Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.sao.ru/hq/lizm/conferences/pdf/2006/2006_p172.pdf Äàòà èçìåíåíèÿ: Wed Apr 21 16:48:56 2010 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 16:55:15 2012 Êîäèðîâêà:

Physics of Magnetic Stars, 2007, pp. 172­183

Magnetic fields in Herbig Ae/Be stars
G.A. Wade1 , C. Catala2 , E. Alecian2 , C. Folsom1,3 , S. Bagnulo4 , T. B¨ ohm5 , J.-C. 6 , J.-F. Donati5 , D. Drouin1 , J.D. Landstreet7 Bouret
1 2 3 4 5 6 7

Department of Physics, Royal Military College of Canada, Kingston, Canada Obs. de Paris LESIA, Meudon, France Department of Physics, Queen's University, Kingston, Canada European Southern Observatory, Santiago, Chile Obs. Midi-Pyr´ ´ Toulouse, France enees, Laboratoire d'Astrophysique de Marseille, Marseille, France Department of Physics & Astronomy, University of Western Ontario, London, Canada

Abstract. Studies of stellar magnetism at the pre-main sequence phase can provide imp ortant new insights into the detailed physics of the late stages of star formation, and into the observed prop erties of main sequence stars. This is esp ecially true at intermediate stellar masses, where magnetic fields are strong and globally organised, and therefore most amenable to direct study. This pap er reviews recent high-precision sp ectrop olarimetric observations of pre-main sequence Herbig Ae/Be stars, which are yielding qualitatively new information ab out intermediatemass stars: the origin and evolution of their magnetic fields; the role of magnetic fields in generating their sp ectroscopic activity and in mediating accretion in their late formative stages; the factors influencing their rotational angular momentum; and the development and evolution of chemical p eculiarity in their photospheres. Key words: stars: Herbig Ae/Be ­ magnetic fields ­ stars: chemically p eculiar

1

Intro duction

Herbig Ae/Be (HAeBe) stars are pre-main sequence stars of intermediate mass (Herbig 1960; Hillebrand et al. 1992). They are characterized by sp ectral typ es A and B with strong, often ubiquitous emission lines. They are distinguished from the classical Ae/Be stars by their IR colours and frequent presence within dust-obscured regions and asso ciation with nebulae (Waters & Waelkens 1998). According to stellar evolution theory, HAeBe stars should not p osses deep outer convection zones which generate the imp ortant quantities of outward-flowing mechanical energy required to p ower an MHD dynamo. Rather, these stars are exp ected to have convective cores surrounded by primarily radiative sub-photospheric envelop es (Ib en 1965; Gilliland 1986). However, since 1980, rep eated observations (e.g. Praderie et al. 1982; Catala et al. 1986; Hamann & Persson 1992; Pogo din et al. 2005) have shown that many HAeBe stars are intensely active. In particular, some stars display characteristics often asso ciated with magnetic activity and the presence of chromospheres or coronae (e.g. Skinner & Yamauchi 1996; Hamaguchi et al. 2000). These prop erties have b een prop osed as indicators that many of these stars or their circumstellar envelop es are intensely magnetically active. This prop osal has imp ortant implications for our picture of how intermediate-mass stars form. In lower-mass pre-main sequence T Tauri stars, it is now generally supp osed that accretion is 172


