Using Adaptive Optics to probe the incidence of Binarity among Be Stars


A sample of 31 nearby Be stars within 100 pc have been observed with the Altair adaptive optics system at Gemini-North. Close companions were found in 4 cases. The sample was selected to include the nearest isolated Be stars in each spectral subtype. The proposed binaries have angular separations ranging from 0.13”-0.83”, corresponding to physical separations 19-78 AU. For the tightest binary, assuming low inclination and a circular orbit, the predicted orbital period is of order 20 years. JHK photometry indicates the companions are A-type stars. The incidence of companions in the range of magnitudes and separations found, is in excess of the random probability by a factor of ?? and in ? cases relative motion was measured. Binarity is a leading contender for explaining the origins and behavior of Be-stars, a unique and fascinating class that exhibit dynamic circumstellar disks and extreme rotation. Adaptive optics has opened a new parameter space for binary star searches not probed by radial velocity (RV) or light-curve studies.


Binarity is a leading contender for explaining the origins and behavior of classical Be-stars, a unique and fascinating class that exhibit dynamic circumstellar disks and extreme rotation.

Adaptive optics opens a new parameter space for binary star searches not probed by radial velocity (RV) or light-curve studies. to measure the incidence of close companions around Be stars and characterize their spectral types and orbits.

Be stars are O or B type main-sequence (dwarf-to-subgiant) stars exhibiting balmer lines in emission, accompanied by a strong infrared excess. Extensive spectral studies have shown that Be-stars possess optically thin circumstellar gas disks in their equatorial plane. The rotating disk gives rise to double peaked emission lines via the doppler effect. Line structure frequently evolves on timescales of weeks-to-years, reflecting changes in the density distribution of the disk.

There are 3 leading theories for Be-star equatorial disk formation. First of all, (Porter 1996) showed that all Be stars are (and normal B stars are not) rotating at or close to their break-up velocities, thus matter could more easily flow from the stellar envelope into the equatorial plane. Secondly, quasi-periodic oscillations have frequently been observed in Be stars, and these have been suggested as a mechanism for coupling the stars’ spin (angular momentum) to the outflow REF Thirdly, binary companions have been suggested to stimulate disk-formation by lowering the effective gravity at the Be-star surface sufficiently for matter to spill off. Binary evolution can also be invoked to spin-up the Be stars and explain their rapid rotation.

In isolation, the above scenarios are insufficient to cause organized acceleration of an equatorial outflow. Furthermore the evolutionary path leading to extreme stellar-spin is not well understood. An important clue lies in the spectral-type distribution of classical Be stars (i.e. the isolated variety) which is uniform, and those in high mass X-ray binaries (HMXB) which cluster strongly around spectral type B0 (Negueruela 1998). The presence of the neutron-star companion in HMXB has been shown to affect both the size and density of the disk (e.g. (R. K. Zamanov 2001)).

Be star disks are very dynamic on timescales of months to years, sometimes exhibiting quasi-periodic fluctuations in both the equivalent width and profile of the H\(\alpha\) emission line. Occasionaly the disk dissipates and later re-forms. Monitoring of spectral line-profiles reveals long-lived density perturbations that precess, giving rise to the quasi-periodic behavior (e.g. (J. S. Clark 2001). However the perturbing entity has yet to be identified. In HMXB the neutron star exerts a tidal influence on the circumstellar disk (A. T. Okazaki 2001), leading to strings of X-ray outbursts correlated with the binary period (For extensive examples see (Galache 2008)). The tidal interaction would also be present in the ”isolated” Be stars if faint companions are present, and is plausibly the driver of quasi-periodic disk behavior. This can be tested by estimating orbital periods from the binary separations and comparing against the disk quasi-periods. There is compelling observational evidence for binarity in classical Be-stars: Carrier conducted spectroscopic (RV monitoring) of 4 Be stars selected for their periodic photometric variations, finding two spectroscopic binaries. The period of one of these binaries (398d) closely matched its photometric period (Carrier 2002). No companions were found for the other 2 stars (one of which was found to exhibit non-radial pulsations). Companions to a few Be-stars are known at much larger separations (for example the famous double-star Albireo). With AO we can search the parameter space not explored by natural imaging and RV studies. VLT/AO observations by (Kervella 2008) have found a companion to the closest Be star ”Achernar”. The separation is 0.15” (6.7 AU at D=44pc) and the orbital period is about 15 years, similar to the quasi-periodicity of the circumstellar disk. The spectrum of Achernar-B shows an A1V-A3V star. The two stars differ in brightness by 3.5 mag.

Of wider significance, the majority of Neutron-star High Mass X-ray Binaries (HMXBs) have a Be star as the mass donor. The Be circumstellar disk is the fuel-reservoir for the accretion powered X-ray outbursts of the neutron star. Be-systems make up 70% of Milky Way HMXBs and 49-of-50 in the Small Magellanic Cloud. The origin and evolution of Be stars is a critical missing-link in explaining the activity cycles and lifetimes of X-ray binaries; in turn the X-ray luminosity function (XLF) of star-forming galaxies is dominated by HMXBs. Thus discovering the role of binarity in making Be stars will improve our quantitative understanding of the relationship between SFR and XLF.

\label{sect:sample}The Sample

This study was designed to sample the closest representatives of each Be spectral type and luminosity class. To control systematics, a volume-limited sample of Be stars is listed in Table 1. These very bright stars have been extensively studied spectroscopically and in the time domain, so that the quasi-periodic behavior of the circumstellar disks is well characterized. Stellar non-radial pulsations do not confuse the picture because they are seen in spectral lines originating on the star’s surface, while the disk emits free-free emission from optically thin plasma.

The sample was constructed from the Be-star catalog of (Jaschek 2004) (in Vizier), by selecting northern (Dec\(>\)0) Be-stars with measured rotational velocities and no known (visual or spectroscopic) binary companion. From this list the nearest exemplars. Our full target list (Table 1) contains 52 Be-stars closer than 1 kpc, and includes all nearby ”isolated” Be-stars across the range of spectral types. Most are brighter than V=7 permitting optimum AO performance in NGS mode. The closest Be star in the sample has spectral type B6V, and V=4.1, for which we estimate D=66pc, a projected separation of 0.1” at this distance corresponds to a physical separation of 6.5AU, it also the fastest rotator in the sample. The brightest star in the sample is a B6III, V=3.62, the physical scale for this object at 0.1” is of order 10 AU. The brightest early-type is B1V with V=4.6, D=398pc, so that 0.1”=40 AU. Using spectral type B8V to represent the lowest luminosity members of the class, we find a distance of 200pc for V=6, at this distance the projected seperation of 0.1” = 20 AU.

Photometric distances were estimated for all Be stars in J82, using the spectral type and apparent V magnitude, referring to AQ (Table 15.7) for the corresponding absolute magnitudes (M\({}_{V}\)). Parallax distances from Hipparcos were available for ?? stars and were adopted wherever the parallax value was at least twice its formal uncertainty. Photometric and parallax distances were in reasonable agreement (within respective errors) as shown in Figure \ref{fig:distances}. The addition of Hipparcos parallaxes significantly reduced the distance for 13 early type (B0-B2) stars, bring them within 1kpc.