EUVE J0317 - 855: a rapidly rotating, high-field magnetic

1997MNRAS.292..205F

Mon. Not. R. Astron. Soc. 292, 205-217 (1997)

EUVE J0317 - 855: a rapidly rotating, high-field magnetic white dwarf

Lilia Ferrario/,2 S. Vennes/,2,3 D. T. Wickramasinghe/,2 J. A. Bailey4 and D. J. Christian3

lAstrophysical Theory Centre, School of Mathematical Sciences, The Australian National University, Canbe"a, ACT 0200, Australia 2Department of Mathematics, Faculty of Science, The Australian National University, Canbe"a, ACT 0200, Australia 3Center for EUV Astrophysics, University of California, Berkeley, CA 94720-5030, USA "Anglo-Australian Observatory, PO Box 296, Epping, NSW 2121, Australia

Accepted 1997 July 28. Received 1997 July 23; in original fonn 1997 April 28

1 INTRODUCTION

ABSTRACT Spectropolarimetric observations of the extreme-ultraviolet (EUV) source and high-field magnetic white dwarf EUVE J0317 - 855 (= RE J0317 - 853) in the wavelength range 4200-7200 A reveal a strongly structured circular polarization spectrum with a peak polarization of 8 per cent. The strongest features in the polarization spectrum are attributed to Hoc 2s0-3p1, 2s0-3pO and Hf3 2s0-4f-1, 2s0-4fO and 2s0-4p1 in a magnetic field of 100-300 MG. The phase-averaged data are modelled using centred and off-centred dipole field structures, and a reasonable fit to the data was obtained for a dipole of strength Bd=450 MG displaced from the centre of the star by 35 per cent of the stellar radius and viewed at a mean angle of 300 -600 to the dipole axis. The wavelength-averaged circular polarization data show variations at a period of 725 ± 10 s, consistent with the photometric period of 725 s reported by Barstow et al. We also report EUV photometric variations on a period of 725.5 ± 0.8 s. The EUV light CUlVe is double-peaked, a possible signature of surface abundance inhomogeneities, in contrast with the optical light curve which is single-peaked on the same period. We show that the optical photometric and polarization variations are most likely caused by changes in the surface-averaged field strength and the line-of-sight projected field structure with rotational phase, as would be expected from an oblique rotator model. The high rotation rate, high mass and age discrepancy with a close DA white dwarf companion (LB 9802) suggest that EUVE J0317 - 855 may be the outcome of a double-degenerate merger.

Key words: stars: individual: EUVE J0317 - 855 (= RE J0317 - 853) - stars: magnetic fields - white dwarfs.

EUVE J0317 - 855 (= RE J0317 - 853) was discovered as an extreme-ultraviolet (EUY) source by the ROSAT Wide Field Camera (Barstow et al. 1995, hereafter B95) and Extreme Ultraviolet Explorer (EUVE) (Vennes 1996) surveys. It was identified as a high-field magnetic white dwarf with a dipole field strength of 340 MG by B95, who also reported photometric variations in white light with a period of 725.4 s. If this period is due to rotation, the period is anomalous for isolated magnetic white dwarfs which typically have periods of a few hours. The estimated value of effective temperature (~40 000-50 000 K) is significantly higher

than that of PG 1658 + 441 (30500 K; Schmidt et al. 1992), PG 1031 + 234 (~1O 000-25000 K; Schmidt et al. 1986), or GD 229 ( ~ 12 000-20 000 K; Schmidt, Latter & Foltz 1990), making it the hottest known magnetic white dwarf.

The observed period of light variations of EUVE J0317 - 855 is consistent with that expected for non-radial pulsations in a white dwarf, but the high temperature of EUVE J0317 - 855 places the star well above the DA instability strip. The most likely explanation is that the optical variations are linked to the rotation of the magnetic white dwarf (B95), although the origin of such a rapid spin rate in an apparently isolated white dwarf remains a mystery. The period is similar to that of the white dwarf in

© 1997 RAS

© Royal Astronomical Society • Provided by the NASA Astrophysics Data System

at California State U

niv/Northrdg on June 7, 2016

http://mnras.oxfordjournals.org/

Dow

nloaded from

http://adsabs.harvard.edu/abs/1997MNRAS.292..205F


1997MNRAS.292..205F

206 L. Ferrario et at.

V471 Tau (see Jensen et al;. 1986 and Robinson, Clemens & Hine 1988), but this star is known to be in a close binary system with on-going accretion.

The hot DAp white dwarf EUVE J0317 - 855 is located some 16 arcsec from the cooler DA white dwarf LB 9802 (Luyten & Anderson 1967). It has been argued that, given the low space density of white dwarfs, these stars are likely to be physically associated. However, the implied separation (~1000 au) excludes any present or past evolutionary interaction of the two stars. B95 noted that the parameters of LB 9802 would then yield a distance of 33-37 pc for EUVE J0317 - 855, and a mass of 1.35 Mo (logg=9.5), close to the Chandrasekhar limit. EUVE J0317 - 855 may therefore also be peculiar in possessing a very high mass, and, like the massive DAp PG 1658 + 441 (Schmidt et al. 1992), may be a magnetic analogue of the white dwarf found to populate the high-mass tail in the mass distribution of isolated white dwarfs (Vennes et al. 1996, 1997; Finley, Koester & Basri 1997; Marsh et al. 1997). This raises the possibility that the high rotation rate may be the result of a merger in a binary system.

