Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://jet.sao.ru/hq/lizm/conferences/pdf/2003/2003_p108.pdf Äàòà èçìåíåíèÿ: Wed Apr 21 16:48:55 2010 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 08:36:59 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: photosphere

Magnetic stars, 2004, 108-113

No weak magnetic field detected in the variable HgMn star Andromedae
Wade G.A.1 , Ab´ ecassis M.2 , Auri` M.2 , Donati J.-F.2 , Lign` ere eres F.2 , Marsden S.2 , 2 2 2 2 3 Mouillet D. , Paletou F. , Petit P. , To qu´ N. , Monin D. , Landstreet J.D.4 , e Bagnulo S.5 , Luftinger T.6 ¨
1 2 3 4 5 6

Royal Military College of Canada, Canada Observatoire Midi-Pyr´ ´ France enees, Universit´ de Moncton, Canada e University of Western Ontario, Canada ESO, Chile Institut fur Astronomie, Universitat Wien, Austria ¨

Abstract. High-S/N, phase-distributed circular p olarisation sp ectra of the variable HgMn star And provide no evidence of any photospheric magnetic field, with an upp er limit for oblique dip olar fields of 280 G (approximately the equipartition field at log 5000 = 0.0). We therefore prop ose that the variability of the Hg 3984 line and the inferred nonuniform distribution of Hg (rep orted by Adelman et al. 2002) most plausibly result from a separation mechanism unrelated to the presence of a magnetic field. Key words: stars: magnetic fields ­ stars: individual: And

1

Intro duction

Although Ap HgMn stars have generally been found to show no clear evidence of variability (e.g. Adelman 1998) or magnetic fields (Shorlin et al. 2002), Adelman et al. (2002) reported convincing evidence that the bright, well-established HgMn star Andromedae A (HD 358) exhibits clear (> 30%) variability of the equivalent width of the Hg ii 3984 line profile. The Hg line variations observed by Adelman et al. (2002) were consistent with a period of 2.38236 ± 0.00011 days, a value in good agreement with the rigid rotation period P = 2.53 ± 0.4 days (Ryabchikova et al. 1999) calculated using the measured v sin i of 52 ± 2 km/s, the inferred rotational axis inclination i = 74 o and the stellar radius R = 2.7 ± 0.4 R . Interpreting the Hg line profile variability as the consequence of rotational modulation of a nonuniform photospheric abundance distribution, Adelman et al. (2002) inverted the variations using the Doppler Imaging technique to recover a map of the Hg surface distribution. The distribution is characterised by a strong ( 4.5 dex) enhancement of the Hg abundance in the rotational equatorial regions, from about -30 o to +30o rotational latitude. The equatorial Hg enhancement appears to be structured into a series of three or four spots of very high abundance, distributed approximately uniformly in longitude, connected by "bridges" of somewhat lower Hg abundance. In the atmospheres of classical Ap stars, magnetic fields seem to be a necessary ingredient in the production of nonuniform surface abundance distributions. Could the inferred nonuniform surface distribution of Hg imply that And is a magnetic HgMn star? In this paper we explore this hypothesis using new high-S/N polarisation observations, well-distributed according to the rotational ephemeris of Adelman et al. (2002).
Based on observations obtained using the MuSiCoS sp ectrop olarimeter at the Pic du Midi Observatory, France.

c Sp ecial Astrophysical Observatory of the Russian AS, 2004


NO WEAK MAGNETIC FIELD DETECTED IN THE...

109

Table 1: Journal of MuSiCoS spectropolarimetric observations of And. Phases are calculated according to the ephemeris of Adelman et al. (2002: JD = 2449279.6898 + (2.38236 ± 0.00011) · E.) UT Date 12 24 26 04 07 09 09 09 10 10 10 11 12 17 17 17 18 18 19 19 19 26 26 26 dec jun jun jul jul jul jul jul jul jul jul jul jul jul jul jul jul jul jul jul jul jul jul jul 00 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 03 HJD 2451891.32 2452815.65 2452817.65 2452825.64 2452828.65 2452830.64 2452830.65 2452830.66 2452831.65 2452831.66 2452831.67 2452832.65 2452833.60 2452838.60 2452838.61 2452838.62 2452839.61 2452839.62 2452840.59 2452840.60 2452840.61 2452847.64 2452847.65 2452847.66 Phase 0.23769 0.22582 0.06532 0.41914 0.68469 0.51790 0.52210 0.52630 0.94185 0.94605 0.95025 0.36160 0.76037 0.85913 0.86332 0.86752 0.28308 0.28727 0.69443 0.69863 0.70283 0.65368 0.65788 0.66208

