X-ray Grating Spectroscopy

1
X-ray Photospheres

• Klaus Werner
• Institute for Astronomy and Astrophysics,
  University of Tübingen, Germany

2
Outline

• Introduction Thermal soft X-ray emission from
  stellar photospheres
• Chandra spectroscopy of the hot DA white dwarf
  LB1919 Implications for vertical chemical
  stratification in WDs
• Chandra spectroscopy of the PG1159 star
  PG1520525 Constraining the GW Vir instability
  strip in the HRD
• Chandra spectroscopy of the naked C/O stellar
  core H150465 The hottest known and chemically
  most extreme white dwarf

3
Introduction

• Only evolved compact stars are hot enough to be
  able to emit thermal (soft) X-radiation from
  their photosphere
• Neutron stars (not covered in this talk)
• (Pre-) white dwarfs (WDs), (some are central
  stars of PNe)
• WDs come in two flavors
• Hydrogen-rich (DA WDs)
• Helium and/or C-O-rich (non-DAs),
• relevant here PG1159 stars, the hottest non-DA
  WDs

4
Introduction

• Hydrogen-rich WDs
• Photospheres of hot DAs are almost completely
  ionized, hence, very low opacity. Observed X-rays
  stem from very deep, hot layers.
• Soft X-rays are detected from objects with Teff
  gt30,000 K
• Famous example Sirius B
• He-C-O-rich (non-DA) WDs
• Opacities of He and metals prevent leakage of
  X-rays from deep hot layers, except for hottest
  objects, where these species are highly ionised
  and more transparent
• Soft X-rays are detected from objects with Teff
  gt140,000 K
• Famous example H150465

5
Outline

• Introduction Thermal soft X-ray emission from
  stellar photospheres
• Chandra spectroscopy of the hot DA white dwarf
  LB1919 Implications for vertical chemical
  stratification in WDs
• Chandra spectroscopy of the PG1159 star
  PG1520525 Constraining the GW Vir instability
  strip in the HRD
• Chandra spectroscopy of the naked C/O stellar
  core H150465 The hottest known and chemically
  most extreme white dwarf

6
Metals as sensitive regulators of X-rays from DA
white dwarfs

• Hydrogen in hot DAs almost completely ionized,
  EUV/soft X-ray opacity strongly reduced ? DAs
  with Teff gt30,000 K can emit thermal soft X-rays
  from deep hot layers
• However, ROSAT All-Sky Survey revealed that X-ray
  emission is the exception rather than the rule ?
  additional absorbers present
• ROSAT and EUVE revealed that metals are the
  origin, EUV spectra are strongly determined by Fe
  and Ni through large number of absorption lines
• Radiative levitation keeps traces of metals in
  the atmosphere (e.g. Chayer et al. 1995)
  Generally, metal abundances increase with
  increasing Teff.
• Consequently, only very few DAs with Teffgt60,000
  K were detected in EUVE and ROSAT All-Sky Surveys.

7

• Breakthrough in understanding DA atmospheres
  development of self-consistent models for
  equilibrium of gravitational settling / radiative
  levitation, yield vertical abundance
  stratification
• Generally, good agreement between observed and
  computed EUV flux distributions (e.g. Schuh et
  al. 2002)
• However, several exceptions are known. Some DAs
  show much larger metal abundances than expected
  from theory, reason ISM accretion or
  wind-accretion from unseen companion
• More difficult to explain objects with
  metallicity smaller than expected

8
The problem of metal-poor DA white dwarfs

• Prominent example HZ43 (Teff 49,000 K),
  virtually metal free, shows no EUV or X-ray
  absorption features
• Even more surprising low metallicity of two DAs
  with even higher Teff. Two of hottest known DAs
  have extraordinarily low metal abundances LB1919
  (69,000 K) MCT0027-6341 (60,000 K)
• These stars could hold the key to understanding
  metal-poor DAs as a class
• We concentrate on LB1919, it is brighter in
  EUV/X-rays
• LB1919 hottest of the 90 DAs detected in EUVE
  all-sky survey (Vennes et al. 1997). Hottest of
  the 20 DAs whose EUVE spectra were analyzed in
  detail by Wolff et al. (1998)
• Chemical composition unknown EUVE resolution
  insufficient to resolve individual lines
  metallicity of fainter DAs usually determined
  relative to G191-B2B (56,000 K) that is well
  studied by UV spectroscopy.

