1765
19 794
±594
19 586
27 032
±811
26 225
18 711
±561
11 146
±334
12 478
HD119608
LB3116
LSIV-04 01
LSIV-12 111
Teff (K)
(B − V)0NAP93
Teff (K)
uvbyHβMO85
20 240+
23 3451a
20 4981b
26 7701d
16 1621c
15 1432a
13 4302b
84863a
-
79003b
19 6604a
36 243∗
±1087
-
24 7184b
56 9784c
IRAS17311-4924
76485
-
IRAS17381-1616
22 4116
-
PHL 1580
LSS5112
Teff (K)
(B − V)GUL89
0
18 7007
15 629∗∗
88408
Teff (K)
loggBAL94
19 020
3.11
23 2891a
±2096
21 4281b
±1929
17 8401c
±1606
16 8292a
±1515
62453a
±562
44013b
±396
20 8124a
±1873
24 0794b
±2167
54 2384c
±4881
36965
±333
79 251
±7133
20 6491e
2.43
15 394
2.40
89813c
1.313c
22 8454d
2.544d
7135
±642
Table 3.1: Photometric temperatures for some of the program stars. The rms scatter in Teff was
given as 3 percent for [u-b] method of NAP93 and 9 percent for GUL89.
•
MO85
•
+
: According to the grid of Moon & Dworetsky(1985). NAP93 used updated version of MO85.
: E(b-y)=0.027 obtained by us via Moon (1985).
• ∗ : Please see related section for the star;∗∗ : E(b-y)=-0.011 obtained by us via Moon (1985).
•
1a
: E(B-V)=0.16 from Pettini & West (1998);1b: E(B-V)=0.14 from Graham (1974);
from Schelgel et al. (1998);1d : E(b-y)=0.137 obtained by us via Moon (1985)
•
2a
•
3a
•
4a
•
3c,4d,1e
•
5
1c
: E(B-V)=0.096
: E(B-V)=0.042 from Schelgel et al. (1998);2b: E(b-y)=0.013 and Hβ =2.607 obtained by us via Moon
(1985).
: E(B-V)=0.394 from Kilkenny (1984);3b : E(B-V)=0.310 from Schelgel et al. (1998)
: E(B-V)=0.16 from Bogdanov (2003);4b : E(B-V)=0.195 from Schelgel et al. (1998);4c : E(B-V)=0.39
from Kilkenny & Pauls (1990)
: Beta parameter was estimated from Moon (1995) on the basis of an assumption that the star has Ib
luminosity class and this beta used with c0 in Balona (1994) for Teff and log gdetermination.
: E(B-V)=0.218 from Schelgel et al. (1998); 6 : E(B-V)=0.541 from Schelgel et al. (1998); 7 : If star
is a super-giant (II); E(b-y)=-0.011 and Hβ =2.593. If it is a bright super-giant(Ib) then E(b-y)=-0.060 and
Hβ =2.580 obtained by us via Moon (1985).; 8 : E(B-V)=0.406 from Schelgel et al. (1998)
47
2. LSIV-12 111 is another interesting star. It has a very high temperature when the (B-V)0
methods of NAP93 and GUL89 applied. This might be due to adopted E(B-V) for the
star. Killkenny & Pauls (1990) give E(B-V)=0.39.
3. It is equally interesting to note that IRAS17311-4924, LSS5112 and LSIV-04 01 show
relatively lower temperatures. Again, since these low temperatures derived from (B-V) 0
methods of NAP93 and GUL89, E(B-V) could be playing important role. Extinction
was estimated using Diffuse Infrared Background Experiment (DIRBE)/IRAS dust maps
from Schelgel et al. (1998) by us. Schelgel et al. (1998)’s maps only give the extinction
in the line of sight since we think that circumstellar extinction rather than interstellar must
dominate the total extinction of these stars.
4. In contrast, photometric methods seem to give photometric temperatures comparable to
results of spectrum synthesis for the standard star HR1765 and LB3116.
In conclusion, our colour temperature analysis showed that it is not a reliable method for B-type
stars and therefore those temperatures were not used in the following chapters to initialize LTE
model atmosphere analyses of the programme stars.
Chapter 4
High-resolution optical spectroscopy of
the B-type abundance standard HR1765
In this chapter, we have analyzed the well-studied early B dwarf HR1765 in order to test the
model atmospheres, atomic data and methods to be used in our programme stars. We also
present a fine analysis of the chemical abundances from a high-resolution optical spectrum of
the star. To the best of our knowledge, we report phosphorus abundance for the star for the first
time. We discuss various agreements between our result and previous work.
4.1 Background
The first spectroscopic studies of the 53 Persei variable HR1765 (22 Ori) go back to the 1950s.
Spitzer (1950) reported the measurements of equivalent widths of CaII H and K lines in the star.
The star was also included by Aller (1950) in a study of abundances of the light elements in
B-type stars.
Sharpless (1952) gives B2 as a spectral type on the basis of a photometric study including
HR1765 while Beals & Oke (1953) classify the star as B3. On the other hand, Abt & Levato
(1977) gives a classification of B2Vs (s:unusually sharp lines).
49
50
Chapter 4. High-resolution optical spectroscopy of the B-type abundance standard
HR1765
HR1765
l
202.63
b
-20.03
m1
0.0832
0.0012
c1
0.1732
0.0042
(b-y)
-0.069
B
4.545
V
4.721
J
5.465∗
0.286
H
5.235∗
0.055
K
5.220∗
0.018
H
12
He
11.01
0.01
11.00
C
7.99
0.22
8.52
N
7.47
0.19
7.92
O
8.27
0.23
8.83
Mg
7.47
0.02
7.58
Al
5.98
0.21
6.47
Si
7.08
0.18
7.55
P
4.95
0.30
5.45
S
7.23
0.37
7.33
±
Star
HR1765
±
SUN
12.00
Ar
6.38
0.37
6.40
• ∗: 2MASS All-Sky Point Source Catalog (PSC)- ADS/IRSA.Gator 2007/0904/102826 26363
• 1: Crawford et al. (1971).
• 2: Hauck & Mermilliod (1998).
Table 4.1: Photospheric chemical abundances of HR1765 given as logn, normalized to logΣµn=
12.15. It is compared to solar abundances (Grevesse & Sauval 1998). Fundamental parameters
for HR1765 with B, V, J, H, K magnitudes were also presented.
HR1765 was classified as a single-lined spectroscopic binary by Abt & Levy (1978) who also
reported marginal spectroscopic orbital elements with an orbital period of 293d. Warren &
Hesser (1977a)’s identification for the star was B2Vs with a rotational velocity of 8 km/s.
Smith (1980) performed high dispersion observations of HR1765 over the period 1976-1978
and reported four different periods each of which was detected at least twice, being 4.5h , 9h ,
14.1h , and 22.9h . Smith (1977) had already shown that large line profile variations and relatively
small radial velocity variations were characteristic of these slowly rotating B stars (53 Persei
variables). 2005 AAT/UCLES data show vsini=5 km/s. Smith (1980) also reports that although
HR1765 shows phase-correlated line-strength variations, they are probably small enough for
the star to still serve as an abundance standard for early B stars.
Gomez & Abt (1982), in a study searching for secondary lines in the visual spectra of 55 binaries
including HR1765, looked for a clue indicating a secondary component. They noted that there
was no evidence for a late-B type sharp-lined secondary. They give K3 as the latest possible
spectral type for the unobserved secondary.
Proffitt & Quigley (2001) calculated rotational velocity and microturbulent velocity from BIII
line at 2066 Å with metallicity ([Fe/H]) as 4.49 km/s, 1.61 km/s and -0.26 respectively.
HR1765 has been studied as an abundance standard in various studies (Hambly et al. 1997;
Munn et al. 2004). Basic parameters for the star is presented in Table 4.1.
Fe
7.19
0.38
7.50
4.2. Model atmospheres analysis and results for HR1765
51
4.2 Model atmospheres analysis and results for HR1765
4.2.1 Determining Teff and log g : Renormalization
Chemical abundances are measured using a model atmosphere with a given Teff , g and nHe .
These parameters can be usually measured from strong hydrogen Balmer and helium lines in
the absorption spectrum. However, in the case of HR1765, this was not possible because these
broad lines are very sensitive to normalization. All hydrogen Balmer lines span more than
one échelle order so that it is difficult to remove the blaze function accurately. Therefore H
lines could not be relied upon. In order to use the information from weak lines in the spectrum of HR1765, we renormalized it using the best model fit to this observed high-resolution
UCLES/AAT spectrum. We obtained the best model fit using SFIT (see Fig. 4.1, 4.2 and 4.3). A
grid of model atmospheres was calculated with Teff = 15 000, 16 000, 18 000, 20 000, 22 000,
24 000, and 25 000 K , log g = 2.5 (0.5) 4.5, and helium abundance by number nHe = 0.10.
This abundance for helium seemed convenient for all helium lines in the spectrum of HR 1765.
The smoothed ratio between model and observed spectrum has been used to correct the normalization. We divided the observed spectrum by the best model fit then smoothed the ratio using a
Gaussian filter. The observed spectrum was multiplied by this ratio to give a renormalized spectrum. In this process, weak lines have not been changed but broad hydrogen lines are affected
and have not been included in the abundance determination.
4.2.2 Radial velocity
The zero-point for velocity was obtained by measuring the accurate positions for 73 lines including N, O, Al, Si, S, Ar, and Fe lines. This indicated to a heliocentric velocity of 27.5±1
km/s.
4.2.3 Metallicity
We calculated metallicity ([Fe/H]) from Fe abundance ([Fe]) in the following form and it was
found to be [Fe/H]≈-0.02 (almost solar). This metallicity corroborates [Fe/H]=-0.09 by Cunha
52
HR1765
Figure 4.1: HR1765 (in black) and best model fit (in red). Both model and observed spectra are
moved to 0 velocity reference frame - part I
53
moved to 0 velocity reference frame - part II
54
HR1765
moved to 0 velocity reference frame - part III
55
Figure 4.4: Determination of microturbulent velocity (v t ) via NII lines for HR 1765. For vt =6
km/s the slope of the regression line vanishes as expected if the choice of vt is correct. Solid
line is the least-squares fit to the data points.
& Lambert (1994). However, 2005 AAT/UCLES spectrum of the star is of better quality so will
indicate to a better value for Fe abundance. The 1/10 of solar metallicity model atmosphere
gave the best model fit and was therefore used in the model atmosphere analysis of the star1 .
Fe H
= log
nFe nH
S T AR
− log
nFe nH
(4.1)
⊙
4.2.4 Microturbulent velocity
Plots of abundance against observed line strength for ions such as NII and OII have been used
in vt determination. Vt was varied until elemental abundances were found to be independent
1
The metallicity of the star is calculated from FeIII lines and it should be noted that the strongest FeIII line
included in the analysis is FeIII 4164.79 Å line with an equivalent width of 34 mÅ . When overall metallicity of
the star is calculated on the basis of its Mg, Al, Si, S abundances, we find average metallicity ([M/H]) of [Mg/H]=0.11, [Al/H]=-0.49, [Si/H]=-0.47, [S/H]=-0.10, and [Fe/H]=-0.02 as metallicity of the star, [M/H]=-0.24 so the
closest metallicity model ([Fe/H]=-1) was chosen and used for the analysis.
56
HR1765
Figure 4.5: Determination of microturbulent velocity (v t ) via OII lines for HR 1765. For vt =6
km/s the slope of the regression line vanishes as expected if the choice of vt is correct. Solid
line is the least-squares fit to the data points.
of line strength (wλ ) (Fig 4.4 and 4.5). The NII and OII lines used for micro-turbulent velocity
determination are presented in Table 4.2.
We carried out a second vt determination. Knowing that in hotter B stars, OII, NII, and SiIII
have well-populated curves of growth which are ideal for determination of vt , abundance measurements are done on a grid from 0 ≤ vt ≤ 23 km/s for ten OII, and nine NII lines in steps of
1 km/s (Fig. 4.6 and 4.7). Sample standard deviation (σ) in logarithmic abundances has been
plotted against microturbulent velocity. The minimum value of this function gives the most
probable value of microturbulent velocity. OII lines give 6 km/s and NII lines give 9 km/s. We
adopted vt =6±3 km/s for HR 1765.
The standard error of OII, NII and SiIII (not given in plots) abundances can be estimated by
√
dividing these minimum values by n, n is the number of the lines used in the calculation. This
would give ≈ 0.02 dex for OII and ≈ 0.04 dex for NII respectively. These standard errors found
0.14
0.14
0.12
0.12
σ
0.16
σ
0.16
57
0.10
0.10
0.08
0.08
0.06
0
5
10
15
Vturb(km/s)
20
25
0.06
0
5
10
15
Vturb(km/s)
20
25
Figure 4.6: Determination of microturbulent velocity ( vt ) with a different method. vt ≈ 9 km/s
is min. of parabola. The effect of [Fe/H] on calculated vt is also represented for [Fe/H]=-1 and
[Fe/H]=1 respectively.
from vt analysis can be confirmed if the line-to-line scatter values presented in Table 4.1 are
√
divided by n. We find ≈0.04 dex for OII and ≈0.05 dex for NII. They agree well.
4.2.5 Rotational velocity
In determination of vsini, we used N, O, and Si lines. The vsini used in the best model fit was
5±1 km/s. Similar vsini was reported by Proffitt & Quigley (2001).
4.2.6 Abundances
Both spectrum synthesis and fine analysis show that the star is an apparently chemically normal
star. The photospheric abundances for each element calculated by the best model fit is presented
in Table 4.1, and abundances for individual lines in Table 4.3, 4.4, and 4.5.
58
HR1765
0.135
0.130
σ
0.125
0.120
0.115
0.110
0.105
0
5
10
15
20
25
Vturb(km/s)
Figure 4.7: Determination of micro-turbulent velocity (v t ) with a different method. vt ≈6 km/s
is min. of parabola.
Table 4.2: NII and OII lines used for micro-turbulent velocity determination.
OII
OII
OII
OII
NII
NII
3954.37 4189.79 4366.89 4638.86
3995 4601.48
3982.72 4294.79 4414.90 4641.82 4035.08 4607.16
4072.16 4303.84 4416.97 4649.14 4043.53 4613.87
4078.84 4317.14 4452.37 4661.63 4227.74 4630.54
4085.11 4319.63 4590.97 4673.74 4236.93 4643.09
4132.8 4325.76 4596.18 4676.24 4432.74 5005.15
4185.45 4349.43 4609.44 4699.22 4447.03 5007.33
5045.09
Notes on individual lines
• SiII 5056.02 line is blend with SiII 5056.35 being a member of multiplet no:5 so that it is
not included in silicon abundance determination.
• SiIII multiplet no:2 members are not included in the analysis instead multiplet no:9 was
preferred.
• The identification for FeIII 4395.78 line is dubious and not used in the analysis. It is
59
Table 4.3: Equivalent Widths and LTE abundances for HR 1765. The abundances are calculated
for vt =5, 6, and 9 km/s.
IONS
SiII
SiII
SiII
OII
CII
CII
ArII
OII
OII
NII
NII
NII
PIII
OII
OII
OII
OII
OII
OII
SiII
SiII
OII
FeIII
SII
SII
FeIII
FeIII
OII
OII
PIII
NII
NII
PIII
SIII
CII
SIII
SII
OII
OII
OII
OII
OII
λ Mult.
EW error(2σ)
(Å)
No. (mÅ)
(mÅ)
3853.65
1
29
±5
3856.02
1
77
±7
3862.59
1
57
±7
3911.96
17
61
±6
3918.97
4
115
±7
3920.67
4
129
±7
3928.62
10
16
±4
3954.37
6
40
±5
3982.71
6
25
±2
3994.99
12
67
±3
4035.08
39
41
±2
4043.53
39
20
±3
4059.27
1
6
±2
4069.63
10
68
±5
4072.16
10
62
±3
4075.86
10
102
±5
4078.86
10
19
±6
4085.12
10
37
±1
4087.16
48
16
±1
4128.05
3
75
±1
4130.88
3
44
±1
4132.80
19
25
±1
4139.37
118
22
±3
4153.09
44
59
±2
4162.69
44
29
±5
4164.79
118
34
±6
4166.86
118
33
±8
4185.45
36
28
±3
4189.78
36
41
±2
4222.15
3
25
±4
4227.74
33
13
±2
4236.93
48
28
±2
4246.68
3
15
±4
4253.59
4
40
±4
4267.02
6
230
±4
4284.99
4
26
±3
4294.43
49
21
±4
4294.82
54
19
±5
4303.82
54
23
±3
4317.13
2
45
±3
4319.63
2
47
±3
4325.77
2
15
±2
Abundances
vt = 5 km/sec.
7.75
7.46
7.47
8.63
8.38
8.25
7.27
8.20
8.19
7.34
7.80
7.23
4.73
8.82
8.33
8.82
8.13
8.50
7.98
7.50
6.92
8.19
7.22
7.65
7.07
7.14
7.62
8.24
8.40
5.26
7.45
7.82
5.28
6.61
7.85
6.74
7.18
8.32
8.17
8.29
8.33
8.26
Abundances
vt = 6 km/sec.
7.73
7.39
7.43
8.60
8.30
8.15
7.26
8.17
8.17
7.30
7.77
7.22
4.73
8.75
8.26
8.72
8.11
8.46
7.96
7.44
6.89
8.17
7.19
7.60
7.05
7.10
7.58
8.20
8.36
5.24
7.44
7.80
5.26
6.56
7.79
6.70
7.17
8.30
8.14
8.24
8.28
8.24
Abundances
vt = 9km/sec.
7.71
7.29
7.37
8.50
8.13
7.94
7.24
8.08
8.12
7.20
7.71
7.19
4.72
8.57
8.10
8.47
8.07
8.37
7.93
7.34
6.85
8.11
7.15
7.52
7.02
7.02
7.50
8.13
8.27
5.19
7.43
7.76
5.24
6.45
7.64
6.55
7.15
8.26
8.09
8.14
8.17
8.21
60
HR1765
IONS
OII
ArII
OII
OII
SIII
OII
C II
FeIII
OII
OII
FeIII
ArII
NII
NII
OII
Al III
OII
Al III
SII
Al III
Si III
Si III
Si III
Ar II
O II
O II
N II
N II
N II
N II
O II
O II
N II
CIII
O II
O II
O II
O II
O II
O II
O II
λ Mult.
EW error(2σ)
(Å)
No. (mÅ)
(mÅ)
4347.42
16
22
±2
4348.11
7
24
±1
4349.42
2
56
±2
4351.26
16
26
±2
4361.53
4
15
±2
4366.89
2
42
±3
4374.28
45
18
±6
4395.78
4
41
±1
4414.90
5
63
±4
4416.97
5
50
±4
4419.59
4
37
±4
4426.01
7
18
±3
4432.73
55
19
±4
4447.03
15
45
±4
4452.37
5
23
±2
4479.89
8
40
±6
4491.25
86
15
±3
4512.53
3
34
±3
4524.94
40
20
±3
4529.17
3
59
±4
4552.65
2
131
±5
4567.87
2
103
±4
4574.77
2
65
±4
4589.93
31
10
±3
4590.97
15
36
±4
4596.17
15
34
±5
4601.47
5
31
±5
4607.15
5
31
±5
4613.86
5
33
±3
4630.54
5
61
±3
4638.85
1
52
±3
4641.81
1
66
±2
4643.08
5
41
±2
4647.42
1
16
±3
4649.13
1
90
±3
4650.84
1
60
±4
4661.63
1
57
±3
4673.75
1
21
±3
4676.23
1
44
±3
4699.21
25
31
±4
4705.36
25
33
±4
Abundances
vt = 5 km/sec.
7.95
6.46
8.09
7.89
6.65
8.16
7.98
7.09
8.14
8.16
6.53
6.62
7.48
7.58
8.35
6.37
7.78
5.95
7.31
6.10
7.32
7.20
7.13
6.53
8.13
8.15
7.45
7.55
7.72
7.48
8.49
8.29
7.65
7.86
8.43
8.65
8.49
8.58
8.38
8.10
8.00
Abundances
vt = 6 km/sec.
7.94
6.45
8.03
7.87
6.63
8.12
7.97
7.06
8.08
8.11
6.51
6.61
7.47
7.55
8.33
6.34
7.77
5.93
7.30
6.05
7.20
7.09
7.06
6.53
8.09
8.12
7.44
7.54
7.70
7.44
8.43
8.22
7.62
7.83
8.34
8.58
8.43
8.56
8.33
8.07
7.97
Abundances
vt = 9km/sec.
7.90
6.42
7.90
7.82
6.59
8.03
7.96
7.00
7.94
8.00
6.46
6.59
7.44
7.49
8.28
6.27
7.73
5.88
7.28
5.94
6.91
6.86
6.91
6.52
8.00
8.06
7.40
7.49
7.66
7.34
8.31
8.07
7.57
7.77
8.13
8.44
8.30
8.52
8.23
7.99
7.89
61
IONS
S II
N II
N II
N II
Si II
λ Mult.
EW error(2σ)
(Å)
No. (mÅ)
(mÅ)
4815.52
9
28
±3
5005.14 19,6
53
±6
5007.32
24
23
±4
5045.09
4
28
±2
5056.02
5
29
±6
Abundances
vt = 5 km/sec.
7.27
7.51
7.45
7.45
6.73
Abundances
vt = 6 km/sec.
7.26
7.47
7.44
7.44
6.72
Abundances
vt = 9km/sec.
7.23
7.38
7.40
7.40
6.70
suspected to be OII 4395.95(26).
• The identification for FeIII 4419.59 is dubious and again not used in the analysis.
It should be noted that some of the weak lines are not included in this table. These lines are as
follows:
• OII 4710.0 multiplet no:24, SII 4716.23 multiplet no:9, SiIII 4716.66 , NII 4779.7
multiplet no:20, NII 4788.13 multiplet no:20, NII 4803.29 multiplet no:20., OII 4890.86
multiplet no:28, OII 4906.83 multiplet no:28, SII 4917.2 multiplet no:15, OII 4924.6
multiplet no:28, OII 4941.07 multiplet no:33., NII 5010.62 multiplet no:4, and SiII
5041.06 multiplet no:5.