MAGNETIC FIELDS IN HERBIG AE/BE STARS

173

mediated by the presence of strong, large-scale magnetic fields (e.g. discussion by Johns-Krull et al. (1999) and references therein). Some authors (e.g. Muzzerole et al. 2004) have suggested that similar "magnetospheric accretion" may o ccur in intermediate-mass PMS stars as well. Insight into the magnetic prop erties of HAeBe stars can likely b e gleaned from the magnetic prop erties of their b etter-studied main sequence descendants. A few p ercent (e.g. Wolff 1968) of all A and B typ e main sequence stars exhibit organised magnetic fields with strengths ranging from a few hundred to a few tens of thousands of gauss (e.g. Borra & Landstreet 1980; Mathys et al. 1997; Bagnulo et al. 1999). Because, like their pre-main sequence progenitors, main sequence A and B typ e stars have radiative envelop es, these magnetic fields are not b elieved to result from a contemp oraneous MHD dynamo. Rather, they are thought to b e fossil fields: the passively-decaying remnants of magnetic fields pro duced at an earlier convective evolutionary stage, or swept up during star formation. The presence of these fields has imp ortant consequences for the structure of the atmospheres of these Ap/Bp stars, suppressing large-scale mixing and leading to strong atmospheric chemical p eculiarities and abundance patches (e.g. Folsom et al. 2006; Lueftinger et al. 2003). Although the separation and mixing pro cesses leading to these chemical p eculiarities and abundance structures are not well understo o d, the p eculiarities are sufficiently unique to the Ap/Bp that they allow for the robust identification of magnetic stars from high- and even mo derate-resolution sp ectroscopy (i.e. without any direct detection of the magnetic field). The presence of the field also app ears to strongly influence stellar rotation, with Ap/Bp stars rotating substantially more slowly than other stars of the same sp ectral typ e (e.g. Abt & Morrell 1995). As prop osed by St¸ n epie ´ (2000), the slow rotation of Ap/Bp stars can b e explained by angular momentum loss at the premain sequence phase, taking into account accretion of matter along the magnetic field lines, the stellar field-disc interaction and a magnetised stellar wind. On the other hand, high-precision, high-resolution sp ectrop olarimetric studies (e.g. Shorlin et al. 2002, Wade et al. 2006a) have demonstrated conclusively that the large ma jority of main sequence A and B typ e stars (i.e. essentially all non-Ap/Bp stars) show no direct or indirect evidence for magnetic fields, often with remarkably small upp er limits (a few gauss). The existence of this "magnetic dichotomy" (i.e. the qualitative difference b etween the magnetic prop erties of the Ap/Bp stars and the other main sequence A and B stars) is further reinforced by the results of Auri` et ere al. (in preparation, but see Auri` et al. 2004), who show convincingly that all sp ectroscopicallyere identified Ap/Bp stars indeed host magnetic fields which are detectable with current techniques, and which are characterized by surface dip ole comp onents stronger than ab out 200 G. Although these magnetic prop erties of A and B typ e main sequence stars are not understo o d in theoretical detail, they are now well-established, providing solid observational results with which to compare their pre-main sequence progenitors. Observations of magnetic fields in Herbig Ae/Be stars can therefore serve to address several imp ortant astrophysical problems: (1) The role of magnetic field in mediating accretion, and the validity of mo dels which prop ose that HAeBe stars are simply higher-mass analogues of the T Tau stars. (2) The origin and prop erties of the magnetic fields of main sequence A and B typ e stars. (3) The development and evolution of chemical p eculiarities and chemical abundance structures in the atmospheres of A and B typ e stars. (4) The loss of rotational angular momentum which leads to the slow rotation observed in main sequence A and B typ e stars. This pap er will review recent magnetic observations of Herbig Ae/Be stars, discussing how these new data help to clarify the issues listed ab ove.


174

WADE ET AL.

25

25

Number of measurements

20 15 10 5 0

Number of measurements

Balmer

Metal
20 15 10 5 0

0

1

2

3

|/ |
z B

4

5

6

7

0

1

2

3

|/ |
z B

4

5

6

7

25

25

Number of measurements

20 15 10 5 0

Number of measurements

Full spectrum

20 15 10 5 0

Balmer w/o H

0

1

2

3

|/ |
z B

4

5

6

7

0

1

2

3

|/ |
z B

4

5

6

7

Figure 1: Histograms of the detection significance z = B z /B of magnetic field measurements of stars observed by Wade et al. (2006b). Each panel corresp onds to one of the 4 measurements obtained using different parts of the sp ectrum. The dashed line indicates the 3 detection threshold. Nearly all measurements corresp ond to non-detections, and formal detections are obtained for only 4 stars.

2

Low-resolution sp ectrop olarimetry with FORS1 at the ESOVLT

Wade et al. (2006b) presented results of a survey of the longitudinal magnetic fields of HAeBe stars conducted using the low-resolution FORS1 sp ectrop olarimeter at the VLT observatory. This work is the largest and most sensitive search for magnetic fields in HAeBe stars published to date. The primary goal of this survey was to determine whether strong, organised magnetic fields, similar to those of main sequence Ap/Bp stars, could b e detected in the HAeBe stars, and if so, to determine the intensities, geometrical characteristics, and statistical incidence of such fields. With this aim, Wade et al. (2006b) obtained 68 high-S/N ( 3000 : 1), low resolution (typically R 1000 - 1500) observations of 50 HAeBe stars in circularly p olarised light. To obtain a high signal-to-noise ratio and to mitigate spurious p olarisation signals, they obtained multiple exp osures of each star, with alternate exp osures corresp onding to a quarter-wave plate orientation of either +45o or -45o . The data were reduced and analysed using the pro cedures of Bagnulo et al. (2006), develop ed and tested in the context of a survey of magnetic fields in 235 magnetic and non-magnetic A and B typ e stars in op en clusters. The longitudinal magnetic field was inferred from the p olarisation sp ectra using a linear leastsquares fit based on the predicted circularly p olarised flux V /I in the weak-field regime, taking into account uncertainties asso ciated with each reduced sp ectral pixel as describ ed by Bagnulo et al. (2006). This pro cedure was rep eated 4 times for each sp ectrum, diagnosing the field from all Balmer lines, all Balmer lines except H , metallic lines, and the full sp ectrum. The mean 1 longitudinal field uncertainty measured from Balmer lines was 66 G, while that measured using the full sp ectrum was 48 G. The histograms of the detection significance z = | B z /B | of the resultant measurements are shown in Fig. 1. To robustly treat cases for which the longitudinal field was detected at ab out the 3 level, and in which minor changes in the data reduction would transform a marginal detection in


