New evidence for weak magnetic fields in Herbig Ae/Be stars

TL;DR

This study provides new spectropolarimetric evidence that Herbig Ae/Be stars have weak magnetic fields, with measurements showing fields as low as 17 G, which are weaker than those in T Tauri stars, impacting theories of star and planet formation.

Contribution

The paper presents the first high-resolution spectropolarimetric measurements of magnetic fields in a sample of Herbig Ae/Be stars, revealing their magnetic fields are significantly weaker than in T Tauri stars.

Findings

Detected magnetic fields as weak as 17 G in Herbig Ae/Be stars.

Observed polarity changes in some stars indicating complex magnetic structures.

Confirmed that Herbig Ae/Be stars have weaker magnetic fields than T Tauri stars.

Abstract

In recent years Herbig Ae/Be stars receive considerable attention as their disks are believed to be the sites of on-going planet formation. Confirming the presence of magnetic fields in these stars is critical for understanding the transport of angular momentum during the protostellar phase. Furthermore, magnetic fields set the conditions for strongly anisotropic accretion. In this study we present the results of our recent observing campaigns of a sample of Herbig Ae/Be stars aimed at measurements of their magnetic fields applying the Singular Value Decomposition method to high resolution spectropolarimetric observations. The strongest longitudinal magnetic field of 209 G is detected in the Herbig Be star HD 58647, whereas the weakest field of 17 G is measured in the Herbig Ae star HD 190073. A change of polarity is detected for HD 58647 and in the Herbig Be star HD 98922. The obtained results provide further evidence that Herbig Ae/Be stars possess much weaker magnetic fields than their lower mass counterpart T Tauri stars with magnetic fields of kG order.

Tables

Table 1. Table 1: Logbook of the observations. The columns give the name of the star, the heliocentric Julian date (HJD) for the middle of the exposures, the signal-to-noise ratio (S/N) of the spectra, the measured longitudinal magnetic field ( ⟨ B z ⟩ delimited-⟨⟩ subscript 𝐵 𝑧 \langle B_{z}\rangle ), and the False Alarm Probability (FAP).

Star	HJD	S/N	$⟨B_z⟩$		FAP
	2 400 000+		(G)
HD 58647	55906.723∗	437	—
	57908.483	392	$-$121$\pm$	11
	57910.467	466	209$\pm$	10
	57911.468	343	—
HD 98922	55706.513∗	379	$-$33$\pm$	4
	57554.588	244	28$\pm$	7
HD 139614	57555.722	218	25$\pm$	3
HD 165133	57911.813	142	97$\pm$	9
HD 190073	57554.886	164	23$\pm$	2
	57555.848	215	34$\pm$	2
	57908.913	367	17$\pm$	1
	57909.861	179	20$\pm$	2
	57910.919	323	19$\pm$	1

Equations

$$⟨B_{\mathrm{z}}⟩=-2.14\times 10^{11}\frac{\int \upsilon V(\upsilon )\mathrm{d}\upsilon }{{\lambda }_0{g}_{0}c\int [I_c-I(\upsilon )]\mathrm{d}\upsilon },$$

Full text

††pubyear: 2019††pagerange: New evidence for weak magnetic fields in Herbig Ae/Be stars– New evidence for weak magnetic fields in Herbig Ae/Be stars

1 Introduction

Herbig Ae/Be stars are pre-main-sequence objects with pronounced emission line features and an infrared excess indicative of dust in their circumstellar disks. It was previously assumed that these stars are intermediate-mass analogs of T Tauri stars, but with convectively stable envelopes that do not support the dynamo action found in the fully convective T Tauri stars. A number of magnetic studies have been attempted in the last years, indicating that about 20 Herbig Ae/Be stars may have globally organized magnetic fields (e.g. Hubrig et al., 2004; Wade et al., 2005; Hubrig et al., 2009; Alecian et al., 2013b; Hubrig et al., 2015; Järvinen et al., 2015, 2018). Magnetic fields in these stars might be fossils of the early star formation epoch, in which the magnetic field of the parental magnetized core was compressed into the innermost regions of the accretion disks (e.g. Banarjee & Pudritz, 2006). Alternatively, Tout & Pringle (1995) proposed a non-solar dynamo that could operate in rapidly rotating A-type stars based on rotational shear energy.

