Resolving subdwarf B stars in binaries by HST imaging

Astronomy & Astrophysics A&A 383, 938–951 (2002) DOI: 10.1051/0004-6361:20020127 c ESO 2002 Resolving subdwarf B stars in binaries by HST imaging⋆,⋆⋆,⋆⋆⋆,† U. Heber1 , S. Moehler1 , R. Napiwotzki1 , P. Thejll2 , and E. M. Green3 1 2 3 Dr. Remeis-Sternwarte, Astronomisches Institut der Universität Erlangen-Nürnberg, Sternwartstr. 7, 96049 Bamberg, Germany Solar-Terrestrial Physics Division, Danish Meteorological Institute, Lyngbyvej 100, 2100 Copenhagen O, Denmark Steward Observatory, University of Arizona, Tucson, AZ 85721, USA Received 4 July 2001 / Accepted 20 December 2001 Abstract. The origin of subluminous B stars is still an unsolved problem in stellar evolution. Single star as well as close binary evolution scenarios have been invoked but until now have met with little success. We have carried out a small survey of spectroscopic binary candidates (19 systems consisting of an sdB star and late type companion) with the Planetary Camera of the WFPC2 onboard Hubble Space Telescope to test these scenarios. Monte Carlo simulations indicate that by imaging the programme stars in the R-band about one third of the sample (6–7 stars) should be resolved at a limiting angular resolution of 0.′′ 1 if they have linear separations like main sequence stars (“single star evolution”). None should be resolvable if all systems were produced by close binary evolution. In addition we expect three triple systems to be present in our sample. Most of these, if not all, should be resolvable. Components were resolved in 6 systems with separations between 0.′′ 2 and 4.′′ 5. However, only in the two systems TON 139 and PG 1718+519 (separations 0.′′ 32 and 0.′′ 24, respectively) do the magnitudes of the resolved components match the expectations from the deconvolution of the spectral energy distribution. These two stars could be physical binaries whereas in the other cases the nearby star may be a chance projection or a third component. Radial velocity measurements indicate that the resolved system TON 139 is a triple system, with the sdB having a close companion that does not contribute detectably to the integrated light of the system. Radial velocity information for the second resolved system, PG 1718+519, is insufficient. Assuming that it is not a triple system, it would be the only resolved system in our sample. Accordingly the success rate would be only 5% which is clearly below the prediction for single star evolution. We conclude that the distribution of separations of sdB binaries deviates strongly from that of normal stars. Our results add further evidence that close binary evolution is fundamental for the evolution of sdB stars. Key words. stars: early-type – stars: binaries: spectroscopic – stars: evolution 1. Introduction Subluminous B (sdB) stars dominate the populations of faint blue stars of our own Galaxy and are found in Send offprint requests to: U. Heber, e-mail: [email protected] ⋆ Based on observations with the NASA/ESA Hubble Space Telescope obtained at the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS 5-26555. ⋆⋆ Based on observations collected at the German-Spanish Astronomical Center (DSAZ), Calar Alto, operated by the Max-Planck-Institut für Astronomie Heidelberg jointly with the Spanish National Commission for Astronomy. ⋆⋆⋆ Based on data obtained at ESO (ESO proposal No. 58.D0478, 65.H-0253(A)). † Some observations reported here were obtained at the MMT Observatory, a joint facility of the University of Arizona and the Smithsonian Institution. both the disk (field sdBs) and globular clusters (Moehler et al. 1997). Observations of elliptical galaxies with the Ultraviolet Imaging Telescope (Brown et al. 1997) and the Hubble Space Telescope (Brown et al. 2000) have shown that these stars are sufficiently common to be the dominant source for the “UV upturn phenomenon” observed in elliptical galaxies and galaxy bulges (see also Greggio & Renzini 1990, 1999). Their space distribution and kinematical properties indicate that the field stars belong to the intermediate to old disk population (de Boer et al. 1997; Altmann & de Boer 2000). However, important questions remain concerning their formation process and the appropriate evolutionary timescales. This is a major drawback for the calibration of the observed ultraviolet upturn in elliptical galaxies as an age indicator. It is now generally accepted that the sdB stars can be identified with models for Extreme Horizontal Branch (EHB) stars burning He in their core, but with a very Article published by EDP Sciences and available at http://www.aanda.org or http://dx.doi.org/10.1051/0004-6361:20020127 U. Heber et al.: Resolving subdwarf B stars in binaries by HST imaging tiny (<2% by mass) inert hydrogen envelope (Heber 1986; Saffer et al. 1994). An EHB star bears great resemblance to a helium main-sequence star of half a solar mass and its further evolution should proceed similarly (i.e. directly to the white dwarf graveyard) as confirmed by evolutionary calculations (Dorman et al. 1993). How stars evolve to the EHB configuration is controversial. The problem is how the mass loss mechanism in the progenitor manages to remove all but a tiny fraction of the hydrogen envelope at precisely the same time as the He core has attained the minimum mass (≈0.5 M⊙ ) required for the He flash. Both non-interacting (scenario i), and interacting (scenarios ii and iii) evolutionary scenarios have been proposed to explain the origin of the sdB stars (see Bailyn et al. 1992). (i) Enhanced mass loss on the red giant branch (RGB) before or during the core helium flash may remove almost the entire hydrogen-rich envelope. This is usually modelled by increasing the η factor in the Reimers (1975) formula to estimate mass loss rates for RGB stars. It has been conjectured that the mass loss rates increase with increasing metallicity, implying that metal rich populations should produce more sdB stars than metal poor ones. Birthrate estimates for sdB stars indicate that only 2% (Heber 1986) or even less (0.25% to 1%, Saffer & Liebert 1995) of the RGB stars need to experience such enhanced mass loss. Evidence that this is possible comes from the existence of RR Lyrae stars of population I which must also have lost half of their mass during evolution. In both cases the physical reason for such strong mass loss is not yet understood. (ii) Mengel et al. (1976) suggest that sdBs could be formed from binaries in which mass transfer starts on the red giant branch and results in a reduction of the hydrogen envelope prior to the helium core flash. Hence all sdBs star are predicted to be found in close binary systems. (iii) An alternative scenario was proposed by Iben (1990), who pointed out that sdBs can be formed from mergers of helium white dwarf