• OII 4609.42Å line is not included in the line list and model atmosphere analysis since
it could not be resolved from nearby ArII line at 4609.60Å. Another important point is
that ArII 4589.93Å line is resolved from OII 4590.97Å which is giving an idea about the
quality of the observations.
• NII lines 4994.36 and NII 4987.38 are blended being members of multiplet no:24 while
NII lines at 5001.47 and 5001.13 are blended being members of multiplet no:19.
• NII line 5002.69 Å is blend with NII(19).
• MgII lines at 4390.60 and 4434.00 Å are weak and not included in the analysis. The
former one is blend with HeI 4388 line.
62
HR1765
OII EW(CL)
Å
(mÅ)
4072
61
4078
24
4085
28
4087
11
4132
24
4185
27
4591
43
4596
41
4609
24
4638
49
4641
64
4649
85
4661
44
4673
13
4676
37
EW(ŞAHİN)
(mÅ ± 2σ)
62±3
19±6
37±1
16±1
25±1
28±3
36±4
34±5
26±2
52±3
66±2
90±3
57±3
21±3
34±3
Table 4.6: Equivalent Widths of OII lines
in the spectrum of HR 1765 are compared
to those measured by Cunha & Lambert
(1992) (CL).
HeI EW(LL)
Å
(mÅ)
4121
300
4387
800
4438
130
4713
275
5015
290
5047
190
EW(ŞAHİN)
(mÅ ± 2σ)
264±12
767±9
125±5
275±3
290±8
178±3
Table 4.7: Measured equivalent widths of
HeI lines in the spectrum of HR 1765 are
compared to those measured by Leone &
Lanzafame (LL), 1998. Please note that
4471 and 4921 are not included in our
atomic list used to calculate LTE abundances.
4.3 Evaluation
4.3.1 Comparison of abundances
HR1765 is often used as standard star in the literature. We compared our abundance analysis
results for HR1765 with Cunha & Lambert (1992) (for OII equivalent widths) and Leone &
Lanzafame (1998) (for HeI equivalent widths). The results are presented in Table 4.6, 4.7,
Figure 4.8, and 4.9. Similar comparison was also performed between HR1765 and γPeg, an
apparently chemically normal star lying within the solar neighbourhood, and presented in Table
4.8. A fourth comparison were performed between Hambley et al. (1997) and our abundance
measurements in Table 4.9. They all agreed within error limits.
4.3. Evaluation
63
Figure 4.8: Comparison of equivalent
widths from Cunha & Lambert (CL)
(1992) and Table 4.8 of this section. There
seems to be no systematic differences between two data sets; a formal least-square
solution gives a gradient of 0.93 and a zero
point difference of 0.001 A.
Ion
Å
CII
NII
OII
AlIII
SiII/III
PIII
SII
SIII
ArII
FeIII
HR1765
(mÅ ± 2σ)
8.05±0.22
7.52±0.19
8.24±0.23
6.11±0.21
7.25±0.18
5.08±0.30
7.28±0.24
6.67±0.05
6.71±0.37
7.09±0.38
No.
4
12
37
3
8
3
4
2
4
5
γPeg
(mÅ ± 2σ)
8.22±0.31
7.79±0.24
8.72±0.32
6.14±0.05
6.90±0.15
−−
7.01±0.31
−−
−−
6.39±0.82
Table 4.8: Photospheric line abundances
for HR 1765 is compared to γPeg (Munn
et al. 2004). Number of the lines used in
the analysis are included only for HR1765.
Figure 4.9: Comparison of equivalent
widths from Leone & Lanzafame (LL)
(1998) and Table 4.9 of this section. There
seems to be no systematic differences between two data sets; a formal least-square
solution gives a gradient of 1.04 and a zero
point difference of 0.002 A.
Ion
Å
CII
NII
AlIII
SII
SIII
ArII
FeIII
HAMB97
(mÅ ± 2σ)
8.38±0.36
7.82±0.20
6.30±0.06
6.95±0.16
6.83±0.11
6.49±0.11
7.23±0.32
No.
10/4
22/12
3/3
18/4
7/4
6/4
6/5
ŞAHİN
(mÅ ± 2σ)
8.05±0.22
7.52±0.19
6.11±0.21
7.28±0.24
6.67±0.05
6.71±0.37
7.09±0.38
Table 4.9: Photospheric line abundances
for HR1765 is compared to Hambley et al.
(1997) (HAMB97). They agreed within error limits.
4.3.2 Testing Teff and log g : ionization equilibrium and
Balmer line fitting
In order for model atmosphere parameters to be tested, we compiled a Teff – log g plot based
on the observed spectrum of the star. The ionization equilibria of SiII/III, CII/III were used to
estimate log g and Teff . As can be seen in Figure 4.10, since these two ionization equilibria run
approximately parallel each other, it was not possible to obtain separate solutions for Teff and
64
HR1765
Figure 4.10: Model atmosphere parameters for HR1765. Error bars for this work have been
also plotted. Filled circles show published model atmosphere parameters for the star.
log g . Line profile fits to Hβ, Hγ, and Hδ were performed and included in this analysis. The intersection with the ioniation equilibrium enabled model atmosphere parameters to be estimated
(Figure 4.10). Loci satisfying model atmosphere parameters are indicated in Figure 4.10. in
which the model atmosphere parameters obtained in previous studies by different groups were
also presented (Table 4.10).
4.3.3 Testing cosmic argon abundance
Model atmosphere analysis of HR 1765 can also provide valuable information about cosmic
argon abundances. According to Brown et al. (1986), main sequence stars normally have atmospheres uncontaminated by the products of interior reactions. Keenan et al. (1990) use weak
interstellar absorption lines in the spectra of main-sequence early type stars. Their assumption
is that, these weak lines lie on the linear part of curve of growth as they will have strengths
which are very sensitive to the element abundance, but not to the assumptions made in the
model atmosphere analysis. These abundances should reflect those of the interstellar material
4.4. Conclusion
65
Table 4.10: Atmospheric parameters of HR1765.
Star
This work
4
5
6
7
9
10
11
12
13
14
15
Teff
(K)
22.000±1300
23.500
18.860
23.500±400
19.850
20.600
19.900
19.500
19.800
20.550
20.550
21.270
log g
(dex)
3.5±0.3
4.0
3.4
4.01±0.05
3.4
3.8
3.7
3.5
3.55
3.74
3.74
3.73
vt
(km/s)
6±3
6
vsini
(km/s)
5±1
1±1
4.2
25
10
10
1.61
9.0
4.1±0.8
4.49
10.0
16±2
Vrad
(km/s)
27.5±1
4 : Munn et al. (2004); 5 : Leone & Lanzafame (1998); 6 : Gummersbach et al. (1998); 7 : Leone et al.
(1997); 9 : Wolff & Heasley (1985); 10 : Keenan et al. (1990); 11 : From Strömgren color indices (Keenan
et al. 1990); 12 : Kaufer et al. (1994); 13 : From BIII resonans line at 2066 Å (Proffitt & Quigley 2001)
;14 : Cunha & Lambert (1994); 15 : Gies & Lambert (1992).
from which the stars formed some 10 6 - 107 years ago, and therefore reliable contemporary
cosmic abundance estimates will be available using this method. They used Ar II 4589.93Å
which is resolved from OII 4590.97 Å. They report an argon abundance of 6.53±0.5 dex. Their
Teff for HR 1765 from Strömgren colours is 19 500 K and log g = 3.5 and they report that error
could be as much as ±2000 K in Teff . In our work, the argon abundance derived from the same
argon line2 with vt = 6 km/s and vsini = 5 km/s is 6.53±0.19. It should be noted that the error in
the argon abundance is given as the error of the mean abundance derived from four argon lines.
4.4 Conclusion
In this chapter, LTE model atmosphere analysis of the well-studied early B dwarf HR1765 was
performed to test the model atmospheres, atomic data and methods to be used in our programme
stars. The findings are as follows:
1. Both spectrum synthesis and fine analysis show that the star is an apparently chemically
normal star.
2
Ar II line at 4609.60 Å which is blend with OII line at 4609.42 Å. This may have an affect on calculated Ar II
abundance since it is blend
66
HR1765
2. The measured equivalent widths of OII and HeI lines in the spectrum are agreed with
Cunha & Lambert (1992) and Leone & Lanzafame (1998) within error limits respectively.
3. The photospheric line abundances for HR1765 is agreed with Hambley et al. (1997)
within error limits.
4. To the best of our knowledge, we report phosphorus abundance of the star for the first
time.
5. We also report a cosmic argon abundance of 6.53±0.19 and it is perfectly agreed with
Keenan et al. (1990).
In conclusion, the test for the model atmospheres, atomic data and methods was successful.
Chapter 5
High resolution spectroscopy of the
rapid rotating hot post-AGB star:
HD119608
In this chapter, we report a detailed abundance analysis of rapid rotating helium rich super-giant
HD119608.
5.1 Background
Morgan (1955) gave an MK spectral type of B1Ib for HD119608, marking it as a supergiant.
Münch (1956) calculated age of the star from the gravitational attraction of the Galaxy as a
whole which gave 32 million years when z=2.7 kpc used as Galactic distance. They note that
this distance was not corrected for interstellar absorption.
HD119608 was studied by Habing (1968) for the detection of 21-cm line profiles in the direction
of the star. The stellar distance and projected height above the galactic plane were reported as
3.4 kpc and 2.4 kpc respectively.
Kilkenny & Hill (1975) performed Hβ photometry of the star. They combined their measure67
68
Chapter 5. High resolution spectroscopy of the rapid rotating hot post-AGB star:
HD119608
ments with UBV photometry of the star by Hill et al. (1974) to study space distribution of 165
early-type stars at intermediate and high galactic latitude including HD119608. The updated
distance of the star was 4.35 kpc with MV =-6.1.
Pettini & West (1982) found rotational velocity vsini=55 km/s. In the same study, they also
give a distance of 4.1 kpc
Edgar & Savage (1989) in a study of determining the density distribution of TiII, CaII, and FeII
away from the Galactic plane report a distance of 4.30 kpc.
Conlon et al. (1993) obtained infra-red observations to compile the spectral energy distribution
of the star. However, due to lack of IRAS data, they could not detect any infra-red excess. On
the basis of 2MASS colours, HD119608 shows infra-red excess.
Martin (2004) calculates the temperature and the gravity from Geneva photometry: Teff =
26 413 K and log g = 4.09 respectively and he trusts the results of Geneva photometry most
(private communication, 2007). Teff and log g were reported to be 20 260 K and 4.90 respectively by Morel et al. (2004).
5.2 Model atmosphere analysis and results for HD119608
A grid of twenty-four model spectra was calculated with Teff = 15 000, 18 000, 20 000, 22 000,
24 000, and 25 000 K ,log g = 2.5 (0.5) 4.0, and nHe = 0.10 in the range of 3900 - 5010 Å . We
used solar metalicity models since there was no reason to believe that the star is metal-poor
rather than solar. Model atmosphere parameters obtained with SFIT were Teff = 23 300 K and
log g = 3.0 and errors in model atmosphere parameters are ± 500 K and ±0.1 dex respectively.
Best model fit was presented in Figure 5.1, 5.2, and 5.3 with TIGER normalized spectrum of
the star. Since we did not have a previously calculated model atmosphere with Teff = 23 000
K in the grid, it was computed by STERNE. Abundances calculated by the best fit model are
presented in Table 5.1.
5.2. Model atmosphere analysis and results for HD119608
69
Figure 5.1: Two model atmosphere fits of Teff =23 000 K and Teff =26 000 K are presented.
The fit in red presents the model atmosphere with Teff =26 000 K while Teff=23 000 K model
fit is shown in green. Both model and observed spectra are moved to zero velocity reference
frame - part I
70
HD119608
frame - part II
71
frame - part III
HD119608
72
HD119608
Geneva Phot.a
Star
HD119608
±
Elements
( Teff , log g)
(23 000,3.0)
±
(26 000,3.0)
±
SUN
l
320.35
U
0.360
Ref.
TSA
C92
MAR04
MOR04
H
12
12
12.00
b
43.13
V
1.038
Teff
( K)
23 000
500
24 000
26 413
20 260
He
m1
0.0332
B1
0.792
log g
(cm/s2 )
3.0
0.1
3.0
4.1
4.9
C
c1
-0.0142
B2
1.579
vt
(km/s)
6
1
Hβ
2.5542
V1
1.730
vsini
(km/s)
55
1
(b-y)
0.0212
G
2.238
vrad
(km/s)
28
B
7.43
VM
7.509
V
7.511
E(B-V)
0.163
15
N
50
74
O
Mg
Al
Si
P
S
Fe
11.24
0.01
11.68
0.01
11.00
8.41
0.01
8.72
0.01
8.52
8.03
0.01
8.53
0.01
7.92
8.72
0.01
9.09
0.01
8.83
7.99
0.03
8.62
0.03
7.58
6.30
0.02
6.90
0.02
6.47
7.50
0.01
7.55
0.02
7.55
5.24
0.04
5.24
–
5.45
7.11
0.02
7.40
0.02
7.33
7.66
0.01
8.30
0.01
7.50
• C92: Conlon, E.S.(1992); MOR04: Morel et al. (2004); MAR04: Martin (2004); 1: Fernie
(1983); 2: Hauck & Mermilliod (1998); 3: Pettini & West (1982); a: Geneva photometric catalogue
http://obswww.unige.ch.
Table 5.1: Atmospheric parameters and photospheric chemical abundances of HD119608 given
as logn, normalized to logΣµn= 12.15, and compared to solar abundances (Grevesse& Sauval
1998). Fundamental parameters are also presented.
5.2.1 Metallicity
The metallicity of the star was calculated from Fe abundance and was found to be [Fe/H]≈0
(solar).
The synthetic spectrum analysis, including several ions, gave vt = 6±1 km/s. From a theoretical
point of view, the vt should be the same for all species. However, it is sometimes possible to
observe a discrepancy between the results measured from different ions or/and ionization stages.
As suggested by McErlean et al. (1998), we used the SiIII triplet around 4550 Å (multiplet no.2)
for the determination of vt as a consistency check. It was found to be vt = 5 km/s (Figure 5.4).
So that vt = 6±1 km/s was used in calculation of the best fit model.
73
Figure 5.4: Determination of microturbulent velocity ( vt ) via SiIII multiplet no.2 for
HD119608. For vt = 5 km/s the slope of the regression line vanishes as expected if the choice
of vt is correct. The dashed line is the least-squares fit to the data points. The equations obtained
from the least-squares fit are also given.
5.2.3 Rotational velocity
The rotational velocity can be measured by comparing synthetic line profiles with the observed
line profiles1 . The Mg II at 4481 Å is a doublet and most suitable for this purpose. Since two
components of this doublet at 4481.13 and 4481.33 Å are not resolved, one can conclude that
HD119608 is likely to be a rapid rotator. The rotational velocity of the star was found to be
v sini = 55±1 km/s by using C, O, N, Al, and Fe lines. As a consistency check, the SiIII triplet
multiplet no.9 around 4810 Å , where normalization related effects were minimum, was used.
It agreed with the above value. The result was also in agreement with Ueugi & Fukuda (1970)
and Pettini & West (1982) (Table 5.1).
1
The synthetic line profiles are calculated for the given model atmosphere parameters ( Teff , log g ) and element
abundance of the star. These synthetic line profiles are folded (convolved) with an assumed rotational velocity.
74
HD119608
5.2.4 Radial velocity
The radial velocity of the star was measured by using C, N, O, Al, Si, and Fe lines. While Al,
Si, P, and Fe lines gave vr ≈ 57 km/s, C, N, and O lines gave ≈53, 49, and 52 km/s respectively.
The mean radial velocity was vr = 55±1. The vr value for the Fe lines, vr = 57±1, was used in
the best model fit. The heliocentric correction for the date of observation was found to be +29
km/s, hence the heliocentric velocity was calculated to be vrad = 28 km/s (Table 5.1).
We have used chi-square(χ2 ) method for testing Teff and log g of the star. In the first part of the
method, a normalized spectrum of the star was solved with SFIT SOLVE code on a grid from
16 000 ≤ Teff ≤ 26 000 K for Teff in steps of 1000 K . Temperatures were fixed at those values
and in turn log g= 3.0 was tested for those temperatures. The χ2 values were plotted against the
temperatures. The minimum of this curve (parabola) gave the most probable interval for Teff
(Fig. 5.5). In the second part of this method, the same procedure was applied to log g on a grid
from 2.5 ≤log g≤ 3.7 in steps of 0.2 dex with Teff = 23 000 K . The χ2 values were plotted
versus gravity. The minimum of this curves gave the most probable log g otherwise an interval
and it is shown in red (Fig. 5.5)
In another check of the model parameters, we masked certain parts of the spectrum. SFIT was
used to calculate a model fit for two different wavelength regions. The first interval was the
region of 3900 - 4300 Å which included both Hδ and Hǫ lines. The second was the region of
4300 - 5010 Å which included both Hβ and Hγ. The χ2 test gave Teff = 23 000 K , log g = 2.9
and Teff = 23 500 K , log g = 3.0 respectively. Abundances of the star were calculated on the
basis of a model atmosphere of Teff = 23 000 K and log g = 3.0.
5.2.6 Testing Teff : ionization equilibrium
We have performed some tests to check the Teff of the star using the ionization equilibrium
of SiIII/SiIV. We used SiIII mutliplet no. 2 (λλ 4552.6, 4567.8, 4574.7 Å ) and 9 (λλ 4813.3,
4819.7, 4829.0 Å) to show whether these two different multiplets gave similar results. SiIII
75
Figure 5.5: χ2 method for checking model atmosphere parameters of HD119608. The red curve
represents the result for χ2 method when whole spectrum used while the results for each half
when the spectrum divided into two halves are shown in black.
Figure 5.6: Ionization equilibrium for SiIII Figure 5.7: Ionization equilibrium for SiIII
multiplet no. 2.
multiplet no. 9.
76
HD119608
Figure 5.8: Ionization equilibrium for SiIII/IV. Both SiIII triplet belong to multiplet
no. 2 and 9. were included.
multiplet no. 2 is known to be subject to non-LTE effects in normal B-type stars (Becker &
Butler 1990). As can be seen from Fig. 5.6, 5.7, and 5.8, SiIII multiplet no.2 gives 20 000 K
while we find 22 200 K by using SiIII(9). The SiIII/SiIV balance indicated Teff lower by ≈800
K.
5.2.7 Testing log g : Helium line fitting
The surface gravity of the star was tested by fitting profiles of hydrogen Balmer lines (Hγ, Hδ)
and diffuse helium lines HeI λ 4026 Å , HeI λ 4387 Å , HeI λ 4471 Å in SPECTRUM. In this
procedure, gravity is determined as a function of effective temperature. As a result of this test
log g has been determined as 3.0.
5.2.8 Testing helium abundance
We have calculated helium abundance as nHe = 0.09 ([He] = 11.02) by using SFIT SYNTH.
A model of nHe = 0.10 helium was used for determination of helium abundance. We inspected
individually all helium profiles whether they agreed with the abundance found. The fit for HeI
4009 Å did not fill the core for nHe =0.09 while the fit profiles for helium lines at λλ 4026 and
4388 Å were in an excellent agreement with an abundance of nHe = 0.12. On the other hand,
while the fit for helium line at 4471 Å gave nHe =0.19, the helium abundance for helium line at
4921 Å was nHe =0.32.
We performed another test to check if the highest helium abundance calculated by SFIT SYNTH
77
was consistent with the model atmosphere parameters. First step was to compile a new model
grid to include models with helium abundances 0.10 and 0.50 as mass fractions. This computation gave us a helium abundance of 0.14 ([He] = 11.24). The second and the last step was to
check whether this new helium abundance had implications on the previously derived Teff and
log g . Those model parameters were re-calculated by fixing [He] at this new value. It is seen
that the previously derived model atmosphere parameters were consistent with this new [He]
within error limits (± 500 K and ± 0.1 dex). The Teff and log g were decreased 10 K and 0.02
dex respectively. [He] was fixed at 11.24 for the rest of the analysis.
5.2.9 Notes on individual lines
The line identifications for HD119608 are given in Table 5.2, 5.3, 5.4, and 5.5. There are some
regions in the spectrum which are not included in the fit. These regions are 4245-4248, 42724300, 4377-4385, 4485-4495, 4550-4580, 4637-4657, 4730-4752 Å .
• CaII K and H lines at 3933 and 3968 Å respectively have interstellar origin so that they
are not included in the synthetic spectrum fitting.
• [4245 − 4248] PIII(3) 4246.7 line was too weak so that not included in the analysis.
• [4272 − 4300] mostly dominated by blend lines of OII(54, 67) and SIII(4) is not successfully fitted.
• [4377 − 4385] OII(102) 4378.030 line appear to be strong in the model spectrum. This is
mostly due to the fact that computed oxygen abundance seems to be high for the line.
• [4485−4495] dominated by the lines of OII(86), A III(7), and FeII(37) so that not included
in determination of oxygen abundance. The model fit gives a higher abundance for FeII
line at 4541.52 Å .
• [4550 − 4580] Model fit for SiIII triplet with multiplet no. 9 (λλ 4813.3, 4819.7, 4829.0
Å) show excellent agreement with observations while SiIII triplet with multiplet no. 2 (λλ
HD119608
78
4552.6, 4567.8, 4574.7 Å) seems to have deeper cores which is in contrast to multiplet
no.9. On the other hand, in normal B-type stars, SiIII multiplet number 2 is subject to
non-LTE effects (Becker & Butler 1990) so that not included in determination of silicon
abundance instead SiIII multiplet number 9 was preferred.
• [4637 − 4657] dominated by blend lines of OII(1), OII(1), and NIII(2) so that not successfully fitted and it is not modelled correctly in LTE. The non-LTE physics of this feature
is not treated in our model calculations for non of the programme stars. It may be seen
in both emission and absorption in some hydrogen-deficient stars e.g. [WC]2 (Leuenhagen et al. 1994). This feature is thought to be unusually strong due to autoionization
broadening of several dielectronic CII(50) lines (DeMarco et al. 1998). So that higher C
abundance is expected.