The compilation of existing magnetic field measurements by Hubrig et al. (2015) showed that only very few Herbig Ae/Be stars have mean longitudinal magnetic fields stronger than 200 G, and half of the sample possesses fields of about 100 G and less. Recent spectropolarimetric observations involving a few Herbig Ae/Be stars with relatively low $v\,\sin\,i$ values confirm the results of this study, revealing the presence of weak, frequently just of the order of a few tens of Gauss, magnetic fields measured with an uncertainty of only a few Gauss (Hubrig et al., 2015; Järvinen et al., 2015, 2018).

In this work, we discuss the results of our recent observing campaigns of a sample of Herbig Ae/Be stars using the High Accuracy Radial velocity Planet Searcher polarimeter (HARPSpol; Snik et al., 2008).

2 Observations and data reduction

The new HARPSpol observations of the two Herbig Ae stars HD 139614 and HD 190073, the two late Herbig Be stars HD 58647 and HD 98922, and the early Herbig Be star HD 165133 were obtained on 2016 June 15 and 16, and on 2017 June 3 to 6. Each observation consisted of subexposures with exposure times varying between about 6 minutes and 47 minutes, depending on the target visual magnitude. After each subexposure, the quarter-wave retarder plate was rotated by $90\degr$. The resolving power of HARPS is about $R=115\,000$, with spectra covering the spectral range 3780–6910 Å, with a small gap between 5259 Å and 5337 Å. The reduction and calibration of the obtained spectra was performed using the HARPS data reduction software available on La Silla. The normalization of the spectra to the continuum level is described in detail by Hubrig et al. (2013). The summary of the HARPSpol observations is given in Table 1 in columns 1 to 3.

3 Longitudinal magnetic field measurements

According to previous studies are the magnetic fields in Herbig Ae/Be stars expected to be weak. Therefore, to study the presence of magnetic fields in these stars, we employed the dedicated Singular Value Decomposition technique (SVD; Carroll et al., 2012). The SVD approach is very similar to that of the Principle Component Analysis (PCA). In this technique, the similarity of the individual Stokes $V$ profiles allows one to describe the most coherent and systematic features present in all spectral line profiles as a projection onto a small number of eigenprofiles. The excellent potential of the SVD method, especially in the analysis of weak fields in Herbig Ae stars, was already presented by Hubrig et al. (2015) and Järvinen et al. (2015, 2018).

We have used the Vienna Atomic Line Database (VALD; e.g. Kupka et al., 2011; Ryabchikova et al., 2015) to construct a line mask for each star based on their stellar parameters. Furthermore, we have checked that each line in the mask is present in the stellar spectra. Obvious line blends and lines in telluric regions were excluded from the line list. The mean longitudinal magnetic field is determined from the SVD spectra by computing the first-order moment of the Stokes $V$ profile according to Mathys (1989):

[TABLE]

where $\upsilon$ is the Doppler velocity in km s*-1*, and $\lambda_{0}$ and$g_{0}$ are the average values for the wavelength (in nm) and the Landé factor obtained from all lines used to compute the SVD profile, respectively. The results of the magnetic field measurements are presented in Table 1 together with the False Alarm Probabilities (FAPs). The FAP is used to classify the magnetic field measurements (Donati, Semel, & Rees, 1992): a profile with ${\rm FAP}<10^{-5}$ is a definite detection, with $10^{-5}<{\rm FAP}<10^{-3}$ is a marginal detection, and with ${\rm FAP}>10^{-3}$ is a non-detection. The older data for HD 58647 and HD 98922 from Hubrig et al. (2013) are used for comparison with the new measurements. In the following we discuss the magnetic field measurements for each star individually.