binary systems. Iben & Tutukov (1992) estimate that 80% of the sdBs could have been formed by mergers. Hence the frequency of sdBs still being in binaries should be at most 20%. Several dozens of objects with composite spectra consisting of an sdB and a dwarf G-K star have been discovered (e.g. Ferguson et al. 1984; Theissen et al. 1993, 1995; Allard et al. 1994) which implies that the binary frequency of sdBs is 50% or more (Allard et al. 1994). The observed large binary frequency rules out the merger scenario (iii) and we are left with scenarios (i) and (ii), i.e. either the sdB binaries are mostly wide systems that did not interact so that the sdB precursors have evolved independently from the companion (i), or they are close systems formed by interaction of the sdB precursor with the companion star (mass exchange, ii). The high spatial resolution of the Planetary Camera (PC) on board the Hubble Space Telescope (HST) allows to perform a crucial test. As we will show in this paper, 939 it should be possible to resolve a significant fraction of the known composite spectrum systems containing an sdB star if scenario (i) is correct, i.e. if the systems have a distribution of separations like normal main sequence binaries (Duquennoy & Mayor 1991). The interacting scenario (ii), however, predicts that all sdB stars reside in short period (P ≤ 100 d) binaries and consequently none of the systems should be resolvable even with the PC. In order to measure their distribution of separations we have imaged 23 sdB binary candidates with the PC by taking advantage of the snap shot mode of HST observations. 2. Observations and data analysis 2.1. Target selection and optical spectroscopy For the snapshot observations a target list of fifty of the brightest sdB star binary candidates was extracted from an updated version of the Kilkenny et al. (1988) catalogue, supplemented by two stars which we discovered in the course of follow-up spectroscopy of hot stars from the Hamburg-ESO survey (see Edelmann et al. 2001a). 23 stars from this target list were actually observed with the Wide Field Planetary Camera 2 (WFPC2) onboard the HST during our snapshot project, i.e. they were scheduled for observation to fill small gaps in the HST schedule. All stars have published photometry (see Tables B.1 and B.2), but only 16 have published optical spectroscopy. Therefore additional spectra were obtained at the Calar Alto and ESO observatories (see Appendix A for details and plots of the spectra in Figs. A.1 and A.2). As can be seen from Fig. A.1 spectral features (Ca i, Ca ii, Mg i and/or Fe i) indicative of a cool star are clearly present in the spectra of PG 1309−078, PG 0942+461, HE 0430−2457, HE 2213−2212, and PG 2148+095 in addition to the Balmer and helium lines of the sdB. Hence these objects are spectroscopic binaries consisting of an sdB star and a cool companion. PG 0942+461 has already been observed by Mitchell (1998), who, however, did not note the binary nature of the star. We do not find any evidence for a cool companion in the spectra of the sdB stars PG 1558−087 and KPD 2215+5037 (see Fig. A.2). We also re-analysed a published spectrum of PG 2259+134 (Theissen et al. 1993) and do not find any spectroscopic evidence for a cool companion. PG 0105+276 turns out to be not an sdB star but a helium-rich sdO star and does not show any spectroscopic evidence for a cool companion. Therefore our sample consists of 19 composite spectrum objects plus four stars showing only photometric evidence for a companion. One of these four stars (PG 0105+276) also does not belong to the programme sample since it is an sdO star. 2.2. WFPC2 data We observed the candidate binary systems with the PC chip of the WFPC2. If the cool companion is a main sequence star, both components should be of comparable brightness in the R band and we therefore used the 940 U. Heber et al.: Resolving subdwarf B stars in binaries by HST imaging Table 1. Programme stars: coordinates, observation dates, and exposure times for the WFPC2 and references for the spectroscopic classification observations. star PB 6107 PHL 1079 HE 0430−2457 PG 0749+658 PG 0942+461 TON 1281 TON 139 PG 1309−078 PG 1421+345 PG 1449+653 PG 1511+624 PG 1601+145 PG 1636+104 TON 264 PG 1656+213 PG 1718+519 PG 2148+095 HE 2213−2212 BD −7◦ 5977 PG 0105+276 PG 1558−007 KPD 2215+5037 PG 2259+134 α1950 δ1950 l b ◦ ◦ 00h 39m 31s +04◦ 53′ 17′′ 118. 59 −57. 64 ◦ ◦ h m s ◦ ′ ′′ +03 23 00 144. 96 −57. 22 01 35 48 ◦ ◦ h m s ◦ ′ ′′ 04 30 59 −24 57 37 223. 49 −40. 55 ◦ ◦ h m s ◦ ′ ′′ 07 49 39 +65 50 13 150. 44 +30. 99 ◦ ◦ h m s ◦ ′ ′′ 09 42 02 +46 08 38 173. 11 +48. 89 ◦ ◦ h m s ◦ ′ ′′ 10 40 57 +23 24 55 213. 62 +60. 89 ◦ ◦ h m s ◦ ′ ′′ 12 53 39 +28 23 31 77. 21 +88. 57 ◦ ◦ h m s ◦ ′ ′′ 13 09 09 −07 49 18 311. 60 +54. 44 ◦ ◦ h m s ◦ ′ ′′ 14 21 29 +34 27 53 58. 36 +69. 01 ◦ ◦ h m s ◦ ′ ′′ 14 49 42 +65 17 58 104. 84 +47. 63 ◦ ◦ h m s ◦ ′ ′′ +62 21 00 99. 21 +47. 96 15 11 25 ◦ ◦ h m s ◦ ′ ′′ 16 01 47 +14 32 58 27. 15 +43. 51 ◦ ◦ h m s ◦ ′ ′′ 16 36 40 +10 24 54 27. 00 +34. 04 ◦ ◦ h m s ◦ ′ ′′ . 16 47 05 +25 15 13 45 16 +37. 12 ◦ ◦ h m s ◦ ′ ′′ . +21 15 05 41 25 +33. 90 16 56 12 ◦ ◦ h m s ◦ ′ ′′ . 17 18 35 +51 55 05 79 00 +34. 94 ◦ ◦ h m s ◦ ′ ′′ . 21 48 41 +09 30 39 66 78 −32. 84 ◦ ◦ h m s ◦ ′ ′′ . 22 13 38 −22 12 26 32 63 −54. 50 ◦ ◦ h m s ◦ ′ ′′ . 23 15 12 −06 44 56 71 55 −59. 65 stars without spectroscopic evidence ◦ ◦ h m s 01 05 32 +27◦ 36′ 53′′ 127. 46 −34. 84 ◦ ◦ 15h 58m 39s −00◦ 43′ 26′′ 9. 34 +36. 51 ◦ ◦ 22h 15m 25s +50◦ 37′ 48′′ 99. 71 −4. 91 ◦ ◦ 22h 59m 16s +13◦ 22′ 31′′ 86. 36 −41. 31 F 675W filter of the WFPC2. We obtained four observations of each target, which were offset relative to the first one by (−11,−5.5), (−16.5,−16.5), (−5.5,−11) pixels. We first rebinned the data linearly to a step size of 0.5 pixels and then aligned them according to the offset pattern mentioned above. We then determined the median value of the four aligned images to avoid cosmic ray hits and hot pixels and used these median-averaged images for visual inspection. All flux measurements are performed on manually cleaned average images to ensure proper flux conservation. The median-averaged images were first inspected by eye to see if any companion could be detected. Only 6 stars (cf. Fig. 1) showed obvious nearby stars and the angular separations and brightness differences can be found in Table 2. The brightness differences were determined using the command INTEGRATE/APERTURE from MIDAS, which performs an aperture photometry with a given radius. Aperture photometry is difficult for TON 1281, TON 139, and PG 1718+519, due to the small distance of the components. The sky background was determined in an empty region using the same aperture as for the stars. To get a more quantitative estimate of possible companions we fitted two-dimensional