• [4730 − 4752] dominated by auto-ionization lines of CII(1). It should be noted that CII(1)
lines at 4736, 4738, 4745, and 4747 have been excluded from the fit as they are not present
in the observed spectrum although they are quite strong in the model fit.
• The region of [4412-4423] dominated by the lines of OII(5). NII(55) 4432.74 line is
blend with FeIII(4) 4430.95. OII(35) line at 4443.05 Å is too weak.
• The feature in the region of [4487-4493] consists of blend lines of OII(86)4489.49 with
A III(7) and OII(86)4489.49 with OII(86)4491.25.
• The feature at 4515 Å is an OII line which is too weak and blend with AIII(3) 4512.54
and FeII(37) 4515.34.
• OII(57) line at 4871.58 Å is too weak.
2
Wolf-Rayet type central stars. [WC] is to differentiate from massive Wolf-Rayet stars.
S PECIES
OII
CII
CII
SiIII
HeI
SIII
CaII
OII
OII
HeI
CaII
OII
OII
SIII
NII
HeI
HeI
HeI
NII
NII
NII
OII
OII
OII
OII
OII
OII
OII
OII
SiIV
OII
NIII
OII
OII
SiIV
OII
HeI
SiII
SiII
OII
HeI
AlIII
OII
Table 5.2: Line identification for HD119608.
λ Multiplet
EW error(2σ)
Notes
(Å)
No. (mÅ)
(mÅ)
3911.96
17
115
±6
3918.97
4
bl
3920.67
4
bl
3924.44
a
bl HeII(4)3923.48
3926.53
bl
3928.62
8
3933
int
3945.048
6
123
±8
3954.37
6
bl
3964.72
bl
3968
int
3973.26
6
3982.71
6
bl SIII(8)3983.77+3985.97
3985.97
8
3994.99
12
249
±6
4009.27
271
±5
4023.97
bl HeI
4026.18
770
±7
bl HeI
4035.08
39
55
±4
4041.31
39
79
±4
4043.53
39
56
±5
4048.21
50
11
±3
4062.94
50
bl OII(97)
4069.89
10
4072.16
10
bl OII(10)
4075.86
10
bl OII(10)
4078.84
10
bl OII(10)
4085.11
10
bl OII(48)
4087.15
48
bl OII(48)
4088.86
48
bl OII(48)4089.29
4092.93
10
4097.33
1
bl Hδ+OII(20,48)4097.26
4105.0
20
bl Hδ
4110.78
20
w + bl Hδ, OII(21)
4116.22
1
73
±5
4119.22
20
bl HeI
4120.84
bl OII(20)
4128.05
3
bl OII(19)
4130.88
3
bl
4132.8
19
bl NII(43.01)
4143.76
474
±9
FeIII(119,232) on bw
4150.0
5
bl AIII(5)
4153.30
19
133
±4
bl SII(44)4153.09
79
80
HD119608
S PECIES
λ Multiplet
EW errors(2σ)
Notes
(Å)
No. (mÅ)
(mÅ)
OII 4156.53
19
43
±4
SII 4162.69
44
bl CIII(21)
FeIII 4164.73
118
52
±5
bl FeIII(232)+FeIII(118)
FeIII 4166.86
118
12
±4
bl FeIII(118) 4164.73
HeI 4168.97
52
99
±4
NII
4171.6
43
27
±3
NII 4176.16
43
bl FeIII(232)
NII 4179.67
50
OII 4185.45
36
102
±4
OII 4189.79
36
110
±4
NIII 4195.76
6
bl OII(42)
HeII 4199.87
bl NIII(6)
PIII 4222.15
3
69
±5
NII 4227.74
33
60
±5
NII 4236.91
48
109
±7
NII 4241.78
48
107
±6
PIII
4246.7
3
18
±4
SIII 4253.59
4
202
±7 bl OII(101)4253.98+4253.74
CII 4267.15
6
295
±6
OII 4273.10
68
∗
OII 4275.99
68
bl + ∗
OII 4276.75
54,67
bl + ∗
OII 4281.32
54
bl
OII 4282.96
54
bl
SIII 4284.99
4
bl OII(78)4285.70
OII 4288.82
54
OII 4294.79
54
bl
OII 4303.82
54
113
±10
OII 4307.23
53
bl
OII 4313.44
78
OII 4317.14
2
OII 4319.63
2
OII 4325.76
2
bw Hβ
SIII 4332.71
4
bw Hβ+OII(65)4332.71
OII 4345.56
2
rw Hβ
OII 4347.42
16
45
±3
bl OII(2)
OII 4349.43
2
rw Hβ
OII 4351.26
16
bl OII(2)
SIII 4361.53
4
OII 4366.89
2
bl
OII 4378.03
102
t + bl
OII 4378.41
102
t + bl
HeI 4387.93
459
±9
FeIII 4395.78
4
id:OII 4395.95(26)
S PECIES
OII
OII
HeI
OII
OII
HeI
MgII
OII
OII
AlIII
AlIII
SiIII
SiIII
SiIII
OII
OII
NII
NII
OII
NII
CII
NII
NII
NII
OII
OII
OII
OII
SiIV
OII
OII
OII
OII
OII
OII
OII
HeI
SiIII
OII
NII
NII
NII
SiIII
SII
λ Multiplet
EW errors(2σ)
Notes
(Å)
No. (mÅ)
(mÅ)
4414.90
5
bl OII(5)
4416.97
5
bl OII(5)+FeII(27)+FeIII(4)
4437.55
50
94
±5
4448.21
35
bl NII 4447.03(15)
4452.37
5
73
±5
4471.50
910
±11
4481.21
259
±7
bl AIII 4479.89(8)
4489.49
86
bl t+AIII(7)+OII(86)
4491.25
86
bl OII(86)
4512.54
3
57
±4
4529.20
3
169
±6 bl AIII(3)+NII(59) 4530.40
4552.6
2
447
±6
4567.8
2
405
±5
4574.8
2
245
±5
4590.97
15
202
±5
4596.18
15
142
±4
4601.48
5
108
±4
4607.16
5
76
±4
bl OII 4609.42(93)
4609.42
93
61
±5
bl NII 4607.16(5)
4613.87
5
93
±5
4619.23
50
id ∗∗
4621.39
5
bl OII 4621.24(92)
4630.54
5
199
±5
4634.14
2
t
4638.86
1
bl CII(12.01)
4641.81
1
bl OII(1)+NIII 4640.64(2)
4649.13
1
bl OII(1)+CIII(1) 4647.40
4650.84
1
bl OII(1)
4654.14
7
4661.64
1
198
±5
4673.73
1
bl OII(1)
4676.24
1
bl OII(1)
4696.4
1
bl OII(25)
4699.22
25
bl
4705.35
25
bl
4710.0
24
bl
4713.17
bl SII(9)
4716.65
b
bl HeI+SII(9)4716.2
4751.3
24
4779.7
20
26
±3
4788.13
20
54
±4
4803.29
20
86
±3
4813.30
9
70
±3
4815.51
9
bl SiIII 4813.30(9)
81
82
HD119608
S PECIES
λ Multiplet
EW errors(2σ)
Notes
(Å)
No. (mÅ)
(mÅ)
SiIII
4819.7
9
87
±3
SiIII 4828.96
9
103
±3
SII
4885.6
15
24
±3
OII 4890.86
28
52
±3
OII 4906.83
28
68
±5
SII
4917.2
15
bl HeI
HeI 4921.93
bl SII(15)
OII
4924.6
28
bl SII(7)
OII 4941.07
33
bl
OII 4943.00
33
bl
OII
4955.8
33
16
±4
• bl : Blend with another line; bw : On the blue wing of another line; rw : On the
red wing of another line; id : Identifiation is dubious; t : Line is too weak; int : Has
interstellar origin.
• (∗) OII 4273.10 , 4275.99 mult. no:68, OII 4276.75 mult. no:54,67 are blend/not
resolved.
• (∗∗) Kilian et al. (1991)
• a, b : According to the Moore (1945), it is unclassified line. ILLSS catalogue does not
give a specific multiplet number for the line.
• NII 4432.73 and 4433.48 are members of mult. no:55 and are blend so that they appear
in the spectrum but not included in the table.
• OII 4443.05 (35) appears in the spectrum but not included in the table and is blend with
NII 4441.99(55)
5.3 Discussion
5.3.1 Discrepancy between photometric and spectroscopic Teff
Photometric temperatures of the star were presented in Table 3.1, Chapter 3. It can be clearly
seen that while the methods based on B-V colour of the star gave photometric temperature ≈
23 000 K others based on Strömgen indices gave ≈ 26 000. Our model atmosphere analysis
shows that HD119608 has a temperature of 23 000 K . It should be noted this temperature
seems to be in agreement with Conlon (1992), who gave 24 000 K .
In order to explain this discrepancy on the basis of chemical abundances, we computed a model
5.3. Discussion
83
atmosphere with Teff = 26 000 K and log g = 3.0 with the same vt and vsini being 6 and
55 km/s respectively. The best model fits for both 23 000 K and 26 000 K temperatures are
presented in the Figure 5.1, 5.2, and 5.3. As can be seen from those figures, most of the regions
in the spectrum match except the Balmer lines. The CII doublet at λλ3920 Å is fitted well
by the former model atmosphere. The [4050-4095] region seems to be well presented by the
same model. The [4140-4230] region including HeI line at λλ4143 Å as well. The Balmer
line profile at 4101 Å is fitted well by the 23 000 K model. HeI lines at λλ 4388 and 4471 Å
are also in excellent agreement with these two different temperature model atmospheres. These
latter helium lines are known to be less sensitive to microturbulent velocity and give relatively
more reliable helium abundance. The wings of the SiIII triplets at λλ4550 Å are fitted well
by both models however cores are not filled up. Hence, the silicon abundance was determined
from silicon triplet multiplet no.9 which is fitted well with the 23 000 K model atmosphere (see
Section 5.2.6 for details). The [4625-4720] region is dominated by blend features of carbon and
oxygen.
Generally speaking, in terms of quality of the fits, we have two different temperature models
which can not be easily distinguished visually (by eye). The more rational way of differentiating
these two models would be making this comparison in terms of chemical abundances which
draw two different picture from evolutionary point of view for the star.
Both model atmospheres give almost solar carbon abundance if we estimate the error in the
abundances due to the adopted model atmosphere parameters as 0.3 dex which was used as
step size in grid interpolation during optimization of the best model fit. On the other hand,
while the 23 000 K model gave a solar nitrogen abundance, the 26 000 K model indicated to
an overabundance in this element (≈0.6 dex). In the same way, silicon and sulfur abundances
seem to be normal for both models. Mg shows overabundancy (≈1 dex) in 26 000 K model.
Sulfur was slightly deficient in the 23 000 K model while Al and Fe found to be overabundant
in the 26 000 K model.
84
HD119608
5.3.2 Evolutionary status of the object
When metalicities are compared, the 23 000 K model indicates a solar metallicity. The 26 000
K model gives a metal-rich composition.
Adopting the former model implies that photospheric abundances do not indicate a CNO processed surface (which is accounted for an overabundancy in nitrogen at expense of carbon).
It does not show helium enriched surface composition either and carbon is solar. We did not
observe any s-process element in the star.
The luminosity of the star - We estimated the mass of the star as 0.598 M⊙ from post-AGB
evolutionary tracks of Schönberner (1983, 1987). The gravity is converted to a radius of ≈4
R⊙ which in turn gave log L/L⊙ ≈ 3.6 with Teff from our model atmosphere analysis. This
luminosity is typical for a post-AGB star.
In contrast to above model atmosphere results, a high metallicity does not match with the star’s
current evolutionary state. So that we believe that our model atmosphere parameters represent
the observed optical spectrum best.
5.3.3 The radial velocity of the star: rapid rotation and binarity
It is interesting to note that Martin (2004) computed a heliocentric velocity of -28 km/s and this
value is significantly different from the value of 26 km/s listed in SIMBAD. Our analysis gave
28 km/s which is in agreement with the other published measurements in the literature.
The difference in the radial velocity might be indicating to a binary nature of the star. This
might be also an explanation for the observed discrepancy in abundances (e.g. [He]) for some
elements in the spectrum.
The spectrum of the star is tested by both spectrosocopic and photometric temperatures. Both
high and low temperature model atmosphere fits are not presenting the observed spectrum of
this rapidly rotating star perfectly. This could be understood if star had a companion.
There are several mechanisms which have been invoked to explain such a rapid rotation among
AGB stars. For instance, in AGB star V Hya, Barnbaum et al. (1995) suggests a common
5.3. Discussion
85
envelope binary evolution while Sahai et al. (2003) and Hirano et al. (2004) indicate to an
unseen compact companion. The latter one needs an accretion disk to be developed and has
been proposed to explain the rapid rotation observed in K648 in the M15 by Alves et al. (2000).
ZNG-1 in globular cluster M10 is another example post-AGB star for such a high rotation (170
km/s). However, when the results of our abundances analysis for HD119608 are compared to
these rapidly rotating post-AGBs, one can conclude that there is no such a unique abundance
pattern valid for all rapid rotating post-AGBs.
Chapter 6
High resolution spectroscopy of two
helium-rich hot post-AGB stars
In this chapter, we report model atmosphere analyses of the B-type giant Hen 3-1428 – IRAS173114924 and rapid rotating super-giant LSS4331 – IRAS17381-1616. Emission profiles for Hβ, Hγ,
and Hδ indicate on-going mass-loss in both stars. Generally both spectra resemble each other
except for some unidentified lines in the spectrum of Hen 3-1428. Our analysis indicates that
both stars are helium rich. The spectrum of LSS4331 indicates a vsini similar to helium-rich
super-giant HD119608.
6.1 Background
6.1.1 Hen 3-1428
Hen 3-1428 was classified as a post-AGB source on the basis of IRAS data by Parthasarathy &
Pottasch (1989) and Parthasarathy (1993a). The star shows 30 µm feature and emission of SiC
at 11.5 µm .
Ultraviolet (IUE) spectra of the star show CII (1335 Å ), SiIV (1394 Å , 1403 Å ), CIV (1550
Å ) and NIV (1718 Å ) lines which are typical for the central stars of planetary nebulae. Hen
87
88
Chapter 6. High resolution spectroscopy of two helium-rich hot post-AGB stars
3-1428 was also suggested as an object in transition from AGB to PN by van der Veen et al.
(1989). A photometric distance of 2.6 kpc was derived by Kozok (1985b). Loup et al. (1990)
detected CO emission in the star which is again typical of circumstellar shells around evolved
objects. They found a distance of 1.1 kpc and an expansion velocity of 11 km/s. In another
work, Nyman et al. (1992) found an expansion velocity of 14.1 km/s. The 30 µm feature was
recognised by Szczerba et al. (1999). Hrivnak et al. (2000), Volk et al. (2000, 2002) managed
to resolve 26 µm and 30 µm features. It was suggested that the 30 µm feature was due to
magnesium sulphide (MgS) (Goebel & Moseley 1985). Hony et al. (2002) also identified MgS
as a carrier of both features. Forrest et al. (1979) report that the 30 µm feature is not seen in
oxygen-rich sources. Since this feature is only seen in carbon-rich objects Volk et al. (2002)
suggested a carbonaceous material as a carrier. According to Gauba & Parthasarathy (2003) the
star suffers circumstellar extinction significantly ( E B−VC.S. = 0.39) and they note the presence
of a dusty disk around the star. They also report that IUE spectra of the star indicate a high
stellar wind velocity which is typical for post-AGB mass-loss. Gauba & Parthasarathy (2004)
modelled Hen 3-1428’s spectral energy distribution by using graphite and silicon carbide (SiC)
and they could not obtain a fit in the ≈20–30 µm region.
The star was classified as B1IIe by Parthasarathy et al. (2000a) and their estimation for Teff and
log g on the basis of this spectral classification were 20 300 K and 3.0 respectively.
Sarkar et al. (2005) obtained a high resolution optical spectrum of the star from 4900 Å to 8250
Å but they could not go shortward of 4900 Å due to the spectrograph setup. Because of the
small numbers of CII, NII, OII and iron lines, they were unable to determine the atmospheric
parameters or chemical composition of the star.
6.1.2 LSS4331
The SIMBAD database associates LSS4331=IRAS17381-1616 with PN G010.2+07.5 in the
Strasbourg-ESO Catalogue of Galactic Planetary Nebulae (Acker et al. 1992). Preite-Martinez
(1988) included the star in a sample of PN candidates selected from the IRAS Point Source
Catalogue on the basis of its far-infrared colours and reported that IRAS17381-1616 was not a
PN but a PN-related object. Parthasarathy et al. (2000a) classified the star as a B1Ibe star and
6.1. Background
89
IRAS
17311-4924
Name
Hen3-1428
l
341.41
b
-9.04
BJ
11.16
(1)
VJ
10.74
(1)
Star
References
Hen3-1428
±
This work
Teff
K
24 000
1000
20 300
log g
cm/s2
2.8
0.1
3.0
vt
km/s
10
vsini
km/s
22
vrad
km/s
40
Star
Hen3-1428
±
SUN
H
12
He
11.22
0.01
11.00
C
8.29
0.01
8.52
N
8.19
0.01
7.92
O
9.07
0.01
8.83
Mg
7.73
0.02
7.58
PAR00a
12.00
Al
6.18
0.04
6.47
Si
7.75
0.01
7.55
P
5.12
0.09
5.45
S
6.82
0.02
7.33
Fe
7.39
0.02
7.50
• PAR00a: Parthasarathy et al. (2000a).
• II: Reference for IRAS colours: NASA Ref. Publ., 1190, 1 (1988)
• Johnson BJ and VJ magnitudes were derived from Walraven VBLUW photometry (van der Veen, 1989).
Table 6.1: Atmospheric and fundamental parameters for IRAS17311-4924 in the literature.
Photospheric chemical abundances of IRAS17311-4924 given as logn, normalized to logΣµn=
12.15. It is compared to solar abundances (Grevesse & Sauval 1998).
noted the presence of Hβ and Hγ emission lines in the spectrum. According to Parthasarathy
et al. (2000a), the star is most likely a post-AGB star similar to LSII+34 26 (Parthasarathy
1993b).
Umana et al (2004) did not detect any radio source at the expected optical position of the star
in a VLA interferometric search for nebulosity around a sample of hot of post-AGB stars but
they dedected a 1.42 ± 0.05 mJy radio source at a radio position of α = 17:41:00.05 and δ
= -16:18:12.45. The corresponding shifts in α and δ were 2.16” and 16” respectively. They
report that radio positions are shifted by only 2.16” in α and 1.15” in δ with respect to IRAS
coordinates, inside the IRAS ellipse error (34” X 8”). So that this radio source is associated to
the IRAS source. The star also is included in the list of sources without detectable water maser
emission by Suarez et al. (2007).
90
IRAS
17381-1616
Name
LSS4331
l
+07.49
b
10.26
BJ
10.849
VJ
10.842
Star
References
LSS4331
This work
PAR03
Teff
K
22 000
19 000
±1 000
log g
cm/s2
2.5
2.5
±0.5
vt
km/s
5
vsini
km/s
50
vrad
km/s
-64.5
Star
LSS4331
±
SUN
H
12
He
11.64
0.01
11.00
C
7.28
0.05
8.52
N
7.77
0.03
7.92
O
9.00
0.01
8.83
Mg
8.04
0.06
7.58
12.00
Si
8.35
0.02
7.55
S
6.23
0.20
7.33
Fe
7.92
0.23
7.50
• PAR03: Parthasarathy (2003a).
Table 6.2: Atmospheric and fundemental parameters for IRAS17381-1616. Johnson BJ and VJ
magnitudes were obtained from Tycho magnitudes of the star. Photospheric chemical abundances of IRAS17381-1616 given as logn, normalized to logΣµn= 12.15. It is compared to
solar abundances (Grevesse & Sauval 1998).
6.2 Analysis of Hen3–1428
24 000, and 25 000 K , log g = 2.5 (0.5) 4.0, and helium abundance nHe = 0.10 in the range
3825 – 4990 Å with SFIT SOLVE. A solar metallicity was used as a start. SFIT SOLVE
gave Teff =24 000 and log g =2.8 when this grid used. A new model atmosphere with the same
Teff and log g was calculated with the stellar atmosphere code STERNE and used in SFIT for
spectrum synthesis as well as optimization of the fit. From the analysis of the SiIII triplet around
4550 Å , we derived the microturbulent velocity v t = 10 km/s. The best model fit was found for
Teff =24 000±1000, log g =2.8±0.1, vt = 10km/s and [Fe/H]=0.0 and presented in Figure 6.1,
6.2 and 6.3 with the TIGER normalized spectrum of the star. Abundances and best model fit
parameters with basic information about the star is presented in Table 6.1.
The details of the method have been already presented in Chapter 5 for HD119608. The results of the test for the model atmosphere parameters are presented in Figure 6.4. As can be
seen, the solutions for Teff and log g are in the range 23 000< Teff <24 000 and 2.7<log g<2.9
6.2. Analysis of Hen3–1428
91
respectively. Hence adopted errors in Teff and log g are ±1000 K and ±0.1 respectively.
In order to test best-fit Teff for further abundance analysis, ionization equilibria for SiIII/IV,
SiII/III and SiII/IV were calculated. Since the silicon triplets at 4810 Å were too weak to
obtain a reliable fit, the SiIII multiplet at 4550 Å was used instead. We used the SiIV line
at 4088 and 4116 Å first separately and then in combination. The Teff values for ionization
equilibria are 22 750, 19 880, and 21 350 K when only SiIV at 4088, 4116 Å and all SiIV lines
used together respectively.
For Figure 6.5, we calculated Teff values for the gravities of 2.0, 2.5 and 3.0. This analysis
defines a wide range for both Teff and log g . Because the hydrogen Balmer lines were mostly
in emission, we could not use the Balmer line fitting technique for the gravity determination. In
addition to synthetic spectrum fit, we used helium lines as a consistency check for gravity.
The spectrum shows HI, HeI and OII absorption lines as the most prominent. Hβ, Hγ, Hδ H8,
and H9 show emissions in the core. Hence their wings were used in the model fit whenever
possible. OII lines around Hδ are reasonably well fitted except 4089 Å . OII blends at 4119 and
4120 Å are also resolved as separate components. The helium line at 3965 Å has a comparable
line strength to nearby Hǫ.