3.1 HD 58647

HD 58647 is a late Herbig Be star exhibiting double-peaked profiles in some hydrogen emission lines, such as H$\alpha$ and Br$\gamma$ (e.g. Grady et al., 1996; Brittain et al., 2007). Linear spectropolarimetric observations of H$\alpha$ showed a large polarization change across the line (Vink et al., 2002; Mottram et al., 2007; Harrington & Kuhn, 2009). The detected change in the linear polarization was interpreted in terms of a compact source of line photons that is scattered off a rotating accretion disc. A first measurement of the longitudinal magnetic field $\langle B_{z}\rangle=218\pm 69$ G in HD 58647 using the moment technique (Mathys, 1994) was published by Hubrig et al. (2013). Kurosawa et al. (2016) showed that a model with a small magnetosphere and a disc wind with its inner radius located just outside of the magnetosphere can well reproduce the observed Br$\gamma$ profile, wavelength-dependent interferometric visibilities, colour differential phases, and closure phases simultaneously.

The line mask is based on $T_{\rm eff}=10500$ K (Merín, 2004) and $\log\,g=3.33$ (Montesinos et al., 2009). The SVD profiles have been calculated using 206 unblended lines with an average Landé factor $\bar{g}_{\rm eff}=1.12$. As we show in Fig. 1a, the spectral lines have a rather complex structure with sharp absorption components in the line cores. Only the observations obtained on 2017 June 3 and 5 reveal the presence of Zeeman features. The measured magnetic field strengths show a change of polarity within one day from $\langle B_{z}\rangle=-121\pm 11$ G to $\langle B_{z}\rangle=209\pm 10$ G. We assume that the sharp cores observed in the line profiles are originating in the stellar photosphere.

3.2 HD 98922

HD 98922 is the second late Herbig Be star in our sample. A first measurement of the longitudinal magnetic field in this star, $\langle B_{z}\rangle=-131\pm 34$ G, was published by Hubrig et al. (2013) using the moment technique, whereas Alecian et al. (2013b) reported a non-detection. As discussed by Hubrig et al. (2013), Fe ii and Ti ii lines in the high resolution HARPS spectra of this star show traces of splitting and lines belonging to different elements exhibit different shapes of their profiles, suggesting that some elements are inhomogeneously distributed on the stellar surface. The interferometric study by Kraus et al. (2008) indicated that the size of the Br$\gamma$-emitting region is consistent with magnetospheric accretion.

Our line mask is based on $T_{\rm eff}\approx 10500$ K (van den Ancker, de Winter, & Tjin A Djie, 1998) and $\log\,g=4.0$ (Alecian et al., 2013b). The SVD profiles have been calculated using 202 unblended lines with an average Landé factor $\bar{g}_{\rm eff}=1.18$. The analysis of the SVD profiles (Fig. 1b) reveals a complex and variable shape of the Stokes $I$ profiles. The longitudinal magnetic field shows different magnetic field polarity when comparing the data from 2011 with $\langle B_{z}\rangle=-33\pm 4$ G to data from 2016 with $\langle B_{z}\rangle=28\pm 7$ G.

3.3 HD 139614

HD 139614 is a late Herbig Ae star for which previously published magnetic field measurements indicated both non-detections and detections (Wade et al., 2005; Hubrig et al., 2004, 2007, 2009; Alecian et al., 2013a). Wade et al. (2005) failed to detect a magnetic field in HD 139614 with quoted uncertainties of 25 G on two consecutive nights, whereas Hubrig et al. (2004, 2007) reported the presence of a weak magnetic field in the range from $-$116 G to $-$450 G.