Gaussians with variable angle of the major axis to all shifted and co-added target images and compared the results to fits obtained for obs. date exp. time [s] reference 990627 3.5 Moehler et al. (1990) 981204 4 Theissen et al. (1995) 980417 8 this paper 990329 1.8 Saffer (1991) 980530 10 Heber et al. (1991) 990623 5 Jeffery & Pollacco (1998) 980103 1.8 Green (1997) 980505 8 Ferguson et al. (1984) 990605 14 Ferguson et al. (1984) 990619 7 Moehler et al. (1990) 990513 14 Moehler et al. (1990) 000613 12 Ferguson et al. (1984) 000612 8 Ferguson et al. (1984) 990529 10 Theissen et al. (1993) 980301 12 Ferguson et al. (1984) 990427 7 Theissen et al. (1995) 990411 4 this paper 981207 8 this paper 981125 0.3 Viton et al. (1991) for a cool companion 980226 990424 961213 000615 14 7 7 10 this paper, new type: He-sdO this paper this paper Theissen et al. (1993), this paper Table 2. Separation and estimated brightness differences for the components of the 6 resolved binaries. The photometric data available for HE 0430−2457 do not allow to estimate a temperature or distance of the sdB. system PG 0105+276 HE 0430−2457 TON 1281 TON 139 PG 1558−007 PG 1718+519 separation angular linear [AU] 3.′′ 37 4.′′ 48 1.′′ 25 0.′′ 22 0.′′ 32 2.′′ 80 0.′′ 24 3700 4900 250 300 2500 230 brightness difference ∆F 675W m 0.9 m 1.6 m 2 .1 m 3.7 m 0.8 m 3.1 m 0.8 archive point-spread functions (PSFs; F 675W filter, PC chip). The archive PSFs define a good correlation between the length of the two axes, which is shared by most target PSFs (see Fig. 2). Besides the resolved binaries (where stray light can affect the determination of the axis ratio) four stars deviate from the main correlation between major and minor axis (see Fig. 3): PG 2148+095 (2.03/1.26), KPD 2215+5037 (2.38/1.61), TON 264 (2.35/1.83), and PG 0749+658 (2.36/1.87). U. Heber et al.: Resolving subdwarf B stars in binaries by HST imaging 941 Fig. 1. The images of the resolved binaries. The bar in each image corresponds to 1′′ . We used DAOPHOT (Stetson 1987) to obtain an average PSF from those target stars that share the axis-relation of the archive PSFs. This “target PSF” was then used to deconvolve all systems that are either resolved by eye or show deviations from the axis-relation defined by the archive PSFs. No additional components were resolved in this process, but we could verify the brightness differences between the components of the resolved systems listed in Table 2, which were reproduced by DAOPHOT also for small separations. For 13 of our target stars a homogeneous set of groundbased RC measurements exist (Allard et al. 1994, see Table B.2). Comparing those data to the instrumental F 675W magnitudes integrated within an aperture of 0.′′ 5 radius F 675W = −2.5 log flux0′′. 5 − sky0′′. 5 exposure time we find that most of the stars lie on a line with slope 1 (except KPD 2215+5037 and PG 1601+145, see Fig. 4). From the 11 stars on the line we determine a zeropoint m m of 21 . 21 ± 0 . 02. From the WFPC2 data handbook we m determine a zeropoint of 21 . 9 (gain 14, including an m . aperture correction of −0 1) that has to be corrected to m Cousins R by adding −0 . 65 (assuming a spectral type of A5 for the combined spectra of our binary stars), yielding a final zeropoint of 21m. 25, in agreement with our empirically determined zeropoint. Since our empirically derived Fig. 2. The major and minor axes of the point spread functions for the target stars (circles, filled symbols mark brightest star of resolved binaries) and of archive point-spread functions (triangles, filled ones mark stars with positions on the PC chip close to our targets). zeropoint automatically takes into account the unusual flux distribution of our binary stars we decided to use it to calculate RHST given in Table B.2. 942 U. Heber et al.: Resolving subdwarf B stars in binaries by HST imaging Fig. 3. The images of the unresolved stars (PG 2148+095, KPD 2215+5037, TON 264, PG 0749+658) which show deviations from the standard PSF shape (see text). The images of PB 6107 and PG 1421+345 are well matched by the standard PSF shape and are displayed for comparison. Note that – in contrast to all other stars displayed here – there is no spectroscopic evidence for binarity of KPD 2215+5037. The bar in each image corresponds to 1′′ . 3. Spectral energy distribution To obtain an upper limit to our resolution we tried to estimate the R brightness of the cool companion by fitting the available photometric data of those stars that have sufficient measurements. In order to disentangle the flux of the hot star from that of the cool star we analyse the composite spectral energy distribution. For this purpose ultraviolet, optical and infrared (spectro-) photometry is collected from literature and archives (IUE, 2MASS). To determine the contribution of the hot star we fit synthetic spectra (Kurucz 1992) to the bluest part of the observed spectral range, i.e. IUE data plus u or u/U plus v/B (if no UV data were available) and determine the effective temperature of the sdB star. In doing so we assume that the companion does not contribute to the flux in this wavelength range (cf. Fig. 5). While this is probably true for the IUE data, some contamination may be present in the u/U - and v/B-band and consequently the temperature determination for the sdB star can be compromised. However, for some stars photometric data are so incomplete that no meaningful fit can be obtained. Aside from the F 675W measurements discussed here PG 0942+461 and HE 2213−2212 have only JHK photometry from 2MASS, which are insufficient for a fit. While HE 0430−2457 has BV R photometry it is still not possible to constrain the sdB star’s temperature with these data as B − V is insensitive to Teff at sdB temperatures. To convert the magnitudes into flux values we used the data given in Table 3. By comparing the measured flux in the R band to the model flux of the sdB star we derive the flux ratio of the hot vs. the cool star in the system. For those systems which should have a rather bright companion according to their photometric data we verified the flux ratio in R between sdB and cool companion from two colour diagrams similar to those used by Ferguson et al. (1984), which is best suited for components of comparable brightness (for details see Ferguson et al. 1984). With this method we found that the companion of TON 1281 is bright enough to affect also the u filter, yielding a temperature of 25 000 K to 27 000 K for the sdB instead of the 22 000 K given in m m Table 4 and a brightness difference ∆R of 0 . 2 to −0 . 1. Also for PG 1601+345 we find a much smaller brightness m difference (0 . 1) and higher temperature (29 500 K) from this method than from our photometric fits. In this case the B filter is already affected by the cool companion. For reasons of consistency we keep the values from the photometric fits for these two stars in Table 4. For all other stars m with brightness differences ≤0 . 