4.1. Evidence for multiple outbursts
Our data implies that, within the recent history of the AFGL 961 region, there have been several phases of activity resulting in the ejection of matter which, after subsequent interaction with ambient cloud material or older ejecta from the sources, have resulted in the creation of the shock-excited H2 emission nebulae seen in Fig. 1. The most distant H2 feature observed (CA 6) is 0.85pc from the binary. Clearly it is a difficult task to estimate outflow velocities without quantitative information. If we assume a typical ejection velocity of 200km s-1, the derived dynamical age of this feature is 4000 years. Support for this assumed ejection velocity comes from proper motion studies of HH knots (e.g. Burrows et al. 1996 for HH 30) where typical values for optical knot velocities of 200km s-1 have been found. Even though the energy source of HH 30 is a low-luminosity object and AFGL 961 is of considerably higher luminosity, we adopt this typical tangential velocity for the sake of our simple dynamical timescale estimates but note that our assume velocity could be in error by a factor of possibly 2-3. For CA 1, the largest and most bow-like of the H2 features, we obtain a dynamical age of 1200 years. For CA 3, the bow shock structure lying to the south of the binary, a dynamical age of 800 years is found. These simple timescale estimates, coupled with the multiple outflow axes found, suggests that multiple outburst events have obviously occurred in AFGL 961 over a period of at least the last 4000 years. Performing the same calculation for all observed H2 nebulae suggests a typical inter-outburst timescale of between 500 and 750 years. In optical HH objects, a typical timescale for the formation of working surfaces/bow shocks is thought to be 500 years between events. In low-mass PMS objects such as T Tauri stars, these outbursts are being linked to disk accretion events resulting in outbursts of the FU Orionis (FUor) type (Reipurth & Aspin, 1997). AFGL 961 is considerably more luminous than either the typical energy source of an optical HH object or a typical FUor, specifically, 7500 whereas a typical HH energy source (HHES) has 1000 and can be as low as a few (e.g. 5). One exception to this is the exciting source of the large-scale HH flow HH 80/81 (Reipurth & Graham, 1988) namely GGD 27 (Aspin et al. 1991) which has 6000. The most luminous FUor found to date is Z CMa (the energy source of a HH 160) which has 600. It seems however, that the inter-outburst timescales for all these classes of objects are perhaps related and not obviously dependent on source luminosity.
Clearly our assumed ejection velocity of 200km s-1 critically effects the derived outburst timescales. Also, the relation between physical ejection velocity (of gas) and shock velocity (producing H2 emission) is not a trivial matter since it is related to the velocity difference between the outflowing material and that of the material upon which it impacts. In the case of impact on previous ejecta from the originating source, the shock velocity can be lower than the ejection velocity. H2 dissociation would occur in regions with shock velocities 50km s-1 so for planar shocks 200km s-1 is a clearly impractical. However, often found in reality are oblique shocks whereby the shock direction is not orthogonal to the `outflow' direction. In these cases only the orthogonal component of the shock velocity need be considered. This would clearly lower the effective shock velocity. Clearly, a through observational study of the velocity structure of both the outflowing gas and the gas upon which it is impacting is required before further conclusive comparative work be undertaken. What we can state here is that eruptive events similar in timescale to those occurring in FUors involving variable accretion rates, the buildup of disk material and eventual mass-dumping from disk to photosphere resulting in a significant luminosity increase and ejection of matter/shocks manifesting as HH objects and jets (and in embedded sources, NIR shock-excited structures), are consistent with what we have found observationally in AFGL 961.
4.2. Evidence for non-axially symmetric outflow axes
What is also obvious from the structure seen in our images is that the repetitive ejection of matter resulting in the shock-excited H2 emission observed has clearly occurred along more than a single axis originating at the binary components i.e. the activity has been non-axially symmetric over time. One axis can be traced north-south from CA 1 through AFGL 961b to CA 3. Another axis is clearly defined from CA 6 through CA 7, through the binary system and onto CA 5. Other single feature axes are also defined from the binary towards both CA 4, and CA 8. Both of these outburst axes have no detectable counterparts on the opposite side of the binary. It is unclear whether the H2 seen closeby AFGL 961 W is associated with AFGL 961 or whether it is resultant from activity in AFGL 961 W itself, however, from the clearly defined axes present around AFGL 961 we can trace four distinct directions along which activity has been directed. At the present time, only AFGL 961b shows direct observational evidence for activity (in the form of the weak CO bandhead emission at 2.294 µm) which when viewed in combination with the thermal infrared excess exhibited (Castelaz et al. 1985) and the fact that the extinction towards the source is relatively low (AV=16.7 magnitudes (Castelaz et al. 1985) indicates the presence of a dense, hot, inclined accretion disk. AFGL 961a was found to be coincident with a compact HII region (0.59" Bally & Predmore 1983) suggesting that it also has significant (ionized) circumstellar material possibly also in the form of an accretion disk. AFGL 961a clearly possess a strong thermal excess (see Fig. 3) and has a line-of-sight extinction similar to that of AFGL 961b (i.e. AV=17 magnitudes; Castelaz et al. 1985). It is not possible to determine the precise origin of the outflow axes within the binary since the H2 features seen are relatively diffuse with no associated jet-like structures as often seen in the optical. Since multiple outflow axes are present, it would seem feasible to assume that at one time both AFGL 961a and AFGL 961b have contributed to the shock-excited emission observed. One possible scenario for the multiple outflow axes is that binary orbital motion coupled with accretion disk axial precession have caused significant deviations of the polar (outflow) axes of the stars with time. Alternatively, if both stars have produced outflow activity in the past then their respective polar/outflow axes are not aligned and not orthogonal to the binary orbital plane.
© European Southern Observatory (ESO) 1998
Online publication: June 26, 1998