Discovery of a magnetic field in $\rho$ Pup

TL;DR

This paper reports the detection of a very weak magnetic field in the pulsating star $ ho$ Pup, marking it as only the second known magnetic $ ho$ Scuti star and suggesting a link to ultra-weak magnetic Am stars.

Contribution

The study presents the first high-precision spectropolarimetric detection of a magnetic field in $ ho$ Pup, expanding the understanding of magnetism in pulsating Am stars.

Findings

Magnetic field detected with longitudinal strength below 1 G

$ ho$ Pup is the second magnetic $ ho$ Scuti star

Possible connection to ultra-weak magnetic Am stars

Abstract

$\rho$ Pup is a $\delta$ Scuti F2 pulsator, known to host a main radial mode as well as non-radial pulsations, with chemical peculiarities typical of evolved Am stars. We present high-precision spectropolarimetric observations of this star, obtained with ESPaDOnS at the Canada France Hawaii Telescope (CFHT) in the frame of the BRITE spectropolarimetric survey. A magnetic field is clearly detected in $\rho$ Pup, with a longitudinal field strength below 1 G. This makes $\rho$ Pup the second known magnetic $\delta$ Scuti discovered, after HD 188774, and a possible cool evolved counterpart of the recently discovered ultra-weakly magnetic Am family.

Tables

Table 1. Table 1: Journal of ESPaDOnS observations of ρ 𝜌 \rho  Pup indicating the Heliocentric Julian Date at the middle of the observations (mid-HJD - 2450000), the exposure time, and the signal-to-noise ratio of the (co-added) spectrum at ∼ similar-to \sim 500 nm.

#	Date	mid-HJD	Texp	S/N
		-2450000	s
1	Feb 10, 2014	6698.86667	2x4x30	1416
2	Oct 31, 2015	7327.14429	8x4x25	3528

Table 2. Table 2: Detection probability in the Stokes V and N profiles, detection status, and longitudinal magnetic field measurements from the V and N profiles with their error bars for the two spectropolarimetric measurements of ρ 𝜌 \rho  Pup.

#	ProbV	ProbN	Detection	$B_l\pm \sigma B$	$N_l\pm \sigma N$
	%	%		G	G
1	11.5	0.9	No	-0.05$\pm$0.72	-0.31$\pm$0.72
2	100	23.8	Definite	-0.29$\pm$0.32	0.35$\pm$0.32

Full text

††pubyear: 2016††pagerange: Discovery of a magnetic field in the $\delta$ Scuti F2m star $\rho$ Pup–References

1 Introduction

$\delta$ Scuti variables are A and F stars pulsating with low radial order pressure modes and mixed modes driven by the $\kappa$ mechanism operating in the He ii ionization zone (Pamyatnykh, 2000) and by the turbulent pressure in the hydrogen ionisation layer (Antoci et al., 2014; Xiong et al., 2016). Their pulsation periods range from approximately 15 min to 8 h. $\delta$ Scuti stars can be on the pre-main sequence (PMS), main sequence (MS), or start of the post-MS. They have masses from 1.5 to 2.5 M⊙ and effective temperatures between 6700 and 8000 K. Between 6900 and 7400 K, $\delta$ Scuti stars can also host gravity modes of $\gamma$ Dor type. These pulsations are driven by convective blocking near the base of their convective envelopes (Guzik et al., 2000; Dupret et al., 2004). The stars showing both types of pulsations at the same time are called hybrid stars.

$\rho$ Pup (HD 67523) is a bright (V=2.81) and long-known $\delta$ Scuti star (e.g. Cousins, 1951; Eggen, 1956; Ponsen, 1963; Mathias et al., 1997; Dall & Frandsen, 2002). Its main pulsation frequency $f=7.0972$ c d*-1* has been identified as the radial fundamental mode, and it also hosts at least one non-radial mode with $f=7.9$ c d*-1* (Mathias et al., 1997; Tkachenko et al., 2013; Antoci et al., 2013; Nardetto et al., 2014). In addition, a search for solar-like oscillations was performed for this star thanks to a spectroscopic multisite campaign, but no such pulsations were detected (Antoci et al., 2013).