In this paper, we present spectropolarimetric observations of EUVE J0317 - 855 covering the wavelength range 4200-7200 A which show the star to be circularly polarized at a peak level of 8 per cent. The intensity and polarization spectra are rich in structure, and show features not previously reported in this star that confirm its high-field nature. An analysis of the phase-dependent polarimetric data shows variability in the wavelength-averaged broad-band circular polarization at a preferred period of 725 ± 10 s. We also present modelling of the spectroscopic and spectropolarimetric data, and investigate possible explanations for the observed light and polarization variations in terms of an oblique rotator model. Finally, we report the discovery of a double-peaked EUV (100 A) photometric variation phased on a period of 725.5 s, in contrast with the single-peaked optical light curve.

2 OBSERVATIONS

2.1 eTIO observations

On 1994 January 25 and 26, S. Vennes and John R. Thorstensen obtained, at the eno 4-m telescope, spectroscopic observations of candidate optical counterparts to a series of newly discovered EUV sources. As part of this programme, EUVE J0317 - 855 was observed on UT 1994 January 26 (Table 1) using the Reticon II CCD at a resolution of ~4.5 A (Vennes et al. 1997). The phase-averaged spectrum obtained with a total exposure of 23 min is shown in Fig. 1; note the highly displaced spectral features indicative of a large magnetic field found only in a few rare white dwarf stars. This spectrum first alerted us to the possible association of a new high-magnetic-field white dwarf with an EUV source.

2.2 AAT observations

Observations of EUVE J0317 - 855 were obtained with the Anglo-Australian Telescope (AAT) spectropolarimeter on 1994 October 10 (Table 1). This instrument consists of the RGO Spectrograph used with a thinned Tektronix CCD

Table 1. enD and AAT observation log.

Date HJD t.xp Instrument (UT) (2449000+) (8) 1994 Jan 26 378.67600 180 CTIO 4m 1994 Jan 26 378.68747 600 CTIO 4m 1994 Jan 26 378.69540 600 CTIO 4m 1994 Oct 10 636.04689 628 AAT4m 1994 Oct 10 636.06142 628 AAT4m 1994 Oct 10 636.06936 628 AAT4m 1994 Oct 10 636.07707 628 AAT4m 1994 Oct 10 636.08484 627 AAT4m 1994 Oct 10 636.09248 647 AAT4m 1994 Oct 10 636.10038 627 AAT4m 1994 Oct 10 636.10802 628 AAT4m 1994 Oct 10 636.11621 628 AAT4m 1994 Oct 10 636.12395 628 AAT4m 1994 Oct 10 636.13213 627 AAT4m 1994 Oct 10 636.13996 628 AAT4m 1994 Oct 10 636.14801 628 AAT4m 1994 Oct 10 636.16408 627 AAT4m 1994 Oct 10 636.17184 627 AAT4m 1994 Oct 10 636.17952 628 AAT4m 1994 Oct 10 636.18728 627 AAT4m 1994 Oct 10 636.19492 627 AAT4m

detector and a modulator consisting of a rotating superchromatic quarter-wave plate. A calcite block analyser, positioned behind the slit, produces two spectra corresponding to orthogonal polarization states. Switching the quarter-wave plate between two positions which are 90° apart modulates these spectra according to the circular polarization. Each pair of exposures at the two place positions was reduced to give intensity and V Stokes spectra using the Starlink TSP package (Bailey 1992). Spectropolarimetric observations were obtained over the wavelength range 4200-7200 A with a spectral resolution of ~ 10 A and a time resolution of about 9.5 min. We present in Fig. 2 the summed polarization and intensity spectra appropriately binned in wavelength.

2.3 Mount Stromlo observations

We complete the optical data set with a spectrum of the companion star, LB 9802, obtained at the Mount Stromlo 74-inch telescope on UT 1997 April 1. We used the Cassegrain spectrograph with a 300 line mm- 1 grating set at 2.78A pixel-I, or a spectral resolution of 5-6A. We exposed LB 9802 for 30 min. We reduced the spectrum using standard lRAF procedures; the spectrum covers the wavelength range from 3714 and 6116 A.

2.4 FUV and EUV observations

We obtained EUV photometric and far-ultraviolet (FUV) spectrophotometric measurements of EUVE J0317 - 855 from the EUVE archives at the Center for EUV Astrophysics (Berkeley) and the International Ultraviolet Explorer (IUE) archives at NASA's Goddard Space Flight Center. EUVE J0317 - 855 was observed with the EUVE Deep Survey (DS) imager in the lexan bandpass (A ~ 100 A) from (UT) 1996 June 3 05:03:04 to 1996 June 5 09:51:35 for an

© 1997 RAS, MNRAS 292, 205-217





Dow

nloaded from



1997MNRAS.292..205F

EUVE J0317 - 855: a rapidly rotating white dwarf 207

10-14

rn '" 8x 10-15

S ()

'-.. rn l>ll ~ 6xl0-15 ~

4xl0-15

2xl0B

1.5xl0B Ui' rn ;::J oj

2- lOB o:l

5x107

0 4500 5000

Wavelengths (1) 5500

Figure 1. (Top) Phase-averaged CTIO spectrum ofEUVE J0317 - 855 obtained on 1994 January 26. (Bottom) Diagram showing the fieldwavelength dependence of Zeeman components of hydrogen.

effective exposure time of 42 000 s; the selected time intervals exclude high-background events and Earth occultations. We co-added the [UE images L WP 28287 and 28288, and the images SWP 50926, 50927 and 50928, and built a FUV energy distribution covering the wavelength range from 1150 to 3250 A; we corrected the spectra for temporal degradation using a template based on contemporaneous spectra of the hot white dwarf G191 B2B.

2.5 Zeeman features

The spectrum of EUVE J0317 - 855 is very steep, consistent with a high effective temperature, and exhibits circular polarization in the continuum up to a maximum of 8 per cent in our measurements, thus confirming the high-field nature of this object. Weak absorption lines are seen in the data, spanning the entire observed wavelength range, and these are associated with structure in the circular polarization spectrum. We have identified these features using the results of Zeeman calculations of hydrogen calculated by the Tiibingen group (Forster et al. 1984; Rosner et al. 1984; Wunner 1990), supplemented by calculations of Kemic (1974) and Henry & O'Connell (1984, 1985). The calculations for the strongest lines are plotted in a magnetic field (B) versus wavelength (A) diagram at the bottom of Figs 1 and 2.

