MNRAS 464, 1294–1305 (2017) doi:10.1093/mnras/stw2335 Optical linear polarization of 74 white dwarfs with the RoboPol polarimeter Michał Zejmo,˙ 1 Aga Słowikowska,1‹ Krzysztof Krzeszowski,1 Pablo Reig2,3 and Dmitry Blinov2,3,4 1Janusz Gil Institute of Astronomy, University of Zielona Gora,´ Lubuska 2, 65-265 Zielona Gora,´ Poland 2Foundation for Research and Technology, 71110 Heraklion, Crete, Greece 3University of Crete, Physics Department, PO Box 2208, 710 03 Heraklion, Crete, Greece 4Astronomical Institute, St Petersburg State University,Universitetsky Pr. 28, Petrodvoretz, 198504 St Petersburg, Russia Accepted 2016 September 14. Received 2016 September 13; in original form 2016 February 1 ABSTRACT We present the first linear polarimetric survey of white dwarfs (WDs). Our sample consists of WDs of DA and DC spectral types in the SDSS r magnitude range from 13 to 17. We performed polarimetric observations using the RoboPol polarimeter attached to the 1.3-m telescope at Skinakas Observatory. We have 74 WDs in our sample, of which almost all are low-polarized WDs with a polarization degree (PD) lower than 1 per cent; only two have a PD higher than 1 per cent. There is evidence that on average isolated DC-type WDs have a higher PD (with a median PD of 0.78 per cent) than isolated DA-type WDs (with a median PD of 0.36 per cent). On the other hand, the median PD of isolated DA-type WDs is almost the same (i.e. 0.36 per cent) as the median PD of DA-type WDs in binary systems with red dwarfs (dM type; i.e. 0.33 per cent). This shows, as expected, that there is no contribution to the PD from the companion if the WD companion is a red dwarf, which is the most common situation for WD binary systems. We do not find differences in the PD between magnetic and non-magnetic WDs. Because 97 per cent of WDs in our sample have a PD lower than 1 per cent, they can be used as faint zero-polarized standard stars in the magnitude range from 13 to 17 of the SDSS r filter. They cover the Northern sky between 13h and 23h in right ascension and between −11◦ and 78◦ in declination. In addition, we found that for low extinction values (<0.04), the best model that describes the dependence of the PD on E(B − V) is given by the equation 0.12 PDmax,ISM[per cent] = 0.65 E(B − V) . Key words: standards – polarization – instrumentation: polarimeters – techniques: polarimet- ric – white dwarfs. & Shearer 2014; Mignani et al. 2015) and magnetars (e.g. Wang 1 INTRODUCTION et al. 2015), as well as neutron stars in high-mass X-ray bina- In recent years interest in optical polarimetry has grown signifi- ries (e.g. Reig et al. 2014, Słowikowska et al. in preparation) and cantly (e.g. Marscher et al. 2010). The reason for this increase is low-mass X-ray binaries (e.g. Baglio et al. 2014). There are also that polarimetric measurements give an invaluable additional con- polarization studies of Gamma-Ray Bursts (Mundell et al. 2013) straint on theoretical models that cannot be provided by photom- and of polarized light from exoplanets, for which a dedicated de- etry nor astrometry nor spectrometry. Optical polarimetry studies tector, namely the Spectro-Polarimetric High-contrast Exoplanet encompass all kinds of astrophysical objects. There are regular REsearch (SPHERE), at Very Large Telescope (VLT) has recently monitoring campaigns to study the polarization changes of active been built.1 galactic nuclei (AGNs) in the optical domain, for example the op- The scientific community has started to use polarimetric measure- tical monitoring of selected blazars with the RoboPol polarimeter ments extensively to study stellar and non-stellar objects. However, (Pavlidou et al. 2014). There are also optical polarization studies attempts to reach fainter objects using infrastructure with larger of isolated neutron stars, including pulsars (e.g. Słowikowska et al. mirrors have encountered a serious problem, namely a lack of faint 2009; Lundqvist et al. 2011; Moran et al. 2013; Moran, Mignani polarization standards of both types – zero-polarized and polarized. E-mail: [email protected] 1 https://www.eso.org/sci/facilities/paranal/instruments/sphere.html C 2016 The Authors Published by Oxford University Press on behalf of the Royal Astronomical Society Optical linear polarization of 74 white dwarfs 1295 Each measurement using a polarimeter or spectropolarimeter needs of the WDs in our sample is that they are even fainter, namely to be properly calibrated. Thus, polarized standards are necessary in the SDSS r magnitude range from 13.2 (WD2149+021) to 17 to establish the intrinsic depolarization caused by the instrument, (WD2213+317), than those currently available in the literature. In while zero-polarized