2

Observations

To search for a magnetic field in And, in June-August 2003 23 new high-S/N Stokes V spectra of And were obtained using the MuSiCoS spectropolarimeter (Baudrand & B¨ ohm 1992, Donati et al. 1999) on-board the 2m TBL at Pic du Midi Observatory. The spectra were reduced using the ESpRIT reduction package (Donati et al. 1997). Least-Squares Deconvolution (LSD; Donati et al. 1997, Wade et al. 2000a) was employed to extract highprecision mean Stokes I and V profiles from the spectra. For the extraction we have employed a speciallydesigned HgMn line mask (see, e.g., Shorlin et al. 2002) calculated for an effective temperature of 14000 K (Ryabchikova et al. 1999). The typical S/N of the extracted Stokes V mean profiles is 10 4 . The journal of observations is reported in Table 1. Fringes are apparent in the reduced Stokes V spectra of And, with amplitudes of order 0.2% and wavelengths larger than about 4 ° Preliminary fringe removal was performed using the wavelet transform A. procedure described by Monin et al. (in preparation). After fringe removal, LSD profiles obtained within ±0.05 days of each other were averaged, finally resulting in 13 sets of Stokes I and V LSD profiles that were employed for all of the following analysis.

3

Constraints on the photospheric magnetic field

We find that no Stokes V signatures are detected within any of the 13 final LSD line profiles, with 3 upper limits from 0.015% to 0.035%. Longitudinal magnetic fields were inferred from each of the Stokes I and V profile sets, using the firstorder moment method described by Donati et al. (1997) and Wade et al. (2000b). The resultant measurements


110

´ ` WADE, ABECASSIS, AURIERE ET AL.

have associated 1 uncertainties from 29­98 G, and a median uncertainty of 45 G. We detect no significant longitudinal magnetic field or field variation. To place an upper limit on the strength of any photospheric magnetic field present, we have calculated synthetic Stokes I and V LSD profiles, using the polarized line synthesis code Zeeman2 (Landstreet 1988, Wade et al. 2001), and assuming a triplet Zeeman pattern and the mean Land´ factor and wavelength e appropriate to the And line mask (the accuracy of such a modelling procedure has been demonstrated by e.g. Donati et al. 2001). We performed calculations for a large grid of magnetic field models and compared the Stokes V profile rotational variation predicted by each model with the observed Stokes V phase variation. Over 3000 individual models were calculated for pure dipolar and pure linear quadrupolar fields of various geometries; equivalently over 27300 comparisons were made between observed and calculated profiles. Note that although very few magnetic upper main sequence stars exhibit obviously quadrupolar magnetic fields, these solutions were calculated in order to test the robustness of the upper limits we determine.

Figure 1: Reduced 2 function of the polar symmetry axis . A correspond to models with the observations

of the model/data comparison for dipole (top) and quadrupolar (bottom) fields, as a field strength Bp and the obliquity between the rotational axis and the magnetic field rotational axis inclination of i = 74o has been assumed for al l models. Dark regions consistent with the observations, while white regions correspond to models incompatible at more than 99.9% confidence.


NO WEAK MAGNETIC FIELD DETECTED IN THE...