© 1997 RAS, MNRAS 292, 205-217

The strongest of the observed features occurs at 6790 A, which we attribute erimarily to Hoc 2s0-3pl, which is nearly stationary at 6800 A for fields in the range 100-300 MG. The Hoc transition 2p + 1-3s0, which is intrinsically weaker in oscillator strength, has a turning point at 6500 A at 130 MG, and is also expected to make some contribution to the profile of this feature. The extended red wing of this feature is seen in the polarization spectrum as being due to a third component centred at 7100-7200 A, and is most likely to be caused by a blend of Hoc 2p-2-3d-2 and Py 3dO-5pO. Blueward of this blend, one encounters a feature centred at 5850 A, which we ident~ with Hoc 2s0-3pO that has a broad turning point at 5500 A. This feature is a blend of several faster moving components of Hoc and Hf3 (in particular the redward moving component of Hf3). A deeper feature occurs at 4700A, which we attribute to Hf3 2s0-4f-l in a field of 100-200 MG. Two further features centred at 4350 and 4220 A can be assigned to Hf3 2s0-4fO and Hf3 2s0-4pl respectively, corresponding to a similar field range. We conclude that all of the weak spectral features seen in the optical spectrum of EUVE J0317 - 855 can be attrributed to Balmer Hoc and Hf3 features in a field range 100-300 MG. A detailed modelling of this spectrum will be presented in Section 3.

At the time of our observations we were unaware of the 725.4-s photometric period (B95). The time resolution that





Dow

nloaded from



1997MNRAS.292..205F

208 L. Ferrario et al.

.§ 0.1 ..., ~ 0.08 .s:: ttl '00.06 0..

~ 0.04 "3 C)

.=: 0.02 u

o ~-+--+-~--~--~~--+--+--+-~--~--r-~--~-T~

s CJ ....... til bIJ r...

10-14

~ 5xl0-15

><: ~

~

0

2xl0B

....... 1.5xl0B til til ~ ttl 0 lOB ~

co

5xl07

0 4000 5000 6000

Wavelengths (A)

Figure 2. Phase-averaged AA T spectra of EUVE 10317 - 855 obtained on 1994 October 10. (Top) Polarization spectrum. (Middle) Intensity spectrum. (Bottom) Diagram showing the field-wavelength dependence of Zeeman components of hydrogen.

we adopted for the spectropolarimetric observations is therefore not optimal for detecting this period in our data. Nevertheless, we were fortunate that the time span from the beginning to the end of each set of two exposures is ~ 630 s, i.e., about 86 per cent of the rotation period, so that each spectrum samples a different part of the rotational phase of the star. The spectra were obtained at regular intervals, on average 680 s apart. We have constructed wavelength-averaged (4200-6500 A) polarization curves and calculated residuals relative to sinusoidal fits for periods between 300 and 1800 s. These results are shown in Fig. 3. A broad peak is seen at a period of 725 ± 10 s (QDAp)' which demonstrates a direct correspondence between photometric and circular polarization variations (Fig. 3, bottom). A secondary peak, near 677.9 s (QAAT)' corresponds to the data sampling frequency. All other periods are convincingly excluded. The wavelength-averaged polarization measurements slowly vary between 4.8 and 6.5 per cent on a time-scale of 0.12 d (Fig. 3, top); the effect is clearly related to the beating frequency (QAAT - QDAp) at 0.0966 mHz (2h53m). The residual of the circular polarization at the best-fitting period (Fig. 3, top) is 0.15 per cent, i.e., a relative error of only 3 per cent; the epochs of polarization minima are HID

2449636.04588 + 0.008 3958E, where E is an integer. Clearly, the polarization is variable, and is consistent with the period of 725 s found by B95 in their SAAO photometric data.

We extracted some information on the variability of the polarization spectra as follows. We assume that the photometric and polarization variability is caused by the rotation of the star, and that the relative calibration (in wavelength and from one spectrum to another) is accurate and not subjected to systematic errors larger than 10 per cent. The star is divided into two hemispheres, A and B, corresponding to maximum and minimum polarization respectively. We calculate the fraction of time spent in each hemisphere in a given exposure (Ai' Bi), and co-add in Sl the spectra exposing mostly hemisphere A, and in Sz the spectra exposing mostly hemisphere B. Typically, A i or Bj are of the order of 0.4-0.6, and there are about an equal number of spectra in each set. We assume that the flux in each spectrum Fi is a linear combination of the fluxes FA and FB from hemispheres A and B respectively:

Fi=AiFA + BiFB·

We then sum in Sl and Sz:

© 1997 RAS, MNRAS 292, 205-217





Dow

nloaded from



1997MNRAS.292..205F

EUVE J0317 -855: a rapidly rotating white dwarf 209

i\ 7 .::: o ." ....., cd 6 N .~

~ o 5

0...

~ v 4

AAT 1994 Oct

636.05 636.1 636.15 636.2 HJD 2449000+

60

°DAP

,.-., 40 C\l I

~

C\l I

20 b

0 1.5 2 2.5 3

v(10-3 Hz)

Figure 3. (Bottom) Periodogram of the AAT circular polarization spectra showing a peak at 725 ± 10 s (QDAp) well aside from the data sampling frequency near 680 s (QAAT). (Top) Wavelength-averaged circular polarization showing variations on a period of 725 s.