9
The problem of metal-poor DA white dwarfs

• Our stratified models successfully describe the
  EUV spectrum of G191-B2B. EUVE spectra of other
  DAs could also be fitted by simply scaling G191s
  relative metal abundance pattern.
• But the EUV spectrum of LB1919 cannot be
  described by chemically homogeneous models scaled
  to G191 relative abundances. Also, radiatively
  stratified models fail spectacularly.

10
The problem of metal-poor DA white dwarfs

• Four processes can disturb equilibrium between
  gravitation and levitation potentially
  responsible for metal-poor hot DAs
• Mass-loss tends to homogenize chemical
  stratification. However, M-dot drops below
  critical limit (10-16 M?/yr) for Tefflt70,000 K
  (Unglaub Bues 1998). So, LB1919 should not be
  affected.
• Wind-accretion calculations (MacDonald 1992) show
  that ISM accretion is prevented for LB1919 since
  its Lgt1L?. Instead, mass-loss rate of 10-18 M?/yr
  will be sustained
• Convection not expected in LB1919
• Rotation could lead to meridional mixing,
  however, WDs are generally slow rotators. In
  particular, LB1919 shows sharp Lyman line cores
  (FUSE), ruling out high rotation rate.
• Currently there is no explanation for the low
  metallicity in LB1919 and similar DAs

11
Chandra observation of LB1919

• Aim Empirical determination of abundance
  stratification using individual lines of
  different ionization stages of Fe, Ni ...
• IF metals are stratified, then they are in
  diffusion/levitation equilibrium. The low
  metallicity might originate in earlier
  evolutionary phases (selective radiation driven
  wind?)
• IF metals are homogeneous, then one of the above
  mechanisms is at work, contrary to our
  understanding

12

• Individual lines can in principle be identified
  in Chandra spectra, as was shown for the hot DA
  GD 246 (Vennes et al. 2002).

GD 246 - Chandra
13
Simulated Chandra observations of
LB1919LETGHRC-S, 120 ksec

• Parameters of LB1919
• Teff70,000 K logg8.2 Fe/H7.510-7
  Ni/H510-8
• (EUVE analysis of Wolff et al. 1998 with
  homogeneous models)

Two models shown a) nickel enhanced by 1 dex
(black line) b) iron and nickel enhanced by 1 dex
(red line) Strong sensitivity of the spectrum
against abundance variations
14
Chandra observation of LB1919

• Low Energy Transmission Grating (LETGHRC-S)
• Integration time 111 ksec, Feb. 02, 2006

15
Chandra observation of LB1919

• Low Energy Transmission Grating (LETGHRC-S)
• Integration time 111 ksec, Feb. 02, 2006
• Line features in model too strong, analysis
  is on-going, no results yet

16
Outline

• Introduction Thermal soft X-ray emission from
  stellar photospheres
• Chandra spectroscopy of the hot DA white dwarf
  LB1919 Implications for vertical chemical
  stratification in WDs
• Chandra spectroscopy of the PG1159 star
  PG1520525 Constraining the GW Vir instability
  strip in the HRD
• Chandra spectroscopy of the naked C/O stellar
  core H150465 The hottest known and chemically
  most extreme white dwarf

