EDP Sciences
Free Access
Issue
A&A
Volume 591, July 2016
Article Number A84
Number of page(s) 22
Section Catalogs and data
DOI https://doi.org/10.1051/0004-6361/201526625
Published online 20 June 2016

© ESO, 2016

1. Introduction

To date almost 1,300 brown dwarfs have been found, classified with spectral types L, T, and Y (Kirkpatrick 2008; Cushing et al. 2011)1. This number has been surpassed by indirect detections in the planetary regime – more than 600 by radial velocity variations and more than 1200 by the transit method2.

While indirect methods are most sensitive to objects in very close orbits around the targets, direct imaging detects objects in wider orbits and thus is complementary. Ideally, direct imaging is combined with simultaneous measurement of the radial velocity variation (Guenther et al. 2005). Young substellar objects still contract and are self-luminous (Burrows et al. 1997) so that their direct detection is less difficult up to ages of a few hundred Myr (Malkov et al. 1998; Neuhäuser & Schmidt 2012). Unlike high-resolution imaging in space (e.g. Marengo et al. 2006), ground-based observations need to correct the seeing that can be done with adaptive optics (e.g. Neuhäuser et al. 2003; Duchêne et al. 2007). Coronagraphy with intransparent (e.g. McCarthy & Zuckerman 2004; Chauvin et al. 2005) or semi-transparent coronagraphs (e.g. Biller et al. 2010; Boccaletti et al. 2013; Guenther et al. 2005; Itoh et al. 2006, 2008a; Neuhäuser et al. 2007; Wahhaj et al. 2011) shade the bright star so that the exposure time and thus the sensitivity of the images can be increased. A number of very sophisticated techniques have been developed recently (see the review by Fischer et al. 2014).

The frequency of low-mass companions to stars gives important clues regarding our understanding of the formation of brown dwarfs and planets (Ida & Lin 2004; Alibert et al. 2005; Broeg 2007). At wide orbits, where direct imaging surveys are sensitive, the frequency of brown dwarfs is of the order of several percent for host stars with spectral types A-M and is less constrained for giant planets (Rameau et al. 2013; Bowler et al. 2015). In young nearby associations Neuhäuser & Guenther (2004) measure a frequency of 6 ± 4% of substellar companions which is not very different from the value of 1 ± 1% obtained for isolated late-type stars (McCarthy & Zuckerman 2004). The frequency of brown dwarfs around Hyades members is not very different from the latter value (Guenther et al. 2005; Bouvier et al. 2008; Lodieu et al. 2014) although the Hyades are still young at an age of 600 Myr. In the younger Pleiades (125 Myr), Yamamoto et al. (2013) confirmed two brown dwarf companions in a sample of 20 stars.

Young moving groups of an age intermediate between the Pleiades and the Hyades offer interesting opportunities to study homogeneous samples of common age and origin. No substellar companions have been found in the Her-Lyr association (Eisenbeiss et al. 2007; Biller et al. 2013) which has an age of 250 Myr, similar to the UMa group (Eisenbeiss et al. 2013). Although a systematic survey of the UMa group at high resolution has been missing, low-mass companions have been detected: GJ 569 Ba, Bb (Martín et al. 2000; Zapatero Osorio et al. 2004), HD 130948 B & C, (Potter et al. 2002), and χ1 Ori B (König et al. 2002)3. The only known planet around a probable UMa group member, ϵ Eri, was found by radial velocity variations (Hatzes et al. 2000; Benedict et al. 2006). Although there are still doubts about its existence (Zechmeister et al. 2013), there are even suggestions that a second planet might exist (Quillen & Thorndike 2002; Deller & Maddison 2005). The planet(s) of ϵ Eri have been subject to many attempts (still unsuccessful) of direct detection (Itoh et al. 2006; Janson et al. 2007; Marengo et al. 2006; Neuhäuser & Schmidt 2012, and the present work).

The goal of the present work is to find additional substellar companions in the UMa group by direct imaging. The search for close and faint companions in the UMa group benefits from a relatively young age of 200–600 Myr4, the small distance of members (50 pc on average; Fig. 1), and thus the availability of precise Hipparcos astrometry (Perryman et al. 1997; van Leeuwen 2007). Furthermore, the proper motion of UMa group members is high on average (150 mas/yr) because of their proximity (Fig. 1) and their peculiar space motion (Ammler 2006). Hence, co-moving companions can already be identified after a short epoch difference. Preliminary results of the present study have already been published (Ammler-von Eiff et al. 2009).