SI = IJi=FA IAi +FB I Bi

S2=IFi=FA IAi+FB IBi

(Ai>B.),

(Bi>AJ

We solve the above linear system for FA and FB, and for the initial phase To. The resulting spectra, FA and FB, and the wavelength-averaged polarization curve are shown in Fig. 4.

The data show dramatic changes in the level of wavelength-averaged continuum polarization, from a peak level of ~ 10 per cent down to ~O per cent. The inferred polarization variations result quite naturally from the linear model adopted: if one is to find a contrast between subsequent exposures that do cover largely overlapping areas, the intrinsic contrast between hemispheres A and B has to be extreme. It remains to be investigated whether more sophisticated descriptions of the white dwarf intrinsic polarization properties would lead to different results. We do not have at our disposal sufficient information to study spectral line variations as a function of phase, but such changes are expected in a rotating magnetic white dwarf which presents different field aspects to the observer at different rotational phases. Changes with rotational phase of the line spectrum have been reported in several magnetic white dwarfs, including the high-field magnetic white dwarf PG 1031 + 234 (Schmidt et al. 1986; Latter et al. 1987) and the intermediate-field magnetic white dwarf PG 1015 + 015 (Wickramasinghe & Cropper 1988). The circular polarization spectrum is expected to be even more sensitive to field geometry than the line spectrum, since it also carries information on field direction.

© 1997 RAS, MNRAS 292, 205-217

2.6 EUV ( ~ 100 A) photometric variations

Photospheric emission from hot white dwarf stars extends beyond the EUV range into soft X-ray wavelengths. EUVE J0317 - 855 is a variable star, and a multiwavelength study of its pulsation properties may help discriminate between a number of models. For example, the 0.5-cycle phase offset between optical and EUV/soft X-ray light curves of the hot DA white dwarf in the binary V 471 Tau bears the signature of surface abundance inhomogeneities. On the contrary, pulsations driven by partial ionization in the envelope are coherent. We have analysed the 1996 June EUVE observation of EUVE J0317 - 855 using the ROSAT/PROS IRAP

package; we searched for periodicity between 250 and 900 s and identified two periods, P=362.9 and 725.5 s. The shorter period is the first harmonic (2.76 rnHz) of the fundamental frequency at 1.38 mHz (P=725.5 s). Fig. 5 (bottom) shows the periodogram of the EUV light curve and (top) the folded light curves on the preferred periods. The EUV light curve has a double-peak appearance on a period characterized by a single peak in optical light curves. In both cases we find the heliocentric epoch of the first minimum To=HJD 2450237.72019.

3 MODELLING

3.1 Stellar parameters of the DA2 white dwarf LB 9802 and DAlp EUVE J0317 - 855

Fig. 6 shows a pure-hydrogen model atmosphere fit to the H I Balmer line series in LB 9802. Following Wood's





Dow

nloaded from



1997MNRAS.292..205F

@

~

0 ~ ~ - > III ..... ., 0 =

0 :3 .... ~ ~ - rJ). 0 ~ .... ~ .....

~ • "'C ., 0 -< .... Q.

~

Q.

r:::;' ~ ~

~ z >

rJ).

>

>

III ..... ., 0 'C

@

=-.....

. ~

\0

\0

III

-.I

.... ~ ~

III ~

~

2 ~

rJ).

~ ~

III .....

N

~

15 :3

~

VI I N ......

-.I

15

/\

Q

0 :.;:::;

10

<tl

N

...... ... <tl

......

5 0 0..

~ v

0

Q

20

o :.;::

:; <tl

N

.~

10

<tl

...... o 0..

~

o

I- 0

-l

0.5

1 1.

5 2

Ph

ase

(P

=

725

".4)

EU

VE

J0

31

7-8

55

Ph

ase

0

.5

~.T·':-

-~:~n

;;t~:~.:

·-;-·T~(

:~.f:-:~

:; i.nr

:~rj ~

Hj ~'r·

;:-: ..

40

00

5

00

0

6000

7

00

0

A(A

) F

igu

re 4

. (T

op)

Wav

elen

gth-

aver

aged

cir

cula

r po

lari

zati

on a

s a

func

tion

of

phas

e. (

Bot

to

m)

Lin

earl

y de

conv

olve

d A

AT

cir

cula

r po

lari

zati

on s

pect

ra (

50-A

bin

s) c

orre

spon

ding

to

min

imum

(B

, pha

se =

0.0)

and

max

imum

(A

, ph

ase

=0.

5) p

olar

izat

ion.

0.3

0.2

~ I rtl

rtl

0.1

....,

Q

0.3

;::J

0 C)

0.2

L

P

=

36

2".

9

0.1

0 0.

5 1

1.5

2 P

has

e

rn

rn

~

CD

4f-

I.()

(\l

(\

l l'-

co

~

I:'J

2 o 1

2 3

4 v(1

0-3

H

z)

Fig

ure

5. (

Bot

tom

) P

erio

dogr

am o

f th

e E

UV

E D

S co

unt

rate

s sh

owin

g va

riab

ility

on

a 72

5.5-

s pe

riod

and

a 3

62.p

-s p

erio

d. (

Top

) E

UV

E D

S lig

ht c

urve

s fo

lded

on

the

two

mos

t si

gnif

ican

t pe

riod

s; t

he l

ight

cur

ve f

olde

d on

the

725

.5-s

per

iod

is d

oubl

e-pe

aked

, in

co

ntra

st w

ith t

he o

ptic

al li

ght c

urve

(B

95)

whi

ch is

sin

gle-

peak

ed o

n th

e sa

me

peri

od;

the

peri

od a

t 36

2.9

s is

sim

py t

he f

irst

har

mon

ics

of th

e 72

5.5-

s pe

riod

.

tv

~

o t"

~

~ ~.