17

• The PG1159 spectroscopic class, a group of 40
  stars
• Very hot hydrogen-deficient (pre-) WDs
• Teff 75,000 200,000 K
• log g 5.5 8
• M/M? 0.51 0.89 (mean 0.62)
• log L/L? 1.1 4.2
• Atmospheres dominated by C, He, O, and Ne, e.g.
  prototype PG1159-035
• He33, C48, O17, Ne2 (mass fractions)
• chemistry of material between H and He burning
  shells in AGB-stars (intershell abundances)

18
late He-shell flash causes return to AGB
Evolutionary tracks for a 2 M? star. Born-again
track offset for clarity. (Werner Herwig 2006)
19

• Loss of H-rich envelope consequence of (very)
  late thermal pulse during post-AGB phase (LTP) or
  WD cooling phase (VLTP) (like Sakurais object
  and FG Sge)
• Hydrogen envelope (thickness 10-4 M?) is ingested
  and burned (VLTP) or diluted (LTP) in He-rich
  intershell (thickness 10-2 M?)
• In any case, composition of He/C/O-rich
  intershell region dominates complete envelope on
  top of stellar C/O core

20
Late He-shell flash
10-4 M?
10-2 M?
CO core material (dredged up)
21
Pulsating (filled circles) and non-pulsating
PG1159 stars
PG1159-035
PG1520525
0.6 M? track (Wood Faulkner 1986)
Blue (Gautschy et al. 2005) and red (Quirion et
al. 2004) edges of GW Vir instability strip
22
The pulsator/non-pulsator pair PG 1159-035 and
PG 1520525

• Very similar atmospheric parameters

PG1159-035 PG1520525
Teff 140,000 150,000
log g 7.0 7.5
He 0.33 0.43
C 0.48 0.38
O 0.17 0.17
Ne 0.02 0.02
Mass/M? 0.60 0.67
23
The pulsator/non-pulsator pair PG 1159-035 and
PG 1520525

• Do both stars confine the blue edge of the
  instability strip?
• To what accuracy is their Teff known?
• PG1159-035 140,000 /- 5,000 K
• from HST/STIS high-resolution UV spectrum, Jahn
  et al. (2007)
• PG1520525 150,000 /- 15,000 K
• from HST/GHRS medium-resolution UV spectrum,
  Dreizler Heber (1998)
• Need a more precise Teff estimate for PG1520525
• Try soft X-ray region PG1520525 is the soft
  X-ray brightest PG1159 star after H150465.
• First attempt with EUVE suggested Teff around
  150,000 K, however, poor-S/N spectrum

24
Werner et al. (1996)
25
Soft X-ray spectral modeling of PG 1520525

• Grid of non-LTE models (hydrostatic, radiative
  equilibrium)
• Ions included He I-III, C III-V, O IV-VII, Ne
  IV-IX, Mg IV-IX
• Particular model shown here tailored to
  PG1520525
• For comparison with Chandra observation model
  flux used to simulate Chandra count rate spectrum
  including expected S/N

26
Chandra observation of PG 1520525

• Low Energy Transmission Grating (LETGHRC-S)
• Integration time 142 ksec, April 04-06, 2006

27
Chandra observation of PG 1520525

• Low Energy Transmission Grating (LETGHRC-S)
• Integration time 142 ksec, April 04-06, 2006

28
Detail of PG 1520525 Chandra spectrum, 100-123 Å
model observation
29
Detail of PG 1520525 Chandra spectrum, 100-123 Å
NeVII OVI
OVI
NeVI
model observation
NeVII NeVI
Identification of OVI and NeVI / VII lines
30
FUV studies of PG1520525

• Element abundances affect soft X-ray flux, but
  due to difficult line identification there,
  constraints must be provided from other
  observations
• Crucial FUSE spectroscopy. All identified
  species in PG1520525 display lines in this
  wavelength region. Besides presence of He, C, O
• First identification of silicon, sulfur,
  phosphorus
• abundance determinations by Reiff et al. (2007)
• First identification of neon and fluorine
• strongly over-solar abundances (Werner et al.
  2004, 2005)