2. Coronagraphic observations and data reduction

As the definition of the UMa group and the list of members is controversial, we compiled the targets in the following way to obtain a meaningful number of reliable UMa group members closer than 30 pc and observable with NACO at the ESO VLT:

We did not observe one system, GJ 569 (=HIP 72944) because it has been extensively studied before (e.g. Simon et al. 2006). Some objects were excluded because of known bright secondaries in the field of view (HD 24916, HD 29875, HD 98712, and HD 134083). HIP 104383 A is also a close binary but fully fits below the coronagraph so that we did observe it (Fig. D.1), i.e. 20 targets in total (Tables A.1 and A.2).

thumbnail Fig. 1

Left: distribution of distance and proper motion. The major concentration of the UMa group is the UMa open cluster in the Big Dipper constellation at distances of about 25 pc. All known members and candidates (see references in the text) are included in the histogram (solid line; distances mostly from Montes et al. 2001 and van Leeuwen 2007). We add the distribution of the southern stars within 30 pc addressed in the present work (hatched; distances from van Leeuwen 2007). Right: image of the coronagraphic mask with the four coronagraphs (black filled circles) taken with the S27 camera. Each of the coronagraphs has an angular diameter of . We placed the star below the lower left (south-eastern) coronagraph, so that a field with a radius of 9′′ is completely covered outside the coronagraph (hatched circle). Incomplete coverage is achieved up to separations of 25′′ (arrow).

Open with DEXTER

The targets were observed with NACO, the adaptive optics imager at the Nasmyth platform of the ESO VLT UT4 (Yepun)5. The stars themselves were used as adaptive optics reference stars for the visual wavefront sensor of NAOS (VIS). First epoch images were taken in 2003/2004 (programmes 072.C-0485, 073.C-0225) and second epoch imaging followed in 2005/2006 (programmes 076.C-777, 077.C-0268) for those stars with identified faint companion candidates.

The semi-transparent coronagraph is a rarely used NACO mode. It has a diameter of and dims the incoming stellar light by about 6 mag in the K band (see Fig. 1). Hence, it blocks most of the light of the star but still allows precise astrometry with the stellar point spread function (PSF).

The observations were obtained in the Ks band (Persson et al. 1998; Tokunaga et al. 2002) further reducing the brightness difference of stars and any low-mass companions. The sensitivity of the observations has been further improved by not using the S13 camera with the smallest pixel scale but instead observing with the S27 camera (27.15 mas/pix, FOV 28′′ × 28′′). This means that a larger field is covered (Fig. 1) and more light is collected in a single pixel. The astrometric precision using the S27 camera is sufficient since the average proper motion of the UMa group of 150 mas per year (Fig. 1) corresponds to 5.5 pixels on the detector.

The data reduction follows the usual steps of sky subtraction, flat-field correction, bad pixel correction, and shift and add6. The sky was subtracted using jittered exposures of nearby sky positions free of bright stars. The flat-field correction is based on coronagraphic night-time flat-field exposures to correct for the variable transmittance of the coronagraphic substrate and standard day-time exposures to correct for pixel-to-pixel variations of the detector below the coronagraph used. Bad pixel frames were taken from standard calibrations or derived individually with ESO tools and standard twilight flat-field exposures.

3. Identification and characterization of candidates

The stellar PSF was subtracted in a box of 500 × 500 pixels centred on the star to facilitate the visual identification of faint companion candidates. The PSF was derived by calculating the average of a set of rotated frames (in steps of 2°) and using sigma-clipping to get rid of bright features, in particular the diffraction spikes. A smoothed image was subtracted to further facilitate detection (Figs. C.1C.5)7.

We noticed that artefacts due to dust particles on the coronagraphic substrate can be mistaken for companion candidates. The flat-field correction fails to remove them since occasional displacements of the coronagraphic substrate can occur between calibration and target exposures. We removed these artefacts from the list of companion candidates guided by a visual cross-match of the scientific exposures with the flat-field exposures. In the most extreme cases (exposures of HD 11171 and HD 22049), several tens of artefacts had to be removed.