~

$::l :-




Dow

nloaded from



1997MNRAS.292..205F


x 1.50 ;:l

r,;: ~ 1.00 :::l ., ~ 0.50

LB 9802

Te"=16360 log g=8.41

0.00 L.....L......L-~I-L-.1-.L_L.-L.I...L-.l..-L-'IL......L..-'--'

8.60

8.50

b()

b() 8.40 -.£

8.30 -

8.20 16.0

-100.0

I

0.0 aA(A)

I

@:~

I I

16.2 16.4 Te" (103 K)

100.0

I

-

-

-

I

16.6 16.8

Figure 6. (Bottom) Confidence contours (66, 90 and 99 per cent) in the (Tef(, logg) plane, and (top) best model atmosphere fit to the Balmer line series in the DA2 LB 9802.

Table 2. Inferred properties of EUVE J0317-855.

Teff (K) 30000 40000 50000

log 9 (c.g.s. ) 9.33 9.52 9.62

M (Me) 1.31 1.35 1.37

Mv (mag) 12.49 12.51 12.52

(1995) mass-radius relations, the measured temperature (Tefl = 16 360 ± 80 K) and surface gravity (logg=8.41 ± 0.02) are characteristic of a high-mass white dwarf (0.85 ± 0.01 Mo). The star therefore lies at a distance of 30 pc (M - M = 2.39). If the DA2 star LB 9802 and the DAlp star EUVE J0317 - 855 are physically associated, our distance estimate toward LB 9802 can be used to infer some properties of EUVE J0317 - 855. Adopting LG 9802's distance modulus, we can infer the absolute magnitude of the DAlp EUVE J0317 -855, Mv=12.51, characteristic of either cool white dwarfs or massive white dwarfs (see Vennes et a1. 1996, 1997). The FUV and EUY continuum measurements do imply a high temperature (Tefl> 30 000 K), and so a high mass is likely. Table 2 summarizes inferred properties of EUVE J0317 - 855 for temperatures between 30000 and 50000 K. The mass is between 1.31 and 1.37 Mo. The inferred age of the white dwarf LB 9802 is 0.4 Gyr, and the age of EUVE J0317 - 855 is far more uncertain, but possibly close to ~0.1-0.3 Gyr following Wood's evolutionary models at 1.1 and 1.2 Mo extrapolated to 1.3 and 1.4 Mo. The DAlp may therefore be only slightly younger than the DA2 star (-0.1 Gyr) or much younger (-0.3 Gyr), depending on its exact mass and temperature.

In the context of single-star evolution, the DAlp star EUVE J0317 - 855, being much more massive, would be

© 1997 RAS, MNRAS 292, 205-217

the evolutionary outcome of a main-sequence star earlier (Mi;;::: 6 Mo) than the progenitor of the white dwarf LB 9802 (Mi=3-6Mo; Weidemann 1987). Accounting for uncertainties in intial mass-final mass relations as well as in the stellar parameters, we find that EUVE J0317 - 855 should have preceded the progenitor of LB 9802 by ~ 30 to 300 Myr on the white dwarf cooling sequence. We found that the DAlp is either younger than DA2, or at best has approximately the same age as the DA2. This age paradox can possibly be resolved if we accept the idea that massive white dwarfs, such as EUVE J0317 - 855, represent the outcome of double degenerate mergers, which effectively involve evolution of low-mass main-sequence stars.

3.2 The DAlp white dwarf EUVE J0317 - 855

We have attempted to model the spectropolarimetric data using the magnetic white dwarf stellar atmosphere code developed by Martin & Wickramasinghe (1979) and Wickramasinghe & Martin (1979). The code and the method of modelling have been described in previous papers (see Wickramasinghe & Ferrario 1988; Achilleos & Wickramasinghe 1989) and reviews (Wickramasinghe 1990), and the reader is referred to these for details. The modelling procedure allows for magnetic field structures that are dipolar or quadrupolar with offsets from the centre of the star, takes into account magneto-optical effects, and allows for field dependent bound-free hydrogen continuum opacity following the Jordan (1988) prescription, magnetobremmstrahlung, and bound-bound hydrogen line opacities (see Wickramasinghe 1995) in the polarized radiative transfer calculations. However, the temperature-pressure structure of the underlying atmosphere is assumed to be unaffected by the magnetic field; this is expected to be a reasonable assumption at the fields under consideration.

Previous studies have shown that offset dipoles are a good first approximation to the field structure of isolated magnetic white dwarfs, particularly at low fields. However, in some cases more general offsets are required, and in systems such as PG 1031 + 254 a single dipolar component appears not to be adequate to represent the data (Schmidt et a1. 1986).

In this paper, we consider only a restricted class of centred and off-centred dipolar field structures, with the offset assumed to be along the dipolar axis. The amount of offset is specified by z =d/~d' where d is the displacement of the dipole from the centre of the star, and RWd is the radius of the white dwarf. The dipole is assumed to be viewed at an angle i to the magnetic axis, and a negative offset corresponds to the stronger pole being in the hemisphere further away from the observer. The models are labelled by B d , the polar field strength of the equivalent centred dipole. We emphasize that in reality the field structures are likely to be more complex than assumed in our analysis.

The zero-field atmospheric structures were calculated using ATLAS (Kurucz 1971). We assumed an effective temperature Te of 50000 K, consistent with the IUE observations of B95 and a pure hydrogen composition. Given the absence of a suitable Stark broadening theory at high magnetic fields, and the additional uncertainties in dealing with field structure and consequent magnetic field broadening, we are unable to estimate the gravity with any accuracy from





Dow

nloaded from



1997MNRAS.292..205F


line profiles, and we have therefore adopted the standard value logg=8.0 in our analysis.