MAGNETIC FIELDS IN HERBIG AE/BE STARS

175

4 2 V/I

x 10

-3

4 2 V/I

x 10

-3

4 2 V/I 0 -2

x 10

-3

4 2 V/I 0 -2

x 10

-3

0 -2 -4 3850 15 10 5 0 3850 3900 3950 4000 Wavelength (angstrom)
5

0 -2 -4 4060 15 10 5 0 4060 4080 4100 4120 4140 Wavelength (angstrom)
5

-4 4300 15 10 5 0 4300

4320 4340 4360 4380 Wavelength (angstrom)
5

-4 4800 15 10 5 0 4800

4850 4900 Wavelength (angstrom)
5

x 10

x 10

x 10

x 10

I (ADU)

I (ADU)

I (ADU)

3900 3950 4000 Wavelength (angstrom)

4080 4100 4120 4140 Wavelength (angstrom)

4320 4340 4360 4380 Wavelength (angstrom)

I (ADU)

4850 4900 Wavelength (angstrom)

Figure 2: Stokes I and V profiles in the FORS1 sp ectrum of HD 101412. The solid curve in the upp er frames represents the prediction of the weak-field mo del. From Wade et al. (2006b).

a null or into a definite detection, Wade et al. (2006b) evaluated the consistency of the longitudinal fields determined from the various sp ectral regions using two detection criteria describ ed by Bagnulo et al. (2006). This allowed them to realistically judge the true significance of apparent detections. Ultimately, Wade et al. (2006b) concluded tentatively that weak longitudinal magnetic fields were detected in sp ectra of two HAeBe stars: HD 101412 (obtaining a measurement of 512 ± 111 G from Balmer lines) and BF Ori (obtaining a measurements of -180 ± 38 G from Balmer lines). Selected regions of the sp ectrum of HD 101412 are shown in Fig. 2, overlaid with the mo del V /I sp ectrum computed using the weak field mo del and the measured value of the longitudinal field. Although no obvious p olarisation signatures are visible in individual absorption lines in Fig. 2, this do es not necessarily indicate the absence of a field. As explained at the end of Sect. 8 and in Fig. 8 of Bagnulo et al. (2006), in their survey of cluster magnetic Ap stars they encountered numerous cases where a simple insp ection of the V /I profiles did not show any obvious p olarisation signal, but in which the magnetic field was detected by the linear regression of V /I . The reliability of this result is supp orted by the histograms of Fig. 7 of Bagnulo et al. (2006), showing that virtually no detections were found in "normal" A stars, where they do not exp ect to find any magnetic field. Magnetic fields were detected only in Ap stars, i.e., where they are exp ected to b e found. Magnetic observations of a small numb er of HAeBe stars have also b een carried out using FORS1 by Hubrig et al. (2004, 2006a, 2006b), who rep orted marginal detections of magnetic fields in several stars. Of these, Wade et al. (2006b) observed HD 144432 and HD 31648, obtaining no detection in either. Hubrig et al. (2006b) also rep orted the presence of strong Stokes V signatures in sp ectral lines of some HAeBe stars, which they interpreted to b e due to "circumstellar" magnetic fields. Wade et al. (2006b) were unable to confirm the detection of such features in their own sp ectra. As a further test, to check if they could confirm the strong p olarisation in the same data employed by Hubrig et al., they extracted from the ESO Science Archive 3 sp ectra of the HAeBe star HD 190073 discussed by Hubrig et al. (2006b), reducing and analysing them according to the pro cedures of Bagnulo et al. (2006). Remarkably, Wade et al. found no evidence for the presence of strong, systematic p olarisation signatures as rep orted by Hubrig et al. (2006b). This implies that the presence of this p olarisation is highly sensitive to the details of the data reduction, suggesting that it is not astrophysical in origin.


176

WADE ET AL.

1.2 1.2 1.8 1 0.8 1.6 0.6 0.4 1.4 0.2 5875 1.2 0.6 1 0.4 6450 6500 6550 6600 Wavelength (angstroms) 6650 6700 5000 5010 5020 Wavelength (angstroms) 5030 5040 5900 0.8 Stokes I Stokes I 1 Stokes V

0.8 6400

Figure 3: profile of -- Detail profile of

Left -- ESPaDOnS Stokes I sp ectrum of the HAeBe star BF Ori, showing the emission H. Absorption profiles of He i 6678 and He i 5876 (inset) are indicated by ticks. Right of the ESPaDOnS Stokes I and V sp ectra of BF Ori showing the complex absorption Fe ii 5018. No circular p olarisation is detected across this line.