standards are necessary to determine the in- this sense our sample complements earlier work. A larger group of strumental polarization (e.g. the RINGO3 polarimeter at the Liver- zero-polarization standards will enable observers to find a visible pool Telescope, see Słowikowska et al. 2016). standard at a convenient time of night and position on the sky. The aim of our work is twofold: (i) to perform a statistical anal- Many polarization studies of large samples of WDs have been ysis of the linear polarization properties of a WD sample; and conducted to date, for example Angel, Borra & Landstreet (1981), (ii) to provide observers with new faint linear polarimetric stan- Schmidt & Smith (1995), Putney (1997), Kawka et al. (2007), dard sources. Jordan et al. (2007), Kawka & Vennes (2012) and Landstreet et al. There are more than 23 000 WDs known to date (see Section 2). (2015). Recently, Bagnulo et al. (2015) published a spectropolari- For many of these, the spectral type is known. Most WD atmo- metric catalogue of 809 objects, obtained with the FORS/VLT spheres are hydrogen-rich atmospheres (DA), and almost all the instrument, which includes 70 WDs. However, there is only one rest are helium-rich (DB). However, a significant fraction of WDs WD common to our list and that of Bagnulo et al. (2015), also contain trace elements in their atmospheres, and these are la- namely WD2149+021. The crucial difference between the above- belled with Z for metals or Q for carbon, for example as DZ or DQ. mentioned studies and our work is that our observations represent There are also objects for which the WD spectrum does not show the first WD linear polarization survey, whereas the others measured any strong lines, but their atmospheres are still helium-rich. Such the circular polarization. WDs are classified as DC-type WDs. WDs with magnetic fields We describe the sample selection method in Section 2, the obser- stronger than 1 MG can be detected using Zeeman splitting, while vations in Section 3, and the data analysis in Section 4. Results and weaker magnetic fields can be detected using spectropolarimetry. conclusions are given in Sections 5 and 6, respectively. However, because DC-type WD spectra do not have strong spectral lines, it is not possible to use the Zeeman effect to measure their magnetic field strength. Previous studies showed that about 10 to 2 SELECTED SAMPLE 20 per cent of all WDs are magnetized with a strong magnetic field Our sample was built up from the following catalogues: the White (Ferrario, de Martino & Gansicke¨ 2015, and references therein). Dwarf Catalogue of Villanova University2 (McCook & Sion 1999, The PD of magnetized WDs is between a fraction of a per cent hereafter VUWDC), the ‘SDSS DR7 White Dwarf Catalog’ (Klein- and a few per cent. For example GD 229 has a linear PD of almost man et al. 2013, hereafter DR7WDC), the ‘Post-Common Envelope 8 per cent in the R band (Berdyugin & Piirola 1999). The popula- Binaries from SDSS-XIV. The DR7 White Dwarf-Main-Sequence tion of magnetized WDs could be even larger than current estimates Binary Catalogue’ (Rebassa-Mansergas et al. 2012, hereafter because the magnetic field of many sources is unknown. DR7WDC-bin). We gathered 23 068 objects from those catalogues. Linear polarimetric population studies can help in the selection However, it should be noted that there are 14 235, 18 913 and of WDs that are good candidates for stable polarimetric calibra- 2248 WDs in VUWDC, DR7WDC and DR7WDC-bin, respec- tion sources. Moreover, for each WD we have information about tively. Some of WDs appear in all three catalogues, and therefore whether it is an isolated WD or a WD in a binary system. In most the final number of WDs, namely 23 068, is smaller than the sim- cases, the companion is a low-mass red dwarf. Magnetized WDs in ple sum of WDs from these catalogues. Once the sky coordinates binary systems with a low-mass star with which they are in contact were extracted, we searched for their photometric brightness mea- are classified as magnetic cataclysmic variables (MCVs, i.e. polars surements in the following sky surveys: SDSS (York et al. 2000; and intermediate polars) and represent one-quarter of the whole CV Ahn et al. 2012), UKIDSS (Lawrence et al. 2007; Casali et al. population (Wickramasinghe & Ferrario 2000). There are also bi- 2007;Hewettetal.2006; Hodgkin et al. 2009; Hambly et al. 2008), nary systems composed of a WD and a low-mass star that are close 2MASS (Skrutskie et al. 2006) and WISE (Wright et al. 2010). Of but not in contact, namely pre-CVs. However, no such system with the 23 068 WDs, 18 877 have a known spectral type. Based on the a magnetic WD is known to date (Liebert et al.