A wide comoving companion will not change its position relative to the star, while a distant and thus non-moving background object will reflect the stellar parallactic and proper motion (Fig. 2). Our assessment of astrometric measurement uncertainties is based on Chauvin et al. (2010) who give a long-term average of the S27 pixel scale of 27.012 ± 0.004 mas/pix and of the detector position angle (true north) of . In the present work, we corrected the position angles of all candidates for the mean deviation of from the true north. From the measurements obtained by Chauvin et al. (2010), we expect uncertainties of 0.016 mas/pix on separation and on position angle. We neglected the errors in fitting Gaussian profiles to measure the positions of the stars and the candidates. It is as precise as a few milliarcseconds and of the order of the Hipparcos errors (Table E.2).

thumbnail Fig. 2

Left: based on Hipparcos data (cf. Tables E.2, A.1, and A.2), this example of HD 22049 shows how the relative motion of non-moving background stars in right ascension (solid lines) and declination (dashed lines) will differ from a comoving companion (respective horizontal lines). The lines encompass the Hipparcos uncertainty, scaled by a factor of 100 for visibility. Possible orbital motion of a hypothetical companion has been neglected. The vertical dotted lines highlight the dates of the 1st and 2nd epoch exposures, respectively. The two candidates found are background objects (circles for right ascension and squares for declination; see Table D.1 and Fig. C.1). Right: dynamic range curves for a 10σ detection as a function of angular separation for all exposures of the survey.

Open with DEXTER

The instrumental magnitudes of the target stars and the candidates were all measured in apertures of 13 pixels across which is more than double the typical FWHM of the PSF. Then, the Ks band magnitude of the candidates were derived from the measured flux ratios taking into account the brightness of the target star (Table A.2) and the attenuation of the coronagraph. While the brightness of the star was measured in the reduced frames, the brightness of the candidates was measured in the PSF-subtracted frames constructed in the way described above.

To calibrate the transmission of the coronagraph, a standard star (HD 1274) was taken during the programme 384.C-0245A, once outside and once below the coronagraph, and was reduced in the same way as the exposures of the science targets. The flux values were integrated within apertures of 19 pixels radius which fits well inside the coronagraph. Relating the measurements inside and outside the coronagraph gives a transmission of 0.47 ± 0.03% in the Ks band which translates to a dimming of 5.83 ± 0.06 mag. The error bar accounts for uncertainties involved in the aperture photometry but not for other systematics like cross-talk, light-leaks, temporal variation of the PSF, improper/variable placement of the star below the coronagraph, or variable residual absorption by the transparent substrate which carries the coronagraphs. The new measurement gives a dimming that is weaker by half a magnitude than the value given in the NACO manual (Ks = 6.3 ± 0.1 mag), a measurement that was obtained in a similar way to the present work but using frames reduced via a different process and using peak counts instead of aperture photometry (ESO, priv. comm.). A conservative estimate of 6.1 ± 0.3 is adopted in the present work since the discrepancy cannot be explained with the data available.

For each of the companion candidates, the average magnitude of two epochs is adopted and the uncertainty is given by half their difference plus the square-added error bar of 0.3 mag of the transmissivity of the coronagraph. For single-epoch images, the transmissivity is the only well-known dominant source of error.

4. Assessment of field coverage and detection limits

We could not take advantage of the full field of view of the NACO mode applied (28′′ × 28′′) since the corners are affected by obstructions. Furthermore, the coronagraphs themselves are insensitive patches in the field of view.

The field of view is covered completely between angular separations of (coronagraphic radius) and which correspond to different linear scales depending on the distance of the star. The parallaxes of the sample stars vary from 37 mas (HIP 104383) to 311 mas (HD 22049) so that the field of view covers very different parts of the stellar environments, e.g. the closest view in the case of ϵ Eri (=HD 22049; 1.1−29 au) and the farthest view for HIP 104383 (9.2−237 AU) (Table B.1). The expected widest bound orbit of low-mass companions has been estimated based on the linear law given in Close et al. (2003) given the central mass of the target systems.

Although the field of view is complete only up to separations of ~9′′, it samples separations of up to ~25′′ corresponding to as much as ~650 au in the case of HD 125451 and HIP 104383. Still the field of view never reaches the widest possible bound orbit in the present sample. We note, however, that inclination is a free parameter in this whole consideration and that measured separations are projected separations. Stars with exoplanets are known where very wide companions (up to one third of a parsec) with common proper motion have been detected (Mugrauer et al. 2014).

To understand the detection limits in the wings of the PSF, we built noise maps from the reduced and PSF-subtracted images. We measured the standard deviation at each pixel using 7 × 7 adjacent pixels. This value has been multiplied by the square root of the number of pixels in an aperture of 13 pixels across in order to compare to the photometric flux measurements described above.

