An extremely fast fading population II dwarf nova candidate: caught spectroscopically on the rise

TL;DR

AT2022kak is a rapidly evolving, blue transient discovered by KNTraP that rose by several magnitudes and faded within days in 2022, with a second outburst observed in 2025 enabling time-resolved spectroscopy. Light-curve analysis places it among the fastest fading dwarf novae, with $t_2$ around $1.16$–$1.28$ d and an approximate orbital period of $P_ ext{orb} \approx 1.4$ h, while spectroscopic data show Balmer absorption and He features consistent with DN outbursts. Distance and vertical location ($D_\odot \approx 6.2$ kpc, $z \approx 1.9$ kpc) position AT2022kak in the Galactic thick disk, making it a strong Population II DN candidate and a valuable probe of accretion in metal-poor environments. The study underscores the value of day-cadence surveys like KNTraP for uncovering fast, faint transients and motivates continued, high-cadence spectroscopic follow-up of similar events with next-generation facilities.

Abstract

We present AT2022kak, a rapidly evolving optical transient discovered by the KiloNova and Transients Program (KNTraP). This interesting burst exhibited extremely fast evolution, with a large amplitude blue outburst of m > 3.3 in a single night, and a rapid fade back to quiescence in the following two nights. We deployed a multi-wavelength follow-up campaign, monitoring the object for the next two months, but saw no recurrent burst. Three years later, while observing to get spectroscopy of the object in quiescence, there was a new outburst, enabling the collection of time-resolved spectra of the rise and fade of the outburst. The light curve properties of the first burst and spectra of the second burst are consistent with a dwarf nova. Its fast evolving behaviour makes it one of the fastest and faintest dwarf novae observed. The estimated distance of AT2022kak from the Galactic centre is ~6.6 kpc, with a scale height of ~2 kpc. This scale height places it in the Galactic thick disk, where only very few dwarf novae have been found, and is therefore a potential Population II dwarf novae system.

Figures (9)

• Figure 1: The optical-infrared light curve and spectroscopic follow up of AT2022kak over $\sim$3 years. KNTraP DECam i and g filter photometry is shown as red and blue filled circles, respectively. Follow-up photometry from GROND, Zadko, and Swift are denoted with filled squares (see legend), with upper limits shown as inverted triangles. Down-pointing arrows indicate the nights during which spectroscopy was acquired with SOAR (dotted arrow), KOALA (solid arrows), and SALT (dashed arrow). Note: light curves include quiescent source flux.
• Figure 2: KNTraP images of AT2022kak during its initial outburst. From left to right: the template image (first night's data), the science image during the peak-burst emission on night 5 (MJD 59627.27), and the difference image. The green circle indicator has a radius of 2 arcseconds.
• Figure 3: Top: Light curve of the February 2022 outburst and best-fit polynomials of fourth order. The polynomial fit is only applied to outburst nights 3 -- 6. Bottom: Light curve (g-i) colour evolution.
• Figure 4: Distribution of dwarf novae plotted as a function of their amplitude of outburst in magnitudes versus $t_2$ in days. The $t_2$ value of AT2022kak in the g-band is indicated with a star marker. Contours indicate 10%, 50%, and 90% density levels. Data is taken from Kawash:2021 and magnitudes are measured in ASAS-SN g-band.
• Figure 5: (Top) Panels from left to right shows KOALA spectra in multiple stacked sets to show nightly evolution from observing nights one, two, and three (MJD 60731.56, 60732.55, 60733.51). Night one and three have three 20 minute exposures per stacked set, except set 1 night one, which has two. Night two has a stacks of two and three exposures for sets 1 and 2 respectively. (Bottom) Stacked and normalised spectra from night one, showing four zoomed in features expected of a DN; Balmer series ($H_{\beta}, H_{\gamma}, H_{\delta}$) and He II. Vertical lines expected to be absorption DN features marked with dashed lines, emission features marked with solid lines.