8 the results from both U. Heber et al.: Resolving subdwarf B stars in binaries by HST imaging Fig. 4. The instrumental F 675W magnitudes compared to the RC data from Allard et al. (1994). The open symbols are KPD 2215+5037 and PG 1601+145. The line marks the reflux ′′ −sky ′′ 0. 5 0. 5 +21.21. lation RC = −2.5 log exposure time Table 3. Flux for a star with mλ = 0. The data are taken from Lamla (1982, p. 59, uvby; p. 82 BV RC IC ), Zombeck (1990, JHKUT98 ) and from the 2MASS Team (priv. comm., JHK2MASS ). filter flux [erg/(cm2 s Å)] u v b y U B V RC IC J2MASS H2MASS K2MASS JUT98 HUT98 KUT98 1.169 × 10−8 8.444 × 10−9 5.826 × 10−9 3.700 × 10−9 4.187 × 10−9 6.597 × 10−9 3.607 × 10−9 2.254 × 10−9 1.196 × 10−9 2.91 × 10−10 1.11 × 10−10 3.83 × 10−11 3.18 × 10−10 1.18 × 10−11 4.17 × 10−11 λc [Å] 3500 4110 4670 5470 3600 4400 5500 6400 7900 12 510 16 280 22 030 12 500 16 500 22 000 methods were the same. To correct for interstellar reddening we used the reddening-to-infinity maps of Schlegel et al. (1998) which give somewhat higher values than the older data of Burstein & Heiles (1982). KPD 2215+5037, PG 1558−007, and PG 2259+134 all lie in regions of quite high reddening according to Schlegel et al. (1998) and show no spectroscopic evidence for a cool companion (see Appendix A). The observed apparent infrared excess can be explained by high interstellar reddening alone, without invoking the presence of a cool companion. We also find no evidence for a companion from available photometry 943 of PG 1656+213, although there is spectroscopic evidence (Ferguson et al. 1984). However there are are no flux measurements redwards of V available and B and V fluxes are inconsistent. Therefore we keep PG 1656+213 as a programme star. Aznar Cuadrado & Jeffery (2001) present an extensive discussion of sdB parameters derived from energy distributions, which also includes some of the stars discussed in this paper. In Table 5 we present the temperatures given in their paper and other values collected from literature in comparison to the ones derived here. As can be seen from Table 5 differences of ±10% in Teff between different authors are quite common. The temperatures derived from the photometric data and from line profile fits for the stars in regions with high reddening agree moderately well (compare Tables 4 and A.1). The discrepancies may be due to small scale variations in reddening that affect the temperatures derived from photometry but not those derived from line profile fits. From the photometric fit we can derive the apparent R magnitudes of the sdB and of the cool star and correct both for interstellar extinction. The uncertainty in Teff of about ±10% evident from Table 5 causes an estimated uncertainty in the derived brightness for both components of m ±0 . 2. Knowing the absolute R magnitude of the sdB stars then allows to determine their distance. We use the mean MV derived by Moehler et al. (1997) for hot subdwarfs in the globular cluster NGC 6752. They found two groups of hot subdwarfs, a cooler one with a mean effective temm perature of 22 000 K and < MV > = 3 . 2 (5 stars), and a m hotter one with <Teff > = 29 000 K and < MV > = 4 . 2 (12 stars). From Kurucz (1992) model atmospheres for m [M/H] = 0 we find V − R = −0 . 120 for Teff = 22 000 K m m and −0 . 152 for 29 000 K. We therefore use MR = 3 . 3 for m stars cooler than 25 000 K and MR = 4 . 4 for hotter stars. Using the archive point spread functions we estimated the minimum separation that we can resolve for a given brightness difference by adding two PSFs with a defined brightness difference and angular separation and examining the resulting image by eye. We find the following resm m olution limits: ∆αlim (∆R) = 0.′′ 2 (2 . 0), 0.′′ 1 (1 . 5), 0.′′ 07 m m ′′ (1 . 0), 0. 05 (0 . 5). Using the distances determined above we can now derive upper limits for the linear separation of the unresolved binaries (cf. Table 4), ranging from 50 AU to 210 AU. Table 2 shows that the brightness differences between the components in TON 1281 and HE 0430−2457 are too large to reproduce the spectral energy distribution of TON 1281 and the photometry of HE 0430−2457, respecm tively. The large brightness difference of 3 . 1 (from the WFPC2 data) for PG 1558−007 agrees with the lack of photometric and spectroscopic evidence for a companion. In the remaining two cases (PG 1718+519, TON 139) the brightness differences in Table 2 are somewhat larger than those derived from the spectral energy distribution. To see whether we can in principle accommodate the HST observations by fits to the photometric data we repeated the 944 U. Heber et al.: Resolving subdwarf B stars in binaries by HST imaging Table 4. Estimated temperature of sdB stars, resulting reddening-free brightness of subdwarf B star (RsdB,0 ) and companion (Rcomp,0 ), distance d, brightness difference ∆R, and upper limit for linear separation alim derived from upper limit of angular separation ∆αlim . The reddening estimates are from the maps of Schlegel et al. (1998) and we used AR = 2.6·EB−V . The three different temperatures for PG 1511+624 result from the three available SWP spectra. If no evidence for a companion can be found from available photometry no entry is given in Col. 4. Star PB 6107 PG 0105+276 PHL 1079 PG 0749+658 TON 1281 TON 139 PG 1309−078 PG 1421+345 PG 1449+653 PG 1511+624 PG 1558−007 PG 1601+145 PG 1636+104 PG 1656+213 TON 264 PG 1718+519 PG 2148+095 KPD 2215+5037 PG 2259+134 BD −7◦ 5977 Teff,sdB [K] AR Rcomp,0 23 000 32 000 25 000 22 000 22 000 20 000 24 000 24 000 28 000 31 000 28 000 33 000 23 000 25 000 20 000 17 000 26 000 27 000 26 000 35 000 30 000 29 000 0 . 086 m 0 . 156 m 0 . 104 m 0 . 125 m 0 . 065 m 0 . 026 m 0 . 138 m 0 . 044 m 0 . 042 m 0 . 047 14 . 4 m 15 . 8 m 14 . 9 m 14 . 4 m 14 . 4 m 13 . 6 m 15 . 5 m 16 . 0 m 14 . 7 m 15 . 7 m 15 . 8 m 15 . 6 m RsdB,0 m m 0 . 468 m 0 . 133 m 0 . 156 m 0 . 172 m 0 . 146 m 0 . 081 m 0 . 169 m 0 . 871 m 0 . 341 m 0 . 093 m 15 . 2 m 14 . 5 m 16 . 0 m 14 . 1 m 14 . 5 m 10 . 2 MR,sdB m 13 . 0 m 14 . 4 m 13 . 4 m 12 . 1 m 13 . 6 m 13 . 2 m 14 . 2 m 14 . 9 m 14 . 0 m 14 . 8 m 14 . 8 m 14 . 9 m 13 . 1 m 14 . 6 m 13 . 7 m 14 . 6 m 14 . 1 m 14 . 3 m 13 . 0 m 12 . 8 m 14 . 4 m 11 . 9 m 3.3 m 4.4 m 4.4 m 3.3 m 3.3 m 3.3 m 3.3 m 3.3 m 4.4 m 4.4 m 4.4 m 4.4 m 3.3 m 4.4 m 3.3 m 3.3 m 4.4 m 4.4 m 4.4 m 4.4 m 4.4 m 4.4 d [pc] ∆R 870 1100 630 580 1150 950 910 2100 830 1200 1200 1260 910 1100 1200 1800 870 950 520 480 1000 320 1.4 m 1.4 m 1.5 m 2.3 m 0.8 m 0.4 m 1.3 0.′′ 1 m 0.7 m 0.9 m 1.0 m 0.7 0.′′ 1 0.′′ 1 0.′′ 1 0.′′ 2 0.′′ 07 0.′′ 05 0.′′ 1 210 0.′′ 07 0.′′ 07 0.′′ 07 0.′′ 07 87 110 63 116 80 48 91 0.6 m 0.8 m 0.′′ 07 0.