HD 139614 is probably possessing a transition disk, where the dust gap has been opened by a single giant planet. Interferometric observations obtained with the MIDI instrument on the Very Large Telescope Interferometer together with temperature-gradient modeling favoured a two-component disk structure with spatially separated inner and outer dust components (Matter et al., 2014). The disk inclination was found to be 20°, signifying that we see the target close to face-on. Later Matter et al. (2016) performed a first multi-wavelength modeling of the dust disk and confirmed a gap structure. The CRyogenic high-resolution InfraRed Echelle Spectrograph (CRIRES) observations by Carmona et al. (2017), made at the European Southern Observatory’s Very Large Telescope (VLT), suggested the presence of an embedded $<2M_{\rm J}$ planet inside the dust gap at a very close distance of about 4 AU.

HD 139614 is the coolest target among this sample. Our line mask is based on $T_{\rm eff}=7600$ K and $\log\,g=3.9$ (Folsom et al., 2012). The SVD profiles (Fig. 1c) were calculated using 2870 lines with an average Landé factor $\bar{g}_{\rm eff}=1.20$. The analysis of the SVD profiles indicates a definite detection of the longitudinal magnetic field with $\langle B_{z}\rangle=25\pm 3$ G (FAP $<10^{-8}$).

3.4 HD 165133

HD 165133 is an early Herbig Be star listed in the catalogue of members and candidate members of the Herbig Ae/Be stellar group by Thé, de Winter, & Pérez (1994). Already in the late 1930s it was reported to belong to the Sagittarius region of the Milky Way (Wallenquist, 1939). The study by Chen, de Grijs, & Zhao (2007) confirmed its membership in the extremely young open cluster NGC 6530 at the age of 1.5–2 Myr, which is the dominant cluster in the Sgr OB1 association and is located in the eastern part of the Lagoon nebula. Fig. 1d shows the detection of a weak longitudinal magnetic field with $\langle B_{z}\rangle=97\pm 9$ G. Our line mask is based on the B2 spectral type indicated in the SIMBAD database. The SVD profiles were calculated using 103 spectral lines with an average Landé factor $\bar{g}_{\rm eff}=1.21$.

3.5 HD 190073

The absorption and emission spectrum of the Herbig Ae star HD 190073 was studied in detail by Catala et al. (2007) and Cowley & Hubrig (2012). This star has a very low projected rotational velocity ($v\,\sin\,i\leq 9$ km s*-1*), which may indicate either a very slow rotation or a very small inclination of the rotation axis with respect to the line of sight. The latter option is supported by interferometric observations (Eisner et al., 2004) as well as by IUE observations (Hubrig et al., 2009).

A first measurement of the longitudinal magnetic field $\langle B_{z}\rangle=84\pm 30$ G in HD 190073 was published by Hubrig et al. (2006) and was based on spectra obtained with the low-resolution FOcal Reducer low dispersion Spectrograph (FORS 1; Appenzeller et al., 1998) at the VLT. Wade et al. (2007) reported a non-detection, but the following studies by Catala et al. (2007), Hubrig et al. (2009), and Alecian et al. (2013b) reported definite detections of longitudinal magnetic fields with strengths ranging from 111 G to $-$35 G.

The observations obtained during 2011 and 2012 (Järvinen et al., 2015) showed a definite detection $\langle B_{z}\rangle=-8\pm 6$ G with a FAP smaller than $10^{-8}$ and a marginal detection $\langle B_{z}\rangle=-15\pm 10$ G with a FAP of $8\times 10^{-4}$. Since the variability time scale of the magnetic field of this star is not yet known, we decided to monitor it on additional epochs. The line mask applied in our study is based on $T_{\rm eff}=9250$ K and $\log\,g=3.5$ (Acke & Waelkens, 2004). The SVD profiles (Fig. 1e) using new data from 2016 and 2017 are calculated using 688 lines with an average Landé factor $\bar{g}_{\rm eff}=1.21$ and show definite detections of positive fields ranging from $17\pm 1$ G to $34\pm 2$ G.