Based on the noise analysis, the limiting magnitudes were derived relative to the magnitude of the primary star in the same way as by Brandeker et al. (2006). We found that a source can be detected at a given location if the flux measured is larger than a detection limit of 5−10σ of the local noise level and so we adopted the conservative and common value of 10σ. The limits derived were corrected for the coronagraphic attenuation of 6.1 mag inferred above. The right panel of Fig. 2 shows the average value of pixels at the same angular distance for each exposure.

Uncertain age is the largest contribution to the error budget when determining the mass of a companion that is still detectable. Using the stellar Ks band magnitude and the measured dynamic range, we calculated the Ks band magnitude of faint objects which could still have been detected at a given separation. The magnitudes were interpolated in evolutionary models to obtain a mass estimate. We used the COND03 models (Baraffe et al. 2003) for effective temperatures lower than 1300 K8 and DUSTY00 (Chabrier et al. 2000) models for temperatures higher than 1300 K. For ages of 100 Myr, 500 Myr, and 1 Gyr, this temperature corresponds to a mass of 10, 31, and 42 MJup, respectively, and is accessible to the present survey. In this range of mass, the difference in K band magnitude between 100 Myr and 1 Gyr varies from 4 at the high-mass end to 8 at the low-mass end. Even if we avoid interpolating the true age range of the UMa group in the evolutionary models, we can assess that the age uncertainty of the UMa group implies an error of several K band magnitudes which is the dominant source of error in the present study.

5. Results and discussion

The noise level is observed to decrease strongly with increasing separation from the primary star (Fig. 2). The longest on-target exposure time was spent on the brightest target, HD 22049, resulting in the highest dynamic range of the present work (17 mag for a 10σ detection). Typically, the dynamic range for a 10σ detection limit ranges between ΔKs = 13 and 15 at a separation beyond 3′′ (Figs. 2, E.1E.3). Given the K band magnitude of the central star, this corresponds to about 12−20 MJup if an age of 500 Myr is assumed (Figs. E.4 and E.5). Objects with 12 MJup and younger than 1 Gyr could have been detected at separations of less than 10 au in the case of HIP 57548 which has the deepest exposure in the present work, and objects with 20 MJup closer than 3 au (Table E.1).

More than 200 candidates have been identified (Figs. C.1C.5, Table D.1). In some fields, not a single object has been detected around the star (HD 11171, HD 26923, HD 38393, HD 63433, HIP 57548, HD 95650, HD 125451, HD 139006, and HD 217813): however, some 160 have been found around HD 165185 which is located in the galactic plane. A number of additional candidates have been found close to HD 11131, HD 22049, HD 26913, HD 41593, HD 60491, HD 61606 A, HD 135599, HD 147584, HD 175742, and HIP 104383 A9.

thumbnail Fig. 3

Left: Ks magnitude distribution of candidates. Based on Table D.1. Errors in magnitude are not considered. Right: residual motion of candidates. For all candidates detected in two epochs, the figure shows the residual shifts in right ascension and declination over the time elapsed between the epochs (Table D.1) after correcting for the proper and parallactic motion of the central stars (Table E.2). The symbols group the candidates according to their central stars. Error bars have been omitted for clarity.

Open with DEXTER

The relative shift of a companion with respect to the star was measured when a second epoch was available (Table D.1). These cases are all non-moving background stars when compared to the predicted shift due to the stellar parallactic and proper motion (Table E.2 and Fig. 3)10. The deviation from these expectations is less than 0.̋1 in all cases and of the order of the astrometric error bars (Table D.1). This shows that the true astrometric uncertainties agree with the assessments by Chauvin et al. (2010) implemented here (Sect. 3).

The detection limits can be compared to the brightness of faint background objects detected in the field of view. The uncertainties of the photometric measurements are below 1 mag in almost all cases. Usually, they are close to 0.5 mag and they increase towards the detection limit while the photometric measurements deviate by less than a magnitude from epoch to epoch. Enhanced error bars can be ascribed to smearing, noise, flux missed by the aperture photometry at the edges of the frame, or background features like diffraction spikes, ghosts, and reflections (indicated in Table D.1).

The faintest object identified (HD 165185, ID0101) advocates an empirical detection limit of Ks ≈ 22 for HD 165185. However, the aperture photometry of this object is affected by insufficient background correction so that it must be somewhat brighter (Fig. C.3). Moreover, the lower envelope to the distribution of most candidates (Fig. E.2) is a bit higher as is also indicated by the brightness distribution of all candidates (cf. Fig. 3). It covers magnitudes in the range Ks = 13−23 with a single brighter object next to HD 11131. As can be expected for a distribution of field stars, there are a few bright objects and a larger number of faint objects. The number of objects decreases rapidly at a brightness level fainter than Ks = 20 indicating that the census is complete down to this value. Since the distribution is almost entirely made up of objects next to HD 165185, this completeness limit is certainly valid for this subset.