′′ 07 77 84 m 0.′′ 2 0.′′ 05 0.′′ 1 174 48 52 m 0.′′ 2 64 m 1.9 m −0 . 2 m 1.5 −1 . 7 ∆αlim alim [AU] 58 84 84 88 Table 5. Effective temperatures for sdB stars derived from energy distributions by various authors. The sources are Aznar Cuadrado & Jeffery (2001, ACJ01), Allard et al. (1994, A94), Theissen et al. (1993, T93; 1995, T95), Ulla & Thejll (1998, UT98). star this paper PB 6107 PG 0105+276 PHL 1079 PG 0749+658 TON 1281 TON 139 PG 1449+653 PG 1511+624 PG 1636+104 TON 264 PG 1718+519 PG 2148+095 KPD 2215+5037 PG 2259+134 23 000 32 000 25 000 22 000 22 000 20 000 28 000 31 000: 20 000 26 000 27 000 26 000 35 000 30 000 Teff [K] derived by ACJ01 T93 A94 T95 UT98 30 000 30 000 25 000 32 000 35 850 26 350 25 050 23 275 23 500 29 500 28 150 28 000 33 000 21 000 28 500 25 000 26 000 24 500 18 000 29 950 22 950 23 500 28 300 28 500 fits, this time enforcing the brightness difference in the R band obtained from the HST data. The results are shown in Fig. 5 (in comparison to the original fits). Obviously the companion of PG 1718+519 is sufficiently bright to affect also the u filter, thereby rendering our assumption that 30 000 25 000 22 500 this filter is unaffected by the cool companion obsolete. The fits for TON 139 do not show much difference. We conclude that the spectral energy distribution of TON 139 and PG 1718+519 are consistent with the R band flux ratio measured with the HST WFPC2 camera. U. Heber et al.: Resolving subdwarf B stars in binaries by HST imaging 945 Fig. 5. Fits of ATLAS9 model spectra (Kurucz 1992, [M/H] = 0) to the photometric data of PG 1718+519 (left panel, including IUE spectra) and TON 139 (right panel). The upper panels show the fits obtained assuming that the bluest photometric data points (IUE spectra and u for PG 1718+519, u and v for TON 139) are not affected by the cool companion. The lower panels show fits that reproduce the brightness differences measured on the WFPC2 images. 3.1. The sdO star PG 0105+276 Since the He-sdO PG 0105+276 does not belong to the programme sample, we discuss it separately. It is the only programme star that is resolved into three components. However, the two companions are quite distant from the primary (3.′′ 37 and 4.′′ 48, respectively). The light of these companions can explain at least qualitatively the IR excess observed by ground based aperture photometry. The spectrum of PG 0105+276, however, does not show any signature of a cool companion, probably because due to the orientation and the small width of the slit no light of the distant companions was included. The diaphragm used in the photometry was large (18′′ ) and included the companions’ light. The brightness differences measured on the WFPC2 m m image (0 . 9, 1 . 6) for PG 0105+276 are smaller than the m one derived from the photometric fit (1 . 4), i.e. one companion is brighter than expected. However, as discussed in Appendix A, the true temperature (from line profile fitting) is much higher than the one obtained from the spectral energy distribution (63 000 K vs. 35 000 K) making the companion’s luminosity obtained from photometry a lower limit only. 4. Simulation of separability in binary systems In order to interpret our results with respect to the different evolutionary scenarios we simulate binary systems containing main sequence (MS) companions and sdBs with period distributions found for normal main sequence binaries (Duquennoy & Mayor 1991). Assuming that the sdB mass is 0.5 M⊙ and the MS companion mass is 1 M⊙ we convert the period distribution published by Duquennoy & Mayor (1991) to physical separations using Kepler’s Harmonic law. The orientation of the axis of the system is then chosen to be random in space and the projected separation, or a sin i, is calculated, given the distance to the system which is found from the apparent and absolute brightness of the system. The orbits are assumed to be circular. 946 U. Heber et al.: Resolving subdwarf B stars in binaries by HST imaging Based on the spectroscopic distances derived above (see Table 4) we then simulate a huge number of such binary systems. For three stars (HE 0430−2457, PG 0942+461, and HE 2213−2212) the magnitude ratio of the components could not be determined and therefore the distances are unknown. We adopted the mean value of m the other stars (∆R = 1 . 1), which is consistent with their spectral appearance (see Fig. A.1). The numerical simulation predicts a mean value of a sin i = 0.′′ 04 and that, out of the 19 observed systems, we should resolve six systems at a resolution limit of 0.′′ 1, one of which should show a separation greater than 1.′′ 0. Since the orbital motion for an eccentric orbit is lower during phases of large separation, the time averaged distance is larger than the semi major axis. Thus eccentric orbits would increase the detectability. Duquennoy & Mayor (1991) also provide a distribution of ellipticities for normal stars. If the sdB systems did not experience phases of binary interaction, the distribution of eccentricity should correspond to that of normal stars. We used Duquennoy & Mayor’s distribution corrected for selection effects. For each eccentricity the ratio of the time averaged distance to a was calculated and finally the mean over the Duquennoy & Mayor distribution was computed. We find the average distance of the companions to increase by 17%. Another mechanism that tends to increase the separation of the components in a sdB binary is mass loss during post-main sequence evolution in order to reduce the mass of the sdB progenitor to its present value of half a solar mass. Assuming that the sdB evolved from a 1 M⊙ main sequence progenitor it must have lost 0.5 M⊙ due to a stellar wind during its post-main sequence evolution. Assuming that the wind emanates in a spherical symmetric manner and does not interact with the companion the increase in separation can be calculated acs cording to ȧa = − MsṀ +Mc (Pringle 1985), with a being the separation and Ms and Mc the masses of the sdB progenitor and that of the cool star, respectively. As a result the separation increases by 33%. We repeated the Monte Carlo simulations for increased separations. Even when we consider both elliptical orbits and evolution of the orbits due to a stellar wind as described above the prediction increased only slightly to 7 resolvable stars in our sample. Hence we predict that 6 to 7 stars should be resolvable in our sample if the systems have separations consistent with the Duquennoy & Mayor (1991) distribution. 5. Chance projections and triple systems In the vicinity of five programme stars we found an additional object within a radius of 3.