The stellar parameters of $\rho$ Pup have been determined to be within $T_{\rm eff}=6500-6850$ K and $\log g=3.3-3.7$ dex, depending on the authors (e.g. Kurtz, 1976; Prugniel et al., 2011; Nardetto et al., 2014). According to Hipparcos measurements, $\rho$ Pup is located at $19.48\pm 0.06$ pc (van Leeuwen, 2007). Its radius is $R=3.52\pm 0.07$ R⊙ (Antoci et al., 2013) and its absolute angular diameter varies by 11 $\mu$as (or 0.7%) due to the pulsations (Nardetto et al., 2014).

Moreover, $\rho$ Pup is an evolved F2 star, showing enhanced metal lines (Kurtz, 1976). Kurtz (1976) and Gray & Garrison (1989) suggested that $\rho$ Pup is a cool, evolved Am star, and Yushchenko et al. (2015) confirmed that the abundance pattern of $\rho$ Pup is similar to Am stars. In fact, stars showing such types of peculiar spectra have been named $\rho$ Puppis stars by Gray & Garrison (1989). Other authors also called them $\delta$ Delphini stars, although this class contained varied types of spectra and the name is therefore deprecated now.

Finally, no companion was detected for this star (Nardetto et al., 2014).

2 Spectropolarimetry

2.1 Observations

The BRITE spectropolarimetric survey aims at acquiring very high signal-to-noise (S/N) Stokes V spectra of all stars brighter than V=4 with high-resolution spectropolarimeters (Neiner et al., 2016). This program provides ground-based support for the BRITE (BRIght Target Explorer) constellation of nano-satellites, which performs photometric observations of bright stars, in particular for seismology (Weiss et al., 2014).

In this frame, we observed $\rho$ Pup with the ESPaDOnS spectropolarimeter at the Canada France Hawaii Telescope (CFHT) in Hawaii. ESPaDOnS covers a spectral range from about 375 to 1050 nm, with a resolving power of $\sim$68000, spread on 40 echelle orders.

We used the circular polarisation mode to measure the Stokes V spectrum together with the intensity spectrum (Stokes I). Each Stokes V sequence consists of the combination of 4 sub-exposures obtained with the polarimetry half-wave Fresnel rhombs set at various angles. The sub-exposures are also destructively combined to produce a null polarisation (N) spectrum to check for pollution by, e.g., variable observing conditions, instrumental effects, or non-magnetic stellar effects such as pulsations. In addition, successive Stokes V sequences have been acquired to increase the total S/N ratio of a magnetic measurement.

$\rho$ Pup was observed a first time on February 10, 2014, for 2 successive sequences of 4$\times$30 s each, and a second time on October 31, 2015, for 8 successive sequences of 4$\times$25 s (see Table 1). Following the ephemeris provided by Nardetto et al. (2014), this corresponds to pulsation phases 0.87 and 0.49, respectively. The duration of each Stokes V sequence, including the detector readout time ($\sim$38 s), is either 234 or 214 s, which is less than 1/50th of the main pulsation period and corresponds to a radial velocity change below 0.2 km s*-1* (i.e. well below the 1.8 km s*-1* resolution of the data). This insures that the spectropolarimetric observations are not influenced by line profile variations or radial velocity shifts due to pulsations.

The data were reduced with the Upena pipeline feeding the Libre-Esprit reduction package (Donati et al., 1999) available at CFHT. This included standard bias removal, flat-fielding, and wavelength calibration. The Stokes I spectra were then normalized to the continuum level using IRAF111IRAF is distributed by the National Optical Astronomy Observatory, which is operated by the Association of Universities for Research in Astronomy (AURA) under a cooperative agreement with the National Science Foundation., and the same normalization was applied to the Stokes V and null N spectra.