Table 1

Identification of candidates with previous detections.

We compared our candidates with previous work for targets in common with other surveys which have not necessarily been dedicated to the UMa group (Macintosh et al. 2003; Itoh et al. 2008b; Metchev & Hillenbrand 2009).We recovered all of the previously known candidates (Table 1) at the positions expected as long as they are covered by our field of view. Photometric measurements agree well within the error bars as long as they have been measured in the same or a similar photometric band. The single exception is ID 4 next to HD 175742 (Metchev & Hillenbrand 2009, object 1). Most certainly, the difference of one Ks magnitude can be explained by the location of ID 4 in the PSF of HD 175742 where the photometric measurement is very sensitive to imperfections of the PSF subtraction. The same object was found by Janson et al. (2013) but no photometric measurement is given there. The southern component of HD 165185 listed by Mason et al. (2014) is out of the field of view of the present work but is listed in Table B.1.

thumbnail Fig. 4

Mean survey detection probability for companions of 12, 20, and 35 MJup as a function of separation and for different age of 100 Myr, 500 Myr, and 1 Gyr as indicated in the legend. The solid line gives the limiting case of a detection probability of 1 in each frame.

Open with DEXTER

The deepest exposure was chosen for each target to derive a survey upper limit on the frequency of stars with a companion of given mass and age. This was done in a basic way similar to that used by Lafrenière et al. (2007), without assuming any prior knowledge on the mass-period distribution of companions.

For this purpose, maps of 10σ detection probability were derived from the maps of limiting magnitude presented in Sect. 4 by comparing them to the signal of a 12, 20, and 35MJup companion at different ages of 100 Myr, 500 Myr, and 1 Gyr (using the evolutionary models described in Sect. 4)11. The mean survey detection probability was assessed by centring, rescaling, and averaging these probability maps (Fig. 4). In the limiting case, where the detection probability equals 1 throughout each frame, the average detection probability is only limited by the respective field of view and approaches zero at separations beyond ~500 au, i.e. these separations are not covered by any exposure of the survey. The age range considered has a stronger effect than the mass range. On average, the survey is most sensitive at separations between 100 and 200 au since the distance of the targets and thus the field of view covered is very different.

We derived the upper limit on the frequency of stars with a companion where the mean 10σ detection probability is high. We followed the Bayesian approach described by Lafrenière et al. (2007) using the Poisson approximation and a confidence level of 95%. In the most sensitive range between 100 and 200 au, we can place an upper limit of ~25% on the frequency of UMa group members with low-mass companions more massive than 35 MJup if the age of the UMa group is close to 500 Myr or younger. The limit is never below 15%, defined by the number of stars, the confidence level, and the case of a detection probability of one everywhere (solid line in Fig. 4).

When interpreting the outcome of the present study it is important to be aware of specific formation environments represented by the significant fraction of targets with stellar companions (cf. Tables A.2 and B.1). In addition we note that the known companions of GJ 569 have not been considered for the frequency estimate since they are closer to the star than the field covered by the present work and we did not account for the companions of HD 130948 since they have masses close to the deuterium burning limit of 12−13 MJup which we are not sensitive to on average. Although the results are not at variance with previous work, it is obvious that the study of the UMa group would benefit from a larger sample size. So far, northern UMa group members have not been studied systematically. These actually comprise the largest part of the UMa group. Although its densest part on the northern hemisphere is within 30 pc, the group extends far beyond.


1

See http://www.DwarfArchives.org for a full account.

2

See http://www.exoplanet.eu for further details and updates.

3

It is worth noting that Bannister & Jameson (2007) identified one T dwarf and three L dwarfs in the UMa group using the moving cluster method.

4

The Ursa Major (UMa) group has been studied extensively under various aspects (e.g. Eggen 1994; Montes et al. 2001; King et al. 2003; Fuhrmann 2004; Ammler 2006) but the precise age and the list of members are still a matter of debate. Age estimates range from 200 to 600 Myr (König et al. 2002; King et al. 2003; Fuhrmann 2004; King & Schuler 2005; Brandt & Huang 2015).

5

NACO consists of the adaptive optics system NAOS (Lenzen et al. 2003) and the camera CONICA (Rousset et al. 2003). The instrument was decommissioned in August 2013 and recommissioned on UT1 in January 2015.