′′ 01 . We have demonstrated above that only in two cases (TON 139 and PG 1718+519) the relative brightnesses are consistent 1 Note that PG 0105+276, which is resolved in three components (see Fig. 1), is an sdO star and does not belong to our sdB sample. with the expectations from the deconvolution of the spectral energy distribution. The remaining three cases must then be chance projections or triple systems. Since the programme stars lie at high galactic latitudes (except KPD 2215+5037, see Table 1), we expect chance coincidences to be rare. Indeed, we do not find any additional object in the PC field (40′′ ×40′′ ) except for the low galactic latitude object KPD 2215+5037. According to Abt & Levy (1976) 16% of multiple systems of normal stars are triples. If the fraction of triple systems is the same for our sample, we expect three programme stars to be triple. Most of these, if not all, should be resolvable. Besides TON 139 and PG 1718+519 we find in three cases companions to the sdB stars which are too faint to match the spectral energy distribution. These could be triple systems consisting of an unresolved sdB binary and a distant third star. 6. Radial velocities Important additional information can be obtained from radial velocity measurements. A systematic search for radial velocity variations of our programme stars is needed. Such projects have already been started by Saffer et al. (2001) and Maxted et al. (2001) who observed six of our programme stars (PB 6107, PHL 1079, PG 0749+658, TON 1281, PG 1449+653 and PG 2148+095). None of them showed significant radial velocity changes. Saffer et al. (2001) find in their survey of 21 composite spectrum sdB stars that the velocity variations of the individual components as well as the velocity difference between the two components are very small (less than a few km s−1 ) or undetectable, and conclude that the binaries have likely periods of many months to several years. Green et al. (2001) estimate from these measurements that the current periods average 3–4 years with separations 540– 650 R⊙ . We have obtained multiple precise radial velocities for TON 139 and a single measurement of PG 1718+519 using the MMT Blue Channel spectrograph at 1 Å resolution from 4000–4930 Å (see Table 6). The radial velocities of the cool companions were determined by cross correlation against super-templates of main sequence spectral types from F6 to K5. The sdB velocities were derived using a preliminary attempt at subtracting out the cool star companion spectrum. For details of the data reduction and analysis see Saffer et al. (2001). Improved sdB velocities using better cool star template spectra for the subtractions will be determined by Green et al. (2002, in prep.). For TON 139 the cool star’s velocity is constant, whereas the sdB velocity is changing by more than 50 km s−1 . This can be explained if an additional companion is orbiting the sdB star. This companion has to be so faint that it does not contribute to the light in the R band. Hence we have to conclude that the resolved system TON 139 is a triple system. A radial velocity study of PG 1718+519, the second resolved system in our sample, is not available yet. The single measurement listed in U. Heber et al.: Resolving subdwarf B stars in binaries by HST imaging 947 Table 6. Heliocentric radial velocities for the sdB- and the cool star components of TON 139 and PG 1718+519. star date UT TON 139 TON 139 TON 139 TON 139 TON 139 TON 139 TON 139 PG1718+519 1996-01-14 1996-03-11 1996-06-09 1997-01-28 1997-07-04 1998-01-22 mean 1997-09-10 HJD-2450000 exposure time [s] S/N vrad [km s−1 ] (sdB component) 96.91396 153.84515 243.75586 476.96939 633.66734 836.03834 600 300 600 1800 500 750 95.6 71.9 66.5 69.4 72.7 82.9 701.71120 1400.0 82.0 −6.3 ± 4.9 −7.4 ± 7.8 −13.1 ± 8.4 −20.2 ± 7.9 32.6 ± 6.7 −22.1 ± 9.1 −3.6 ± 20.2 −69.2 ± 10.1 Table 6 gives identical radial velocities for the sdB and the cool companion. This argues against a third faint component orbiting the sdB star in a narrow orbit as was found for TON 139. Additional radial velocity measurements are urgently needed to clarify the nature of PG 1718+519. Assuming that PG 1718+519 is not triple, this would be the only resolved binary system in our sample of 19 objects. 7. Conclusions In total we have resolved six systems out of a sample of 23 stars. Of those 23 stars, however, four do not really belong to the intended sample of sdB stars showing evidence for a cool companion: PG 1558−007, KPD 2215+5037, and PG 2259+134 show no photometric or spectroscopic evidence for a companion. The observed infrared excess can be explained by interstellar reddening rather than by a cool companion. PG 1558−007 does have a resolved near-by star (linear separation 1500 AU), which, however, is too faint to contribute detectably to the combined light in the R band. PG 0105+276 is a helium-rich sdO star (with two possible distant companions at 3700 AU and 4900 AU). Of the remaining four resolved systems the nearby stars are in two case (TON 1281, HE 0430−2457) too faint to reproduce the photometric and/or spectroscopic observations of the stars. Only in the two systems TON 139 and PG 1718+519 (separations 0.′′ 32 and 0.′′ 24, respectively) do the magnitudes of the resolved components match the expectations. These two stars could be physical binaries whereas in the other cases the nearby star may be a third component or a chance projection. Radial velocity measurements indicate, however, that the resolved system TON 139 is also triple. Hence, the observed sdB binary sample was reduced to 19 objects with two bona-fide resolved systems, which have apparent separations of 0.′′ 24 and 0.′′ 32. From the numerical simulations we would expect to resolve six to seven systems if sdB stars have the same binary characteristics as normal stars, out of which one system is expected to have a sin i > 1′′ and two should have separations between 0.′′ 1 and 0.′′ 2. The discrepancy becomes even more pronounced if one recalls that our photometric fit procedure tends to underestimate the brightness of the companion (and thus to overestimate the limiting angular separation vrad [km s−1 ] (cool companion) 19.9 20.2 21.8 20.2 22.4 20.7 20.8 −68.0 ± ± ± ± ± ± ± ± 0.6 0.7 0.8 0.9 0.7 0.6 1.0 0.9 that can still be resolved). In addition we expect three triple systems to be present in our sample. Most of these, if not all, should be resolvable. Such systems could explain some of the more distant companions as well as the radial velocity measurements of TON 139. This success rate (1 resolved binary out of 19 candidates) is clearly below the prediction of numerical simulations assuming single star evolution (about 30%), using the distribution of binary separations given by Duquennoy & Mayor (1991). This indicates that the distribution of separations of sdB binaries strongly deviates from that of normal stars. If, on the other hand, all sdB stars were produced by close binary evolution, none of the binary