There are unidentified sharp absorption lines at 3957, 3962, 4229, 4232, 4237, and 4300 Å . We
checked the possibility that these are an artifact of data-reduction but they are present in the raw
spectrum. Second we suspected that these lines could be of interstellar origin like CaII 3933,
3968 Å . However, as can be seen in Figure 6.1, they are narrow compared to those interstellar
CaII lines. The probability for those unidentified lines of being formed in a different density
environment stands still1. We could not fit these lines with an LTE model atmosphere.
It is also note worthy that CII line at 4267 Å is in emission so it was not included in the analysis.
The HeI lines at 4026, 4143, 4253, and 4388 Å show an asymmetry and excess absorption in
blue wings. Hence only wings of these lines were used for fitting.
1
Vienna Atomic Line Database (VALD) at http://ams.astro.univie.ac.at shows a CaII line at 4300.45, but this
does not seem to be a likely identification.
92
Figure 6.1: Hen3–1428 (in black) and best model fit (in red). Model and observed spectra were
corrected for radial velocity.
6.2. Analysis of Hen3–1428
93
94
6.3. Analysis of LSS 4331
95
Figure 6.4: χ2 method for checking model atmosphere parameters of IRAS17311-4924. Teff
in x-axis presented in 1000 K units. The red represents the result when the whole spectrum is
used while the results when the spectrum divided into two halves are shown in black.
The magnesium abundance is calculated from the MgII line at 4481 Å which is blended with
an AlIII line in the blue wing.
The emission feature at 4815 Å is a SII line. The members of SiIII triplet multiplet no.9 are
weak and so are not used in the silicon abundance determination.
6.3 Analysis of LSS 4331
24 000, 25 000 K , log g = 2.5 (0.5) 4.0, and helium abundance by number nHe = 0.10 in the range
of 3920 - 5187 Å with SFIT SOLVE. A solar metallicity was used as a start. SFIT SOLVE gave
Teff = 22 000 and log g = 2.6 when this grid used. The best model fit was calculated by using
the closest model atmosphere in the grid, Teff = 22 000, log g = 2.5, vt = 5 km/s and [Fe/H]=0.0
96
Figure 6.5: Ionization equilibrium for Si II/III, Si II/IV, and Si III/IV.
and is presented in Figure 6.6, 6.7 and 6.8 with the TIGER normalized spectrum of the star.
6.3.1 Radial Velocity
Since the number of the metal lines in the spectrum was relatively low and the Balmer lines
are in emission, we used the best model fit to cross-correlate with the observed spectrum. The
heliocentric radial velocity of the star was found to be v rad =-64.5 km/s (Table 6.2).
6.3.2 Rotational Velocity
The rotational velocity was found to be 55 km/s from SiIII triplet multiplet no.2. around 4550
Å.
The CaII line at 3933 Å is strong and broad and of interstellar origin. The hydrogen Balmer
lines at 3970, 4100, 4340 and 4860 Å show P-Cygni like emissions in their cores. The red
6.4. Discussion
97
wings of the Balmer lines were fitted better. The [SII] line at 4069 and [FeII] line at 4287 Å are
forbidden lines and probably arise in a nebula. The OII lines around 4350 Å are not resolved
because of the relatively high rotation velocity. The OII doublet (multiplet no.2) at 4317 and
4319 Å corroborates this. The emission feature at 4359 Å is suspected to be [FeII], a member
of multiplet no.7.
Magnesium abundances was calculated from the MgII line at 4481 Å . Aluminum and phosphorus abundances could not be determined. The AlIII line at 4479 is not resolved. The AlIII lines
at 4512 and 4929 Å are not present in the spectrum.
6.4 Discussion
6.4.1 Abundances
Hen3–1428: The slightly N, O rich and C deficient structure of the photoshere supports the
idea that the star might have experienced third-dredge-up. The observed underabundances in
Al, P, and S with Mg and Si being underabundance might be indicating to a gas-dust fractination
in the stellar atmosphere.
LSS 4331: The star shows large deficiency in C and S. Mg and Si are overabundant. Fe is
≈0.4 dex above solar. However, since the spectrum used in the model atmosphere analysis did
not have a high signal-to-noise ratio, the overabundance of Fe is unreliable.
6.4.2
Teff derived from photometry
Hen3–1428: As presented in Table 3.1 in Chapter 3, the photometric Teff = 7648 and 3696
K are not in agreement with that obtained from the spectrum. These temperatures were calculated using broadband photometry. These methods rely on knowing the extinction so an error
due, for example, to non-standard extinction could have caused such a discrepancy. The presence of HeI clearly shows the star is much hotter. If we approach the problem in an opposite
way and assume that our Teff is correct, then GUL89 would indicate E(B-V)=0.66. Using
NAP93 would have given E(B-V)=1.2. In either case, we conclude that the star might be sur-
98
Figure 6.6: LSS 4331 (in black) and best model fit (in red). Model and observed spectra were
corrected for radial velocity. Observed spectrum was binned to a pixel size of 0.15 Å .
6.4. Discussion
99
Figure 6.7: LSS 4331 and best model fit.
100
Figure 6.8: LSS 4331 and best model fit.
6.4. Discussion
101
rounded by a thick circumstellar envelope. A dusty disk around the star was detected by Gauba
& Parthasarathy (2003) with a circumstellar E(B-V)=0.39.
Sarkar et al. (2005) identified a diffuse interstellar band (DIB)2 at 5780.410 Å . They estimated
interstellar E(B-V) = 0.20 from DIB intensity (Herbig 1993) in agreement with E(B-V)≈0.22
Schelgel et al. (1998).
LSS 4331: The Teff found from photometric analysis using NAP93 agrees well with our
analysis.
6.4.3 Evolutionary status of the objects
The question is whether these stars are true post-AGB stars, or something else.
Hen3–1428:
The question of whether the star is a Luminous Blue Variable (LBV) rather than a hot post-AGB
star remains unclear since both classes have very similar spectra. LBVs are generally found in
the Galactic disk and known to be associated with star forming regions. The similarity of the
spectrum of IRAS17311-4924 to the hot post-AGB stars IRAS18062+2410 (Parthasarathy et
al. 2000b) and IRAS01005+7910 (Klochkova et al. 2002) and its moderately high galactic
latitude support the idea that IRAS17311-4924 is a hot post-AGB star. Moreover, it does not
belong to a star forming region.
The presence of P-Cygni like emission features in the spectrum, together with CO emission
lines (Loup et al. 1990) and the IRAS infra-red excess fluxes strengthen the argument that
IRAS17311-4924 is a post-AGB star with an outflow and a dusty disk.
Although Gauba & Parthasarathy (2004) found a carbon-rich circumstellar dust in the ISO
spectrum, 2005 AAT/UCLES observations indicate a slight underabundance of carbon.
LSS 4331:
With Teff = 22 000 K and log g = 2.5 this star lies close to the post-AGB evolutionary tracks of
Wood & Faulkner (1986) for a post-AGB mass of 0.79 M⊙ . The high radial velocity of the star
2
DIBs are absorptions features in the spectra of reddened stars. They may formed in either circumstellar or
interstellar medium.
102
is typical for a post-AGB star.
Chapter 7
High resolution spectroscopy of the hot
post-AGB stars:
LSIV-04 01, LSIV-12 111, LSS5112, and
LB3116
In this chapter, we present abundance analyses for the hot post-AGB stars LSIV-04 01, LSS5112,
LB3116 and LSIV-12 111 as a very metal-deficient star , a N-normal star , an extremely C-poor
star and a young halo planetary nebula respectively. We start by reporting previous studies of
these stars. Our new data are substantially better than those on which previous abundances were
based. These are reported in Tables 7.1 – 7.4. All stars show metal deficiency in their spectra.
7.1 Background
7.1.1 LSIV-04 01
Nassau & Stephenson (1963) gave a spectral type B7Ib for LSIV-04 01. McCausland et al.
(1992) obtained a high-resolution spectrum of the star as part of a continuing program to identify
103
Chapter 7. High resolution spectroscopy of the hot post-AGB stars:
104
LSIV-04 01
l
14.40
Star
References
LSIV-04 01
This work
MCDK92
KP90
Star
LSIV-04 01
±
SUN
H
12
12.00
b
22.86
m1
-0.023∗
c1
1.012∗
(b-y)
0.262∗
Teff
( K)
15 000
11 000±1000
10 000
11 000:
11 000
log g
(cm/s2 )
2.5
2.0±0.2
1.75
<2.0?
2.0
vt
(km/s)
5
15
vsini
(km/s)
5
He
8.68
0.02
11.00
C
<5.79
0.09
8.52
N
<5.30
2.15
7.92
vrad
(km/s)
90
1
2
3
0?
0
O
<5.0
fixed
8.83
Notes
Mg
6.32
0.01
7.58
Si
5.99
0.01
7.55
Ca
5.18
0.02
6.36
Fe
5.45
0.30
7.50
• * : Kilkenny & Busse (1992); KP90 : Kilkenny & Pauls (1990); MCDK92: McCausland et al. (1992);
1 : From Kiel diagrams; 2 : from line profile fitting - Hγ; 3 : from line profile fitting- Hδ .
Table 7.1: Atmospheric and fundamental parameters of LSIV-04 01. Photospheric chemical
abundances of the star are given as logn, normalized to logΣµn= 12.15. These are compared
to solar abundances (Grevesse & Sauval 1998). Broadband photometry in the literature is also
included.
young early-type high galactic latitude stars.
In a study of Reticon spectroscopy with a 100 Å mm−1 linear resolution and Strömgen uvby
photometry of high-latitude, apparently normal early type stars selected from Luminous Stars
in Northern Milky Way catalogue, Kilkenny & Pauls (1990) gave a range of stellar distances
9 to 18 kpc and of z distances 3.5 to 7 kpc if the star was assumed to be a super-giant. They
reported that LSIV-04 01 had sharp Balmer lines. The CII line at 4267 Å was present in the
low-resolution spectrum of the star. It was also noted that the HeI line was absent indicating a
late B-type star. The vsini for LSIV-04 01 could not be estimated because these measurements
were based on HeI lines.
7.1.2 LSIV-12 111
LSIV-12 111 was originally classified as an emission-line B(Be)-type star by Wackerling (1970).
Kilkenny & Pauls (1990) classified the star as a B0e-type star on the basis of a low resolution (≈
3.5 Å ) spectrum. Since the resolution of the spectrum was too low to determine the evolutionary
status of the object accurately, they concluded that LSIV-12 111 might be either a central star
7.1. Background
105
of a planetary nebula or a distant young halo B0e-type star. They noted that Hβ and Hγ were
in emission and the forbidden line of [OII] at 3727 Å was present, indicating a weak planetary
nebula. OII 4639-42 Å and CIII 4647-51 Å suggested a spectral type of ≈ B0. They searched
Palomar Observatory sky survey plates for a sign of nebulosity but it could not be seen.
In a study of Reticon spectroscopy with a 100 Å mm−1 linear resolution and Strömgen uvby
photometry of high-latitude and apparently normal early type stars selected from the Luminous
Stars in Northern Milky Way catalogue, Kilkenny & Pauls (1990) assumed that if the star was
a main-sequence B0 star then the absolute magnitude would be ≈ -4 (Schmidt-Kaler 1982). If
E(b-y)≈0.34 from the (b-y)/c1 calibration in the latter paper then the distance would be ≈ 6 kpc
and the distance from the galactic plane (z), z≈-2 kpc. In the case of being a planetary nebula
central star, they indicated that the stellar distance would be much less and in a agreement
with the galactic latitude of the star l=-21.26. Visual magnitudes obtained by the group on
two consecutive nights showed that the star was variable by at least a tenth of a magnitude
(V=11.28 and V=11.41). If the star was a Be star then this variation might not be significant
but for a planetary nebula this would indicate binarity.
The star is associated with IRAS source IRAS19590-1249 and classified as a young halo planetary nebulae with a dust temperature of 100 - 200 K (Conlon et al. 1992). According to
McCausland et al. (1992) and Conlon et al. (1993a,b), the chemical composition and atmospheric parameters of the star are consistent with a post-AGB evolutionary state. However,
the carbon abundance given by McCausland (1992) seems to be different from that of Conlon
et al. (1993) since the same group in the latter paper found carbon severely under-abundant.
According to Conlon et al. (1993) this difference was due to CII 4267 Å which is known to
be sensitive to non-LTE effects in the atmosphere. This sensitivity was partly corroborated by
Eber & Butler (1988) in their non-LTE model atmosphere calculations. Conlon et al. (1993)
also report that the strength of the observed line might have been reduced by nebular emission
because of the fact that the line is a well known recombination emission line. They add that
an abundance derived from this line is generally higher than those found from UV lines (Aller
& Czyzak 1983). Parthasarathy (1990, 1993a) classified the star as a hot post-AGB star. Both
Conlon et al. (1993) and Garcia-Lario et al. (1997) suggest that LSIV-12 111 is a hot post-
106
LSIV-12 111
l
29.18
b
-21.26
m1
-0.071∗
c1
-0.043∗
(b-y)
0.180∗
Star
Ref.
LSIV-12 111 01
This work
RYA03
MCDK92
Teff
( K)
20 000
20 000
23 750
log g
(cm/s2 )
2.5
2.5
2.7
vt
(km/s)
201
20
18
vsini
km/s
60
vrad
(km/s)
74
Star
LSIV-12 111
±
SUN
H
12
He
10.72
0.01
11.00
C
7.61
0.02
8.52
N
7.46
0.01
7.92
O
8.82
0.01
8.83
Mg
7.21
0.02
7.58
12.00
Al
5.42
0.04
6.47
Si
7.22
0.01
7.55
S
6.0
−−
7.33
Fe
<6.50
−−
7.50
• *: Reference for Stromgen colours: Kilkenny & Busse (1992); **: 2MASS All-Sky Point Source Catalog
(PSC) - ADS/IRSA.Gator 2007/0906/094224 8503; Gar97: Garcia-Lario et al. (1997)
• 1 : vt=20 km/s is adapted from Ryans et al. (2003); MCDK92: McCausland et al. (1992); RYA03: Ryans
et al. (2003).
Table 7.2: Atmospheric and fundamental parameters of LSIV-12 111. Photospheric chemical
abundances of the star are given as logn, normalized to logΣµn= 12.15. These are compared
to solar abundances (Grevesse & Sauval 1998). Broadband photometry in the literature is also
included.
AGB star evolving to a low excitation planetary nebula. The recent LTE and non-LTE studies
(Mooney et al. 2002, Ryans et al. 2003) confirmed LSIV-12 111 to show carbon deficiency.
Parthasarathy (2000a) gives a spectral classification of B1Ibe. A dust temperature of 120 K
was estimated by Gauba & Parthasarathy (2004) in good agreement with Conlon et al. (1993b)
as well as Bogdanov (2003).
In a study of H2 emission from 51 proto-planetary candidates (PPN), Kelly & Hrivnak (2005)
observed Balmer lines in emission. They did not detect H2 emission from the star. They suggest
that LSIV-12 111 may have undergone a He thermal pulse in the post-AGB phase in which case
it has looped back to the PPN phase (Blöcker 1995).
7.1.3 LSS 5112
LSS5112 was originally included in the Stephenson-Sanduleak OB star survey (see Stephenson
& Sanduleak 1971) as an OB+ star. OB+ stars are hot luminous stars. Their low-resolution
spectra are generally characterized by very weak or absent hydrogen Balmer lines. In addition
7.1. Background
107
to almost all of the optically identifed X-ray sources lying in the Milky Way, some hydrogen
deficient, very hot subluminous O-type stars, some dwarf novae, and FG Sagittae have been
classified as OB+ stars.
Martinez (1988), in a search for new planetary nebulae among unidentified sources of the IRAS
Point Source Catalogue, found no nebula.
Parthasarathy et al. (2000a) classify the star as B1IIIe and conclude it is most likely a postAGB star. The presence of CII(1335 Å), SiIV(1394 and 1403 Å), CIV(1550 Å), NIV(1718Å)
and MgII(2800) in the IUE spectrum of the star, which is also a high galactic latitude star,
supports this. Venn et al. (1998) found it to be a B2.5Ia star, and reported Hδ to be filled in or
to have a weak P-Cygni type profile. Gauba & Parthasarathy (2003) analyzed the IUE spectrum
and reported LSS5112 as a hot post-AGB star with a negligible circumstellar extinction. Their
model atmosphere analysis gave Teff = 19 000±1000 K and log g = 2.5±0.5.
Umana et al (2004) performed VLA interferometric observations of a sample of hot post-AGB
stars and did not detect any radio emission at the optical coordinates of LSS5112, in agreement
with the negligible circumstellar extinction found by Gauba & Parthasarathy (2003).
Gauba & Parthasarathy (2004) derived parameters of a circumstellar dust shell for LSS5112.
They combined optical, near- and far-infrared data of the star to reconstructed its spectral energy
distribution. They noted that the star is variable in the near-infrared based on J, H, K magnitudes
from Garcia-Lario et al. (1997b) and the 2MASS catalogue.
Kelly & Hrivnak (2005) report that two-thirds of the bipolar nebulae show H2 emission. There
seems to be a relation between bipolar morphologies and H 2 emission of PPNe. They include
LSS5112 in their list as a Be-type PPN (B star with Balmer emission). LSS5112 seems to have
a bipolar morphology (Hrivnak et al. 2004) on the basis of the spatial extent of its H2 emission
in the star.
7.1.4 LB 3116
This Luyten blue star was studied by Hill & Hill (1966) as a part of a project involving the
measurement of photo-electric colours of faint blue stars in the Southern hemisphere (Table
108
LSS
5112
Name
l
16.50
b
-5.42
BJ
11.977
VJ
11.977
Star
Ref.
LSS5112
This work
Teff
( K)
25 000
log g
(cm/s2 )
3.0
vt
(km/s)
13
vsini
(km/s)
0
vrad
(km/s)
-142
Star
LSS5112
±
SUN
H
12
He
10.91
0.01
11.00
C
7.33
0.02
8.52
N
7.42
0.02
7.92
O
7.95
0.01
8.83
Mg
7.19
0.03
7.58
12.00
Al
5.60
0.08
6.47
Si
6.79
0.02
7.55
P
5.05
0.09
5.45
Ca
6.74
0.03
6.36
Fe
7.00
0.05
7.50
Table 7.3: Atmospheric and fundamental parameters of LSS5112. Photospheric chemical abundances of the star are given as logn, normalized to logΣµn= 12.15. These are compared to solar
abundances (Grevesse & Sauval 1998). Broadband photometry in the literature is also included.
3.15). They gave a spectral type of B4 on the basis of its spectrum. The star was first analyzed
by Kilkenny & Lydon (1986) and re-analyzed by Quin et al. (1992) in UV via IUE observations.
According to Kilkenny & Lydon (1986), LB3116 is a super-giant star located at a distance of 25
kpc. They calculated Teff and log g from optical spectroscopy (Table 3.18). Quin et al. (1992)
report that UV energy distribution of the star matches the atmospheric parameters adopted by
Kilkenny & Lydon (1986). They also show that LB3116 is a metal deficient star in transition
between AGB and PNe.
Quin et al. (1992) suggest to use C/N ratio to test whether LB3116 is a massive super-giant. If
so, its C/N ratio is expected to be either normal (i.e. about solar) or smaller than normal. In most
of the massive population-I super-giants, the C/N ratio is small. On the other hand, post-AGB
stars are expected to show an enhancement in C/N ratio as helium burning products reach the
surface following thermal flashes and 3rd dredge-up (Iben 1984).
7.2. Analysis of LSIV-04 01
LB3116
Star
LB3116
±
Star
LB3116
±
SUN
l
331.56
References
This work
KL
QL
H
12
12.00
b
-27.23
Teff
( K)
16.000
16.500
1500
16.000
He
10.98
0.01
11.00
109
m1
0.060
log g
(cm/s2 )
2.5
2.5
0.35
2.7
C
6.10
0.44
8.52
c1
0.26
vt
(km/s)
0
(b-y)
-0.038
vsini
(km/s)
25
120
25
B-V
-0.12
vrad
(km/s)
75
N
7.84
0.17
7.92
O
8.25
0.29
8.83
Mg
6.50
0.01
7.58
U-B
-0.71
Notes
III
A
5.86
0.01
6.47
Ref.
I,II
Si
6.20
0.10
7.55
P
6.16
0.47
5.45
S
5.97
0.36
7.33
Fe
6.57
0.28
7.50
• I : Reference for Stromgen colours; Kilkenny et al. (1977); II : Reference for B-V and U-B colours; Hill
& Hill (1966); III : From UV spectra; KL : Kilkenny & Lydon (1986); QL : Quin & Lamers (1992).
Table 7.4: Atmospheric and fundamental parameters of LB3116. Photospheric chemical abundances of LB3116 given as logn, normalized to logΣµn = 12.15. These are compared to solar
abundances (Grevesse & Sauval 1998). Broadband photometry in the literature is also included.
7.2 Analysis of LSIV-04 01
A standard grid of model spectra was calculated with Teff = 15 000, 18 000, 20 000, 22 000,
24 000, 25 000 K , log g = 2.5 (0.5) 4.0, and the helium abundance by number nHe = 0.10 in
the range of 3850 - 5010 Å with SFIT2. The best model fit is calculated for Teff = 15 000
K , log g = 2.5, vt = 5 km/s and [Fe/H] = 0.0 and presented in Figure 7.1, 7.2, and 7.3 with
the TIGER normalized spectrum of the star. Abundance and the best model fit parameters are
presented in Table 7.1. The complete line list of the star is presented in Table 7.5.