6

Using scripts and tools provided by ESO, including ESO-eclipse v5.0.0, the jitter recipe (Devillard 1997, 2011), and esorex using the recipe naco_img_twflat to derive bad pixel frames.

7

Smoothing was performed by the application of a Gaussian filter (FILTER/GAUSS in ESO MIDAS version 13SEP; Banse et al. 1983; Warmels 1992) using all adjacent pixels within a radius of 18 pixels weighted by a Gaussian with a width of σ = 3 pixels.

8

According to Baraffe et al. (2003), the COND03 models are more appropriate below 1300 K for predicting infrared colours, for methane dwarfs and extrasolar giant planets at large orbital separation. The difference between the Ks and the K band has been neglected in the present study.

9

HIP 104383 A is a special case since it appears as a visual binary below the coronagraph (Fig. D.1). According to Balega et al. (2004), it is a binary with a separation of 0.̋3 and a magnitude difference of 1.67 in the R band. In the Ks band we measured a magnitude difference of 0.45. The astrometric measurements presented here were done with respect to the brighter component while the co-added signal was used for photometric measurements.

10

The parallactic motion was calculated assuming a value of the obliquity of the ecliptic plane of and expressing the solar longitude by with the time T given by Julian centuries since 1900, January 0, 12 h (Kovalevsky 1995). The eccentricity of Earth’s orbit was neglected.

11

The probability has been computed that a signal exceeds the 10σ detection threshold in the presence of noise, assuming a normal distribution.

Acknowledgments

M.A. thanks Eike W. Guenther for fruitful discussions. We thank the anonymous referee for the constructive comments. A.B. acknowledges support from DFG in grants NE 515/13-1 and 13-2. M.A. was supported by a graduate scholarship of the Cusanuswerk, one of the national student elite programs of Germany, and an individual fellowship granted by the Fundação para a Ciência e a Tecnologia (FCT), Portugal (reference SFRH/BPD/26817/2006). M.A. acknowledges research funding granted by the Deutsche Forschungsgemeinschaft (DFG) under the project RE 1664/4-1. M.A. further acknowledges support by DLR under the projects 50OW0204 and 50OO1501. R.N. acknowledges general support from the German National Science Foundation (Deutsche Forschungsgemeinschaft, DFG) in grants NE 515/13-1, 13-2, and 23-1. T.O.B.S. would like to thank Evangelisches Studienwerk e. V. Villigst, the state of Thuringia as well as DFG for support in program NE 515/30-1. R.N. and R.E. would like to thank DFG for support in the Priority Programme SPP 1385 on the First ten Million years of the Solar System in project NE 515/34-1 and 34-2. Use was made of the CDS services SIMBAD, Vizier, and the NASA/ADS abstract service. This research has made use of the Washington Double Star Catalog maintained at the US Naval Observatory. This publication makes use of data products from the Two Micron All Sky Survey, which is a joint project of the University of Massachusetts and the Infrared Processing and Analysis Center/California Institute of Technology, funded by the National Aeronautics and Space Administration and the National Science Foundation.

References

Appendix A: Sample

Table A.1

The sample of UMa group members.

Table A.2

Basic stellar data of the sample of UMa group members.

Appendix B: Field of view and nearby visual companions

Table B.1

Characterization of the NACO field of view according to Fig. 1.

Appendix C: Images with candidates

thumbnail Fig. C.1

Candidates for stars with images in two epochs: HD 22049, HD 26913, and HD 41593. The images display the central star under the coronagraph. Faint companion candidates in the field of view are indicated by circles and are enumerated (cf. Table D.1). The scale is indicated in the image. North is at the top and East is to the left. The PSF of the central star has been removed in a rectangular area centred on the star and a smoothed image has been subtracted. The residual speckle pattern and the diffraction spikes of the mount of the secondary mirror remain visible (several of those due to the addition of several exposures). In addition, there are reflections, ghosts, a vignetted region to the lower left, and the shadows of the other three coronagraphs.

Open with DEXTER

thumbnail Fig. C.2

Candidates for stars with images in two epochs: HD 60491, HD 61606, and HD 135599. The layout is the same as in Fig. C.1.

Open with DEXTER

thumbnail Fig. C.3

Candidates for stars with images in two epochs: HD 147584, HD 165185, and HD 175742. The layout is the same as in Fig. C.1. Stellar residuals are present in the second epochs since jittering of the sky exposures did not work in those cases.