111

The values of the summed reduced 2 for each model/data comparison are shown in Fig. 1 as a function of the model magnetic field polar strength Bp and obliquity . Calculated Stokes V profiles consistent with the observations are obtained only for dipolar fields weaker than 600 G and for quadrupolar fields weaker than 400 G. Moreover, the strongest field solutions are consistent with the observations only for magnetic fields nearly aligned with the stellar rotational axis (obliquities close to zero or 180 o ). Aligned magnetic fields tend to be characteristic of magnetic A stars with long rotational periods (Landstreet & Mathys 2000), whereas oblique magnetic fields tend to be characteristic of magnetic A stars with short periods (like And). If we consider only solutions with significant obliquity (30o < < 150o ), we find that any magnetic fields present in the photosphere of And should be weaker than about 280 G if the field is dipolar, and weaker than 380 G if the field is quadrupolar. Note that these values are strict upper limits at > 99.9% confidence for this range of . Let us now restrict our discussion to the more plausible case of the predominantly dipolar magnetic field. In the atmosphere of And, at log 5000 = 0.0, the equipartition field strength (for which the magnetic and thermodynamic energy densities are approximately equal) is Beq = 12 nk T 260 G. This is approximately equal to the upper limit derived for any dipolar field that might be present, and so we can state that our data rule out the presence of any oblique dipolar field stronger than the equipartition field strength. Furthermore, if we believe that a magnetic field of at least this order is required to produce and support horizontal abundance nonuniformities (as appears to be the case for Ap stars, Auriere et al., this conference), then an oblique dipolar magnetic field cannot be responsible for the nonuniform distribution of Hg of And, at least in atmospheric layers around log 5000 = 0.0. Parenthetically, we also comment that although we have performed no quantitative modeling of tangled magnetic fields in this study (fields such as those exhibited by active late-type stars such as HR 1099, e.g. Petit et al. 2003), the fact that such fields are detected in LSD profiles of similar quality in a sample of late-type stars using the MuSiCoS spectropolarimeter implies that similar fields are not present in the photosphere of And. This result is fully consistent with the conclusions of Shorlin et al. (2002), who using high-S/N Stokes V profiles found no evidence for any tangled magnetic fields in a sample of 74 A and B type stars. A comparison between the sample of the final observed LSD profiles and the predictions of 500 G dipolar magnetic field with = 90o is shown in Fig. 2.
Stokes I
065 238 0.95 361 419 0.9 522 684 760 0.8 863 946 0.75 -0.25 -0.2 -0.1 -0.05

Stokes V
0

1

0.85

-0.15

-150 -100

-50 0 Velocity (km/s)

50

-150 -100

-50 0 Velocity (km/s)

50

Figure 2: Stokes I and V profiles for = 90o and Bd = 500 G. Thick lines -- observed LSD profiles. Thin lines -- calculated profiles.


112

´ ` WADE, ABECASSIS, AURIERE ET AL.

4

Implications

We stress that And represents the only convincing example of a HgMn star which shows intrinsic line profiles variability, making it an important testbed for theories describing the formation of abundance nonuniformities. If the nonuniform surface Hg distribution does not result from the presence of a magnetic field, what is its origin? One possibility is that the nonuniform abundance distribution is supported via interaction with the star's binary companion. As reported by Ryabchikova et al. (1999), the HgMn star And is orbited by a lower-mass companion, in a highly eccentric orbit with a period of 96.7 days. At apastron, the stars are separated by 1.12 AU (approximately 45 diameters of the HgMn star), whereas at periastron they are separated by only 0.32 AU (approximately 12 stellar diameters). We speculate that variable tidal interaction (with a period equal to one-half of the orbital period, or 48.4 days) between the two components might excite nonradial waves in the surface layers of the stars, producing mixing at selective longitudes (perhaps extrema of the surface wave), and relative stability at other longitudes (possibly nodes). Although this hypothesis is totally speculative, we point out that it naturally explains the concentration of Hg in the rotational equator, if the rotational and orbital axes of the system are aligned. A low-order surface wave might also explain the apparently equal 90o spacing in longitude of the Hg spots (although only three spots are obvious in the maps of Adelman et al. (2002; at longitudes of 90, 180 and 270 degrees), taking into account nonuniform phase sampling and uncertainties in interpreting the abundance scaling in the published maps, a fourth spot could plausibly exist at longitude 360o ). Because the orbital period of the And system is not an integral multiple of the rotational period of the HgMn star, one would expect that a tidally-induced surface inhomogeneity would change with time, probably on timescales of order the orbital period. Such changes might correspond to a precession of the surface abundance pattern. Oleg Kochukhov has recently reported (Kochukhov 2003) that in fact tentative evidence for variations in the structure of the Hg map do exist. A second possibility (suggested by John Landstreet (2003)) is that And does indeed host a magnetic field, but which is weaker than the local equipartition field in layers around log 5000 = 0.0. However, because of the exponential decrease of the gas density with height in the outer layers of the star, such a field could in fact be stronger than the equipartition field in atmospheric layers around (say) log 5000 = -3 (where Beq is only a few tens of G). If the Hg concentration and nonuniformities were confined to these high layers (as, for example, in the cases of other HgMn stars like 3 Cen A, Sigut et al. 2000), then even a very weak magnetic field could be responsible. Although plausible, such a hypothesis raises some interesting questions regarding the stability of weak magnetic fields. In particular, could a magnetic field weaker than the equipartition field in layers with log 5000 = 0.0 (and therefore sub ject to distortion by flows in those layers) survive disruption to produce nonuniformities in the upper photosphere?