$\rho$ Pup is known to host a chromosphere, detectable as emission in the UV Mg ii h and k lines (Fracassini et al., 1983). Weak emission was also claimed in the Ca ii K line by Dravins et al. (1977) but remained unconfirmed by subsequent observations (Mathias et al., 1997; Mathias et al., 1999), possibly indicating a transient nature due to a shock wave. In the ESPaDOnS spectra presented here, we detect no emission, neither in the Ca ii K line nor in the H$\alpha$ line (see Fig. 1).

2.2 Magnetic field detection

To check for the presence of a magnetic field, we used the Least Squares Deconvolution (LSD) technique (Donati et al., 1997). LSD requires a mask listing the lines in the spectrum, their wavelength, depth, and Landé factor. To produce this line mask, we first extracted a line list from the VALD3 atomic database (Piskunov et al., 1995; Kupka et al., 1999; Ryabchikova et al., 2015) for the typical values $T_{\rm eff}=6750$ K and $\log g=3.5$ of $\rho$ Pup. We only used lines with a depth larger than 10% of the continuum. We then removed from the mask all lines that are not visible in the intensity spectra, hydrogen lines because of their Lorentzian broadening, the lines blended with H lines or interstellar lines, as well as lines in regions affected by telluric absorption. Finally, the depth of each line in the LSD mask was adjusted to fit the observed line depth. The final mask tailored for $\rho$ Pup contains 12515 lines and was used to produce the LSD Stokes profiles.

Consecutive sequences of spectra were co-added to produce one single magnetic measurement per night. Our LSD procedure forces the LSD mean intensity weight and mean polarisation weight to be equal to 1. As a consequence, the mean Landé factor is 1.187 for the first observation and 1.188 for the second observation. The mean wavelength is 548.94 nm for the first observation and 541.08 nm for the second observation. Adopting the standard wavelength of 500 nm, this corresponds to an equivalent Landé factor of 1.383 for the first observation and 1.377 for the second observation.

We find that both observations show a Zeeman signature in their Stokes V profile, indicating that the star is magnetic (see Fig. 2). The two signatures, obtained in 2014 and 2015, are very similar in shape and shifted with the intensity profile. As expected, thanks to the short duration of the polarimetric sequence, the N profiles are flat and show only noise, i.e. the measurements have not been polluted by the stellar pulsations.

The formal statistical detection of a magnetic field is evaluated by the False Alarm Probability (FAP) of the signature in the LSD Stokes V profile inside the LSD line, compared to the mean noise level in the LSD Stokes V profile outside the line. We adopted the convention defined by Donati et al. (1997): if the FAP is below 0.001% the magnetic detection is definite, if the FAP is between 0.001% and 0.1% the detection is marginal, otherwise there is no magnetic detection. The same procedure was also applied to the N profiles.

We obtain no formal detection in the first observation, which has a low S/N. Nevertheless, our experience is that the FAP criterion is relatively conservative. We judge the signature in the line profile obtained on February 10, 2014, to be significant, and it is corroborated by our subsequent definite detection of the field in the second observation from October 31, 2015, thanks to its higher S/N. When applied to the N profiles instead of the V profiles, we get no detection in both measurements (see Table 2).

2.3 Magnetic field measurements

From the Stokes V and I profiles of each night of observations, we can calculate the longitudinal magnetic field value, i.e. the strength of the magnetic field in the direction of the observer. We use the center-of-gravity method (Rees & Semel, 1979; Wade et al., 2000) over a velocity range of $\pm$33 km s*-1* around the line centroid (48 and 42.5 km s*-1*, respectively). Results are shown in Table 2. The longitudinal magnetic field of $\rho$ Pup is very weak, below 1 G.