Open with DEXTER

thumbnail Fig. C.4

Candidates found in single-epoch observations (HD 11131 and HIP 104383) and observations of stars without any candidates (HD 11171, HD 26923, HD 38393). The layout is the same as in Fig. C.1. HIP 104383 A is a double star below the coronagraph (see Fig. D.1 and footnote to Sect. 5).

Open with DEXTER

thumbnail Fig. C.5

Observations of stars without any candidates: HD 63433, HD 95650, HIP 57548, HD 125451, HD 139006, and HD 217813. The layout is the same as in Fig. C.1.

Open with DEXTER

Appendix D: List of candidates

Table D.1

Astrometry and photometry of candidates for each target and each epoch of observation.

thumbnail Fig. D.1

The visual binary HIP 104383 A below the semi-transparent coronagraphic mask. The binary has already been characterised (Balega et al. 2004). The full frame exposing companion candidates is shown in Fig. C.4. The magnitudes of the binary components differ by 0.45 mag in the Ks band.

Open with DEXTER

Appendix E: Detection limits

thumbnail Fig. E.1

Dynamic range curves are shown for stars with candidates and observations in two epochs: HD 22049, HD 26913, HD 41593, HD 60491, HD 61606, and HD 135599. The lines show the 10σ detection limits for the 1st and 2nd epoch as is indicated in the legend. The adoption of a 5σ limit to describe the visual detection of candidates would lower the curves by 2.5 log  2 = 0.75 mag. Furthermore, the curves can vary by 0.3 mag due to the uncertainty of the transmission of the coronagraph. Filled circles indicate candidates found in the 1st epoch, open circles those in the 2nd epoch, and filled squares those seen in both epochs.

Open with DEXTER

thumbnail Fig. E.2

Similar to Fig. E.1 for HD 147584, HD 165185, and HD 175742. Some data points are below the detection limits owing to unreliable magnitude measurements (discussed in the text).

Open with DEXTER

thumbnail Fig. E.3

Dynamic range curves are shown for stars with candidates detected in a single epoch (HD 11131 and HIP 104383) and for all the other stars without any candidates.

Open with DEXTER

thumbnail Fig. E.4

Top row: detection limits on mass as a function of angular separation. Bottom row: the same as a function of separation in linear scale. Based on the dynamic range curves presented in Figs. 2 and E.1E.3 and on evolutionary models (DUSTY00 and COND03) assuming an age of 500 Myr. The layout follows Figs. E.3b–d.

Open with DEXTER

thumbnail Fig. E.5

Same as in Fig. E.4 for the remaining objects.

Open with DEXTER

Table E.1

Detection limits derived from exposures of different epochs.

Table E.2

Apparent motion in right ascension and declination predicted for all stars with at least one candidate observed in two epochs.

All Tables

Table 1

Identification of candidates with previous detections.

Table A.1

The sample of UMa group members.

Table A.2

Basic stellar data of the sample of UMa group members.

Table B.1

Characterization of the NACO field of view according to Fig. 1.

Table D.1

Astrometry and photometry of candidates for each target and each epoch of observation.

Table E.1

Detection limits derived from exposures of different epochs.

Table E.2

Apparent motion in right ascension and declination predicted for all stars with at least one candidate observed in two epochs.

All Figures

thumbnail Fig. 1

Left: distribution of distance and proper motion. The major concentration of the UMa group is the UMa open cluster in the Big Dipper constellation at distances of about 25 pc. All known members and candidates (see references in the text) are included in the histogram (solid line; distances mostly from Montes et al. 2001 and van Leeuwen 2007). We add the distribution of the southern stars within 30 pc addressed in the present work (hatched; distances from van Leeuwen 2007). Right: image of the coronagraphic mask with the four coronagraphs (black filled circles) taken with the S27 camera. Each of the coronagraphs has an angular diameter of . We placed the star below the lower left (south-eastern) coronagraph, so that a field with a radius of 9′′ is completely covered outside the coronagraph (hatched circle). Incomplete coverage is achieved up to separations of 25′′ (arrow).

Open with DEXTER
In the text
thumbnail Fig. 2

Left: based on Hipparcos data (cf. Tables E.2, A.1, and A.2), this example of HD 22049 shows how the relative motion of non-moving background stars in right ascension (solid lines) and declination (dashed lines) will differ from a comoving companion (respective horizontal lines). The lines encompass the Hipparcos uncertainty, scaled by a factor of 100 for visibility. Possible orbital motion of a hypothetical companion has been neglected. The vertical dotted lines highlight the dates of the 1st and 2nd epoch exposures, respectively. The two candidates found are background objects (circles for right ascension and squares for declination; see Table D.1 and Fig. C.1). Right: dynamic range curves for a 10σ detection as a function of angular separation for all exposures of the survey.