3

High-resolution sp ectrop olarimetry with ESPaDOnS at the CFHT

Wade, Catala and collab orators have also undertaken a large circular p olarisation survey of HAeBe stars using the new high-resolution ESPaDOnS sp ectrop olarimeter at the Canada-France-Hawaii Telescop e. Although complete, the general results of this survey have not yet b een published. However, first results and results for a few individual stars of great interest have b een rep orted by Wade et al. (2005), Catala et al. (2006), Folsom et al. (2007) and Alecian et al. (2007). The ESPaDOnS survey comprises over 100 circular p olarisation sp ectra of ab out 50 HAeBe stars. Some of the stars observed by Wade et al. (2006b) and Hubrig et al. (2004, 2006a,b) were also observed using ESPaDOnS. The sp ectra cover the range 3800-10400 ° with a resolving p ower of A approximately 65000, and a median p eak S/N of ab out 300:1 p er 2.6 km s -1 pixel (equivalent to a ° S/N of ab out 1450:1 p er A at 5000 ° This is ab out one-half the mean S/N of the FORS1 sp ectra A). of Wade et al. (2006b); it furthermore varies more strongly with wavelength, decreasing to ab out 1/2 its p eak value at 4000 ° and 9000 ° However, the excellent results obtained with ESPaDOnS A A. so far (see Donati et al. 2005, 2006ab; Wade et al. 2006; Catala et al. 2006, 2007) demonstrate that the advantage of substantially higher resolving p ower (45-65 times b etter resolution) easily comp ensates for this difference. In particular, the ESPaDOnS measurements are characterised by a substantially higher magnetic field detection rate, more robust detections of weak longitudinal fields, straightforward confirmation of field detections, a lower ultimate detection threshold, and sensitivity to more complex field top ologies. In addition, ESPaDOnS data provide high-resolution line profiles, allowing detailed studies of the circumstellar environment, photospheric prop erties and structures, chemical abundances and stellar rotation. Fig. 3 illustrates an ESPaDOnS sp ectrum of somewhat less than typical quality (p eak S/N of 180:1) corresp onding to the HAeBe star BF Ori, a UX Ori-typ e star rep orted by Hillenbrand et al. (1992) to have Teff = 6750 K, and by Grinin et al. (2001) to have T eff = 8750 K. In the lefthand frame of Fig. 3, we show the H region of the sp ectrum. H displays strong, double-p eaked emission, and He i 6678 is clearly visible in absorption. The detection of this absorption line is supp orted by identification of He i 5876, also in strong absorption (also rep orted by Wade et al. 2006b). The app earance of these features in the sp ectrum of this star suggest that the published photospheric temp eratures are seriously underestimated. Also, at high resolution many sp ectral


MAGNETIC FIELDS IN HERBIG AE/BE STARS

177

2500 2000 longitudinal magnetic field (G) 1500 1000 500 0 0.98 -500 -1000 0.97 1.02

1.05 Epsilon UMa 50 x Stokes V/I 1 Stokes I HD 78362 250 x Stokes V/I

1.01

1 Stokes I

0.95

0.99

0.9

0.85

-0.4

-0.2

0

Phase

0.2

0.4

0.96

-100

-50

0 Velocity (km/s)

50

100

-100

-50

0 Velocity (km/s)

50

100

Figure 4: Left -- ESPaDOnS verification observations for magnetic and non-magnetic standard stars. Left frame -- Longitudinal magnetic field variation of the co ol magnetic Ap star HD 12098 obtained with the ESPaDOnS predecessor MuSiCoS (op en symb ols; Ryab chikova et al. 2005). Sup erimp osed are 3 ESPaDOnS measurements (filled symb ols) with error bars of ab out 25 G (also shown are their corresp onding LSD Stokes V profiles). Each ESPaDOnS measurement agrees with the MuSiCoS field variation to within 1 . Centre frame -- ESPaDOnS Stokes I and V LSD profiles of the weak-field Ap star UMa. The LSD profiles, detected with ab out 16 confidence, corresp ond to a longitudinal magnetic field of 97 ± 6 G. Right frame -- ESPaDOnS Stokes I and V LSD profiles of the non-magnetic Am star HD 78362. The LSD profiles corresp ond to a longitudinal magnetic field of 0 ± 1 G. lines resolve into complex profiles, which app ear to b e comp osed of two rotationally-broadened stellar comp onents (separated by approximately 140 km s -1 ), up on which are sup erimp osed sharp, p ossibly circumstellar absorptions (e.g. Na i D in Fig. 3, left panel inset, and Fe ii 5018 in Fig. 3, right panel). Clearly the new sp ectra are rich in sp ectroscopic information, in addition to their unique value for diagnosing the magnetic field. In Fig. 3 (right panel), we show the Stokes V circular p olarisation across the Fe ii 5018 line of BF Ori. No Stokes V signature is detected in this line (or in any of its individual absorption comp onents, for that matter), or in other absorption or emission lines of this star, notwithstanding the tentative detection of a magnetic field by Wade et al. (2006b). The ESPaDOnS observations are further analysed using the Least-Squares Deconvolution (LSD; Donati et al. 1997) multi-line pro cedure, extracting p olarisation information from hundreds or thousands of lines in the observed sp ectrum to obtain the magnetic field diagnosis. The resultant highS/N mean profiles allow the diagnosis of the magnetic field directly from the presence (or absence) of resolved Stokes V signatures (see Fig. 4). The longitudinal field can also b e diagnosed: it is inferred from the first moment of the Stokes V LSD profile (e.g. Wade et al. 2000b). However, b ecause ESPaDOnS is a high-resolution sp ectrop olarimeter, the inferred errors are very sensitive to the line profile characteristics -- in particular v sin i and emission contamination. This results in a substantially broader distribution of longitudinal field uncertainties, the largest of which (for stars with the highest v sin i or lines most contaminated by emission) reach several hundred gauss. However, for the ma jority of the sample, the longitudinal field uncertainties are comp etitive with those obtained using FORS1, and for those stars which are most suitable sp ectroscopically, the longitudinal field uncertainties are substantially b etter than those from FORS1. For example, the LSD profiles of BF Ori yield a longitudinal magnetic field of 25 ± 32 G, and those rep orted by Catala et al. (2006) have uncertainties as small as 10 G. The longitudinal field precision and accuracy obtainable with ESPaDOnS are illustrated in Fig. 4. Similar null magnetic field results are obtained for the ma jority of stars observed, including HD 144432, HD 31648 and HD 139614, in which Hubrig et al. (2004, 2006a) have claimed marginal magnetic field detections using FORS1, and in HD 36112, in which FORS1 data were rep orted


