The 1995 Superoutburst of AL Com

The section "Early VSNET Postings!" has been moved to the later part of this page.

Poster Presented at Keele CV Conference

(The object was selected as "the star of the year" in this conference!)

Observation of WZ Sge-type Dwarf Nova AL Comae Berenices

T. Kato (1), D. Nogami (1) , H. Baba (1), K. Matsumoto (2), J. Arimoto (2), K. Tanabe (2), K. Ishikawa (2) (1. Department of Astronomy, Kyoto University, Sakyo-ku, Kyoto 606-01, Japan, 2. Astronomical Institute, Osaka Kyoiku University, Asahigaoka, Kashiwara, Osaka 582, Japan)

1. Introduction

Although presently classified as a dwarf nova, WZ Sagittae (WZ Sge) is peculiar in many respects. The star exhibited three historical outbursts separated by 33 years. The light curves of its ourbursts resemble a fast nova in that they show (1) rather rapid initial decline, (2) subsequent slower decline which sometimes lasts 60 to 100 days, and (3) large (about mag) outburst amplitudes. From these features, this object had been long believed to be a recurrent nova.

However, spectroscopic and photometric observations in 1978 outburst confirmed the object to be a dwarf nova, not a classical nova (Patterson et al., 1981, and references therein). Most striking was the discovery of superhumps which are one of the defining characteristics of SU UMa-type dwarf novae.

There are two major competing theories to explain the peculiar nature of WZ Sge. One is the mass-transfer burst theory, which is supported by the discovery of large-amplitude periodic humps early in its 1978 outburst (Patterson et al. 1981). The period of the humps was equal to the orbital period, and they were considered to reflect the hot spot enhanced by the mass-transfer burst, although the hump maxima occured 0.17 orbital phase prior to the orbital humps in quiescence.

The other is the extension of thermal and tidal instability theory of SU UMa-stars towards the lowest mass-transfer rate (Osaki 1995). Numerical simulations have shown that thermal instability of the accretion disk occurs rarely under combination of very low mass-transfer rate and very low viscosity of the accretion disk during quiescence. It is shown that a normal outburst in such condition always leads to the tidal instability to trigger a superoutburst.

We here report new findings and its implications obtained through the intensive observation of the very recent spectacular outburst of AL Com, which has proven to be an almost perfect twin to WZ Sge in many respects.

2. 1995 Superoutburst of AL Com

AL Com was discovered by Rosino during its outburst in 1962. Its dwarf nova nature was confirmed by subsequent spectroscopy during outburst, and by discovery of historical outbursts back to 1892. The existence of a deep dip in the middle of the fading stage was noticed during this outburst (Bertola 1964). The next impressive outburst was observed in 1975, when the object showed a similar long-lasting outburst and possible re-brightening after the decline to near quiescence. Its exceptionally large outburst amplitude (about 8 mag) for a dwarf nova has attracted a number of investigators. For example, Howell and Szkody (1987) showed a large amplitude variation with a period near 40 min. They suggested from this period that AL Com may belong to either DQ Her-type magnetic systems or AM CVn-type double degenerate systems. Further spectroscopic observation by Mukai (1990) precluded the latter possibility. More extensive photometry by Abbott et al. (1992) showed two distinct periodicities of 41 min and 87 -- 90 min, the latter was suggested to be the orbital period, and the former the rotation period of the white dwarf. From these periodicities they concluded that AL Com bears the properties of both an enigmatic dwarf nova WZ Sge and a unique intermediate polar EX Hya.

After 20 years' dormancy, AL Com was again observed by D. York (AAVSO) during its rise to an outburst in early April 1995. The news was soon relayed through Internet, accompanied by a burst of e-mails full of astonishment. The object reached its historical record of mv=12.2 at maximum, and started fading. We then started time-resolved CCD photometry during the full course of the outburst. Most of the observations were done by a 60-cm telescope at Ouda Station (Kyoto University) and by a 50-cm telescope at Osaka Kyoiku University. The overall light curve is given in Fig. 1.

(General V-band light curve of AL Com during the 1995 outburst. Large dots represent CCD observations by us, and small dots visual observations reported through VSNET. Note the exceptionally long duration of the outburst interrupted by a deep dip.)