31
First discovery of fluorine in hot post-AGB
stars F VI 1139.50 Å F abundance in
PG1520525 200 times solar PG1159-035 10
times solar
32
140,000 K model too cool, good fit with 150,000 K
model.
33
Pulsating (filled circles) and non-pulsating
PG1159 stars
PG1159-035
PG1520525
0.6 M? track (Wood Faulkner 1986)
PG1159-035 and PG1520525 indeed confine the
blue edge of the GW Vir instability strip
34
Outline

• Introduction Thermal soft X-ray emission from
  stellar photospheres
• Chandra spectroscopy of the hot DA white dwarf
  LB1919 Implications for vertical chemical
  stratification in WDs
• Chandra spectroscopy of the PG1159 star
  PG1520525 Constraining the GW Vir instability
  strip in the HRD
• Chandra spectroscopy of the naked C/O stellar
  core H150465 The hottest known and chemically
  most extreme white dwarf

35
Properties of H150465

• 1983 H1504 is the 7th brightest X-ray source in
  the 0.25 keV band (HEAO1 survey, Nugent et
  al.)
• 1986 Optical identification Extremely hot
  white dwarf, lacking H and He lines (Nousek et
  al.)
• 1991 NLTE analysis of optical spectra (Werner)
  
• It is the hottest WD known (Teff close to 200
  000 K)
• H1504 is devoid of hydrogen and helium
• Dominant photospheric species C and O (5050)
• 1999 Analysis of EUVE Keck data (Werner
  Wolff)
• High neon abundance 2-5 (gt20 times solar)
• H1504 is an extreme member of the PG1159
  spectroscopic class

36

• Chandra LETGHRC-S observation of H150465
• Sept. 27, 2000, integration time 7 hours
• Richest absorption line spectrum ever recorded
  from a stellar photosphere
• (Werner et al. 2004, AA 421, 1169)
• Examples for spectral fitting

37
Model fit to H150465 Chandra spectrum 80-110 Å
Relative flux
Wavelength / Å
38
Model fit to H150465 Chandra spectrum 80-110 Å
Relative flux
Wavelength / Å
39
Model fit to H150465 Chandra spectrum 80-110 Å
Relative flux
Wavelength / Å
40
Model fit to H150465 Chandra spectrum 110-140 Å
Relative flux
Wavelength / Å
41
Model fit to H150465 Chandra spectrum 110-140 Å
Relative flux
Wavelength / Å
42
Model fit to H150465 Chandra spectrum 110-140 Å
Relative flux
Wavelength / Å
43
Strong Fe-group line blanketing
44
Strong Fe-group line blanketing
45
Strong Fe-group line blanketing
46

• Origin of unique C/O/Ne surface composition of
  H1504 remains unknown. Obviously, H1504 is a bare
  C/O core of a former AGB giant.
• Detection of Mg?2 in Chandra spectrum even
  suggests
• H1504 could be a bare O/Ne/Mg white dwarf, i.e.
  first observational proof for existence of such
  objects
• Approved HST UV-spectroscopy (2005) Search for
  Na, but
• failure of STIS just before observations should
  be done

47
Summary

• Hottest WDs have detectable photospheric soft
  X-ray emission
• X-ray grating spectroscopy important (and often
  essential) to derive stellar parameters and
  details of photospheric processes
• Results are relevant for our understanding of
  late phases of stellar evolution
• In detail Chandra spectroscopy of a hot DA white
  dwarf and of two PG1159 stars
• Analysis of LB1919 will provide clues to answer
  the question why some hot DAs show a lower
  metallicity than expected from radiative
  levitation theory
• PG1520525 confines the blue edge of the GW Vir
  instability strip (Teff140,000150,000 at
  logg7)
• H150465 could turn out to be the first
  definitive proof for the existence of (single)
  O/Ne/Mg white dwarfs