Abstract

We are continuing a study of hot subdwarf stars (sdB or sdO stars) that, on a statistical basis, appear to be members of wide binaries (common-proper-motion pairs). These c.p.m. candidates make up several percent of the approximately 1300 catalogued subdwarfs.

We present what is currently known about these pairs from literature and our own investigations, including: magnitude differences and colors, distribution of angular separation vs. magnitude, 2MASS colors, spectra, etc. Confirmation of these candidate binary systems awaits radial velocity and proper motion studies.

This sample cannot be regarded as complete in any rigorous way, and is certainly biased in the selection process. It nevertheless is showing promise that a purified sample can be used to learn about the luminosities, ages, and original metallicities of the hot subdwarf stars.

1. Introduction
   
   Hot subdwarfs are defined as having effective temperatures exceeding 20,000 K, with surface gravities higher and luminosities lower than main sequence stars of the same temperature. Hot subdwarfs fall in the HR Diagram at fainter visual magnitudes and bluer colors than the horizontal branches of most globular clusters. However, field subdwarf B (sdB) stars in particular are consistent with belonging to the extended horizontal branch (EHB) populations seen in some globular clusters (Humason & Zwicky 1947). Thus, the sdB stars are understood as core helium burning objects with very low envelope mass and total masses of ~0.5-0.55 M_sun (Saffer et al. 1994). It has been proposed that a binary mechanism is responsible for the formation of sdB stars in particular (Mengel et al. 1976). Were this the case than we would expect all sdB stars to be found in close binaries.
   
   During a project to revise the coordinates of spectroscopically identified hot subdwarfs (Kilkenny et el. 1988, 1992), we observed that an interesting number of these stars (several percent of all hot subdwarfs) appear to have resolved companions. Possibly these systems are triple so that the hot subdwarf is also in a close binary (Maxted 2001). In any case resolved companions, if verified, could be used to access the distances, luminosities, and perhaps also the ages and original metallicities of the hot subdwarfs.
   
   We statistically demonstrated the high likelihood of physical association of these pairs using Digitised Sky Survey images and USNO photographic magnitudes (Wade & Stark 2001). But, to better establish the physical association of these candidate pairs, accurate photometric, proper motion, and radial velocity data are needed.
   
   
2. What We Know...
   • USNO photographic magnitudes (Monet et al. 1999; these were fully explored in Wade & Stark 2001)
     • Some close pairs were not resolved by USNO, thus were blended or "thrown out" as an extended source
     • Colors and magnitudes of the remaining companions are consistent with main sequence stars
   • A few Teff and log g measurements for the subdwarfs in the pairs (from Kilkenney 1988, 1992):
     • Median log T_eff = 4.59 (Fig 1a)
     • Median log g = 5.5 (Fig 1b)
   • Angular separations:
     • Observed separations, median separation = ~10'' (Fig 2a)
     • Separations scaled to a distance of 1 kpc using 2MASS J_sd, and assuming M_V,sd = 5.5, (V-J)_sd = -0.5, gives a median separation = ~7.7'' (Fig 2b)
   • J, H, and K_s magnitudes for pair components (from 2MASS - Sec. III)
   • Radial velocity differences for a few pairs (Sec. IV)
   
   
3. a. Two Micron All Sky Survey (2MASS)
   • 2MASS coverage of 33 pairs out of 46 "good candidates"
   • The distribution of J mag differences of the pair components (J_sd-J_comp) has a a median value of J_sd-J_comp = +0.97, showing that there is not a significant difference in magnitude between the subdwarfs and their companions - this is likely a selection effect (Fig 3a)
   • In J-K_s color the subdwarfs are, in general, bluer then their associated companions. The median values of the two distributions are: (J-K_s)_sd = -0.12 and (J-K_s)_comp = +0.48 (Fig 3b)
   • Some of the subdwarfs in resolved doubles have IR colors of J-K_s >~ +0.1 which are redder than what would be expected for hot subdwarfs. This suggests that they may contain a third unresolved late-type companion and therefore are potential hierarchical triple systems