7.2.1 Metallicity
Overall metalicity of the star is calculated on the basis of its Si and Fe abundances. We used
average ([M/H]) of [Si/H]=-1.56 and [Fe/H]=-2.05 as metallicity of the star, [M/H]=-1.81. By
comparison, magnesium and silicon are depleted by ≈-1.3 and ≈-1.5 dex respectively, indicating
an overall metallicity [M/H]≈-1.5
The determination of vt was not a straightforward process for LSIV-04 01 due to lack of sufficient lines of given species in the spectrum. For this reason, McCausland et al. (1992) as-
110
summed vt = 15 km/s and claimed their choice of vt was consistent with that given in Underhill
& Fahey (1973) for low gravity stars. We adopted v t = 5 km/s. Adoption of a larger vt would
indicate lower abundances. Indeed, when vt was assumed to be 15 km/s, our model atmosphere
analysis indicated 0.02, 0.04, 0.09, and 0.05 dex depletions in He, C, Mg, and Si abundances
respectively while N abundance was enhanced by ≈0.2 dex and the Ca abundance was found to
be lowered by ≈ 0.4 dex.
The results of the tests for both Teff and log g are presented in Figure 7.4. The solutions for Teff
and log g are in the range 15 000< Teff <17 000 and 2.3<log g<2.5 respectively. Hence adopted
errors in Teff and log g are ±1000 K and ±0.1 respectively.
7.2.4 Chemical composition
Carbon, nitrogen, oxygen - 1999 AAT observations of the star indicate that this star is depleted
by He burning products such as C, N, O since there are no detectable lines of C, N, and O.
CII line at 4267 Å is not present (too weak) in the spectrum although it was reported to be
present by Kilkenny & Pauls (1990) in their low-resolution (100 Å /mm−1 ) spectrum hence the
presence of this feature with sharp Balmer lines was reported to be a result of being a high
luminosity star. Because of the fact that our 1999 AAT/UCLES spectrum of the star has a
much better resolution (≈ 0.1 Å per pixel) compared to their low-resolution spectrum, it is a
possibility that they mis-identified this feature in their spectrum or disappearance of this line
indicates to a real physical event in the star. Our analysis results corroborate the presence of
sharp Balmer lines. We did not detect any C, N, O lines in the spectrum. Since they were too
weak or absent we only set an upper limit for those abundances.
Sulfur - S is not present in the spectrum so we could only fix S abundance as an upper limit.
Silicon - There are a couple of Si II lines in the spectrum. However, the Si III triplets multiplet
no.2 and 9 at λλ 4550 and 4810 Å respectively are absent which may be telling us a different
story. Silicon abundance is determined on the basis of weak Si II lines.
Figure 7.1: LSIV-04 01 (in black) and best model fit (in red).
111
112
Figure 7.2: continued - part II.
113
Figure 7.3: continued - part III.
114
Figure 7.4: χ2 method for checking atmospher parameters of LSIV-04 01.
Magnesium - Mg abundance is measured from the MgII 4481 Å line by matching the line with
SFIT SYNTHE profile. The amount of depletion is ≈ 2 dex. In fact, this MgII line is a multiplet
given by the transitions at λλ 4481.126, 4481.150, and 4481.352 Å and blended with Al III.
The element shows ≈ 2 dex depletion.
7.3 Analysis of LSIV-12 111
Our standard model grid was used to calculate model atmosphere parameters using SFIT . The
best fit gave Teff =20 000 K , log g = 2.5, vt = 20 km/s and [Fe/H] = 0.0 (Figure 7.5, 7.6 and 7.7).
Abundances and best model fit parameters are presented in Table 7.2. The overall metallicity
[Fe/H] = -0.07 based on iron abundance should be considered as an upper limit due to the
weakness of those lines. By comparison, aluminum and sulfur are depleted by ≈-1.1 and ≈1.3 dex respectively, suggesting an overall metallicity [M/H]≈-1.2. We used average([M/H])
of [Al/H] = -1.05, [Si/H] = -0.33, [S/H] = -1.33 and [Fe/H] = -1.00 as metallicity of the star,
S PECIES
SiII
SiII
SiII
HI(H8)
CaII
HeI
CaII
HI
HeI
HI
SiII
115
λ
(Å)
3853.66
3856.02
3862.60
3889.05
3933.66
3964.73
3968.47
3970.07
4026.36
4101.74
4128.05
S PECIES
SiII
FeII
HI
HeI
HeI
MgII
FeII
FeII
FeII
HI
λ
(Å)
4130.88
4233.17
4340.47
4387.93
4471.48
4481.13
4481.33
4508.28
4522.63
4583.83
4861.33
Table 7.5: Line identification for LSIV-04 01. SiII line at 3853.66 Å is weak. HeI at 4026.36
and FeII at 4508.28 and 4522.63 Å are too weak.
[M/H] = -0.93.
Balmer hydrogen lines show P-Cygni like profiles which indicates that LSIV-12 111 might be
experiencing a strong stellar wind and mass-loss. The emission features in the line centers were
problematic, we managed to fit these lines by giving higher weights to unaffected part of the
line profiles while for the regions affected by emission low weights were used. In addition to
P-Cygni like profiles of hydrogen Balmer lines, [SII] 4068.70 Å ,[FeII] 4287.33 Å and CII
4267.18 Å were also observed to be in emission.
It is interesting to note that most of the neutral helium lines are asymmetric. λλ HeI 4026,
4143, 4388, 4471, 4713, and 4921 Å lines show extended blue wings, so that they could not be
modelled succesfully. These blue wings may also be associated with mass-loss.
In additon to [SII], SII 4360.49 Å (multiplet no.1) seems to be in emission.
The [4630 - 4680 Å ] region is mostly dominated by carbon and oxygen lines so that line cores
in this region could not be fitted well. It requires a higher abundance in the corresponding
elements.
The [4690 - 4730 Å ] region is dominated by CII lines which are known to give abundances
different from overall carbon abundances in the spectrum.
116
Figure 7.5: LSIV-12 111 (in black) and best model fit (in red). Both model and observed spectra
are moved to zero velocity reference frame - part I.
117
118
7.4. Analysis of LSS 5112 (IRAS 18379-1707)
119
7.4 Analysis of LSS 5112 (IRAS 18379-1707)
We used our standard grid of model spectra and assumed solar metallicity model to start. SFIT
SOLVE gave Teff = 24 500 and log g = 2.7. The closest model in the grid to these parameters
was used. The best fit for Teff = 25 000, log g = 3.0, vt = 13 km/s and [Fe/H] = 0.0 is presented
in Figure 7.8, 7.9 and 7.10 with the TIGER normalized spectrum of the star. Abundances and
best model fit parameters are given in Table 7.3.
7.4.1 Metallicity
Using Fe abundance, we find [Fe/H]=-0.5. By comparison, Al, and Si are depleted by ≈-0.9
and ≈-0.8 respectively, indicating to an overall metallicity [M/H]=-0.9 and average of [Fe/H],
[Al/H], and [Si/H] gives -0.7 dex.
The indirect affects of mass-loss can be seen in HI lines at 4340 and 4860 Å as P-Cygni like
line profiles while HI lines at 3970 and 4101 Å do not have comparable emission line strengths.
The line profiles for HeI lines at 4143, 4388 and 4471 Å show asymmetry, mostly in the blue
wings while 4921 and 4713 Å HeI lines show interesting profiles.
120
Figure 7.8: LS5112 and best model fit. Both model and observed spectra are moved to zero
velocity reference frame. Observed spectrum was binned to a pixel size of 0.15 Å- part I.
7.4. Analysis of LSS 5112 (IRAS 18379-1707)
121
122
7.5. Analysis of LB3116
123
Figure 7.11: χ2 method for checking model atmosphere parameters of LB3116.
7.5 Analysis of LB3116
A grid of thirty-six model spectra was calculated with Teff = 14 000, 16 000, and 18 000 K ,
log g = 2.0 (0.5) 3.0, and helium abundances nHe = 0.001, 0.05, 0.10, and 0.50 in the range
of 3825 - 5145 Å . A solar metallicity model was used as a start. SFIT SOLVE gave Teff =
16 600 and log g = 2.7 when this grid used. The best model fit is calculated for Teff =15 000
and log g =2.5 and it is presented in Figure 7.12, 7,13, and 7.14 with the TIGER normalized
spectrum of the star. The adopted errors in model atmosphere parameters are ± 1000 Kand ±
0.2 dex. Abundances and the best model fit parameters are given in Table 7.4.
7.5.1 Metallicity
The [Fe/H] of the star is -0.93 dex. Al, Si, and S show depletions by ≈-0.6, ≈-1.4, and ≈-1.4
respectively, indicating to an overall metallicity [M/H]=-1.1. Average metallicity for the star is
[M/H]=-1.1 and calculated using [Fe/H], [Al/H], [Si/H], and [S/H] ratios.
124
The red wing of the HI line at 3889 Å is not fitted well. The model fits to HeI lines at 3926,
4009, 4026, 4121, 4144, 4169, 4388, 4438, 4471, 4713 and 5016 seem to be well fitted. MgII
at 4481 Å is present in the spectrum. Interestingly, there are no CNO lines observable in the
spectrum. For abundances of those not observed an upper limit set and spectrum inspected for
the this choise.
The chi-square(χ2 ) method was employed to test Teff and log g . As first step, SFIT SOLVE
code was used to solve a normalized spectrum of the star on a grid from 12 000 ≤ Teff ≤ 19 000
K for Teff in steps of 500 K . Temperatures were fixed at those values and in turn log g= 2.5
was tested for those temperatures. The χ2 values were plotted against the temperatures. The
minimum of this curve (parabola) gave the most probable value for Teff (Fig. 7.11). As a
second step, the same procedure was applied to log g on a grid from 1.0 ≤log g≤ 4.0 in steps of
0.5 dex with Teff = 15 000 K . The χ2 values were plotted versus gravity. The minimum of this
curves gave the most probable log g otherwise an interval and it is shown in red (Fig. 5.5)
7.6 Discussion
7.6.1 Abundances
LSIV-04 01: The star shows deficiency in C, N, O, Mg, Si, Ca, He and Fe.
LSIV-12 111: This metal poor star is O normal and shows deficiency in C and N as well as
Al and S.
LSS 5112: This helium normal star is C, N, and O deficient. Mg, Al, Si, P are subsolar. Ca
is found to be overabundant, ≈0.4 dex.
LB3116: This He normal star shows deficency in C and O and has a solar N abundance.
While Mg, Al, and S appear to deficient in the photosphere, P is found to be overabundant by
7.6. Discussion
125
Figure 7.12: LB3116 (in black) and best model fit (in red). Both model and observed spectra
are moved to 0 velocity reference frame - part I.
126
Figure 7.13: continuing part-II.
7.6. Discussion
127
Figure 7.14: continuing part-III.
128
≈0.7±0.47 dex. This may be the first example of a post-AGB star which is rich in phosphorus.
7.6.2
Teff derived from photometry
LSIV-04 01
As seen from Table.3.1, the photometric temperatures derived for LSIV-04 01 from Johnson
colours are relatively low while Teff from Strömgen colours gave 11 146±334 K which is
close to our analysis result (10 000 K ).
LSIV-12 111
As seen from the Table.3.1, the Strömgen photometric temperature for LSIV-12 111 is quite
high (36 243±1087). This high temperature for the star was also reported by McCausland et
al. (1992) by using the calibrations of Lester, Gray, & Kurucz (1986). Our model atmosphere
analysis of the star gave Teff = 20 000 K . However, the temperature derived from Johnson UBV
colours ( Teff =19 660 K ) for the star seems to agree with our temperature when E(B-V) from
Bogdanov (2003) is used. The star also has a clear infra-red excess which can be seen in IRAS
colours. That might be indicating a cool companion or the presence of a circumstellar material.
A reason for such a high photometric temperature could be the selection of total extinction.
Conlon et al. (1993) deduced E(B-V)=0.37 from observed (b-y) and intrinsic colours (b-y) 0
(Relyea & Kurucz 1978). They used the relation between E(b-y) and E(B-V) of Crawford
(1978).
7.6.3 Evolutionary status of the objects
Again, the question is to check that the targets are true post-AGB stars.
LSIV-04 01
The radial velocity of the star is 90 km/s so that it is a high velocity star. It is also very metalpoor, making it likely to be a Pop-II star and most probably a post-AGB star.
7.6. Discussion
129
M10 membership - LSIV-04 01 is a member of the globular cluster, M10 which has a distance
of ≈ 4 kpc (Harris 1996). We performed a consistency check for the derived values of the
atmospheric parameters by calculating a distance estimate for the star and comparing it with
that obtained by Harris (1966). For this, we positioned the star in the HR diagram on the basis
of its model atmosphere parameters. Assuming that the star is in post-AGB stage, the mass of
the star can be estimated using evolutionary tracks of Schönberner (1983, 1987). LSIV-04 01
lies close the 0.598M⊙ track.
We used Mendez et al. (1988) distance formula obtained from high-resolution spectrum of 25
PNes to determine a distance of 4.6±1.2 kpc so we think that our results are reliable.
The lower gravity given by McCausland et al. (1992) would translate to a distance ≈8.2±2.1
kpc, which contradicts the cluster’s distance.
Luminosity of the star - When this distance is used in the gravity-radius relation, we obtained
a radius of ≈5x109 m which corresponds to a luminosity of ≈ 2.3x103 L⊙ . We also wanted to
check whether this luminosity could be corroborated by an external method. The luminosity
– core-mass relation given by Jeffery (1988) for the AGB stars with 0.5M ⊙ core-mass gives a
luminosity of 2.5x103 L⊙ which agrees perfectly. However, 0.6M ⊙ core-mass corresponds to a
luminosity of 6.8x103 L⊙ . When the luminosity – core-mass relation of Boothroyd & Sackman
(1985) used, we find 7.4x103 L⊙ .
LSIV-12 111
Distance and luminosity of the star - We estimated the mass of the star by using evolutionary
tracks of Schönberner (1983, 1987) and performed similar type of analysis as we did to LSIV-04
01. LSIV-12 111 lies close the 0.644 M⊙ evolutionary track. This core-mass can be converted
to a distance of 4.45 kpc. Gravity – core-mass relation gave a radius of 7.45 R⊙ . By using
luminosity – effective temperature – radius relation, we found a luminosity of ≈ 104 L⊙ which
is typical for a post-AGB star. The luminosity – core-mass relation of Boothroyd & Sackmann
(1988) indicated to a luminosity of 9.8x103 L⊙ . So that in contrast to LSIV-04 01, this last
relation gave an almost similar result for the luminosity of the star.
130
Radial velocity - The heliocentric radial velocity of the star is 74 km/s. Such a relatively high
radial velocity is a typical characteristic of post-AGB stars.
LSS 5112
The model atmosphere analysis of the AAT/UCLES 1999 spectra of the star does not corroborate the model atmosphere parameters given by Venn et al. (1998). Their analysis was based on
the SiII λλ4130 Å multiplet. They compared the strength of the line to a group of standard stars
studied by Lennon et al. (1992). Crude comparison of our spectrum to Venn et al.(1998) clearly
show that 1999 AAT/UCLES data has a better resolution and lower S/N ratio. This can be easly
seen when Hǫ and Hδ hydrogen Balmer lines are compared to Fig.6 in Venn et al. (1998).
The distance of the star - Assuming that LSS5112 is a post-AGB star, in a similar way to LSIV04 01 and LSIV-12 111, we found a core-mass of 0.598 M⊙ which corresponds to a distance of
≈ 3.8 kpc. The star is found to have a radius of ≈4 R⊙ and a luminosity of log(L/L⊙)=3.76. The
luminosity calculated from Boothroyd & Sackmann (1988) is log(L/L⊙ )=3.87.
Radial velocity - LSS5112 has the biggest heliocentric radial velocity star in the our sample
with a velocity of -142 km/s and this is agreed with the post-AGB classification.
LB 3116
LB3116 is a relatively high radial velocity star. With its high galactic latitude, extreme deficieny
observed in C, Mg, Si, and S indicate LB3116 to be a post-AGB star.
Chapter 8
Conclusions And Future Work
In this thesis, a high-resolution LTE model atmosphere analysis of early B-type post-AGB stars
has been presented. This analysis included obtaining high-resolution spectra, data preparation
including pre-reduction, reduction, and continuum normalization, and using LTE model spectra
computed with Armagh LTE Stellar Atmosphere Code STERNE.
We have used the method of fine analysis to measure the abundances of the standard star
HR1765 and proto-type post-AGB star HD119608, and to estimate the internal errors for these
abundances. Abundance analyses were then carried out for six additional post-AGB stars. In
order to minimise systematic errors, the results are presented as relative abundances.
The model atmosphere parameters of the programme stars, effective temperature ( Teff ), surface
gravity (log g ), and helium abundance (nHe ) have been determined spectroscopically by χ2 minimization. The results for Teff being in the range 15 000 – 25 000 K agrees with the Teff
range for B spectral type being 10 000 - 30 000 K . The log g for those hot post-AGB stars is
in the range 2.5 – 3.0, while nHe is generally solar except Hen3-1428 and LSS4331 which are
helium-rich and LSIV-04 01, which is helium poor. A compilation of the LTE best model fits
with the continuum normalized spectra is presented as a temperature-sequence.
The determination of Teff and log g enabled us to place these early B-type hot post-AGB stars
on the log g – Teff diagram. Their positions agree with the post-AGB classification. It should
131
132
Chapter 8. Conclusions And Future Work
be noted that LSIV-12 111 was already reported to be a young PN (Ryans et al 2003).
The spectra of the programme stars were searched for Sr, Y, Ba, and Hf. Although observations of these were reported in the literature for the cooler counterparts, we did not detect
those s-process elements in the programme stars. This expected finding is mainly due to high
temperature of the programme stars.
In addition, archival photometry has been collected and used to distinguish variability in two
groups of post-AGB stars (see Appendix-B). This work included 12 post-AGB and 7 EHe stars.
We found variability in five of these post-AGB stars. We did not detect variability in EHes. Relatively large variations detected in HR4049, HD213985, and HD52961 appear to be correlated
with the binary period. In the same study, we also conclude that these large variations may be
due to variable absorption by or illumination of circumstellar material. The Hipparcos data for
HD137569 and HD172324 do not show evidence of periodicity.
8.1
Teff – log g diagram
The calculated model atmosphere parameters were used to place the programme post-AGB
stars, one being a young PN, in a Teff – log g diagram with other groups of post-AGB stars,
namely EHes, RCrB, and H-deficient Wolf-Rayet central stars of planetary nebulae. As can
be seen RCrB star are placed far away from the programme post-AGB stars. The similarity of
model atmosphere parameters between post-AGB and EHe stars are interesting to note. The
positions of LB3116, LSIV-04 01 and IRAS17381-1616 coincide with the three EHes within
error limits on the model atmosphere parameters (Figure 8.1).
The positions of these stars are compared with post-AGB evolutionary tracks from Schönberner
(1983) and Blöcker (1995), which can be used to estimate masses1 .
IRAS17311-4924 implies a current mass of 0.625M⊙ while IRAS17381-1616 seems to be close
a post-AGB mass of 0.696M⊙. HD119608 and LB3116 are on and close to the 0.565M⊙ track
1
The post-AGB evolutionary tracks of Schönberner (1983) for post-AGB masses of 0.546, 0.565M⊙ and the
evolutionary tracks of Blöcker (1995) for 0.605, 0.625, 0.696, 0.836, and 0.940M⊙ were kindly provided by
Schönberner.
8.1. Teff – log g diagram
133
3.8
3.9
Post-AGB
WC-PN
Hot post-AGB
4.1
4.2
4.3
it
4.4
4.5
7
4.6
6
5
4
3
2
1
lim
n
to
ing
d
d
E
log g
0
log Teff(K)
ri
Ho
Programme stars
EHe
RCrB
Post-AGB / Group-I
Post-AGB / Group-II
ta
n
zo
M
40
Hydrogen-Main Se
quence
l
a
Br
nc
h
SUN
sun
5 M n
0
6
0.
Msu
65
5
.
0
n
Msu
46
5
.
0
un
9 Ms
0.51
0.9
4.0
B
AG
sun
3.7
3.6
3.5
Figure 8.1: The log g – Teff diagram for post-AGB stars showing the positions of the programme stars with R CrB (Asplund 1997; Lambert et al. 1997) and EHe (Drilling et al. 1998;
Pandey et al. 2001) stars. Hydrogen main sequence, horizontal branch, AGB, and the classical Eddington limit for radiative stability are included. Evolutionary tracks of post-AGB
(Schönberner 1983 for post-AGB masses of 0.546 and 0.565M ⊙ ; Blöcker 1995 for 0.605,
0.625, 0.696, 0.836, and 0.940M⊙; Gingold (1976) for a 0.519M⊙ star) are also shown. 1.
LSS5112; 2.IRAS17311-4924; 3.HD119608; 4.IRAS17381-1616; 5.LSIV-12 111; 6.LB3116;
7.LSIV-04 01.
134
respectively. LSS5112 implied an evolutionary track of 0.605M⊙.
8.2 Helium Abundance
Two of the program stars namely Hen3-1428 and LSS4331 are found to have helium-enriched
photospheres by 0.2 and 0.6 dex respectively. LSS5112 and LB3116 appear to have solar helium
abundances while the metal-deficient post-AGB LSIV-04 01 is found to be helium-poor by ≈2
dex.
A possible scenario for such unusual helium enrichment is as follows: the third dredge-up
is triggered by consecutive thermal pulses which occurs in a helium-rich environment (He
intershell-region) so that an efficient dredge-up process in addition to He products would also
bring some amount of He to the surface. This idea cannot be tested for cooler post-AGB stars
because He is only visible at spectral type B and earlier.
Hen3-1428 shows a slightly enriched N and O abundances but otherwise solar if error in the
abundances due to adopted model atmosphere parameters is assumed to be 0.3 dex. So this
result does not provide evidence for the occurence of a third dredge-up. However, when the
same analysis is performed for LSS4331, within the same error margin (0.3 dex), we find C to
be extremely deficient (≈0.9 dex) and N and O are solar. So that above scenario might be used
in explaining helium enhancement in LSS4331.
8.3 Phosphorus Abundance
One objective of this thesis was to measure post-AGB phosphorus abundances and to compare
this with other groups of post-AGB stars. In particular, EHes are known to show an excess of
phosphorus. We did not detect a significant overabundance in phosphorus for our programme
post-AGB stars, specifically:
1. HD119608: The PIII lines (multiplet no. 3) at 4222.15, 4246.70, and PII line (multiplet
no. 25, 35) 4530.81 Å are observable. The measured equivalent widths for 4222.15
and 4246.70 Å are 69 and 18 mÅ which translates to abundances of 5.82 and 5.27 dex
8.3. Phosphorus Abundance
135
respectively. The 4530.81 is a blend so not used in the analysis. The line analysis for
those phosphorus lines gave a mean abundance of 5.55±0.39. The error given presents
the line-to-line scatter. So we conclude that HD119608 has a solar phosphorus surface
composition.