2.1 Early Stage of Outburst

Within two days of the outburst maximum, periodic modulation was discovered independently from many observatories. The first night of the Ouda data showed doubly humped modulation (Fig. 2, Apr. 7).

(Time-resolved light curve obtained at the earliest stage of the outburst. Low-amplitude, doubly-peaked periodic modulation is evident.)

This feature showed a continuous decaly until Apr. 12. A period analysis using PDM (Phase Dispersion Minimization) gives a very stable period of 0.05666 +/- 0.00002 day (Fig. 3).

(Folded light curve of the early stage modulation. We interpret the best estimate of the period (0.05666 day) as the orbital one (this modulation therefore may be called "orbital superhumps"). Very similar features have been also observed in WZ Sge and HV Vir.)

This period was later confirmed to be exactly twice (within the error of each estimates) of 41-min periodicity observed during quiescence. From this perfect agreement and later discovery of typical superhumps with a significantly longer period, we have identified this period as the orbital period. A similar feature was discovered in WZ Sge, which was assumed to originate from the enhanced hot spot (Patterson et al. 1981). The suggested orbital period differs from that of WZ Sge only by 2 sec!.

2.2 Appearance of Superhumps

After the decay of initial periodic modulation, a new distinct feature appeared. The Ouda data on Apr. 17 showed a large amplitude modulation (Fig. 4)

(Time-resolved light curve obtained 12 days after the onset of the outburst. Large-amplitude hump features characterstic to superhumps are evident.)

whose profile was very characteristic to superhumps in usual SU UMa-type dwarf novae. A period analysis yielded a period of 0.05722 +/- 0.00010 day, which is 1.0 +/- 0.2% longer than the "orbital superhumps" observed in the early stage of outburst. The superhump profile is shown in Fig. 5.

(Folded light curve of superhumps. The superhump period (0.05722 day) is longer than the suggested orbital period by 1.0 \%. These observations first firmly established the SU UMa-type nature of AL Com.)

This observation first firmly established the SU UMa-type nature of AL Com. After the development of superhumps, the object entered a stage of more gradual decline. The outburst amplitude and historical behavior, the course of the outburst, orbital and superhump periods, and the time lag before development of superhumps already made us convincing that AL Com is a perfect twin of WZ Sge.

2.3 The "Dip"

The "dip" is one of the unique characterstics of the outburst light curve of WZ Sge. The dips occur in the middle stage of the outburst, and are characterized by a short (a few days) excursion to the faint state, although still enough brigter than usual quiescence. There seems to be some evidence that such dips are rather common among SU UMa-type dwarf novae with very large outburst amplitudes. The same feature was naturally sought in the case of AL Com. The dip actually occurred as expected (Fig. 1).

Although our observations were unfortunately hindered by the unfavorable weather, the latter half of the main dip was clearly visualized. Fllowing the dip, AL Com brightened rather slowly to the second maximum, then started a rapid fading. During this stage, we could no longer find any evidence of superhumps or orbital humps. After this "second dip", the object showed a recovery which led to a plateau stage.

2.4 The Plateau Stage

After recovery from dip episodes, the time-resolved light curve of AL Com became flat again. However, within a week from recovery, the object again started to show a low-amplitude modulation. An analysis gives a period consistent with the superhump period observed earlier (Fig. 6, Osaka Kyoiku University data).

(Folded light curve during the "plateau" phase, after the dip. The best period agrees with the superhump period within the error of the estimate. From the fact that the superhumps growed after the dip, we interpret the "plateau" phase as a new superoutburst.)

This finding first clearly demonstrate that "after-the-dip" behavior traces the usual course of a superoutburst.

2.5 The Final Decline

After the plateau phase with superhumps, AL Com started to show rather irregular variability with a time-scale of hours to a day. The superhumps seemed to have completely disappeared (Fig. 7).

(Time-resolved light curve just before the final decline. The superhumps have completely disappeared.)

Soon after this stage, the object showed a precipitous decline. The quiescent magnitude was not yet reached during our observations. AL Com stayed at about V=18.6 at least for some days after the rapid decline.

