| Magnetic fields are effective in shaping the AGB wind into axi-symmetric geometries (e.g., Garcia-Segura et al. 2005 ). However, magntic fields drain the differential stellar rotation that is needed to sustain them. Nordhause et al. (2007) determined that the time-scale for the magnetic field to extinguish itself is of the order of 100 years. This is too short a time for the magnetic field to shape the AGB mass-loss. The presence of a companion not too far from the primary is an effective way to resupply angular momentum and the best way to shape the AGB mass-loss which is at the origin of PN shapes. On the other hand, if all (or almost all) PN are shaped by binary interactions we should see a sign of this in the binary fraction of central stars: A large fraction of them should be close binaries, even if, presumably, some interaction led to mergers and others were with companions too small and sub-luminous to be easily detected. |
What we know about binary central stars |
|
Howard Bond has spent about 30 years monitoring central stars for photometric variability indicative of binarity (from illumination of one side of a cool close companion, from eclipses or from ellipsoidal variation). In Bond (2000) one can find a list updated in De Marco (2006) [see also astro-ph]. A total of 17 close binary systems are known today with an additional 4 showing composite spectra. There are also 13 visual binaries where the two stars are so far
apart that the companion is unlikely to have affected the shaping - in a scenario where binary interactions are needed to shape a PN, these objects would be triples, with an
unknown close companion and a distant tertiary. About 10% of G type main sequence stars are known to be triple systems.
The binary fraction remains however unconstrained. Bond (2000) argues that 10-15%, the binary fraction from their survey, is a lower limit because the effects leading to photometric variability are smaller the larger the separation. However it is not clear what exactly the period bias of his survey should be. If it is 3 days then indeed this is the reason why all the systems detected by Bond (2000) have periods less than 3 days. However if the inherent bias is, say, 2 weeks then Bond (2000) should have detected more systems with periods 3 days < P < 2 weeks. Since they have not, then one might deduce that there are only very few systems with periods in that range. The verdict will have to await results of a new on-going survey (PI Hillwig) and the results of new radial velocity surveys (PI De Marco; see also below) Finally, one could argue that since 30% of post-AGB stars are single lined spectroscopic binaries with 100 days < P < 1500 days (van Winckel 2003), and assuming that all the post-AGB observed will make a PN, then ~30% of all central stars should be binaries with periods in that range. This would mean that the short/intermediate period binary fraction is >40-45%. |
Why is it so hard to detect binary central stars? |
| Mendez (1991), Sorensen and Pollacco (2004), De Marco et al. (2004) and Afsar and Bond (2005) all tried to carry out radial velocity (RV) surveys for binaries. The RV variability fraction is extremely high (~100% once biases are included) but no periods have as yet emerged. De Marco et al. (2007) showed that bright central stars are subject to winds and pulsational variability with strong RV variability (RV amplitudes up to 60 km/s). This would make it next to impossible to detect all but the most extreme central star binaries. Intrinsically dimmer stars, unlikely to have winds and large amplitude pulsations are being targeted now (PI De Marco). |
The known binary central stars of PN | |||
| PN name | Type of binary | Period (days) | Comments |
| PNG145 (BE UMa) | S2,EC,I | 2.29 | |
| A46 | S2,Ec,I | 0.47 | |
| A63 | Ec,I | 0.46 | |
| HFG1 | S2,I,El | 0.58 | |
| K1-2 | S,I | 0.36 | |
| DS1 | S2,I | 0.36 | |
| Sp1 | I | 2.91 | |
| NGC6337 | I | 0.17 | |
| Hf2-2 | I | 0.40 | |
| A65 | I | ~1 | |
| A41 | I? | 0.11 | Primary model: companion main sequence star |
| -- | El? | 0.23 | Alternative model: companion sdOB |
| HaTr4 | I? | 1.74 | |
| PNG136 | El,other | 0.16 | WD companion |
| SuWt2 | S2,I | 4.9 | A+A binary!!! |
| NGC2346 | S1 | 15.99 | A5V spectrum; semi-periodically eclipsed by dust |
| NGC6026 | El | 0.53 | Companion: WD/SdO |
| A35 | C | - | Bright evolved companion |
| LoTr1 | C | - | Bright evolved companion |
| LoTr5 | C | - | Bright evolved companion |
| NGC1514 | C | - | |
| Legend: Ec: eclipsing. El: Ellipsoidal variation. C: composite spectrum. S1: single lined spectroscopic binary. S2: Double lined spectroscopic binary. I: irradiated secondary. | |||