2. Hen 3-1428: The above phosphorus lines are not observable in the spectrum except PII
4530.81, being a blend and not reliable for phosphorus abundance measurement. The star
shows deficiency in the phosphorus ≈0.2. dex.
3. LSS4331:
The low signal-to-noise ratio in the spectrum of the star did not allow us
to determine the phosphorus abundance from the line analysis. The spectrum synthesis
method allowed us to put an upper limit in abundance of this element as 3.0 dex. So the
observed deficiency in phosphorus is ≥2.45 dex for LSS4331.
4. LSIV-04 01 & LSIV-12 111: The former is a very metal-deficient star. We did not detect
any phosphorus line in the spectrum of this star. The latter is reported to be a young PN.
We fixed the phosphorus abundance at 5.45 dex, the solar abundance, which appears to
be an upper limit for phosphorus in LSIV-12 111.
5. LSS5112: The single- and double-ionized phosphorus lines are not distinguishable from
the noise in this particular region of the spectrum, hence not detectable. The phosphorus
abundance given as 5.05 dex for the star should be considered as an upper limit. So the
observed deficiency in phosphorus abundance would be ≥0.40 dex.
6. LB3116: The error in surface abundance of phosphorus is ±0.47 dex. Those particular
regions including phosphorus are not resolved well from background noise so that the
abundance of phosphorus is only an upper limit, even though it is 0.7 dex above solar.
Since most of all our post-AGB stars do not show phosphorus overabundance, we can draw two
conclusions.
1. The overabundances of phosphorus measured in the some EHes by Jeffery et al. are
probably real, because they used the same lines and same methods in stars of similar Teff
and log g .
136
Star
He
C
N
O
Al
Si
P
S
Fe
HD
119608
+0.23
+0.42
+0.56
+0.45
+0.32
+0.42
+0.29
-0.12
+0.47
Hen
3-1428
+0.21
+0.30
+0.72
+0.80
+0.20
+0.67
+0.17
-0.41
+0.20
LSS
4331
+0.63
-0.71
+0.30
+0.73
–
+1.27
<-1.95
-1.00
+0.73
LSIV
-04 01
-2.33
<-2.2
<-2.17
<-3.27
–
-1.09
<+0.50
–
<-1.74
LSIV
-12 111
-0.29
-0.38
-0.01
+0.55
-0.56
+0.14
<+0.50
-1.23
<-0.69
LSS
5112
-0.10
-0.66
-0.05
-0.32
-0.38
-0.29
<+0.10
–
-0.19
LB
3116
-0.03
-1.89
+0.37
-0.02
-0.12
-0.88
<+1.21
-1.26
-0.62
HR
1765
11.01
7.99
7.47
8.27
5.98
7.08
4.95
7.23
7.19
SUN
11.00
8.52
7.92
8.83
6.47
7.55
5.45
7.33
7.50
Table 8.1: Logarithm of abundances relative to the standard star HR1765. Abundances of
HR1765 (this work) and the Sun (Grevesse & Sauval 1998) are also shown in their usual form.
2. The origin of the excess phosphorus in EHes cannot be directly explained by previous
evolution on the AGB, although it is possible that phosphorus created in the intershell is
not dredged-up on the AGB, but may reach the surface later.
8.4 Relative Abundances
In this section, we compare the abundances of the programme stars to a well-known standard
star HR1765 (see Chapter 4). We have chosen this standard star as spectroscopically similar to
the sample stars as possible (Gray 1992). By doing this, the effect of uncertainties in the atomic
data will be reduced. The results are presented in Table 8.1.
A comparison of our findings to other groups of post-AGB stars (e.g. Group-I and Group-II in
the Table 8.2.) enables us to distinguish the photospheric characteristics of the programme stars
studied in this thesis. We have calculated [X/Fe]2 ratios for this comparison and presented them
in Table 8.2.
Our findings for these calculated ratios of [C/Fe], [N/Fe], and [O/Fe] are as follows:
1. HD119608: The star appear to be slightly N deficient. The observed underabundances
in C and O are ≈0.3 dex.
2. Hen3-1428: It is normal in C and enriched in N and O.
3. LSS4331: This star is extremely C deficient. N and O also show deficiency.
2
[X/Fe] = log(n x/nFe )S T AR − log(n x /nFe )⊙
8.5. Temperature Sequence for the programme stars
137
4. LSIV-04 01: This star is very metal poor and may be a group-I post-AGB.
5. LSIV-12 111: It is extremely overabundant in O. N show overabundance and C seems
to be solar. This could be a group-II post-AGB star.
6. LSS5112: LSS5112 has a normal N abundance. Otherwise C and O are deficient.
7. LB3116: This extremely C-poor star shows overabundance in N and O, so also appear
to be a group-I post-AGB star.
It should be noted that this analysis is based on Fe abundance and those ratios will be highly
dependent on the quality of the lines used in Fe abundance determination. For some of the
programme stars, there were few and/or weak Fe lines available due to low ossillator strength
and high extitation potential of FeIII lines in the blue-visual. It is better to study these star in
UV. The weakness of the Fe lines in B-type star was also reported by Lyubimkov et al. (2005).
8.5 Temperature Sequence for the programme stars
We present LTE best model fits with normalized spectra of the programme stars as a temperaturesequence in Figure 8.1, 8.2, 8.3, and 8.4.
8.6 Stellar Wind Signatures
We have encountered evidence of wind in post-AGB stars.
Helium line profiles - In several cases, the determination of helium abundance was not straightforward due to asymmetry in line profiles and emissions observed in the line centers.
Ciurla (1966), in his study of absorption lines in a moving atmosphere, computed the line profiles in the atmosphere of a B2 main-sequence star with X=0.68, Y=0.32, Z=0.00, log g =4.0,
and Teff =16 000 K . In this study he assumed that the velocity gradient in the atmosphere was
small enough to leave the structure unchanged. An expanding atmosphere caused a shift in the
direction of shorter (blue) wavelength and would enhance the long-wavelength wing while a
contraction would cause an opposite effect.
138
Post-AGBs
Hen 3-1312
HR 6144
HR 7671
SAO 173329
HD95767
HD 107369
HD 108015
HD 131356
ROA 24
Post-AGB stars/Group-I
HR4049
HD52961
BD+39 4926
HD44179
Post-AGB stars/Group-II
89 Her
SAO 239853
HD 133656
HD 161796
Programme post-AGB stars
HD119608
Hen3-1428
LSS4331
LSIV-04 01
LSIV-12 111
LSS5112
LB3116
EHes
RCrBs
[Fe/H]
-1.1
-0.4
-1.1
-0.8
+0.1
-1.1
-0.1
-0.6
-2.1
[C/Fe]
+0.1
+0.3
-0.3
+0.3
-0.1
<-0.2
+0.1
+0.3
+0.6
[N/Fe]
+1.6
+0.9
+0.1
+0.3
-0.2
+0.4
+0.2
+0.3
+1.8
[O/Fe]
+0.9
+0.3
-0.3
+0.1
-0.5
0.0
-0.1
-0.1
+1.3
[C/O]
-0.8
0.0
0.0
+0.2
+0.4
<-0.2
+0.2
+0.4
-0.7
Ref.
1
5
5
6
6
6
6
6
7
<-3.2
-4.8
-2.85
-3.3
+3.0
+4.40
+2.45
+3.30
–
–
–
–
>2.7
+4.20
+2.75
+2.90
>0.30
+0.20
-0.30
+0.40
8
9
10
2
-0.4
-0.8
-0.7
-0.3
+0.3
+0.4
+0.2
+0.3
+0.6
+0.6
+0.5
+1.1
+0.1
+0.8
+0.6
+0.4
+0.2
-0.4
-0.4
-0.1
5
6
6
5
+0.16
-0.11
+0.42
-2.05
-1.00
-0.50
-0.93
-0.27
-0.12
-1.66
–
+0.09
-0.69
-1.49
1.4
1.3
-0.05
+0.38
-0.57
–
+0.54
0.00
+0.85
1.0
1.6
-0.27
+0.35
-0.25
–
+0.99
-0.38
+0.35
0.0
0.3
0.00
-0.47
-1.91
-2.46
-0.9
-1.07
-1.84
1.4
1.0
tsa
tsa
tsa
tsa
tsa
tsa
tsa
Table 8.2: Comparison of CNO abundances and metallities for the programme stars and other
groups of post-AGB stars in the literature.
• 2: Luck & Bond (1989); Luck, R. E., & Bond, H. 1989, ApJS, 71, 559
Luck & Lambert (1985); Luck, R. E., & Lambert, D. 1985, ApJ, 298, 782
• 3: van Winckel & Reyniers (2000); van Winckel, H., & Reyniers, M. 2000, å, 354, 135
• tsa: This work.
• 5: Luck et al. (1990); Luck, R. E., Bond, H., & Lambert, D. L. 1990, 357, 188
• 6: van Winckel (1997); van Winckel, H. 1997, å, 319, 561
• 7: Gonzalez & Wallerstein (1994); Gonzalez, G., & Wallerstein, G. 1992, MNRAS, 254, 343
• 8: Lambert et al. (1988); Lambert, D. L., Hinkle, K. H., & Luck, R. E. 1988, ApJ, 333, 917
• 9: Van Winckel (1995); Van Winckel, H. 1995, Ph.D. Thesis, Katholieke Univ. Leuven, The Netherlands
• 10: Kodaira (1973); Kodaira, K. 1973, å, 22, 273
• tsa: This work.
8.6. Stellar Wind Signatures
139
Figure 8.2: Temperature-sequence for the programme stars. The Teff increases from top to
bottom. Each spectrum is vertically offset by 0.5 continuum units. The observed spectra of
LSS5112, IRAS17311-4924, IRAS17381-1616, LB3116, and LSIV-04 01 were binned to pixel
sizes of 0.25, 0.15, 0.25, 0.15, and 0.15 Å respectively - I.
140
sizes of 0.25, 0.15, 0.25, 0.15, and 0.15 Å respectively - II.
141
sizes of 0.25, 0.15, 0.25, 0.15, and 0.15 Å respectively - III.
142
sizes of 0.25, 0.15, 0.25, 0.15, and 0.15 Å respectively - IV.
LSIV
-12 111
Hβ
Hγ
Hδ
Hǫ
143
λrest
λ1 (Å)
∆λ(Å)
v∞ (km/s)
(Å)
4861.33
4340.46
4101.76
3970.07
AAT – ESO
4857.99 / 4860.38
4337.99 / 4339.48
4098.91 / 4099.32
3967.92 / 3968.70
AAT – ESO
-3.34 / -0.95
-2.47 / -0.98
-2.85 / -2.44
-2.15 / -1.37
AAT – ESO
206.12 / 58.63
170.72 / 67.73
208.45 / 178.46
162.47 / 103.52
EW (mÅ)
Absorption
AAT – ESO
369 / 31
410 / 86
468 / 211
238 / 136
EW (mÅ)
Emission
AAT – ESO
1580 / 6896
239 / 2770
407 / 688
170 / 228
• ∆λ = λ1 - λrest ; v∞ = c × (∆λ / λ)
Table 8.3: Measurement results for emission and absorption line profiles for both AAT and ESO
runs. ESO spectrum was kindly provided by Dufton at QUB.
We observed such blue-shifted line profiles in the spectra of two programme stars, namely
Hen3-1428 and LSIV-12 111. In the latter, we detected emission in Balmer lines as well as
[FeII] and [SII]. The former showed emission only in Balmer lines. These observational findings with presence of blue-shifted absorption line profiles of HeI indicate an expanding atmosphere in these stars. Parthasarathy (private communication, 2008) commented that these blue
shifted helium lines are clearly due to stellar wind.
We measured terminal velocities (v∞ ) for Balmer lines in LSIV-12 111. In this calculation, the
wavelength difference (∆λ) between blue absorption edge of those balmer lines (λ1 ) and their
laboratory central wavelengths (λrest ) were used (Table 8.3). These velocities are compared to
1994 ESO/CASPEC spectrum of the star obtained by Dufton at QUB. ESO 1994 spectrum of the
star has been presented in Chapter 2 for a region around Hδ for the evaluation of the normalization process. This comparison of the v∞ between 1999 AAT/UCLES and 1994 ESO/CASPEC
indicates to a change in emission line equivalent widths and v∞ values for the Balmer lines.
Same observational finding was also seen in blue wings of HeI lines. This finding need to be
investigated further whether it is caused by a change in the mass-loss rate for the star.
144
8.7 Future Directions
From the programme stars listed in Table 8.1, two stars namely Hen3-1428 and LSS4331 are
helium-rich. Further high signal-to-noise ratio observations of these are highly desirable. LSIV12 111 probably is a binary system. Comparison of the spectra obtained at different epochs
reveal the observed difference in Balmer line profiles and this might be indicating to changing
physical conditions (e.g. mass-loss rate) in an expanding stellar atmosphere. Radial velocity
measurements can provide direct proof of binarity (de Ruyter et al. 2005) for the star. Another
interesting programme star which needs to be investigated further by radial velocity observations is the rapidly rotating hot post-AGB star HD119608. As stated in Chapter 6, in order for
the possible mechanisms to be tested for the cause of the binarity for the star such as presence
of a disk and/or an accompanied jet, one would need interferometric studies in the infra-red,
Hα , and/or CO mapping respectively.
During our 2005 AAT run, we obtained high-resolution and high signal-to-noise data for a
binary post-AGB namely IRAS19157-0247 but it could not be fitted by the Armagh Stellar
Atmosphere Code. Preliminary model fits were not successful for abundance analysis to be
performed.
We have obtained low-resolution spectra of HD119608, IRAS 17311-4924, LS5112, and IRAS173811616 at SAAO in 20073. These data will be used to corroborate the model atmosphere parameters and check the hydrogen and helium abundances.
Another star included in the list of programme stars in Chapter 2 was CPD-53 5736 (IRAS144885405). Due to presence of diffuse interstellar band, the normalized high signal-to-noise ratio
spectrum of the star could not be fitted successfully. Preliminary inspection showed that star is
helium-poor and overabundant in Mg. Balmer lines are in emission. CPD-53 5736 is waiting to
be analyzed.
In addition to above, the tables of broadband photometric data presented in Appendix-C will
enable us to perform surface flux fitting method for the program stars. By doing this, we aim to
3
SAAO 2007 data were not available at the time this work was carried out and analyses were already and/or in
a great extent finalized.
8.7. Future Directions
145
better check temperatures. It will also provide new estimates for E(B-V).
We collected UV data for some programme stars from the FUSE and IUE archives. These data
will be used to further constrain on Teff and E(B-V) using the flux fitting method. It will also
enable us to determine any obscuration of the hot central stars due to circumstellar dust. C,N,O,
and Fe abundances will be also determined via their UV model atmosphere analysis. By doing
this, we aim to determine/corroborate metalicity of the programme stars as well as their CNO
and refractory element abundances.
We also aim to model wind properties of some of the programme stars. This will enable us to
determine mass-loss rate in these stars.
Bibliography
Abia, C., & Isern, J. 2000, ApJ, 536, 438
Abt, H. A., & Levato, H. 1977, PASP, 89, 797
Abt, H. A., & Levyi S. G. 1978, ApJ, 36, 241
Aller, L. H. 1950, AJ, 111, 434
Aller, L. H., & Czyzak, S. J. 1983, ApJS, 51, 211
Aller, L. H. 1994, ApJ, 432,427
Alves, D., Bond, H., & Livio, M. 2000, ApJ, 120, 2044
Balona, L. A., 1994, MNRAS, 268, 119
Barnbaum, C., Morris, M., & Kahane, C. 1995, ApJ, 450, 862
Beals, C. S., & Oke, J. B. 1953, MNRAS, 113,530
Beichmann, C. A., Neugebauer, G., Habing, H. J., Clegg, P. E., & Chester, T. J. 1985, Infrared Astronomical Satellite (IRAS) catalogs and atlases, Explanatory Supplement, Jet Propulsion Lab., California
Institute of Technology, Pasadena, California
Bidelman, W. P. 1951, ApJ, 113, 304
Bidelman, W. P. 1953, ApJ, 117, 25
Blanco, V. M., Demers, S., Douglass, G. G., Fitzgerald, M. P.:1968, Publ. of the United States Naval
Observatory, 2nd series, Vol.21
Blöcker, T. 1995, A&A, 299, 755
Bogdanov, M. B. 2003, Astronomy Reports, vol. 47, Issue 6, p.480
Bolton, C. T., & Thomson, J. R. 1980, ApJ, 241, 1045
Blocker, T., & Schönberner, D. 1997, A&A, 324, 991
Bond, H. E., Luck, R. E., & Newman, M. J. 1979, ApJ, 233, 205
Bond, H. E., Luck, R. E. 1987, in: Azzopardi, M., Mateucci, F. (eds.) Proceedings of the ESO Workshop
on Stellar Evolution and Dynamics in the OUther Halo of the Galaxy. p.103
147
148
BIBLIOGRAPHY
Bonsack, W. K., & Greenstein, J. L. 1956, PASP, 68, 402, 249
Boothroyd, A. I., & Sackmann, I. J. 1988c, ApJ, 328, 671
Cameron, A. G., Fowler, W. A. 1971, ApJ, 164, 111
Cannon, A. J., & Pickering, E. C.:1918-1924, The Henry Draper Catalogue, Ann. Astron. Obs. Harvard
College, 91-99
Clayton, G. C. 1996, PASP, 108, 721
Coluzzi, R. 1993, Bull. Inf. Centre Donnees Stellaires, Vol. 43, p.7
Conlon, E. S., Dufton, P. L., Keenan, F. P., & Leonard, P. J. T. 1990, A&A, 236, 357
Conlon, E. S:, Dufton, P. L., Keenan, F. P., & McCausland, R. J. H. 1991, MNRAS, 248, 82
Conlon, E. S. 1992b, IAU Symposium 155
Conlon, E. S., Dufton, P. L., Keenan, F. P., & McCausland, R. J. H., Little, J. E., 1993, A&A, 272, 243
Conlon, E. S., Dufton, P. L., & Keenan, F. P. 1994, A&A, 290, 897
Cramer, N., & Maeder, A. 1979, A&A, 78, 305
Crawford, D. L., Barnes, J. V., & Golson, J. C. 1971, AJ, 76, 1058
Cunha, K., & Lambert, D. L. 1992, ApJ, 399, 586
Cunha, K., Lambert, D. L. 1994, ApJ, 426, 170
Cunha, K., & Lambert, D. L. 1994, ApJ, 426, 170
De Marco, O., & Crowther, P. A. 1998, MNRAS, 296, 419
Dieckvoss, W., & Heckmann, O.:1975, AGK3 Catalogue
Dimitrijevic, M. S., & Sahal-Brechot, S. 1984, JQSRT, 31, 301
Eber, F., & Butler, K. 1988, A&A, 202, 153
Edgar, R. J., & Savage, B. D. 1989, ApJ, 340, 762
ESA:1997, The Hipparcos Catalogue, ESA SP-1200
Epchtein N., de Batz B., & Copet E., et al. 1994, Ap&SS, 217, 3
Feautrier, P. 1964, in: E.H. Avrett, O. Gingerich & C.A. Whitney (eds.), SAO Special Report No. 167 80
Fernie, J. D. 1983, ApJS, 52, 7
Forresti W. J., McCarthyi J. F., & Houck, J. R. 1979, ApJ, 233, 611
Garcio-Lario, P., Manchado, A., Pottasch, S. R., Suso, J., & Olling, R. 1990, A&AS, 82, 497
Garcia-Lario, P., Manchado, A., Pychi W., & Pottasch, S. R. 1997, A&A, 126, 479
Gauba, G., & Parthasarathy, M. 2003, A&A, 407, 1007
Gauba, G., & Parthasarathy, M. 2004, A&A, 417, 201
BIBLIOGRAPHY
149
Gautschy A. 1993, MNRAS, 265, 340
Gies, D. R:, & Lambert, D. L. 1992, ApJ, 387, 673
Gieske, H. A., & Griem, H. R. 1969, ApJ, 157, Gieske, 963
Giridhar, S., Lambert, D. L., & Gonzales, G. 2000, ApJ, 531, 521
Giridhar, S., & Ferro, A. A. 2005, A&A, 443, 297
Glatzel W.:1998, in Bradley P. A., Guzik J. A., eds, ASP Conf. Ser. Vol. 135, A Half Century of Stellar
Pulsation Interpretations: Attribute to Arthur N. Cox. Astron. Soc. Pac., San Francisco, p. 89
Goebel, J. H., & Moseley, S. H. 1985, ApJ, 290, 35
Gomez, A. E., & Abt, H. A. 1982, PASP, 94, 650
Green, R. F., Schmidt, M., & Lievert, J. 1986, ApJS, 61, 305
Grevesse, N., & Sauval, A. J. 1998, Space Sci. Rev., 85, 161
Groenewegen M. A. T. 2004, to be published in ’Planetary Nebula Beyond the Milky Way’ ed. J.R.