3. Discussion

The present observations clearly show that AL Com is a genuine member of SU UMa-type dwarf novae, and that this object is almost a twin of an enigmatic dwarf nova WZ Sge in many respects. We here try to explain several peculiar features common to these objects (hereafter "WZ Sge-type dwarf novae", cf. Bailey 1979).

3.1 Origin of Early Stage Modulation

As described in Chap. 2.1, we consider the period of this modulation has the same period with the orbital period. In the case of WZ Sge, the singly humped profile suggested the hot spot as the origin. However, in the case of AL Com (and HV Vir in our study), the profile of this modulation is doubly peaked, and sometimes dominated by dip features rather than humps. This finding seems to preclude the hot spot as the source of light. A possibility of reprocessed light by the secondary is also precluded, because this interpretation should predict a sinusoidal light variation. We therefore suspect that this modulation originates from the accretion disk itself (either by changing energy output from the disk or the effect of geometrical obscuration). In either cases, we have no need to think of enhanced mass-transfer as was assumed in explaining WZ Sge-type phenomenon. One attractive possibility is that this modulation represents "immature" superhumps. Numerical simulations have shown that during the growing stage of the tidal instability, the period of superhumps tends to be close to the orbital period. It is, however, a mystery why such growing stages are observed only in WZ Sge-type dwarf novae.

3.2 Superhumps

Fully grown superhumps were observed in AL Com. The good seasonal condition allowed us unambiguously determine the period. The superhump period of AL Com is 1.0% longer than the suggested orbital period. This value of superhump excess is the second smallest among SU UMa-type dwarf novae (the smallest: WZ Sgr 0.8% *1). Since the superhump excess is known to correlate with the binary mass ratio, the mass ratio of AL Com should be the second extreme one among established SU UMa-type dwarf novae. A small difference from WZ Sge in this respect may be a cause of small difference in the outburst behavior between these two systems.

(*1: Note that this value has been later found to be incorrect. cf. vsnet-alert 6285)

3.3 The Dip

We don't have enough observational evidence about the origin of the dip. However, the reported rate of decline of AL Com from reported observations (mostly from visual observers of AAVSO) suggests that the dip may be produced by a thermal transition wave starting from the outer accretion disk. The magnitudes during the dips were well above quiescence in both AL Com and WZ Sge. This implies that the inner part of the accretion disk remained hot even during the dip. The recovery from the dip was rather slow. This implies that the recovery represents the "inside-out" type propagation of the thermal transition wave. Nonexistence of orbital humps during this stage strongly supports that the recovery from the dip was not powered by a mass-transfer event, but by the thermal instability of the accretion disk.

3.4 The Plateau Phase

After the dip, AL Com showed a short-living maximum followed by a rapid decline. After this "secondary dip" the star entered a new stage of the "plateau phase". The same features can be traced in the light curve of WZ Sge during 1978 outburst. As already described in Chap. 2.4, the superhumps growed during this phase. We can therefore identify this stage as a start of a new superoutburst. A short-living maximum after the main dip may represent a normal outburst which triggered this superoutburst.

3.5 The Final Decline

The final decline can be again considered to represent the propagation of a thermal transition wave through the accretion disk. The disappearance of superhumps a few days before this decline implies that the cessation of the tidal instability is responsible for the decline. After the decline, AL Com stayed about 2 mag brighter than the reported quiescence. Although the observational materials are limited due to its faintness, it is likely that some inner part of the accretion disk remained hot even after the final decline. This mechanism may be responsible for the long fading tail of WZ Sge, which still remains poorly understood. A similar phenomenon was also observed after the main (or super-) outburst of GRO J0422+32, a USXT (ultrasoft X-ray transient) which bears a number of similarities with WZ Sge-type dwarf novae. Measurement of the contribution of the accretion disk and the hot spot to the observed light may be a clue in understanding this poorly understood unique feature of WZ Sge-type dwarf novae and USXTs. The persitent "superhump" periodicity observed in HV Vir, another object very similar to WZ Sge, during the same stage by Leibowitz et al. (1994) supports a significant contribution to the light from the accretion disk this stage. [Although this modulation was originally ascribed to an orbital modulation (Leibowitz et al. 1994), our new analysis strongly favors the persistent superhumps.]