178

WADE ET AL.

Figure 5: ESPaDOnS Stokes I and V LSD profiles of V380 Ori (Left) and HD 200775 (Right). Strong, ordered magnetic fields are detected in b oth of these stars within the context of the ESPaDOnS survey. Mo deling of the magnetic field geometries of these stars is presented by Alecian et al. (2007). to b e "suggestive" of magnetic fields (Wade et al. 2006b). Nor do the ESPaDOnS sp ectra yield evidence for the systematic presence of p olarisation signatures in lines diagnostic of the circumstellar environment, such as Ca ii H and K, Mg ii 4481, the H emission line, the O i 6300, 6364 emission lines, the IR calcium triplet, etc. On the other hand, magnetic fields are clearly detected in photospheric metal lines of 4 stars: HD 190073 (Catala et al. 2006), HD 72106 (Folsom et al., 2007), and HD 200775 and V380 Ori (Alecian et al., 2007). All of these stars have b een detected in multiple observations during different observing runs, and two of these stars have b een shown to exhibit coherent, p erio dic variations of their Stokes V profiles (Alecian et al., 2007, Folsom et al., 2007). LSD profiles illustrating the detection of fields in these stars are shown in Figs. 5 and 6. Although Hubrig et al. (2006b) suggest that LSD is inappropriate for magnetic field diagnosis of HAeBe stars, the results rep orted by Alecian et al. (2007), Folsom et al. (2007) and Catala et al. (2006) demonstrate that LSD can b e used effectively for field diagnosis in HAeBe stars, even those stars with rather complex sp ectra. This conclusion is supp orted by the analysis of the large b o dy of data accumulated within the context of the ESPaDOnS survey. A final remarkable result comes from noting that 3 of the 4 stars detected in the ESPaDOnS survey (HD 72106, HD 190073 and V380 Ori) were also observed by Wade et al. (2006b), with no magnetic field detected. This is certainly partly due to the relative weakness of the longitudinal fields, but it also results from the inability of FORS1 to resolve line profiles into their (p olarised) absorption comp onents and (unp olarised) emission comp onents, and to thereby isolate and detect the p olarisation. This illustrates a clear advantage of high-resolution sp ectrop olarimetry for studies of magnetism in HAeBe stars.

4

General magnetic characteristics of Herbig Ae/Be stars

For the large ma jority of stars studied by Wade et al. (2006b) and in the ESPaDOnS survey, no magnetic fields are detected. Wade et al. (2006b) rep ort the results of Monte Carlo simulations aimed at mo deling their FORS1 measurements. They compute synthetic distributions of longitudinal field


MAGNETIC FIELDS IN HERBIG AE/BE STARS

179

1.04 1.03 1.02 1.01 1 0.99 0.98 0.97

HD 72106AB

20 Feb 05 21 Feb 05

1.2

50 x Stokes V/I

1.1

LSD Stokes V x 25
1 Stokes I

LSD Stokes I

0.9

0.8

HD 190073 -50 0 Velocity (km/s) 50

-100

-50

0 50 velocity (km/s)