Walsh & L. Stanghellini, astro-ph/0407282
Gulati, R. K., Malagnini, M. L., & Morossi C. 1989, A&A, 80, 73
Gummersbach, C. A., Kaufer, A., Schaefer, D. R., Szeifert, T., & Wolf, B. 1998, A&A, 338, 881
Habing, H. J. 1968, Bull. Astron. Inst. Netherlands, 20, 120
Hambly, N. C., Dufton, P. L., Keenan, F. P., & Lumsden, S. L. 1996, MNRAS, 278, 811
Hambley, N. C., Rolleston, W. R. J., Keenan, F. P., Dufton, P. L., & Saffer, R. A. 1997, ApJ, 111, 419
Hauck, B., & Mermilliod, M. 1998, A&AS, 129, 431
Herbig, G. H. 1993, ApJ, 407, 142
Herwig, F., Blöcker, T., Langer, N., & Driebe, T. 1999, A&A, 349, L5
Herwig, F. 2005, ARA&A, 43, 435
Hill, P. W. 1964, MNRAS, 127, 113
Hill, P. W., & Hill, S. R. 1966, MNRAS, 133, 205
Hill, P. W., Kilkenny, D., & Van, B. I. G. 1974, MNRAS, 168, 451
Hill, P. W., Lynas-Gray, A. E., & Kilkenny, D. 1984, MNRAS, 207, 823H
Hirano, N., Shinnaga, H., & Dinh-V-Trung, et al. 2004, ApJ, 616, L43
Hoffleit, D.:1982, ”The Bright Star Catalogue”, 4th Edition (in collaboration with C. Jaschek), Yale
Observatory (catalog V/25)
Hog E., et al. 2000, A&A, 355, 27
Horne, K. 1986, PASP, 98, 609
150
BIBLIOGRAPHY
Howarth I.D., & Murray J. 1991, Starlink User Note, Rutherford Appleton Laboratory, 50.13
Howarth I.D., Murray J., Mills D., & Berry D.S. 1995, Starlink User Note, 50.16 (Rutherford Appleton
Laboratory)
Hrivnak, B. J., Kwok, S., & Volk, K. M. 1988, ApJ, 331, 832
Hrivnak, B. J., Kwok, S., & Volk, K. M. 1989, ApJ, 346, 265
Hrivnak, B. J:, Volk, K., & Kwok, S. 2000, ApJ, 535, 275
Iben, I. Jr. 1975, ApJ, 196, 525
Iben, I. Jr. 1976, ApJ, 208, 165
Iben, I. Jr. 1981, ApJ, 246, 278s
Iben, I. 1982, ApJ, 259, 244
Iben, I. Jr., & Renzini, A. 1983, ARA&A, 21, 271
Iben, I. Jr., Kaleri J. B., Truran, J. W., & Renzini, A. 1983, ApJ, 264, 605
Iben, I.Jr. 1984, ApJ, 277, 333
Iben, I. Jr., & Tutukov, A.V. 1984, ApJS, 55, 335
Iben, I. Jr. 1985, QJRAS, 26, 1
Iben, I. Jr., & Tutukov, A. V. 1985, ApJS, 58, 661
Iben, I. Jr., & Tutukov, A. V. 1986, ApJ, 311, 753
Iben, I. Jr. 1990, ApJ, 353, 215
Iben, I. Jr. 1995, Phys. Rep., 250, 1
Iben, I. Jr., & MacDonald, J. 1995, in White Dwarfs, ed. D. Koester, & K. Werner (Springer-Verlag), 48
IRAS Explanatory Supplement to the Point Source Catalogue, eds. C. A. Beichman, G. Neugebauer, H.
J. Habing, P. E. Clegg, & T. J. Chester, JPL D-1855
IRAS Point Source Catalogue, 1985, US Government Publication Office
Jeffery, C. S., Woolf, V. M., & Pollacco, D. L. 2001, A&A, 376, 497
Jeffery, C. S. & Heber, U. 1992, A&A, 260, 133
Jeffery, C. S., Heber, U., Hill, P. W., et al. 1996, in Hydrogen Deficient Stars, ed. C.S.Jeffery & U.Heber,
ASP Conf. Ser. Vol. 96, 471
Jeffery, C. S. 1996, Hydrogen-Deficient stars, pp.151-161, eds. C. S. Jeffery, & U. Heber, ASP
Jeffery, C. S., Harrison, P. M., & Drilling J. S. 1997a, A&A, 125, 501
Jeffery, C. S. & Simon, T. 1997b, MNRAS, 286, 487
Jeffery, C. S. 1998, MNRAS, 294, 391
BIBLIOGRAPHY
151
Jeffery C. S., & Saio H. 1999, MNRAS, 308, 221
Jeffery, C. S., Woolf, V. M., & Pollacco, D. L. 2001a, A&A, 376, 497
Jeffery, C. S., & Aznar Cuadrado, R. 2001b, 378, 936
Johnson, H. L., & Morgan, W. W. 1951, ApJ, 114, 522
Johnson, H. L. 1966, ARA&A, 4, 193
Jeffery, C. S:, Skillen, I., Hill, P. W., Kilkenny, D., Malaney, R. A., & Morrison, K.:1985, MNRAS217,
701
Jeffery, C.S., Drilling, J. S., & Heber, U.:1987, MNRAS226, 317
Johnson, H. L., & Morgan, W. W.:1953, ApJ 117, 313
Jura, M. 1986, ApJ, 303, 327
Katz, D., Soubiran, C., Cayrel, R., Adda, M., & Cautain, R. 1998, A&A, 338, 151
Kaufer, A., Szeifert, Th., Krenzin, R., Baschek, B., & Wolf, B. 1994, A&A, 289, 740
Keenan, F. P., Bates, B., Dufton, P. L., Holmgren, D. E., & Gilheany, S. 1990, ApJ, 348, 322
Keenan, F. P. 1992, QJRAS, 33, 325
Keenan, F. P., Brown, P. J. F., Dufton, P. L., & Holmgren, D. E.: Proceedings of the 3rd International
Colloquium of the Royal Netherlands Academy of Arts and Sciences, Amsterdam, August 28-31,
1989, Amsterdam: North-Holland, 1990, edited by J. E. Hansen., p.44
Kelly, D. M., & Hrivnak, B. J. 2005, ApJ, 629, 1040
Kilkenny, D., & Hill, P. W. 1975, MNRAS, 171, 353
Kilkenny, D., Hill, P. & W, Brown, A. 1977, MNRAS, 178,123
Kilkenny, D., & Lydon, J. 1986, MNRAS, 218, 279
Kilkenny, D., & Pauls, L. 1990, MNRAS, 244, 133
Kilian, J., Montenbruck, O., & Nissen, P. E. 1991, A&AS, 88, 101
Kilkenny, D., & Busse, J. 1992, MNRAS, 258,57
Kiss, L. L., Derekas, A., Szabo, Gy. M., Bedding, T. R., & Szabados, L. 2007, MNRAS, 375, 1338
Klochkova, V. G. 1995, MNRAS, 272, 710
Kodaira, K. 1973, A&A, 22, 273
Koen, C., & Eyer, L.:2002, MNRAS, 331,45
Kozok, J. R. 1985, A&AS, 62, 7
Kudritzki, R. P. 1979, in Stellar Atmospheres: Beyond Classical Models, NATO ASI Ser., 341, 441
Kwok, S. 1993, ARA&A, 31, 63
152
BIBLIOGRAPHY
Lambert, D. L., Hinkle, K. H., & Luck, R. E. 1988, ApJ, 333, 917
Lambert, D. L., & Rao, N. K. 1994, A&A, 15, 47
Landolt, A. U., & Grauer, A. D. 1986, PASP, 98, 346
Lattanzio, J. C. 1987, ApJ, 313, L15
Lattanzio, J. C., & Wood, P. R. 2004, in Asymptotic Giant Branch Stars, ed. H. J. Habing & H. Olofsson
(New York: Springer), 23
Lattanzio, J. C., & Lugaro, M. A. 2005, Nuclear Physics A, 758, 477
Leone, F., Catalano, F. A., & Malaroda, S. 1997, A&A, 325, 1125
Leone, F. & Lanzafame, A. C. 1998, A&A, 330, 306
Leuenhagen, U., Heber U., & Jeffery C.S. 1994, A&AS, 103, 445
Lemke, M. 1997, A&A, 122, 285
Loup, C., Forveille, T., Nyman, L. A., & Omont, A. 1990, A&A, 227, 29
Luck, R. E., Bond, H. E., & Lambert, D. L. 1990, ApJ, 357, 188
Luck, R. E. 1993, The chemical composition of luminous high-latitude stars post-AGB stars, ed. D. D.
Sasselov, in Luminous High-Latitude Stars, 87
Lucy, L. B. 1964, in Proceedings First Harvard-Smithsonian Conference on Stellar Atmospheres, Cambridge, Massachusetts, SAO Special Report, 167, 93
Lynas-Gray, A. E., Kilkenny, D., Skillen, I., & Jeffery, C. S. 1987, MNRAS, 227, 1073
Lyubimkov, L. S., Rostopchin, S. I., Rachkovskaya, T. M., Poklad, D. B., & Lambert, D. L. 2005,
MNRAS, 358, 193
Maas, T. 2003, Dissertation, Univ. of Leuven, Leuven
MacConnell, D. J., & Frye, R. L. 1972, PASP, 84, 388
Manchado, A., Pottasch, S. R., Garcio-Lario, P., Esteban, C., & Mampaso, A. 1989, A&A, 214, 139
Martin, J. C. 2004, AJ, 128, 2474
Martin, J. C.:2006, AJ, 131,3047
McCausland, R. J. H., Conlon, E. S., Dufton, P. L., & Keenan, F. P. 1992, ApJ, 394, 298
Mills D., & Webb J., 1994, Rutherford Appleton Laboratory, SUN 152.1
Moehler, S., & Heber, U., 1998, A&A, 335, 985
Moon, T. T. 1985, Communications of University of London Observatory, No. 78
Mooney, C. J., Rolleston, W. R. J., Keenan, F. P., Dufton, P. L., Pollacco, D. L., & Magee, H. R. 2001,
MNRAS, 326, 1101
BIBLIOGRAPHY
153
Mooney, C. J., Rolleston, W. R. J., Keenan, F. P., & Dufton, P. L., et al. 2002, MNRAS, 337, 851
Mooney, C. J., Rolleston, W. R. J., Keenan, F. P., Dufton, P. L., Smoker, J. V., Ryans, R. S. I., Aller, L.
H:, & Trundle, C. 2004, A&A, 419, 1123
Moore, Charlotte Emma, A Multiplet Table of Astrophysical Interest (The Princeton University Observatory, Princeton, NJ, 1933, 1945)
Morgan, W. W., Code, A. D., & Whitford, A. E. 1955, ApJ, 2, 41
Morel, T., Marchenko, S. V., Pati, A. K., Kuppuswamy, K., Carini, M. T:, Wood, E., & Zimmerman, R.
2004, MNRAS, 351, 552
Mowlavi, N., & Meynet, M. 2000, A&A, 361, 959
Munch, G. 1956, PASP, 68, 351
Munn, K. E., Dufton, P. L., Smartt, S. J., & Hambly, N. C. 2004, A&A, 419, 713
Napiwotzki, R., Schnberner, D., & Wenske, V. 1993, A&A, 268, 653
Napiwotzki, R. 1997, A&A, 332, 256
Nassau, J. J., & Stephenson, C. B. 1963, Luminous Stars in the Northern Milky Way, Vol. 4 (HamburgBergedorf: Hamburger Sternwarte and Warner Swasey Obs.)
Nyman, L. -A., Booth, R. S., Carlstrom, U., Habing, H. J., Heske, A., Sahai, R., Stark, R., Van Der Ween
W. E. C. J., & Winnberg, A. 1992, A&AS, 93, 121
Oudmaijer, R. D., van der Veen, W. E. C. J., Waters, L. B. F. M., Trams, N. R., Waelkens, C., &
Engelsman, E. 1992, A&AS, 96, 625
Oudmaijer, R. D. 1996, A&A, 306, 823
Pandey, G., Rao, N. K., Lambert, D. L., Jeffery, C. S., & Asplund, M. 2001, MNRAS, 324, 937
Pandey, G., Lambert, D. L., Rao, N. K., & Jeffery, C. S. 2004, ApJ, 602, L113
Pandey, G., Lambert, D. L., Jeffery, C. S., & Rao, N. K.:2006, ApJ, 638, 454
Parthasarathy, M., & Pottasch, S. R. 1986, A&A, 154, L16
Parthasarathy, M., & Pottasch, S. R. 1989, A&A, 225, 521
Parthasarathy, M. 1993, Luminous High-Latitude Stars, ASP Conferences Series, Vol.45, 173
Parthasarathy, M., Vijapurkar, J., & Drilling, J. S. 2000, A&AS, 145, 269
Pettini, M., & West, K. A. 1982, ApJ, 260, 561
Pottasch, S. R., & Parthasarathy, M. 1988, A&A, 192, 182
Preite-Martinez, A. 1988, A&AS, 76, 317
Proffitt, C. R., & Quigley, M. F. 2001, ApJ, 548, 429
154
BIBLIOGRAPHY
Quin, D. A., & Lamers, H. J. G. L. M. 1992, A&A, 260, 261
Ryans, R. S. I., Dufton, P. L., & Mooney, C. J., et al. 2003, A&A, 401, 1119
Reddy, B. E., Parthasarathy, M., & Sivarani, T. 1996, A&A, 313, 191
Reddy, B. E., Lambert, D. L., Gonzalez, G., & Yong, D. 2002, ApJ, 564, 482
Renzini, A. 1979, in: Stars and star systems, Westerlund B. E. (ed.). Reidel, Dordrecht, p. 155
Renzini, A. 1981, in Physical Processes in Red Giants, ed. I. Iben, & A. Renzini (Dordrecht: Reidel),
431
Rufener, F.:1981, A & AS 45,207
Sahai, R., Morris, M., Knapp, G. R., Young, K., & Barnbaum, C. 2003, Nature, 426, 261
Saio, H., & Jeffery, C.S. 1988, ApJ, 328, 714
Saio, H., & Jeffery, C. S. 2000, MNRAS, 313, 671
Saio, H., & Jeffery, C.S.:2002, MNRAS, 333, 121
Sarkar, G., Parthasarathy, M:, & Reddy, B. E. 2005, A&A, 431, 1007
Schlegel, D. J., Finkbeiner, D. P. & Davis, M. 1998, ApJ, 500, 525
Schmidt-Kaler, T. 1982, in Landolt-Brnstein New Series, Group 6, Vol. 2b, Stars and Star Clusters, ed.
K. Schaifers & H.-H. Voigt (Berlin: Springer)
Schönberner, D. 1977, A&A, 57, 437
Schönberner, D. 1981, A&A, 103, 119
Schönberner, D. 1983, ApJ, 272, 708
Schöning, T., & Butler, K. 1989, A&A, 78, 51
Schwarzschild, M., & Härm, R. 1967, ApJ, 150, 961
Sharpless, S. 1952, ApJ, 116, 251
Skrutskie, et al. 1997, in The Impact of Large Scale Near-IR Sky Surveys, ed. F. Garzon et al. (Netherlands: Kluwer), 25
Shortridge, K. 1986, Starlink User Note 86.4, Rutherford Appleton Laboratory
Slijkhuis, S. 1992, Ph.D. thesis, University of Amsterdam
Smith, M. A. 1977, ApJ, 215, 574
Smith, M. A. 1980, ApJ, 42, 261
Spitzer, L. 1950, MNRAS, 110, 216
Strömgen B., 1966, ARA&A, 4, 433
Struve, O. & Elvey, C. T. 1934, ApJ, 79, 409
BIBLIOGRAPHY
155
Suarez, O., Gomez, J. F., & Morata, O. 2007, A&A, 467, 1085
Szczerba, R., Henning, Th, Volk, K., Kwok, S., & Cox, P. 1999, A&A, 345, 39
Takeda, Y. et al. 2002, PASJ, 54, 765
Trams, N. R., Waters, L. B. F. M., Lamers, H. J. G. L. M., Waelkens, C., & Gabelle, T. R., The P. S.
1991, A&APS, 87, 361 357
Trams, N. R. 1991, ”Optically bright pos-AGB stars”, Thesis University of Utrecht
Thompson, G. I., Nandy, K., Jamar, C., Monfils, A., Houziaux, L., Carnochan, D. J., & Wilson, R.:1978,
Catalogue of Stellar Ultraviolet Fluxes, UK Science Research Council (for the magnetic tape version)
Thompson, H. M. A., Keenan, F. P., Dufton, P. L., Ryans, R. S. I., Smoker, J. V. 2006, MNRAS, 368,
1749
Turon, C., Egret, D., Gomez, A., Grenon M., et al. 1993, Bull. Inf. CDS 43, 5
Uesugi A. & Fukuda I. 1970, Contr. Astroph. Kwasan Obs. Univ. Kyoto, 189.
Umana, G., Cerrigonei L., Trigilio, C., & Zappala, R. A. 2004, A&A, 428, 121
van der Veen, W. E. C. J., Habing, H. J., Geballe, T. R. 1989, A&A, 226, 108
Van Winckel, H., Mathis, J. S:, & Waelkens, C. 1992, Nature, 356, 500
Van Winckel, H., & Waelkens, C. 1995, ApSS 224, 581
Van Winckel, H., Waelkens, C., Fernie, J. D., & Waters, L. B. F. M. 1999, A&A, 343, 202
Van Winckel, H., Waelkens, C., & Waters, L. B. F. M. 2000, IAU Symp. 177, 285
Venn, K. A., & Lambert, D. L. 1990, ApJ, 363, 234
Venn, K. A. 1993, ApJ, 414, 316
Venn, K. A., Smartt, J. S., Lennon, D. J., & Dufton, P. L. 1998, A&A, 334, 987
Vidal, C.R., Cooper, J., & Smith, E.W. 1973, , 214, 37
Volk, K. M., & Kwok, S. 1989, ApJ, 342, 345
Volk, K. M., Kwok, S., Hrivnaki B. J. 1999, ApJ, 516, 99
Volk, K:, Xiong, G. -Zh., Kwok, S. 2000, ApJ, 530, 408
Volk, K., Kwok, S., Hrivnaki B. J., Szczerba, R. 2002, ApJ, 567, 412
Waelkens, C., Lamers, H. J. G. L. M., Waters, L. B. F. M., Rufener, F., Trams, N. R., Le Bertre, T., Ferlet,
R., & Vidal-Madjar, A. 1991a, A&A, 242, 433
Waelkens, C., Van Winckel, H., Bogaert, E., & Trams, N. R. 1991b, A&A, 251, 495
Waelkens, C. Van Winckel, H., Trams, N. R., & Waters, L. B. F. M. 1992, A&A, 251, 495
Warner, B. 1967, MNRAS, 137, 119
156
BIBLIOGRAPHY
Warren, W. H:, & Hesser, J. E. 1977a, ApJ, 34, 115
Webbink, R.F.:1984, ApJ 277, 355
Weigert, A. 1966, Z. Ap., 64, 395
Whitelock, P.A., Menzies, J. W., Catchpole, R. M., Feast, M. W., Carter, B. S., Marang, F., Roberts, G.,
& Sekiguchi, K. 1989, MNRAS, 241, 393
Wolff, S. C., Pilachowski, C. A., & Wolstencroft, R. D. 1974, ApJ, 194, L83
Wolff, S. C., & Heasley, J. N. 1985, ApJ, 292, 589
Woolf, V. M., & Jeffery, C. S. 2002, A&A, 395, 535
Wood, P. R., & Faulkner, D. J. 1986, ApJ, 307, 659
Wright, D. J., Laynas-Gray, A. E., & Kilkenny, D. 2006, MNRAS, 369, 2049
Zombeck, M. 1990, Handbook of Space Astronomy and Astrophysics (2d ed. ; New York: Cambridge
Univ. Press)
Appendix A
Log files for observations
Table A.1: AAT log file for the 19990728 run.
Run
31
32
33
37
38
41
44
45
48
49
55
56
57
58
59
60
Object
HD 119608
CPD-53 5736
CPD-53 5736
LS IV -4 1
LS IV -4 1
Comment
LS IV -12 111
LS IV -12 111
LS IV -12 111 (blue)
LS IV -12 111 (blue)
PHL 1580
PHL 1580
PHL 1580
PHL 1580
PHL 1580
PHL 1580
RA (J2000)
13 44 31.08
14 52 28.76
14 52 28.44
16 56 27.74
16 56 27.64
After
20 01 50.04
20 01 49.94
20 01 49.87
20 01 49.75
21 30 25.35
21 30 25.24
21 30 25.26
21 30 25.26
21 30 25.16
21 30 25.26
Dec (J2000)
-17 56 15.8
-54 17 43.5
-54 17 45.5
-04 47 25.8
-04 47 25.4
LN2refill
-12 41 18.3
-12 41 19.7
-12 41 20.0
-12 41 20.9
-19 22 38.3
-19 22 37.2
-19 22 35.8
-19 22 35.6
-19 22 34.7
-19 22 35.6
157
UT start
09:55:53.3
10:03:20.3
10:30:52.5
11:40:36.8
12:07:58.5
Airmass
1.242
1.132
1.167
1.159
1.206
Seeing
1.5
1.5
1.5
1.4
1.4
Exposed
240.0
1500.0
600.0
1440.0
1440.0
13:56:13.6
14:24:27.9
14:51:29.3
15:35:19.7
16:58:12.8
17:30:23.0
18:01:46.0
18:34:38.3
19:06:13.9
19:38:51.3
1.057
1.071
1.099
1.179
1.122
1.200
1.310
1.473
1.698
2.057
0.9
0.9
1.2
1.7
1.7
1.7
2.0
2.0
2.0
2.0
1500.0
600.0
2400.0
2100.0
1800.0
1800.0
1800.0
1800.0
1800.0
13.0
158
Chapter A. Log files for observations
Run
30
30
32
33
34
36
37
38
42
43
43
44
45
45
47
50
51
51
Object
HD 119608
Comment
CPD -53 5736
CPD -53 5736
CPD -53 5736
LS IV -4 1
LS IV -4 1
LS IV -4 1
Comment
LSS 5112
Comment
LSS 5112
LSS 5112
Comment
Comment
LS 5112
LS 5112
Comment
RA (J2000)
13 44 31.03
Clouds most
14 52 28.59
14 52 28.39
14 52 28.15
16 56 27.81
16 56 27.64
16 56 27.57
Seeing
18 40 48.45
a little
18 40 48.50
18 40 48.57
very cloudy!
After
18 40 48.47
18 40 48.47
more cloud
Dec (J2000)
-17 56 14.0
gone away.
-54 17 43.9
-54 17 45.1
-54 17 45.9
-04 47 22.8
-04 47 23.4
-04 47 24.7
going off?