3. b. Comparison With Composites
   • The J, H, and K_s colors of ~500 subdwarfs (not visual doubles) were obtained from the 2MASS 2nd Incremental Data Release and their distribution in color space, particularly in J-K_s color, was explored in Stark & Wade 2001. Subdwarfs are bimodally distributed in J-K_s color: with one peak comprising ~60% of the subdwarfs centered at J-K_s = -0.15, this blue peak represents "single" subdwarfs; and the second peak containing ~40% of the subdwarfs centered at J-K_s = +0.3, representing "composite-spectrum" subdwarfs.
   • We suspect that the resolved doubles are simply wider-spaced versions of the composite-spectrum systems. We combined the J-K_s colors of the resolved doubles to compare them to the colors of the unresolved ones. For those pairs without J-K_s measurements for the subdwarf, we assumed the color of the subdwarf to be J-K_s = -0.15 (the central value of the "blue" peak). Figure 4 shows the combined colors for the visual doubles compared to the observed distribution of 413 subdwarfs in J-K_s color. The combined colors for most of the visual doubles are indistinguishable from the colors of composite-spectrum subdwarfs. Thus the visual doubles are likely resolved versions of the composite-spectrum subdwarfs systems. So, by studying these resolved doubles, we can gain insight into the composite-spectrum doubles as well.

3. c. 2MASS Color-Magnitude
   • Assuming reasonable parameters for the subdwarf (M_V = 5.5, V-J = -0.5), a color-magnitude diagram was constructed for the resolved subdwarf companions based on 2MASS photometry (Fig 5). Plotted for comparison are lines representing the Pop I main sequence (Zombeck 1990) and several metal poor main sequence populations ([M/H]= -2.0, - 1.5, -1.0; from Baraffe et al. 1997).
   • In the color-magnitude diagram, a majority of the companions are consistent with main sequence stars of spectral types later than mid-F (Pop I). This is consistent with the types of unresolved composite-spectrum companions observed with subdwarfs. However there is a vertical spread that is larger than what can be explained by photometric errors or uncertainties in M_V,sd and (V-J)_sd. This spread is not due to reddening, since reddening has only a minor effect on the IR colors, and reddening would tend to shift the points down and to the right of the Pop I main sequence line. However, it is likely that this spread can be explained by metallicity differences between the companions, in this case, most of the subdwarfs and their companions belong to metal poor populations.
   
   
4. Radial Velocities
   • Spectra of both components of three visual doubles have been obtained to determine the difference in the radial velocities of the two stars. We expected the two stars to have delta RV < ~|5| km/s if they truly are c.p.m. pairs.
   • The spectra are from: The Hobby-Eberly Telescope (High Resolution Spectrograph), and the William Herschel Telescope, La Palma (ISIS)
   • For the three pairs, the radial velocity differences (delta RV = RV_sd-RV_comp) are: -1.0 km/s, 2.4 km/s, -2.0 km/s
   • The radial velocities of the members of these three pairs are consistent with being the same. This holds great promise for future studies. We have a continuing program to obtain radial velocity measurements of resolved doubles using the Hobby-Eberly Telescope.
   
   
5. What is Still Needed...
   • Presently available data is consistent with physical association
   • Further information is still needed to verify that these visual doubles are in fact c.p.m. binaries:
     • Radial velocities of more pairs to determine common space velocity
     • Proper motion studies of both components to search for common proper motion
   • To fully exploit the properties of the resolved companions to determine the properties of the hot subdwarf (such as its distance, luminosity, original population, etc. - assuming they truly are associated) we need:
     • Accurate visual photometry of both components
     • Spectral analysis of the resolved companions for:
       • Temperature
       • Surface gravity (luminosity class)
       • Metallicity (especially now that there seems to be indications that the metallicities are not all solar)

E-mail me