100

150

Figure 6: ESPaDOnS Stokes I and V LSD profiles of HD 72106 (Left) and HD 190073 (Right). Strong, ordered magnetic fields are detected in b oth of these stars within the context of the ESPaDOnS survey. Further discussion of results for these stars is rep orted by Folsom et al. (2007) and Catala et al. (2006). measurements assuming p opulations of stars with various dip olar magnetic field characteristics. Comparing the observed and computed distributions (excluding measurements of HD 101412 and BF Ori), they conclude that their observations have the following prop erties: they are consistent within statistical uncertainty with a distribution of non-magnetic stars; they are inconsistent with a uniform p opulation of magnetic stars with aligned magnetic fields, if their dip ole intensities B d > 300 G; and they are inconsistent with a uniform p opulation of magnetic stars with p erp endicular magnetic fields, if Bd > 500 G. In addition to the large sample of stars in which magnetic fields are not observed, they have identified a small numb er of stars in which strong ( 1 kG) magnetic fields are detected, which must b e organised on large scales (e.g. quasi-dip olar). Although no similar analysis of the ESPaDOnS data has yet b een p erformed, it is clear that the ESPaDOnS measurements are qualitatively consistent with all of the general conclusions of Wade et al. (2006b). In particular, the ESPaDOnS data confirm that magnetic fields are not detected in most stars studied, but that strong, ordered magnetic fields exist in a small fraction of HAeBe stars. In addition, a few apparently non-magnetic stars in the ESPaDOnS survey have b een observed several times. This may allow much firmer upp er limits to b e placed on any magnetic fields that remain undetected in these stars. Finally, unlike the low-resolution FORS1 data, the ESPaDOnS profiles are sensitive to rather complex magnetic top ologies (see, e.g., Petit et al. 2005). The absence of detectable circular p olarisation in the high-resolution LSD profiles therefore also allows the general exclusion of a variety of more exotic field geometries.

5

Magnetic prop erties of HAeBe stars vs. main sequence A and B stars: testing the fossil-field hyp othesis

The primordial fossil-field hyp othesis, which prop oses that the magnetic fields of Ap/Bp stars are the slowly-decaying remnants of interstellar magnetic field swept up during star formation, requires that the magnetic prop erties of HAeBe stars b e qualitatively the same as those of the main sequence


180

WADE ET AL.

A and B stars. It is therefore instructive to compare the known magnetic prop erties of the main sequence A and B typ e stars with the observed prop erties of HAeBe stars. In brief: · Wade et al. (2006b) rep ort that Monte Carlo simulations demonstrate that their complete sample of FORS1 observations of HAeBe stars is consistent with a magnetic field distribution similar to that of the main sequence A and B typ e stars. · The total numb er of confirmed magnetic stars in the sample of Wade et al. (2006b) is 4, and the numb er of susp ected magnetic stars is 1 (if we exclude BF Ori based on the results of the ESPaDOnS survey). This corresp onds to a bulk incidence of (detected) magnetic stars of b etween 8% and 10% (taking into account the p ossibility that HD 101412 is not magnetic). This range is fully consistent with the canonical incidence of 5-10% (e.g. Wolff 1968) of magnetic Ap and Bp stars on the main sequence. · The results rep orted by Wade et al. (2005), Folsom et al. (2007) and Alecian et al. (2007) for HD 72106A, V380 Ori and HD 200775, and by Catala et al. (2006) for HD 190073, show that the magnetic fields of all these stars are organised on large scales, and that the fields of V380 Ori, HD 72106A and HD 200775 have imp ortant dip olar comp onents. The strong longitudinal field measured by Wade et al. (2006b) for HD 101412 also suggests that that its field is organised on large scales. It therefore app ears likely that most, and p ossibly all, of the detected stars have magnetic fields that are structured on global scales, with imp ortant dip ole comp onents. · The mean rms longitudinal field measured for the small numb er of magnetic HAeBe stars detected so far is roughly 200 G, with an intrinsic disp ersion of p erhaps 100 G. This is consistent with the value estimated for a complete magnitude-limited sample of Ap stars by Bohlender & Landstreet (1990; 330 G with an intrinsic disp ersion of ab out 170 G). Moreover, a weaker mean field strength of HAeBe stars is exp ected if fossil magnetic flux is conserved during stellar evolution (Wade et al. 2006b). All of these conclusions and observations show that the magnetic characteristics of HAeBe stars (incidence, top ology, intensity) are analogous to those of the Ap/Bp stars. Although further investigations are clearly required, we view this as strong evidence for the general correctness of the primordial fossil field hyp othesis.

6

Activity and magnetospheric accretion in Herbig Ae/Be stars

Some investigators have interpreted the strong sp ectroscopic, p olarimetric and photometric activity of HAeBe stars to indicate that these stars are magnetically active (e.g. Catala et al. 1986). The magnetic results describ ed in this pap er do not supp ort the prop osal the the activity of most HAeBe stars is of magnetic origin. Some investigators have also suggested that HAeBe stars are the higher mass analogues of classical T Tauri stars (CTTS), and that all of the phenomena asso ciated with CTTS are also op erating in HAeBe stars, including magnetospheric accretion (see for example Muzerolle et al. 2004). Magnetospheric accretion requires the presence of strong, large-scale magnetic fields at the stellar surface, and in the case of CTTS the predicted field strengths range up to several kG for sp ecific stars. Wade et al. (2006b) use their data to evaluate three magnetospheric accretion mo dels (see Johns-Krull et al. 1999). Estimating typical values of the mass, radius, rotation p erio d and accretion rate, they compute that a dip ole magnetic field with an intensity of ab out 500 G is required for magnetospheric accretion according to the mo dels of K¨ onigl (1991) and Shu et al. (1994), and of


MAGNETIC FIELDS IN HERBIG AE/BE STARS

181

ab out 100 G for the mo del of Cameron & Campb ell (1993). Based on the results of their Monte Carlo simulations, they conclude that magnetospheric accretion according to the theories of K¨ onigl (1991) and Shu et al. (1994) is not generally o ccurring in HAeBe stars. The ESPaDOnS data should allow this issue to b e studied further. First, the firmer upp er limits achievable for undetected stars with multiple, high-precision observations should allow testing of the Cameron & Campb ell mo del for several stars. In addition, characterisation of the magnetic strengths and geometries of the detected magnetic stars should allow a detailed comparison of their prop erties with the predictions of all three mo dels, and p erhaps with more sophisticated numerical simulations (e.g. Yelenina et al. 2006).