-17 04 40.0
cloud around
-17 04 38.7
-17 04 38.4
LN2 fill
-17 04 37.7
-17 04 36.9
UT start
09:16:22.1
Airmass
1.144
Seeing
1.2
Exposed
300.0
09:54:40.6
10:21:06.0
10:47:44.3
11:11:09.4
11:42:47.0
12:14:09.4
1.128
1.159
1.200
1.131
1.167
1.228
1.2
1.2
1.2
1.2
1.2
1.2
1500.0
1500.0
1200.0
1800.0
1800.0
1800.0
14:24:59.5
now
14:56:22.0
15:27:43.8
1.181
1.8
1800.0
1.281
1.420
1.8
1.8
1800.0
900.0
16:02:06.1
16:46:00.7
1.648
1.931
1.8
1.8
1800.0
1200.0
Run
16
17
18
23
51
Object
LSIV-14 116
LSIV-14 116
LSIV-14 116
Comment
HR1765
RA (J2000)
20 57 38.97
20 57 38.86
20 57 38.80
After
05 21 45.56
Dec (J2000)
-14 25 47.3
-14 25 47.4
-14 25 48.3
LN2 fill
-00 23 00.2
UT start
11:54:38
12:25:32
12:56:26
Airmass
1.065
1.047
1.047
Seeing
1.4
1.4
1.4
Exposed
1800.0
1800.0
1800.0
19:43:34
1.242
2.0
15.0
AAT 2005-08-26 UCLES + EEV2 Session : A Top End : F/36
Observers : AHMAD, SAHIN (JEFFERY)
Run
49
50
64
65
72
Object
IRAS 17311-4924
IRAS 17311-4924
IRAS 19157-0247
IRAS 19157-0247
Comment
RA (J2000)
17 35 01.72
17 35 01.65
19 18 22.32
19 18 22.19
after
Dec (J2000)
-49 26 33.8
-49 26 34.5
-02 42 18.0
-02 42 16.6
LN2filling
UT start
10:45:40
10:59:55
13:38:44
13:51:19
Airmass
1.100
1.118
1.473
1.541
Seeing
1.9
1.9
1.8
1.8
Exposed
800.0
800.0
700.0
700.0
159
Run
71
72
Object
LB3116
LB3116
RA (J2000)
19 18 48.06
19 18 47.65
Dec (J2000)
-64 35 33.6
-64 35 36.2
UT start
11:33:58
11:59:53
Airmass
1.203
1.217
Seeing
1.3
1.3
Exposed
1500.0
1500.0
AAT UT Date : 2005-08-28 UCLES + EEV2 Session : A Top End : F/36
Run
15
16
Object
IRAS 17381-1616
IRAS 17381-1616
RA (J2000)
17 40 59.57
17 40 59.52
Dec (J2000)
-16 18 16.1
-16 18 16.3
UT start
10:02:04
10:24:38
Airmass
1.054
1.078
Seeing
1.3
1.3
UT Date : 2005-08-29 UCLES + EEV2 Session : A Top End : F/36
Exposed
1300.0
1300.0
Appendix B
Photometry
B.1 Variability and Evolution in Various Classes of Post-AGB
Stars
Published in T. Şahin, C. S. Jeffery, 2007, Astronomische Nachrichten, 328, No.8, 848.
B.1.1 Abstract
In a seperate work, we aimed to compare properties of early-type post-asymptotic giant-branch
(post-AGB) stars, including normal first-time B-type post-AGB stars, and extreme helium stars
(EHes). Hipparcos photometry for 12 post-AGB stars and 7 EHe stars has been analyzed; Five
post-AGB stars are clearly variable. The Hipparcos data are not sufficiently sensitive to detect
variability in any of the EHes.
B.1.2 Introduction
Post-asymptotic giant-branch (post-AGB) stars have luminosity classes from I (super-giant) to
III (giant), spectral types from B to K, and effective temperatures in the range 5 000 – 30 000
K. Initial masses of these objects are in the range 1 – 8 M⊙ . A post-AGB star is generally
believed to be a star which has completed its AGB evolution and started to contract to become
a white dwarf, thus its evolution runs at roughly constant luminosity towards a high effective
temperature. The time-scale for the post-AGB phase predicted by models ranges between 101
– 105 years (Blocker 1995). When T eff reaches about 30 000 K, the star becomes hot enough
161
162
Chapter B. Photometry
to ionize any remaining circumstellar matter which is left over from the AGB stage and then
may appear as a planetary nebula. Cool post-AGB stars are often enshrouded by dust shells,
which are dispersed as the star heats. Their surfaces are dominated by material brought to the
surface by third dredge-up on the AGB. Being very luminous, they have a tendency to pulsate
with unexpected additional modes (strange-mode instability, Gautschy 1993 ; Glatzel 1998)
while they quickly cross the instability strip. EHes are carbon-rich B- and A-type super-giants
with extremely low surface abundances of hydrogen (Jeffery et al. 1987; Pandey et al. 2006).
Their temperatures lie in the range 8 000 – 32 000 K and they also show pulsations due to
strange-mode instability (Saio & Jeffery 1988). Their unusual chemical compositions (helium
abundances ≥ 99 %) are thought to be due to complete mixing of the star’s outer layers. One
idea is that they have undergone a ’late thermal pulse (LTP)’, in which the helium-burning shell
is re-ignited to produce a deep convection zone which mixes the entire stellar envelope (Iben et
al. 1983; Iben & Tutukov 1984). The preferred idea is that EHes are formed from the merger
of two white dwarfs, in which the current surface represents a mixture of the original surface of
the more massive white dwarf and its entire companion (Webbink 1984; Saio & Jeffery 2002).
The two routes produce a quite different chemical signature.
Our idea is that diversification in observed abundances in both group of stars could be due to
differing pulsational characteristics as well as the route they follow, since they were both on
AGB at some stage of their evolution. This could be analyzed by performing both quantitative
spectral analyses and by searching for similar patterns of photometric variability. The latter is
the starting point of this paper.
B.2 Hipparcos Light Curves
Photometry covering almost 3.5 years for all observed A- and B-type post-AGB and EHe stars
was extracted from the Hipparcos database (ESA, 1997). Availability of Hipparcos data and
SIMBAD spectral type have been used as selection criteria (Table 1.). We used the σ/s ratio
(standard deviation over mean error) to test for variability. To be confident of identifying only
true variables, we required σ/s > 3, although a smaller value might have been possible. The
B.2. Hipparcos Light Curves
SpType
Mean
σ Nobs
post-AGB
BD+39◦ 4926 B81
9.329 0.028 115
BD+33◦ 2642 B22
10.7836 0.060 165
HD 137569 B5III:3
7.926 0.041 142
HD 172324 A0Iab4
8.185 0.087 110
HD 52961
A01
7.476 0.061 74
HR 4049
B9.5Ib/II11 5.573 0.068 115
HD 213985 A0III12
8.903 0.186 119
HD 105262 B91
7.113 0.012 107
HD 148743 A7Ib5
6.561 0.009 83
HD 119608 B1Ib12
7.494 0.015 63
EHes
BD+10◦ 2179 B06
9.908 0.035 74
BD+37◦ 1977 sdO:7
10.092 0.026 64
HD 124448 B38
9.962 0.034 82
HD 160641 O9.5Iap...9 9.860 0.050 53
HD 168476 B5 8
9.295 0.028 77
BD−9◦ 4395 B10
10.545 0.054 56
BD+13◦ 3224 B21
10.499 0.034 127
163
s
σ/s JD − 2440000
0.022 1.305
0.038 1.597
0.013 3.296
0.012 7.295
0.010 6.262
0.006 11.988
0.014 13.300
0.008 1.532
0.006 1.449
0.009 1.686
7884 - 9018
7861 - 9025
7904 - 9028
7858 - 9027
7964 - 9058
7859 - 8970
7881 - 8965
7899 - 9018
7911 - 9034
7908 - 9031
0.022
0.024
0.024
0.022
0.017
0.028
0.030
7881 - 8810
7975 - 8975
7914 - 9038
7994 - 9063
7924 - 9045
7912 - 9034
7906 - 9061
1.605
1.100
1.433
2.236
1.596
1.922
1.155
1: Henry Draper (HD) Catalogue, 2: Blanco et al. (1968), 3: Thompson Catalogue, 4:
Bonsack et al. (1956), 5: Johnson and Morgan (1953), 6: Hoffleit (1982), 7: Wolff et
al. (1974), 8: Hill (1964), 9: Turon et al. (1993), 10: MacConnell and Frye (1972), 11:
Michigan Spectral Survey,Vol.3, 12: Michigan Spectral Survey,Vol.4.
Table B.1: Statistics of Hipparcos photometry. Sources for spectral type are given.
largest Hipparcos value of σ/s in our sample for which the target is known to be non-variable
was 1.6 (BD+10◦ 2179: Hill et al.1984).
We were unable to confirm variability in any EHe on the basis σ/s ≥ 3. Two previously known
variables have σ/s ≥ 1.9. These are HD160641 and BD−9◦ 4395.
HD 160641 : Multi-periodic pulsations with periods in the range 0.7 − 1.1 d were reported by
Lynas-Gray et al. 1987. Follow-up observations showed periodicity near half a day. Fourier
analysis failed to find coherent periodic behaviour (Wright et al. 2006).
BD − 9◦ 4395 : Multi-periodic pulsations with periods in the range 3 − 9 d were reported by
Jeffery et al. (1985). The multicolor photometry of this hydrogen-deficient star by Landolt &
Grauer (1986) indicate that any possible period which may exist must be longer than one hour.
Hipparcos photometry shows no evidence of variability in either HD160641 or BD−9◦ 4395
and is too sparse to allow any unique period to be established.
Five post-AGBs are variable at the σ/s ≥ 3 level, being HR4049, HD213985, HD52961,
164
HD137569, and HD172324.
We have investigated the light curves of the clearly variable post-AGB stars. Knowing that Btype EHes show quasi-periods in the range 1 – 21 d, periodograms and window functions were
computed over the range 0 − 1 d−1 , at a resolution of 10 per cent of the resolution frequency
(1/T = 10−3 ) for the five definitely variable post-AGB stars (Fig.1-2). The upper limit was
chosen because higher pulsation frequencies are physically unrealistic in post-AGB stars which
have dynamical time-scales of several days. The apparently non-variable HD119608 was also
examined for comparison.
HR 4049 : Waelkens et al. (1991a) report an orbital period of 434 d. They also report that
photometric variation was due to orbital modulation of the light. A binary period of 429±2 d was
reported by Van Winckel & Waelkens (1995). Koen & Eyer (2002) report a period of 282 d for
the Hipparcos photometry. The highest peak in our periodogram occurs at 0.0021 ± 0.0007 d−1
( log f = −2.68) which corresponds to a period of P=472+242
−115 d (Fig.3).
HD 213985 : Whitelock (1989) reported a period of 254 d in K – band. Van Winckel, Waelkens
& Waters (2000) measured an orbital period of 259 d from long-term radial velocity measurements. They supported the idea that the light variations are due to orbital brightness modulation,
either from variable extinction or light scattering.
Kiss et al. (2007) analysed ASAS1 V – band time-series photometric observations of the star
and reported cycle-to-cycle change of its light curve shape by suggesting that this behaviour was
indicating on-going changes in the circumstellar shell. The Hipparcos data show considerable
power at low frequency, with one peak ( log f = −2.43) at 270+58
−100 d (Fig.4) consistent with the
orbital period.
HD 52961 : This is a binary system (Waelkens et al. 1992) with an orbital period of 1310 ± 8 d
(Van Winckel et al. 1999). Waelkens et al. (1991b) reports a pulsation period of 72 ± 1 d. The
peak at 0.014 d−1 ( log f = −1.85) in the Hipparcos periodogram of the star corresponds roughly
to the pulsation period, the first harmonic is also strong (0.029d−1 ) (Fig.1). We find P=69 ± 4 d
(Fig.5).
Waelkens et al. (1991b) report a spectral type of F6I from the presence of a G – band feature.
1
All Sky Automated Survey – ASAS
B.2. Hipparcos Light Curves
165
Figure B.1: Hipparcos periodograms for post-AGB variables and HD119608. Abscissa is log10
frequency (day−1 ) and ordinate is amplitude.
Figure B.2: Hipparcos window functions for post-AGB variables and HD119608.
The HD catalogue gives a spectral type of A0.
HD 137569 : This was reported as a Pop II ultraviolet-bright star by Bolton & Thomson (1980).
They found that the star was a single-line binary with a period of 529.8 d. Giridhar & Ferro
(2005) report that large depletion of only refractory elements in the star gives strong support
to the post-AGB candidature of HD137569. This old evolved star has also been reported as
166
Figure B.3: Phased light curve for post-AGB
star HR4049 at a period of 472 d . The crosses
show the phased data binned with a width of 0.1
in phase. The average magnitude and scatter in
magnitude were calculated for each phase bin.
Figure B.4: Phased light curve for post-AGB
star HD213985 at a period of 270 d . The crosses
show the phased data binned with a width of 0.1
in phase. The average magnitude and scatter in
magnitude were calculated for each phase bin.
Figure B.5: Phased light curve for post-AGB star HD52961 at a period of 69 d . The crosses
show the phased data binned with a width of 0.1 in phase. The average magnitude and scatter
in magnitude were calculated for each phase bin.
a long-period binary by Martin (2006). Our periodogram analysis does not indicate a single
coherent period in the star. Variation may be due to aperiodic or short-period pulsation. Despite
having σ/s ≥ 3, this bright star has one of the smallest values of σ in our sample.
HD 172324 : This high galactic latitude star was reported as a possible variable star in the
catalog of Rufener (1981). The star was suggested to be either a pulsating variable or a binary
on the basis of doubling in the O I 7774 Å triplet lines by Giridhar & Ferro (2005). Our analysis
does not indicate a single coherent period. Variation may be due to aperiodic or short-period
B.3. Conclusion
167
pulsation.
B.3 Conclusion
We have examined Hipparcos data for early-type post-AGB stars. Variability is clearly found in
five of these. In three cases, relatively large variations appear to be correlated with the binary
period and may be related to variable absorption by or illumination of circumstellar material.
In two cases, HD137569 and HD172324, the Hipparcos data indicate the stars to be variable,
but do not show evidence of periodicity. The variability may be due to aperiodic or short-period
variations. Higher cadence observations will be required to investigate this possibility further.
Appendix C
Broadband photometry in the literature
for the programme stars
169
170
Chapter C. Broadband photometry in the literature for the programme stars
HD119608
Phot. Band
Tycho-2 BT
Tycho-2 VT
DENIS J
DENIS Ks
DENIS Gunn-i
2MASS J∗
2MASS H∗
2MASS K∗s
Mag.
7.405
7.499
18.41
7.590
7.594
7.621
∆mag
0.015
0.012
0.20
0.030
0.026
0.034
• ∗: 2MASS All-Sky Point Source Catalog (PSC)- ADS/IRSA.Gator 2007/0904/102021 24311
Table C.1: Broadband photometry for HD119608 in th literature.
LSIV-04 01
Phot. Band
Tycho-2 BT
Tycho-2 VT
DENIS J
DENIS Ks
DENIS Gunn-i
2MASS J
2MASS H
2MASS K s
Mag.
9.843
9.290
11.279
11.034
11.702
11.270
11.194
11.128
∆mag
0.025
0.020
0.15
0.13
0.12
0.023
0.025
0.025
Table C.2: Broadband photometry for LSIV-04 01 in th literature.
IRAS17381-1616
Phot. Band
Tycho-2 BT
Tycho-2 VT
DENIS J
DENIS Ks
DENIS Gunn-i
DENIS J
DENIS Ks
DENIS Gunn-i
2MASS J
2MASS H
2MASS K s
IRAS - 12(Jy)
IRAS - 25
IRAS - 60
IRAS - 100
Mag.
10.851
10.843
12.219
11.968
12.627
12.338
12.003
12.625
12.306
12.157
12.046
0.25L
5.76
1.25
13.71L
∆mag
0.071
0.113
0.08
0.11
0.04
0.09
0.10
0.02
0.024
0.022
0.023
7
9
-
Ref.
Table C.3: Broadband photometry for IRAS17381-1616 in th literature.
171
IRAS17311-4924
Phot. Band
Tycho-2 BT
Tycho-2 VT
DENIS J
DENIS Ks
DENIS Gunn-i
2MASS J
2MASS H
2MASS K s
Gar97 J
Gar97 H
Gar97 K s
IRAS - 12(Jy)
IRAS - 25
IRAS - 60
IRAS - 100
Mag.
11.411
10.921
9.73
9.05
10.21
9.793
9.543
9.203
9.74
9.54
9.19
18.34
150.70
58.74
17.78
∆mag
0.085
0.085
0.04
0.06
0.04
0.024
0.027
0.023
0.03
0.03
0.03
I
I
I
II
II
II
II
• I: Reference for Johnson J, H and K magnitudes: 2003yCat.2246....0C - CDS/ADC Collection of Electronic
Catalogues, 2246, 0 (2003)
• II: Reference for IRAS colours: NASA Ref. Publ., 1190, 1 (1988)
Table C.4: Broadband photometry for IRAS17311-4924 in th literature.
LSIV-12 111
Phot. Band
Tycho-2 BT
Tycho-2 VT
DENIS J
DENIS Ks
DENIS Gunn-i
2MASS J∗
2MASS H∗
2MASS K∗s
Gar97 J
Gar97 H
Gar97 K s
IRAS - 12
IRAS - 25
IRAS - 60
IRAS - 100
Mag.
11.391
11.453
11.129
10.877
11.193
10.986
10.941
10.792
11.08
10.96
10.78
0.29:
10.26
6.45
1.77
∆mag
0.085
0.137
0.09
0.11
0.07
0.024
0.026
0.019
0.12
0.14
0.10
0.10
0.07
0.10
0.09
• *: 2MASS All-Sky Point Source Catalog (PSC) - ADS/IRSA.Gator 2007/0906/094224 8503; Gar97:
Garcia-Lario et al. (1997)
Table C.5: Broadband photometry for LSIV-12 111 in th literature.
172
Chapter C. Broadband photometry in the literature for the programme stars
LSS 5112
Phot. Band
Tycho-2 BT
Tycho-2 VT
DENIS J
DENIS Ks
DENIS Gunn-i
2MASS J
2MASS H
2MASS K s
Gar97 J
Gar97 H
Gar97 K
IRAS - 12(Jy)
IRAS - 25
IRAS - 60
IRAS - 100
Mag.
11.977
11.977
10.908
10.398
11.349
10.661
10.433
10.155
10.76
10.55
10.33
1.67
23.76
7.12
3.66L
∆mag
0.153
0.265
0.07
0.07
0.03
0.024
0.022
0.023
0.13
0.10
0.10
6
6
10
-
Table C.6: Broadband photometry for LSS 5112 in th literature.
LB3116
Phot. Band
Tycho-2 BT
Tycho-2 VT
DENIS J
DENIS Ks
DENIS Gunn-i
2MASS J
2MASS H
2MASS K s
Mag.
12.307
11.948
12.834
12.826
12.674
12.813
12.927
12.895
∆mag
0.175
0.180
0.14
0.17
0.03
0.023
0.027
0.028
Table C.7: Broadband photometry for LB3116 in th literature.
Appendix D
SPECTRAL ATLAS of HR1765
173
174
Chapter D. SPECTRAL ATLAS of HR1765
Figure D.1: SPECTRAL ATLAS for HR1765-I.
175
Figure D.2: SPECTRAL ATLAS for HR1765-II.
176
Figure D.3: SPECTRAL ATLAS for HR1765-III.
177
Figure D.4: SPECTRAL ATLAS for HR1765-IV.
178
Figure D.5: SPECTRAL ATLAS for HR1765-V.
Appendix E
SPECTRAL ATLAS of HD119608
179
180
Chapter E. SPECTRAL ATLAS of HD119608
Figure E.1: SPECTRAL ATLAS for HD119608-I.
181
Figure E.2: SPECTRAL ATLAS for HD119608-II.
182
Chapter E. SPECTRAL ATLAS of HD119608
Figure E.3: SPECTRAL ATLAS for HD119608-IIII.
183
Figure E.4: SPECTRAL ATLAS for HD119608-IV.
Appendix F
SPECTRAL ATLAS of LSIV-12 111
185
186
Chapter F. SPECTRAL ATLAS of LSIV-12 111
Figure F.1: SPECTRAL ATLAS for LSIV-12 111-I.
187
Figure F.2: SPECTRAL ATLAS for LSIV-12 111-II.
188
Figure F.3: SPECTRAL ATLAS for LSIV-12 111-III.
189
Figure F.4: SPECTRAL ATLAS for LSIV-12 111-IV.
190
Figure F.5: SPECTRAL ATLAS for LSIV-12 111-V.
191
192
193
Appendix G
Some comments on capabilities of
ECHOMOP
• Spectral order location and tracing (special care must be taken during this stage).
• Cosmic-ray removal, detection of bad image rows and columns as well as saturated (algorithm is not successive for cosmic-ray removal. This may have to be done prior to
ECHOMOP).
• Determination of object channels (dekker size should be determined correctly, this can be
checked by inspecting each frame individually).
• Generation of flat-field balance models (although it is not always necessary to produce a
balance frame unless you have highly cosmic-ray contaminated data frames).
• Modelling of scattered light and optimal spectrum extraction.
• Échelle blaze correction (ECHOMOP is not successive for blaze correction and algorithm for blazing is also sampling real science data in some stage).
• Scrunching extracted orders to a (generally) linear wavelength scale is not a straightforward process. Algorithm is not successive although the same method is being used as
195
196
Chapter G. Some comments on capabilities of ECHOMOP
FIGARO (Shortridge 1986) task IS CRUN.
• Merging and normalization can not be performed in ECHOMOP unless current interface
is modified.
• Plotting of data used for a reduction outputting of the final product (although some of the
options of outputting of final product are not working efficiently).

An Observational Study of Post-Asymptotic Giant Branch Stars

Transcription

Similar documents

Tenderflake Linzer Cookies

Chris Geary - Harness Racing NSW

to the Stars and Constellations Fact

Kimbertal Kennels Star Dog Breeder

January 11 - Black Rider Press

Scanned Document

NASA Space Place

Registration form copy

“Baby Stars in the Universe”

Page 6b Entertainment Page - The Fayetteville Press Newspaper