As noted in Section 2, the field strengths represented in this star appear to cover the range 100-300 MG similar to what was found for Grw + 70°8247. The spectrum of Grw + 70°8247 was well represented by a centred dipole with a polar field strength of 320 MG viewed near i = 0° (Wickramasinghe & Ferrario 1988). There are gross similarities

(a) Bd=200 MG. Z= 0.0

4500 5000 5500 6000 Wavelengths (1)

(c) Bd=340 MG. Z= -0.20

4500 5000 5500 6000

Wavelengths (1)

between the spectra of the two stars, although they appear quite dissimilar in detail; this must indicate an intrinsically different field structure.

The intensity spectra of a series of such models corresponding to different values of the dipolar magnetic field strength and offset parameter z are presented in Figs 7(a)(d). The models correspond to: Bd = 200 MG, Z = 0, i = 30°, 60°,90°; Bd = 250 MG, Z= -0.10, i=30°, 60°, 90°; Bd =340

(b) Bd=250 MG. Z= -0.10

4500 5000 5500 6000

Wavelengths (1)

(d) Bd=450 MG. Z= -0.35

4500 5000 5500 6000

Wavelengths (1)

Figure 7. Intensity spectra of a series of models corresponding to different values of the dipolar magnetic field stength and offset parameter z. The CTIO data are also shown in each panel.

© 1997 RAS, MNRAS 292, 205-217





Dow

nloaded from



1997MNRAS.292..205F

EUVE J0317 - 855: a rapidly rotating white dwarf 213

0.06

0.06

0.04

0.02

1=90'

O~~~~~~~~~~~~=+~ 0.06

0.06

0.04

0.02

1=60'

O.~ ~-I--~+--+---+-+-~+--+--+-+-~+--+--+''--t---'l 0.06

0.06

0.04

0.02

O.G! i=-+-l-+---=+---+---+-!-+-+---+---;-;t--!r-+-+---+---+L....f-3

4000 5000 6000 7000 Wavelengths (1)

0.06

0.06

0.04

0.02

" o ~~L+~-+~~~~~~~+-+-+-+-~~~

:3 0.06 .. .~ 0.06 .. '0 0.04 0..

1;; 0.02 '3

~ o. ~ 1=-+---1-+-=-+---+---+-\-+-+---+---+-\-+-+---+----11--\-3 0.06

0.06

4000 5000 6000 7000 Wavelengths (1)

0.06

0.06 i=90o

0.04

0.02

0

" :3 0.06 .. .~ 0.06 .. '0 0.04 0..

:. 0.02 '3 0 0 .. U 0.1

0.06

0.06

0.04

0.02

O.G!

0.04

0~-4~00-0~~~~~5~00-0~~-~~6~00-0~~-~~7~00~0~

Wavelengths (1)

0.06

0.06

0.04

0.02

" 0l=-+---I-~~~'--~+--+-_I___+_--I--+-_+--f~'--il---~ :3 0.06 .. '5 0 .06

'0 0.04 0..

1;; 0.02 '3 e 0 u 0.1 k-lf-~+_--+---+-If-+-_+____+__-+-lf_+-_+____+__--'fI-lH

0.06

0.06

0.02

O.G 4---1....Lr-1.l+--+--Jl----I-I--+---+--:-rI--l-I--+---+----\l-l-3

0.08

0.06

0.02

4000 5000 6000 7000 Wavelengths (1)

Figure 8. Circular polarization spectra of a series of models corresponding to different values of the dipolar magnetic field strength and offset parameter z. The AAT phase-averaged circular polarization data are also shown in each panel.

MG, Z= -0.20, i=30°, 60°, 90°; and Bd=450 MG, Z= -0.35, i=30°, 60°, 90°. The corresponding circular polarization curves are presented in Figs 8(a)-(d). These models show, as expected, that the spectrum is very sensitive to field structure.

It is evident from Fig. 7(a) that centred dipole models are inadequate to explain the spectrum of EUVE J0317 - 855. In particular, the spectrum of EUVE J0317 -855 shows

© 1997 RAS, MNRAS 292, 205-217

many features that appear as rapidly moving components in the magnetic field-wavelength diagram, which would normally be obliterated by magnetic broadening given the field spread of a centred dipolar field distribution. Off-centred dipole models provide a better representation of the data if the hemisphere with the weaker pole dominates the phaseaveraged spectrum. With such a viewing geometry, the field is more uniform than a dipole field, and as a result, the fast-





Dow

nloaded from



1997MNRAS.292..205F


moving components have a greater impact on the spectrum, providing, for instance, an identification for the features at 4700 and 5200 A (Figs 7b-d).

The models closest to the phase-averaged intensity spectrum are those with the following parameters: Bd = 450 MG, Z= - 0.35, i =30°-60°. We show in Fig. 9 the i=30° model, together with the CTIO and AAT observations. The choice of these models is based primarily on positions and profiles of lines, but also takes into consideration the circular polarization spectrum (see below).

At the high temperatures appropriate to EUVE J0317 - 855, the continuum emission and polarization arise mainly from magneto-bremsstrahlung. The complex structure observed in the circular polarization spectrum over and above the level of continuum polarization is caused by the presence of Zeeman absorption lines. The sign of circular polarization exhibited by a given line «(J - ,

(J +) depends on various factors. In general, the sign of the projection of the local magnetic field on to the line of sight plays a central role. For a general field distribution, this sign changes across the visible stellar surface, giving rise to reversals in the sign of circular polarization relative to the level of continuum polarization. The extent to which these reversals are seen strongly constrains field structure. Therefore, if a (J component is spread in wavelength because of

the field spread across the stellar surface, it can exhibit both positive and negative circular polarization. In addition, Faraday repolarization can cause a n component to be circularly polarized (Achilleos & Wickramasinge 1995). These effects are clearly seen in the circular polarization spectra shown in Figs 8(a)-(d). We note in particular that the reversals in the sign of circular polarization relative to the continuum across a (J component are strongly dependent on field geometry.