3.6 Spectral variability on short and long-term scales

Similar to the spectroscopical behaviour of other Herbig Ae/Be stars (e.g. Hubrig et al., 2011a, 2012; Järvinen et al., 2016, 2018, 2019), the Herbig stars with multiple exposures in our sample show variable line profiles. As we show in Fig. 2, the strongest changes in the SVD Stokes $I$ profiles are detected in HD 98922. Since a number of Herbig Ae stars are known to exhibit $\delta$ Scuti-like pulsations (e.g. Zwintz, Guenther, & Kallinger, 2008), we also checked the variability on short time scales in the spectropolarimetric subexposures. Only HD 139614 shows clear differences in the observed line profiles recorded in subexposures separated by 37 min (see Fig. 3).

4 Discussion

Magnetic fields play a key role in the processes leading to the formation of stars and planets. Analytical models and MHD numerical simulations of the evolution of star forming cores show that the magnetic field is critical for transporting angular momentum during the protostellar phase and sets the conditions for strongly anisotropic accretion. The presented magnetic field measurements – with the strongest longitudinal magnetic field of 209 G detected in the Herbig Be star HD 58647 and the weakest field of 17 G measured for HD 190073 – provide further evidence that Herbig Ae/Be stars possess much weaker magnetic fields than their lower mass counterpart T Tauri stars, with magnetic fields of kG order.

While for strongly magnetic T Tauri stars the model of magnetically driven accretion and outflows successfully reproduce many observational properties, the mechanism of mass transfer within the Herbig Ae/Be star + disk system is unclear due to the limit of weak or even absent magnetic fields. Yet the small number of detected magnetic Herbig Ae/Be stars can be explained not only by the weakness of their magnetic fields but also by too large measurement errors. Alecian et al. (2013b) reported that about 90% of their sample of 70 stars do not show detectable magnetic fields. However, the uncertainty of the magnetic field determinations in their study was worse than 200 G for 35% of the measurements and for 32% of the measurements it was between 100 and 200 G, i.e. only 33% of the measurements showed a measurement accuracy below 100 G. As we have shown in our previous work (Hubrig et al., 2015; Järvinen et al., 2015, 2018), the uncertainties of the measurements using the SVD technique are dramatically better than those achieved with the more commonly used Least Squares Deconvolution (Donati et al., 1997) technique. This can be explained by the way the SVD procedure builds a weighted mean SVD line profile. As described by Carroll et al. (2012), the resulting mean profiles are calculated by the projection of the observed line profiles onto the basis of identified signal eigenprofiles (Eq. 13 in Carroll et al., 2012). The projection coefficients (the inner products) are the weights that each line from the line list contributes to the resulting SVD profile. In observed spectra with varying noise level over the wavelength range it has the effect that the regions with higher noise have a lower weight according to their smaller projection coefficients. This is different from the noise weighting in the least-squares solution of the LSD method.

Furthermore, single snapshot observations are not sufficient to judge whether a Herbig Ae/Be star is magnetic or not. The longitudinal magnetic field is defined as the disk-integrated magnetic field component along the line of sight and therefore shows a strong dependence on the viewing angle of the observer, i.e. on the rotation angle of the star. The limitations set by the strong geometric dependence of the longitudinal magnetic field are usually overcome by repeating observations several times, so as to sample various rotation phases, hence various aspects of the magnetic field. Unfortunately, the rotation periods of Herbig Ae/Be stars are poorly known (e.g. Hubrig et al., 2011a, b). The observed light variations in these stars are likely of stochastic nature and caused by fluctuating disk accretion. Multi-epoch rotation-modulated longitudinal magnetic field measurements are frequently used to determine rotation periods, but such monitoring with HARPSpol is possible only in the framework of a large programme.

Acknowledgements

We thank the referee, G. Mathys, for his valuable comments. Based on observations made with ESO Telescopes at the La Silla Paranal Observatory under programme IDs 097.C-0277(A) and 099.C-0081(A). This work has made use of the VALD database, operated at Uppsala University, the Institute of Astronomy RAS in Moscow, and the University of Vienna. This research has made use of the SIMBAD database, operated at CDS, Strasbourg, France.