7

Development of chemical p eculiarities

A characteristic observational feature of magnetic A and B typ e stars on the main sequences is their strong photospheric chemical p eculiarity. Although the basic mechanism resp onsible for the pro duction of this phenomenon is known, the roles of the various separation and mixing pro cesses, and the influence of the magnetic field, are understo o d only schematically. The new observations of HAeBe stars provide the p otential for the study of these phenomena at their earliest stages, allowing us to investigate the conditions and timescales required for the development of chemical p eculiarity. Although this asp ect has not yet b een investigated in much detail, the ESPaDOnS observations suggest that chemical p eculiarity is detected in only one of the magnetic stars identified so far: HD 72106A. As is rep orted by Folsom et al. (2007), this star shows b oth strong p eculiarities analogous to those of the Ap/Bp stars, as well as line profile variability indicative of chemical abundance patches.

8

Angular momentum evolution of magnetic intermediate-mass stars

As discussed earlier, magnetic A and B typ e stars on the main sequence rotate significantly more slowly than non-magnetic stars of the same sp ectral typ e. This suggests that the magnetic field plays an imp ortant role in the shedding of rotational angular momentum in the magnetic stars, probably at the pre-main sequence stage (e.g., see St¸ n 2000). epie ´ We are b eginning to explore this asp ect in some detail. First, we are able to study the rotational characteristics of magnetic HAeBe stars, first by using their measured v sin i, and more accurately using their measured rotational p erio ds and their inferred radii. All detected stars in b oth surveys app ear to b e slow rotators -- hence significant angular momentum already seems to have b een shed by these magnetic stars, even at these relatively young ages. It could b e argued that the low pro jected rotational velo cities of the detected magnetic stars result from a selection effect, b ecause most of these stars were detected using high-resolution sp ectrop olarimetry which b ecomes less sensitive to magnetic fields as v sin i increases. We counter this argument in two ways: first, the FORS1 observations, which are essentially insensitive to v sin i, do not detect fields in any rapidly-rotating stars. In fact, the only star in which a reliable field detection seems to have b een obtained with FORS1 (HD 101412) has very low v sin i ( 7 km s -1 , Guimaraes et al. 2006), fully consistent with the ESPaDOnS result. Secondly, recent ESPaDOnS observations of HAeBe stars in young op en clusters (aimed at studying the influence of environment and evolution of magnetism and rotation) demonstrate that magnetic fields can b e detected in rapidly-rotating HAeBe stars (see Fig. 7).


182

WADE ET AL.

W601 (NGC 6611) 1.02 25 x Stokes V/I

1.01 Stokes I 1

0.99

0.98 -300 -200 -100 0 Velocity (km/s) 100 200 300

Figure 7: ESPaDOnS Stokes I and V LSD profiles of the rapidly-rotating (v sin i 180 km s -1 ) Herbig Be star W601 in NGC 6611, showing a definite detection of circular p olarisation across the mean line. This detection demonstrates that magnetic fields are present in some HAeBe stars with high v sin i, and that LSD is capable of detecting them.

9

Conclusion

In this pap er we have reviewed the results of recent large-scale surveys of the magnetic prop erties of Herbig Ae/Be stars, demonstrating the clear detection of strong, organised magnetic fields in a small sample of the observed stars, and the apparent absence of magnetic fields in the large ma jority. We have reviewed the observed characteristics of these fields, and evaluated their consistency with the primordial fossil field hyp othesis. Finally, we have explored the implications of these new data for our understanding of activity and magnetospheric accretion in PMS intermediate-mass stars, for the evolution of rotational angular momentum, and for the development of photospheric chemical p eculiarity.
Acknowledgements. This research has been partially funded by the Natural Sciences and Engineering Research Council of Canada (NSERC), and the Canadian Department of National Defence Academic Research Programme (ARP).

