Gaia22ayj: outburst from a deeply eclipsing 9.36-min binary?

Gaia22ayj is a transient detected by the Gaia Photometric Science Alerts Team¹¹1 $<$http://gsaweb.ast.cam.ac.uk/alerts/alert/Gaia22ayj/$>$. . The object was detected at 17.01 mag on 2022 March 3. Five Gaia observations between March 3.112 and 4.038 UT showed significant short-term variations between 16.55 and 17.01. The object had been registered as a variable ZTF19aagmvuk²²2 $<$https://lasair.roe.ac.uk/object/ZTF19aagmvuk/$>$. by the Zwicky Transient Facility (ZTF: Masci et al. (2019)). The same outburst was detected in outburst by ZTF and the Transient Alert Broker Lasair (Smith et al. 2019) reported $r$=16.09 on March 3.399 UT. The Gaia EDR3 position, magnitudes and parallax 08^h 25^m 26$.\mkern-4.0mu^{\rm s}$528, $-$22^∘ 32^′ 12$.\mkern-4.0mu^{\prime\prime}$34 (J2000.0), $BP$=19.570, $RP$=18.461 and $\varpi$=0.43(22) mas, respectively (Gaia Collaboration et al. 2021).

Using the ZTF public data³³3 The ZTF data can be obtained from IRSA $<$https://irsa.ipac.caltech.edu/Missions/ztf.html$>$ using the interface $<$https://irsa.ipac.caltech.edu/docs/program_interface/ztf_api.html$>$ or using a wrapper of the above IRSA API $<$https://github.com/MickaelRigault/ztfquery$>$. , I found coherent short-period variations in time-series ZTF observations on 2019 Jan. 10 (T. Kato, vsnet-alert 26674⁴⁴4 $<$http://ooruri.kusastro.kyoto-u.ac.jp/mailarchive/vsnet-alert/26674$>$. ; figure 2). Phase dispersion minimization (PDM: Stellingwerf (1978)) analysis yielded a period of 0.006500(7) d with two different maxima and minima in one cycle (figure 3). I have confirmed that this period can express the entire ZTF light curve as well. The orbital period was refined using the Markov-Chain Monte Carlo (MCMC)-based method introduced in Kato et al. (2010). The resultant ephemeris is

$${\rm Min\;I\;(BJD)}=2458699.17927(1)+0.00649910257(13)E.$$

(1)

The phase-folded light curve of the entire ZTF data (excluding the Lasair data) is shown in figure 4. This period express all the observations very well.

The resultant period of 0.0064991026 d = 9.36 min is very short. Although this could be the spin period of a binary containing a white dwarf, such as in the white dwarf pulsar AR Sco (Marsh et al. 2016; Stiller et al. 2018) and candidate white dwarf pulsars ZTF J185139.81$+$171430.3 = ZTF18abnbzvx (Kato and Kojiguchi 2021; Klingler and Pavlov 2021; Pavlenko and Sosnovskij 2021), ASASSN-V J205543.90$+$240033.5 (Kato 2021; Kato et al. 2021; Wagner et al. 2021), and more recently GLEAM-X J162759.5$-$523504.3 (Katz 2022). The profile of variations in Gaia22ayj and the constancy of the profile, however, do not look similar to those of white dwarf pulsars.

The phase-folded light curve rather suggests a $\beta$ Lyr-type or W UMa-type eclipsing binary with different minima. The slightly brighter second maximum is suggestive of the O’Connell effect (O’Connell 1951). The only such short-period (9.36 min) binaries known to us are AM CVn stars composed of white dwarfs [For a review of AM CVn stars see e.g. Solheim (2010)]. Among AM CVn stars, two short-period systems HM Cnc (5.36 min) and V407 Vul (9.48 min) are considered to be direct impact accretors without forming an accretion disk (Marsh and Steeghs 2002; Ramsay et al. 2002). The third shortest system ES Cet (10.3 min) has recently been shown to have an accretion disk (Bąkowska et al. 2021). Aside from these accreting systems, an eclipsing binary with a period of 6.91 min (ZTF J153932.16$+$502738.8) is also known (Burdge et al. 2019).

If Gaia22ayj is indeed an accreting 9.36-min eclipsing binary, this object is the shortest period known such an object (the record up to now was ES Cet). The presence of eclipses in ES Cet was confirmed only by rotational disturbance in the profiles of emission lines (Bąkowska et al. 2021), and not yet by photometry. Gaia22ayj would be very important case due to the very large amplitudes of photometric variations. Following the interpretation of an eclipsing binary, the large depths of the eclipses require that the system is seen almost edge-on and the brightness of the two components do not differ greatly. A simplified calculation suggests that a luminosity ratio of $\sim$0.4 could reproduce the observed difference in depths of the minima. With this orbital period, the Doppler beaming is expected to be an order of 1%(=0.01 mag) (Zucker et al. 2007) and would be detectable by high-precision photometry.

The presence of an outburst lasting nearly 1 d in Gaia22ayj suggests the presence of instability in the accretion disk, rather than a short-lived flare or an accretion event as seen in in intermediate polars. The detection in X-rays by the Swift X-ray telescope (1SWXRT J082526.4$-$223212: D’Elia et al. (2013)) is compatible with an accreting binary. The presence of an accretion disk makes Gaia22ayj similar to ES Cet. The disks in short-period AM CVn stars, like ES Cet, are considered to be thermally stable due to the high mass-accretion rate considering the standard evolutionary path in which the mass transfer is driven by gravitational wave radiation and the mass-losing white dwarf fills the Roche lobe (Tsugawa and Osaki 1997; Kotko et al. 2012). The case of Gaia22ayj appears to be different. If Gaia22ayj is indeed an AM CVn-like object, it might be in the turn-on phase of the mass-transfer before reaching the period minimum (Deloye et al. 2007; Solheim 2010).