Our circular polarization spectrum does not show evidence of dramatic changes in the sign of circular polarization relative to the continuum across lines, therefore indicating that one sign of projected field dominates in the averaged spectrum. The spectrum is characterized by (J

components of negative circular polarization, although it should be noted that the component that we have identified with the n component Ha 2s0-3pO is also negatively circularly polarized. These results are generally consistent with the predictions of the favoured models as shown in the comparison between theory and observations in Fig. 10. The polarization spectrum perhaps favours a mean viewing angle closer to i = 60°. We note that the level of continuum polarization is well matched by the theory. However, the lack of detailed agreement with the structure seen in the lines may point to an underlying field distribution that is

Bd=450 MG. Z= -0.35

4000 5000 6000

Wavelengths (1) 7000

Figure 9. eno and AAT intensity spectra shown with the calculated spectrum corresponding to a dipolar magnetic field strength of 450 MG and an offset of 35 per cent of the star radius along the dipole axis away from the observer.

© 1997 RAS, MNRAS 292, 205-217





Dow

nloaded from



1997MNRAS.292..205F


0.1

0.08

0.06

0.04

0.02

O.ID

s:: :3 0.08

<IS N

.~ 0.06 '0 0.. ..... 0.04 <IS

'3 ~ 0.02 G

O.ID

0.08

0.06

0.04

0.02

0 4000 5000 6000 7000

Wavelengths (A)

Figure 10. AAT phase-averaged circular polarization spectrum with (superimposed) the calculated spectrum corresponding to a dipolar magnetic field strength of 450 MG and an offset of 35 per cent of the star radius along the dipole axis away from the observer.

fundamentally different from what we have considered. Again we emphasize that we are modelling phase-averaged data, and, given the complex nature of the circular polarization spectrum, closer agreement with observations may not necessarily be expected.

The photometric variations seen in EUVE 10317 -855 remained unexplained in B95. Our modelling provides an explanation in terms of changes in the continuum opacity (magnetic dichroism) and associated changes in flux as a function offield strength across the stellar surface. We show in Fig. 11 a series of spectra corresponding to viewing angles to the dipole axis varying from 0° to 180°. We have also plotted in Fig. 12 the theoretical light curve in white light. Each of the points of the light curve corresponds to one of the spectra of Fig. 11. The predicted light amplitude is more than adequate to explain the observations, strongly supporting an oblique rotator model for this star. However, the double-peaked EUV light curve variations cannot be explained by magnetic dichroism. On the other hand, other effects, such as hotspots located near the magnetic poles or, more likely, heavy-element trace abundances (He or heavier elements) correlated with the poles, could induce

© 1997 RAS, MNRAS 292, 205-217

the observed EUV variations over a rotation cycle (725.3 s).

Finally, we turn to the IUE spectrum of EUVE 10317 - 855, which shows a strong feature at 1300 A. This feature can be identified as LylX Is0-2p-l (B95), and agrees with our model prediction for i = 30° as shown in Fig. 13. It is tempting to identify the associated feature at 1300 A also with LylX at a slightly higher field. The calculations show that this line shifts from 1300 to 1330 A as the field increases from 130 to 180 MG. If these two fields dominate at opposite rotational phases, it is possible that a doublet-like feature may appear in the phase-averaged spectrum.

4 DISCUSSION AND CONCLUSIONS

Our observations have shown that EUVE 10317 - 855 exhibits strong continuum and line circular polarization consistent with the high field attributed to this system. The broad-band circular polarization appears to vary with a period of 725 ± 10 s, providing strong support to the suggestion that the 725-s photometric period is caused by rotation of the magnetic white dwarf. EUVE 10317 - 855 is probably





Dow

nloaded from



1997MNRAS.292..205F


5000 6000 Wavelengths (1)

7000

Figure 11. Predicted spectra as a function of viewing angles between, from top to bottom, 180° and 0°.

I I I 000

0.6 f- o 0 -o o

o

>< ..:: 0.55 l-

0 -0

e 0

0 lD C\l 0.5 f- -

0

0 0

0.45 I-I I

I 0 0-'

o 50 100 150 Viewing angle (0)

Figure 12. Maximum range of variations that could occur assuming that the offset dipole is observed at all angles to the dipole axis.

associated with a normal DA2 white dwarf (LB 9802), and the pair is at a distance of 30 pc; the distance and apparent magnitude imply a mass of 1.31-1.37 Mo for the DA1 star EUVE J0317 - 855. Vennes et al. (1996, 1997) suggest that the large number of massive white dwarfs detected at EUV wavelengths could represent a population of double degenerate mergers. Its high rotation rate, high mass, and age discrepancy with the close DA white dwarf companion (LB 9802) suggest that EUVE J0317 - 855 itself may be the outcome of a double degenerate merger.

All spectral features in the optical band have been identified with Balmer lines of hy'drogen at fields of 100-300 MG. The feature at 1300-1330 A has been attributed to LylX. We have investigated centred and off-centred dipole models (restricted to offsets along the dipole axis) for this star, and

10-12

~ rn 8x 10-13

S CJ

"- 6x 10-13 rn QI) I-. (I)

'-" 4x 10-13 >< ;:l

r;:: 2xl0-13

1200 1300 1400 1500 1600 Wavelengths (1)

Figure 13. IUE intensity spectrum obtained by B95 showing the region between 1150 and 1600A. with (superimposed) the calculated spectrum corresponding to a dipolar magnetic field strength of 450 MG and an offset of 35 per cent of the star radius along the dipole axis away from the observer.

have shown that the best model to interpret our spectropolarimetric observations requires a dipole that is offset by ~ 35 per cent of the stellar radius. The model parameters that we deduce are in good agreement with the parameters of the best-fitting model presented by B95 based on their phase-averaged spectrum covering bluer wavelengths.