3 Discussion

HD 188774 was the only confirmed magnetic $\delta$ Scuti star so far (Neiner & Lampens, 2015). Kurtz et al. (2008) and Hubrig & Scholler (2016) claimed that the Ap star HD 21190 is magnetic and pulsating, however Bagnulo et al. (2012) showed that the star is probably not an Ap star and the magnetic detection is likely spurious (as are most of the few-$\sigma$ detections obtained with FORS). Alecian et al. (2013) also claimed a possible magnetic detection in the $\delta$ Sct star HD 35929. However, this object is a Herbig star and the spectropolarimetric results were unclear. Therefore, the clear detection of a magnetic field in the well known $\delta$ Scuti star $\rho$ Pup makes it the second confirmed magnetic $\delta$ Scuti star, after HD 188774.

The magnetic field measured in $\rho$ Pup is very weak. The weakness of the surface field can only be partly explained by the evolutionary state of $\rho$ Pup. The current radius of $\rho$ Pup is 3.52 R⊙. Assuming magnetic flux conservation, the surface field when the star was on the MS was only about 5 times stronger than the current field.

The magnetic field inferred directly from the Stokes V profile corresponds only to the line of sight component. Thus it is possible that $\rho$ Pup possesses a relatively strong dipole magnetic field that was viewed near the magnetic equator (at which $B_{l}$ is null) in both observations. This is qualitatively consistent with the “crossover” morphology of the Stokes V signatures observed in both measurements. However, given the very high precision of the observations, it seems unlikely that the dipole field would be stronger than a few tens of G. Another possible (although admittedly unlikely) explanation for the similar signatures observed in both measurements is that $\rho$ Pup possesses a surface magnetic field with a dominant toroidal configuration. Testing these various scenarios will require monitoring and modeling of the phase variation of the Zeeman signature.

Very weak fields have been discovered in several Am stars (Petit et al., 2011; Blazère et al., 2016a; Blazère et al., 2016b). In these stars, the longitudinal field strength is usually below 1 G, except for Alhena for which it is a few gauss (Blazère et al., 2016a). Yushchenko et al. (2015) showed that the chemical peculiarities observed in the spectrum of $\rho$ Pup are similar to Am stars. Therefore, it is likely that $\rho$ Pup is a cooler, evolved counterpart of magnetic Am stars, as also proposed by Kurtz (1976) and Gray & Garrison (1989). The magnetic field of Am stars is thought to be of fossil origin. The simplicity and similarities in the Stokes V profiles of the two spectropolarimetric measurements presented here would corroborate this hypothesis. In the case of a dynamo field, we would expect the Zeeman signature to have changed more drastically between the two epochs.

Finally, Nardetto et al. (2014) showed that $\rho$ Pup does not fit their expected atmospheric velocity gradient curve (their Fig. 9). They proposed that this may be related to the low peak-to-peak amplitude of the radial velocity curve. Following the discovery of a magnetic field in this star, we propose that magnetic effects could explain this discrepancy. Indeed, the action of the magnetic field could modify the velocity gradients.

4 Conclusions

$\rho$ Pup is a new magnetic $\delta$ Scuti star, and only the second confirmed member of this class. Its rather simple and stable Zeeman signatures point towards a likely fossil origin of its magnetic field. Its very weak field strength and chemical peculiarities link it to the recently discovered family of ultra-weakly magnetic Am stars.

Following the discovery of a magnetic field in $\rho$ Pup, a complete spectropolarimetric follow-up of this star over its rotation period should be performed to characterise its field strength and topology in detail. Then, this star will be an excellent target for magneto-asteroseismology: its radial and non-radial pulsations combined to the magnetic information will allow one to tidely constraint its seismic models. Moreover, it would be interesting to check how the magnetic field varies during the chromospheric cycle and if it is correlated with a possible shock wave.

Acknowledgements

This research has made use of the SIMBAD database operated at CDS, Strasbourg (France), and of NASA’s Astrophysics Data System (ADS). GAW acknowledges Discovery Grant support from the Natural Sciences and Engineering Research Council of Canada.