5

Conclusions

High-S/N, phase-distributed circular polarisation spectra of the variable HgMn star And have been obtained using the MuSiCoS spectropolarimeter at Pic du Midi Observatory. No evidence of any magnetic field has been detected, leading to the following conclusions: And does not host a dipolar or quadrupolar photospheric magnetic field substantially stronger than the equipartition threshold. Nor does it host tangled magnetic fields with characteristics similar to those of active late-type stars (e.g. RS CVn systems or FK Com stars). We therefore conclude that the variability of the Hg 3984 line, and the inferred nonuniform distribution of Hg, most plausibly result from a separation mechanism unrelated to the presence of a magnetic field.
Acknowledgements. GAW acknowledges Discovery Grant support from the Natural Sciences and Engineering Research Council of Canada.

References
Adelman S.J., 1998, Astron. Astrophys. Suppl. Ser., 132, 93 Adelman S.J., Gulliver A.F., Kochukhov O., Ryabchikova T.A., 2002, Astrophys. J., 575, 449 Baudrand J., B¨ ohm T., 1992, Astron. Astrophys., 259, 711 Donati J.F., Semel M., Carter B.D., Rees D.E., Cameron A.C., 1997, Mon. Not. R. Astron. Soc., 291, 658 Donati J.-F., Catala C., Wade G.A., Gallou G., Delaigue G., Rabou P., 1999, Astron. Astrophys. Suppl. Ser., 134, 149


NO WEAK MAGNETIC FIELD DETECTED IN THE...

113

Donati J.-F., Wade G.A., Babel J., Henrichs H.F., de Jong J.A., Harries T., 2001, Mon. Not. R. Astron. Soc., 326, 1265 Kochukhov O., 2003, private communication Landstreet J.D., 1988, Astrophys. J., 326, 967 Landstreet J.D. & Mathys G., 2000, Astron. Astrophys., 359, 213 Landstreet J.D., 2003, private communication Petit P., Donati J.-F., Wade G.A., Auriere M. et al., 2003, submitted to Mon. Not. R. Astron. Soc. Ryabchikova T.A., Malanushenko V.P., Adelman S.J., 1999, Astron. Astrophys., 351, 963 Shorlin S.L.S., Wade G.A., Donati J.-F., Landstreet J.D., Petit P., Sigut T.A.A., Strasser S., 2002, Astron. Astrophys., 392, 637 Sigut T.A.T., Landstreet J.D., Shorlin S.L.S., 2000, Astrophys. J., 530, 895 Wade G.A., Donati J.-F., Landstreet J.D., Shorlin S.L.S., 2000a, Mon. Not. R. Astron. Soc., 313, 823 Wade G.A., Donati J.-F., Landstreet J.D., Shorlin S.L.S., 2000b, Mon. Not. R. Astron. Soc., 313, 851 Wade G.A., Bagnulo S., Kochukov O., Landstreet J.D., Piskunov N., Stift M.J., 2001, Astron. Astrophys., 374, 265