Small-amplitude photometric variations (0.1 mag in V) have been reported previously in the rotating magnetic white dwarf Feige 7 (Achilleos et al. 1992), which has a rotation period of 132 min. This star has a dipolar field Bd = 36 MG off-centred by 15 per cent, and the photometric variations were attributed to changes in the mean field strength over the visible stellar surface with rotational

© 1997 RAS, MNRAS 292, 205-217





Dow

nloaded from



1997MNRAS.292..205F

EUVE J0317 - 855: a rapidly rotating white dwaif 217

phase, and the field dependence (magnetic dichroism) of continuum opacity. In Feige 7, the effect was amplified by inhomogeneities in chemical composition across the stellar surface.

A similar interpretation appears plausible for EUVE J0317 - 855. In particular, the field is higher, and the effects of magnetic dichroism are therefore more apparent. We have investigated the expected change in light intensity for a rotating offset dipole model with the parameters that we have deduced for this system. Our calculations have shown that an oblique rotator model can readily explain the optical light (and polarization) variations seen in this star. Furthermore, the EUY light curve suggests the presence of surface abundance inhomogeneities (caused by He or heavier elements) at both magnetic poles which would give rise to the additional periodicity at half the rotation period of the white dwarf.

We conclude that our observations of EUVE J0317 - 855 strongly point towards an oblique rotator model. The predictions of the nature of the phase-dependent variations, particularly in circular polarization, are quite specific to assumptions on field geometry, and further observations should reveal if the field geometry is as simple as that of an offset dipole, or whether intrinsically more complex field structures are implicated.

ACKNOWLEDGMENTS

We thank the AAT staff for their technical support. This work is supported in part with NASA grant NAG5-2636. We are indebted to John R. Thorstensen for his assistance with the CTIO observations.

REFERENCES

Achilleos N., Wickramasinghe D. T., 1989, ApJ, 346, 444 Achilleos N., Wickramasinghe D. T., 1995, MNRAS, 272, 24 Achilleos N., Wickramasinghe D. T., Liebert J., Saffer R. A.,

Grauer A. D., 1992, ApJ, 396, 273 Bailey J. A., 1992, Starlink User Note 66 Barstow M. A., Jordan S., O'Donoghue D., Burleigh M. R.,

© 1997 RAS, MNRAS 292, 205-217

Napiwotzki R., Harrop-Allin M. K., 1995, MNRAS, 277, 971 (B95)

Finley D., Koester D., Basri G., 1997, ApJ, in press Forester H., Strupat W., Rosner W., Wunner G., Ruder H., Herald

H., 1984, J. Phys. B: At. Mol. Phys., 17, 1301 Henry R. J. W., O'Connell R. F., 1984, ApJ, 282, L97 Henry R. J. W., O'Connell R. F., 1985, PASP, 97, 333 Jensen K. A., Swank J. H., Petre R., Guinan E. F., Sion E. M.,

Shipman H. L., 1986, ApJ, 309, L27 Jordan S., 1988, PhD thesis, lnstitut fur Theoretische Physik und

Sterwarte der Universitat, Kiel Kemic S. B., 1974, JILA REP. No. 113 Kurucz R., 1971, Smith. Astr. Obs. Spec. Rep. 309 Latter W. B., Schmidt G. D., Green R. F., 1987, ApJ, 320, 308 Luyten W. J., Anderson J. H., 1967, A search for Faint Blue Stars

XLIV. Univ. of Minnesota, Minneapolis Marsh M. C. et aI., 1997, MNRAS, 286, 369 Martin B., Wickramasinghe D. T., 1979, MNRAS, 189, 69 Robinson E. L., Clemens C. J., Hine B. P., 1988, ApJ, 331, L29 Rosner W., Wunner G., Herold H., Ruder H., 1984, J. Phy. B: At.

Mol. Phys., 17, 29 Schmidt G. D., West S. c., Liebert J., Green R. F., Stockman H. S.,

1986, ApJ, 309, 218 Schmidt G. D., Latter W. B., Foltz C. B., 1990, ApJ, 350, 758 Schmidt G. D., Bergeron P., Liebert J., Saffer R. A., 1992, ApJ,

394, 603 Vennes S., 1996, in Bowyer S., Malina R., eds, Astrophysics in the

Extreme Ultraviolet. Kluwer, Dordrecht, p. 193 Vennes S., Thejll P. A., Wickramasinghe D. T., Bessell M. S., 1996,

ApJ, 467, 782 Vennes S., Thejll P. A., Genova-Galvan R., Dupuis J., 1997, ApJ,

480, 714 Weidemann V., 1987, A&A, 188,74 Wickramasinghe D. T., 1990, in Proc. Am. lnst. Phys. 216, Spectral

Line Shapes. Am. lnst. Phys., New York, p. 574 Wickramasinghe D. T., 1995, in Adelman S. J., Wiese W. L., eds,

ASP Conf. Ser. 78, Astrophysical Applications of Powerful New Databases. Astron. Soc. Pac., San Francisco, p. 319

Wickramasinghe D. T., Cropper M., 1988, MNRAS, 235,1451 Wickramasinghe D. T., Ferrario L., 1988, ApJ, 327, 222 Wickramasinghe D. T., Martin B., 1979, MNRAS, 188, 165 Wood M. A., 1995, in Koester D., Werner K., eds, Lecture Notes in

Physics 443, White Dwarfs. Springer, Berlin, p. 41 Wunner G., 1990, in Frommhold L., Keto J. W., eds, AIP Conf.

Proc. 216 Vol. 6, Spectral line shapes. Am. lnst. Phys., New York, p.563





Dow

nloaded from



Documents

EUVE J0317 - 855: a rapidly rotating, high-field magnetic