systems should have been resolved (even at the high spatial resolution of the WFPC2 camera). Our low success rate is thus closer to that predicted by the close binary evolutionary scenario. Recent radial velocity surveys (Saffer et al. 2001; Maxted et al. 2001) revealed that a large fraction of single-lined sdB stars are indeed close binaries with periods below 10 days. Our results could be explained if most of the programme stars were close binaries. Therefore, our study provides further evidence that close binary evolution indeed is fundamental to the evolution of sdB stars. A survey for radial velocity variations in all of our programme stars will be tale telling. Acknowledgements. This work was supported by the DLR under grant 50 OR 96029-ZA. We thank Klaas de Boer (Bonn), Heinz Edelmann (Bamberg) and Heinz Lehnhart (Tübingen) for taking most of the optical spectra for us and Martin Altmann (Bonn) for providing us with his photometric measurements prior to publication. Thanks go also to Anna Ulla and Klaas de Boer for helpful comments and encouragement. 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. We also made extensive use of the Simbad database, operated at CDS, Strasbourg, France. Appendix A: Spectroscopic observations and data reduction The observational setups and observing dates for the new spectra are given in Table A.1. The reduction of the 948 U. Heber et al.: Resolving subdwarf B stars in binaries by HST imaging Fig. A.1. Comparison of normalized spectra of four programme stars to PG 1309−078, which is already known to be a spectroscopic binary containing an sdB. The spectral features indicative of a cool companion are marked. Fig. A.2. Optical spectra of the sdB stars PG 1558−007 and KPD 2215+5037 as well as of the sdO star PG 0105+276. The spectra of the former are dominated by hydrogen lines, that of the latter by He ii lines. spectra of PG 0105+276, HE 0430−2457, PG 0942+461, HE 2213−2212, and KPD 2215+5037 are described by Edelmann et al. (2001b). PG 2148+095 was observed and reduced as described by de Boer et al. (1995), the reduction of PG 1309−078 and PG 1558−007 was performed in the same way as described in Moehler et al. (1997). Fig. A.3. Spectral fit for the sdB star KPD 2215+5037. Hǫ is excluded from the fit because of contamination by interstellar Ca ii. Figure A.2 shows the spectra of the stars that show no spectroscopic or photometric evidence for a cool companion (PG 1558−087, KPD 2215+5037, and PG 0105+276). The Ca ii absorption lines in the spectra of these stars (see Fig. A.2) are probably of interstellar nature. Our spectrum clearly shows that PG 0105+276 is a helium rich sdO star (see Fig. A.2) inconsistent with the photometric classification as sdB+K7 by Allard et al. (1994, where all three stars seen in Fig. 1 were included in the measurements) but in accordance with the early spectroscopic classification by Green et al. (1986). We derived the atmospheric parameters Teff , log g and helium abundance simultaneously for the single stars by matching a grid of synthetic spectra derived from H and He line blanketed NLTE model atmospheres (Napiwotzki 1997) to the data. For temperatures below 27 000 K we used the metal line blanketed LTE model atmospheres of Heber et al. (2000). The synthetic spectra were convolved beforehand with a Gaussian profile of the appropriate FWHM to account for the instrumental profile. Results are given in Table A.1 and Fig. A.3 displays the fit for KPD 2215+5037 as an example. Appendix B: Photometric data for our programme stars In Tables B.1 and B.2 we compile the photometric data collected from literature and used in the photometric deconvolution. U. Heber et al.: Resolving subdwarf B stars in binaries by HST imaging 949 Table A.1. New optical spectroscopy and atmospheric parameters of single programme stars. star PG 0105+276 HE 0430−2457 PG 0942+461 PG 1309−078 PG 1558−007 PG 2148+095 HE 2213−2212 KPD 2215+5037 PG 2259+134 telescope and spectrograph CA 3.5m TWIN ESO 1.5m B&C CA 3.5m B&C ESO 1.5m DFOSC ESO 1.5m DFOSC ESO 1.5m B&C ESO 1.5m B&C CA 3.5m TWIN wavelength spectral range resolution [Å] [Å] 3600–7400 3.1 3600–7450 5.5 3860–5560 5.0 3860–6780 5.4 3860–6780 5.4 3730–4970 3.0 3600–7400 5.5 3260–7450 3.1 Theissen et al. (1993) obs. date 1997/08/31 1996/10/22 1989/01/23 2000/06/21 2000/06/21 1991/07/10-15 1996/10/23 1997/08/29 Teff log g log(He/H) [K] 63 000 [cgs] 5.4 +0.5 20 300 5.0 −2.6 29 400 31 900 5.6 5.9 −2.2 −1.7 Table B.1. Strömgren photometry and UV spectrophotometry for our programme stars. Strömgren photometry is taken from Green (1980, G80), Kilkenny (1984, K84; 1987, K87), Moehler et al. (1990, M90), Theissen et al. (1993, T93), Wesemael et al. (1992, W92). The IUE data were obtained from the IUE final archive (http://archive.stsci.edu/iue/). Star PB 6107 PG 0105+276 PHL 1079 PG 0749+658 TON 1281 TON 139 PG 1309−078 PG 1449+653 PG 1511+624 PG 1558−007 PG 1636+104 PG 1656+213 TON 264 PG 1718+519 PG 2148+095 KPD 2215+5037 PG 2259+134 PG 2259+134 BD −7◦ 5977 y b−y m 12 . 897 m 12 . 889 m 12 . 89 m 12 . 907 m 14 . 481 m 13 . 278 m 12 . 135 m 13 . 371 m 12 . 796 m 14 . 11 m 13 . 580 m 14 . 421 m 13 . 528 m 14 . 090 m 14 . 070 m 13 . 686 m 13 . 694 m 13 . 037 m 13 . 739 m 14 . 478 m 14 . 545 m +0 . 032 m +0 . 026 m +0 . 01 m +0 . 018 m +0 . 022 m +0 . 003 m −0 . 032 m +0 . 094 m +0 . 111 m +0 . 07 m +0 . 041 m +0 . 049 m −0 . 011 m +0 . 169 m +0 . 008 m +0 . 102 m +0 . 131 m +0 . 028 m −0 . 026 m −0 . 038 m −0 . 069 u−b m +0 . 112 m . +0 10 m . −0 194 m +0 . 106 m +0 . 131 m +0 . 175 m +0 . 364 m +0 . 06 m +0 . 047 m −0 . 002 m +0 . 244 m +0 . 426 m −0 . 053 m +0 . 307 m . +0 087 m +0 . 034 m . −0 011 m1 c1 Ref. m +0 . 052 m +0 . 092 m +0 . 05 m +0 . 093 m +0 . 023 m −0 . 094 m −0 . 109 m −0 . 109 m +0 . 087 m +0 . 065 m +0 . 055 m +0 . 18 m +0 . 034 m +0 . 005 m +0 . 091 m +0 . 056 m +0 . 070 m +0 . 084 m +0 . 094 m +0 . 066 m +0 . 068 m +0 . 082 m +0 . 088 References Abt, H. A., & Levy, S. G. 1976, ApJS, 59, 229 Allard, F., Wesemael, F., Fontaine, G., Bergeron, P., & Lamontagne, R. 1994, AJ, 107, 1565 Altmann, M., & de Boer, K. S. 2000, A&A, 353, 135 Aznar Cuadrado, R., & Jeffery, C. S. 2001, A&A, 368, 994 Bailyn, C. D., Sarajedini, A., Cohn, H., Lugger, P., & Grindlay, J. E. 1992, AJ, 103, 1564 Brown, T. M., Ferguson, H. C., Davidsen, A. F., & Dorman, B. 1997, ApJ, 482, 685 Brown, T. M., Bowers, C. W., Kimble, R. A., Sweigart, A. V., & Ferguson, H. C. 2000, ApJ, 532, 308 Burstein, D., & Heiles, C. 1982, AJ, 87, 1165 de Boer, K. S., Schmidt, J. H. K., & Heber, U. 1995, A&A, 303, 95 m −0 . 095 m −0 . 