References
Abt, H.A. & Morrell, N.I., 1995, ApJS, 99, 135 Alecian E., Wade G. A., Catala C., Bagnulo S., et al., 2007, in: Proceed. of Intern. Conf. "Physics of Magnetic Stars" (this issue), Special Astrophysical Observatory, Nizhnij Arkhyz Eds: D.O. Kudryavtsev and I.I. Romanyuk, p.184, (astro-ph 12186A) Auriere M., Silvester J., Wade G. A., Bagnulo S., et al., 2004, in: Proceed. of The A-Star Puzzle, J. Zverko, J. Ziznovsky, S.J. Adelman, and W.W. Weiss (eds.), IAU Symposium, No. 224, Cambridge University Press, p. 633 Bagnulo S., Landolfi M., degl'Innocenti M. Landi, 1999, A&A 343, 865 Bagnulo S., Landstreet J. D., Mason E., Andretta V., Wade G.A., 2006, A&A, 450, 777 Bohlender D. A. & Landstreet J. D., 1990, MNRAS, 247, 606 Borra E. F., Landstreet J. D., 1980, ApJS, 42, 421 Cameron A.C. & Campbell C.G., 1993, A&A, 274, 309


MAGNETIC FIELDS IN HERBIG AE/BE STARS

183

Catala C., Felenbok P., Czarny J., Talavera A., Boesgaard A. M., 1986, A&A, 308, 791 Catala C., Alecian E., Donati J.-F., Wade G.A. et al, 2006, A&A, in press (astro-ph 10499C) Catala C., Donati J.-F., Shkolnik E., Bohlender D., Alecian E., 2007, MNRAS, 374, 42 Donati J.-F., Semel M., Carter B. D., Rees D. E., Cameron A. C., 1997, MNRAS, 291, 658 Donati J.-F., Paletou F., Bouvier J., Ferreira J., 2005, Nature, 48, 466 Donati J.-F., Howarth I. D., Jardine M. M., Petit P., et al., 2006a, MNRAS, 370, 629 Donati J.-F., Forveille T., Cameron A.C., Barnes J.R., et al., Science, 311, 633 Folsom C., Wade G.A., Hanes D.A., et al., 2007, in: Proceed. of Intern. Conf. "Physics of Magnetic Stars" (this issue), Special Astrophysical Observatory, Nizhnij Arkhyz, Eds: D.O. Kudryavtsev and I.I. Romanyuk, p.189 (astro-ph 12231F) Folsom C., Wade G.A., Bagnulo S., Landstreet J.D., 2006, MNRAS, in press (astro-ph 12284F) Gilliland R. L., Kemper E., Suntzeff N., 1986, ApJ, 351, 252 Guimaraes M.M., Alencar S.H.P., Corradi W.J.B. Vieira S.L.A., 2006, A&A, 457, 581 Grinin V. P., Kozlova O. V., Natta A., Ilyin I., 2001, A&A, 379, 482 Hamaguchi K., Terada H., Bamba A., Koyama K., 2000, ApJ, 352, 1111 Hamann F., Persson S. E., 1992, ApJS, 82, 285 Herbig G. H., 1960, ApJS, 4, 337 Hillenbrand L. A., Strom S. E., Vrba F. J., Keene J., 1992, ApJ, 397, 613 Hubrig S., Sch¨ller M., Yudin R. V., 2004, A&A, 428, L1 o Hubrig S., Yudin R. V., Sch¨ oller M., Pogodin M. A., 2006a, A&A, 446, 1089 Hubrig S., Sch¨ oller M., Yudin R. V., Pogodin M. A., Schoeller M., Schnerr R. S., 2006b, A&A, in press (astro-ph 10439H) Iben I., 1965, ApJ, 141, 993 Johns-Krull C., Valenti J., Koresko C., 1999, ApJ, 516, 900 K¨ onigl A., 1991, ApJ, 370, 39 Lueftinger T., Kuschnig R., Piskunov N.E., Weiss W.W., 2003, A&A, 406, 1033 Mathys G., Hubrig S., Landstreet J.D., Lanz T., Manfroid J., A&AS, 123, 353 Muzzerole J., D'Alessio P., Calvet N., Hartmeann L., 2004, ApJ, 617, 406 Petit P., Donati J.-F., Auri+re M., Landstreet J. D., et al., 2005, MNRAS, 361, 837 Pogodin M. A., Franco G. A. P., Lopes D. F., 2005, A&A, 438, 239 Praderie F., Felenbok P., Czarny J., Talavera A., Boesgaard A. M., 1982, ApJ, 254, 658 Ryabchikova T., Wade G. A., Auriere M., Bagnulo S., 2005, A&A, 429, 55 Shorlin, S.L.S., Wade, G.A., Donati, J.-F., Landstreet, et al., 2002, A&A 392, 637 Shu F., Na jita J., Ostriker E., Wilkin F., et al., 1994, ApJ, 429, 781 Skinner S. L. & Yamauchi S., 1996, ApJ, 471, 987 St¸ n K., 2000, A&A, 353, 227 epie´ Wade G. A., Drouin D., Bagnulo S., Landstreet J. D., et al., 2005, A&A, 442, 31 Wade G. A., Auriere M., Bagnulo S., Donati J.-F., 2006a, A&A, 451, 293 Wade G. A., Bagnulo S., Drouin D., Landstreet J. D., Monin D., 2006b, MNRAS, (in press) Waters L. B. F. M. & Waelkens C., 1998, ARAA, 36, 233 Wolff S.C., 1968, PASP, 80, 281 Yelenina T. G., Ustyugova G. V., Koldoba A. V., 2006, A&A, 458, 679