Open with DEXTER
In the text
thumbnail Fig. 3

Left: Ks magnitude distribution of candidates. Based on Table D.1. Errors in magnitude are not considered. Right: residual motion of candidates. For all candidates detected in two epochs, the figure shows the residual shifts in right ascension and declination over the time elapsed between the epochs (Table D.1) after correcting for the proper and parallactic motion of the central stars (Table E.2). The symbols group the candidates according to their central stars. Error bars have been omitted for clarity.

Open with DEXTER
In the text
thumbnail Fig. 4

Mean survey detection probability for companions of 12, 20, and 35 MJup as a function of separation and for different age of 100 Myr, 500 Myr, and 1 Gyr as indicated in the legend. The solid line gives the limiting case of a detection probability of 1 in each frame.

Open with DEXTER
In the text
thumbnail Fig. C.1

Candidates for stars with images in two epochs: HD 22049, HD 26913, and HD 41593. The images display the central star under the coronagraph. Faint companion candidates in the field of view are indicated by circles and are enumerated (cf. Table D.1). The scale is indicated in the image. North is at the top and East is to the left. The PSF of the central star has been removed in a rectangular area centred on the star and a smoothed image has been subtracted. The residual speckle pattern and the diffraction spikes of the mount of the secondary mirror remain visible (several of those due to the addition of several exposures). In addition, there are reflections, ghosts, a vignetted region to the lower left, and the shadows of the other three coronagraphs.

Open with DEXTER
In the text
thumbnail Fig. C.2

Candidates for stars with images in two epochs: HD 60491, HD 61606, and HD 135599. The layout is the same as in Fig. C.1.

Open with DEXTER
In the text
thumbnail Fig. C.3

Candidates for stars with images in two epochs: HD 147584, HD 165185, and HD 175742. The layout is the same as in Fig. C.1. Stellar residuals are present in the second epochs since jittering of the sky exposures did not work in those cases.

Open with DEXTER
In the text
thumbnail Fig. C.4

Candidates found in single-epoch observations (HD 11131 and HIP 104383) and observations of stars without any candidates (HD 11171, HD 26923, HD 38393). The layout is the same as in Fig. C.1. HIP 104383 A is a double star below the coronagraph (see Fig. D.1 and footnote to Sect. 5).

Open with DEXTER
In the text
thumbnail Fig. C.5

Observations of stars without any candidates: HD 63433, HD 95650, HIP 57548, HD 125451, HD 139006, and HD 217813. The layout is the same as in Fig. C.1.

Open with DEXTER
In the text
thumbnail Fig. D.1

The visual binary HIP 104383 A below the semi-transparent coronagraphic mask. The binary has already been characterised (Balega et al. 2004). The full frame exposing companion candidates is shown in Fig. C.4. The magnitudes of the binary components differ by 0.45 mag in the Ks band.

Open with DEXTER
In the text
thumbnail Fig. E.1

Dynamic range curves are shown for stars with candidates and observations in two epochs: HD 22049, HD 26913, HD 41593, HD 60491, HD 61606, and HD 135599. The lines show the 10σ detection limits for the 1st and 2nd epoch as is indicated in the legend. The adoption of a 5σ limit to describe the visual detection of candidates would lower the curves by 2.5 log  2 = 0.75 mag. Furthermore, the curves can vary by 0.3 mag due to the uncertainty of the transmission of the coronagraph. Filled circles indicate candidates found in the 1st epoch, open circles those in the 2nd epoch, and filled squares those seen in both epochs.

Open with DEXTER
In the text
thumbnail Fig. E.2

Similar to Fig. E.1 for HD 147584, HD 165185, and HD 175742. Some data points are below the detection limits owing to unreliable magnitude measurements (discussed in the text).

Open with DEXTER
In the text
thumbnail Fig. E.3

Dynamic range curves are shown for stars with candidates detected in a single epoch (HD 11131 and HIP 104383) and for all the other stars without any candidates.

Open with DEXTER
In the text
thumbnail Fig. E.4

Top row: detection limits on mass as a function of angular separation. Bottom row: the same as a function of separation in linear scale. Based on the dynamic range curves presented in Figs. 2 and E.1E.3 and on evolutionary models (DUSTY00 and COND03) assuming an age of 500 Myr. The layout follows Figs. E.3b–d.

Open with DEXTER
In the text
thumbnail Fig. E.5

Same as in Fig. E.4 for the remaining objects.

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.