089 W92 M90 G80 K87 W92 K84 W92 W92 W92 G80 W92 W92 W92 W92 W92 W92 T93 W92 W92 M90 W92 IUE SWP LWP 56 271 42 338 21 098 56 384 34 298 39 370, 57 359, 57 361 18 491 39 422 39 422 41 571 18 542 18 542 20 308 56 148 44 821, 56 182 23 244 31 030 10 815 de Boer, K. S., Aguilar Sánchez, Y., Altmann, M., et al. 1997, A&A, 327, 577 Dorman, B., Rood, R. T., & O’Connell, R. W. 1993, ApJ, 419, 596 Duquennoy, A., & Mayor, M. 1991, A&A, 248, 485 Edelmann, H., Heber, U., Napiwotzki, R., Lemke, M., & Christlieb, N. 2001a, Spectral Analysis of sdB stars from the Hamburg ESO survey, in Proceedings of the Twelfth European Workshop on White Dwarf Stars, ed. H. L. Shipman, J. L. Provencal, J. MacDonald, & S. Goodchild (Astronomical Society of the Pacific: San Francisco), ASP Conf. Ser., 226, 165 Edelmann, H., Heber, U., Napiwotzki, R., et al. 2001b, A&A, in prep. 950 U. Heber et al.: Resolving subdwarf B stars in binaries by HST imaging Table B.2. BV RI (Allard et al. 1994), U BV I (Ferguson et al. 1984), HST R (this paper) and infrared broadband photometry (UT98: Ulla & Thejll 1998, 2MASS: 2MASS 2nd incremental data release, http://irsa.ipac.caltech.edu/applications/2MASS/BasicSearch/) for our programme stars. Star PB 6107 PG 0105+276 PHL 1079 HE 0430−2457 PG 0749+658 PG 0942+461 TON 1281 TON 139 PG 1309−078 PG 1449+653 PG 1511+624 PG 1558−007 PG 1601+145 PG 1636+104 TON 264 PG 1718+519 PG 2148+095 HE 2213−2212 KPD 2215+5037 PG 2259+134 BD −7◦ 5977 Star PG 1421+345 PG 1601+145 PG 1656+213 1 2 V B−V m . 12 881 m 14 . 448 m 14 . 1551 m 12 . 121 m . m 13 439 13 . 611 m 14 . 527 m 13 . 541 m 14 . 424 m 14 . 039 m 14 . 074 m 13 . 733 m 13 . 021 m 13 . 664 m . 2 −0 038 m −0 . 087 m −0 . 0461 m −0 . 106 +0 070 m +0 . 086 −0 . 035 m −0 . 022 m −0 . 064 m +0 . 028 m +0 . 193 m −0 . 083 m +0 . 113 m −0 . 024 m m +0 . 015 −0 . 093 m . V B−V m . −0 14 m +0 . 01 m −0 . 20 R−I m . +0 096 m +0 . 127 m +0 . 0851 m +0 . 021 +0 094 +0 51 14 78 m 14 . 50 m 14 . 88 m . m . 10 55 m . V −R m . m +0 . 072 m . m . +0 156 +0 176 +0 . 073 m +0 . 113 m +0 . 012 m +0 . 180 m +0 . 191 m +0 . 066 m +0 . 156 m +0 . 060 m +0 . 110 m +0 . 142 m +0 . 110 m +0 . 347 m +0 . 196 m +0 . 136 m +0 . 132 m +0 . 096 m +0 . 052 m m 2 U −B m . −0 89 m −0 . 92 m −0 . 73 V −I m . +0 63 m +0 . 41 RHST J H K Ref. m . 12 80 m 14 . 36 m 13 . 24 m 14 . 07 m 12 . 14 m 13 . 96 m 13 . 27 m 12 . 65 m 14 . 05 m 13 . 57 m 14 . 38 m 13 . 55 m 14 . 37 m 13 . 85 m 14 . 02 m 13 . 53 m 12 . 98 m 14 . 01 m 13 . 91 m 14 . 70 m 10 . 05 RHST m . 14 59 m 13 . 821 m 12 . 23 m 13 . 315 m 13 . 172 m 12 . 503 m 11 . 92 m 13 . 259 m 13 . 813 m 13 . 578 m 12 . 716 m 12 . 34 m 13 . 292 15 . 795 m 8 . 97 m 9 . 017 m J 14 . 347 m 12 . 55 m 13 . 619 13 . 612 m 12 . 758 m 12 . 10 m 13 . 558 14 . 114 13 . 918 13 . 008 m 12 . 18 m 13 . 686 m . 14 035 m 13 . 721 m 12 . 04 m 13 . 208 m 13 . 084 m 12 . 448 m 11 . 93 m 13 . 162 m 13 . 883 m 13 . 600 m 12 . 664 m 12 . 06 m 13 . 236 15 . 220 m 8 . 48 m 8 . 526 m H m . 13 716 m 2MASS UT98 2MASS m 2MASS 2MASS UT98 2MASS m 2MASS m 2MASS m 2MASS UT98 2MASS 14 . 523 m 8 . 38 m 8 . 448 m 2MASS UT98 2MASS K Ref. m . 13 678 2MASS m 14 . 73 Altmann (priv. comm.). m m Derived from Tycho photometry (VT = 10 . 6, (B − V )T = +0 . 6) using the transformation given in Perryman (1997). Ferguson, D. H., Green, R. F., & Liebert, J. 1984, ApJ, 287, 320 Green, E. M. 1997, Proceedings of the Third Conference on Faint Blue Stars, ed. A. G. D. Philip, J. Liebert, & R. A. Saffer (Schenectady: L. Davis Press), 417 Green, E. M., Liebert, J., & Saffer, R. A. 2001, On the Origin of Subdwarf B Stars and Related Metal-Rich Binaries, in Proceedings of the Twelfth European Workshop on White Dwarf Stars, ed. H. L. Shipman, J. L. Provencal, J. MacDonald, & S. Goodchild, (Astronomical Society of the Pacific: San Francisco), ASP Conf. Ser., 226, 192 Green, R. F., Schmidt, M., & Liebert, J. 1986, ApJS, 61, 305 Green, R. F. 1980, ApJ, 238, 685 Greggio, L., & Renzini, A. 1990, ApJ, 364, 35 Greggio, L., & Renzini, A. 1999, Mem. S.A.I., 70, 691 Heber, U. 1986, A&A, 155, 33 Heber, U., Jordan, S., & Weidemann, V. 1991, White Dwarfs, ed. G. Vauclair, & E. Sion (Dordrecht, Kluwer), 109 Heber, U., Reid, I. N., & Werner, K. 2000, A&A, 363, 198 Iben Jr., I. 1990, ApJ, 353, 215 Iben Jr., I., & Tutukov, A. V. 1992, Close Binaries with Evolved Components, in Frontiers of Stellar Evolution, ed. D. L. Lambert (Astronomical Society of the Pacific: San Francisco), ASP Conf. Ser., 20, 403 Jeffery, C. S., & Pollacco, D. L. 1998, MNRAS, 298, 179 Kilkenny, D. 1984, MNRAS, 211, 969 Kilkenny, D. 1987, MNRAS, 228, 713 Kilkenny, D., Heber, U., & Drilling, J. S. 1988, SAAOC, 12, 1 Kurucz, R. L. 1992, Model Atmospheres for Population Synthesis, in The Stellar Populations of Galaxies, ed. B. Barbuy, & A. Renzini, IAU Symp. 149 (Kluwer: Dordrecht), 225 Lamla, E. 1982, in Landolt-Börnstein, Numerical Data Functional Relationships in Science and Technology, vol. VI/2b, ed. K. Schaifers, & H. H. Voigt (Springer Verlag: Berlin-Heidelberg-New York), Sect. 4.2 Maxted, P. F. L., Heber, U., Marsh, T., & North, R. C. 2001, MNRAS, 326, 1391 Mengel, J. G., Norris, J., & Gross, P. G. 1976, ApJ, 204, 488 Mitchell, K. J. 1998, ApJ, 494, 256 Moehler, S., Richtler, T., de Boer, K. S., Dettmar, R. J., & Heber, U. 1990, A&AS, 86, 53 Moehler, S., Heber, U., & Rupprecht, G. 1997, A&A, 319, 109 Napiwotzki, R. 1997, A&A, 322, 256 Perryman, M. A. C. 1997, The Hipparcos and Tycho Catalogues, ESA SP-1200, 1, 57 U. Heber et al.: Resolving subdwarf B stars in binaries by HST imaging Pringle, J. E. 1985, in Interacting binary stars, ed. J. E. Pringle, & R. A. Wade (Cambridge University press), 1 Reimers, D. 1975, in Problems in Stellar Atmospheres and Envelopes, ed. B. Baschek, W. H. Kegel, & G. Traving (Springer Verlag: Berlin-Heidelberg-New York), 229 Saffer, R. A. 1991, Ph.D. Thesis, University of Arizona Saffer, R. A., Bergeron, P., Koester, D., & Liebert, J. 1994, ApJ, 432, 351 Saffer, R. A., & Liebert, J. 1995, The Spatial and Kinematic Distribution of Field Hot Subdwarfs, in White Dwarfs, ed. D. Koester, & K. Werner (Springer Verlag: BerlinHeidelberg-New York), Lect. Notes Phys., 443, 221 Saffer, R. A., Green, E. M., & Bowers, T. P. 2001, The Binary Origins of Hot Subdwarfs: New Radial Velocities, in Proceedings of the Twelfth European Workshop on White Dwarf Stars, ed. H. L. Shipman, J. L. Provencal, J. MacDonald, & S. Goodchild (Astronomical Society of 951 the Pacific: San Francisco), ASP Conf. Ser., 226, 408 Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ, 500, 525 Stetson, P. B. 1987, PASP, 99, 191 Theissen, A., Moehler, S., Heber, U., & de Boer, K. S. 1993, A&A, 273, 524 Theissen, A., Moehler, S., Heber, U., Schmidt, J. H. K., & de Boer, K. S. 1995, A&A, 298, 577 Ulla, A., & Thejll, P. 1998, A&AS, 132, 1 Viton, M., Deleuil, M., Tobin, W., Prevot, L., & Bouchet, P. 1991, A&A, 242, 175 Wesemael, F., Fontaine, G., Bergeron, P., Lamontagne, R., & Green, R. F. 1992, AJ, 104, 203 Zombeck, M. V. 1990, Handbook of Space Astronomy and Astrophysics (Cambridge University Press: Cambridge), 100