An eclipsing 8.56 minute orbital period mass-transferring binary

TL;DR

AM CVn binaries test ultra-compact binary evolution and helium-donor physics while supplying millihertz gravitational waves for LISA. This work reports the discovery of ATLAS J1013-4516, an 8.56-minute eclipsing, disk-accreting AM CVn, characterized with multiwavelength data and a decade-long timing baseline. The measured orbital-period derivative is $\dot{P}=(-1.60\pm0.07)\times10^{-12}\;\mathrm{s\,s^{-1}}$, and modeling under gravitational-wave angular-momentum loss with conservative mass transfer yields component masses $M_1\approx0.87^{+0.36}_{-0.25}\,M_\odot$ and $M_2\approx0.10^{+0.03}_{-0.02}\,M_\odot$, a donor temperature $T_2\approx6550^{+70}_{-40}$ K, and a distance $d\approx700^{+35}_{-30}$ pc. The characteristic gravitational-wave strain is $h_c\approx(2-4)\times10^{-20}$, giving a four-year LISA SNR of $\mathrm{SNR}\approx20$–$30$. ATLAS J1013-4516 thus provides a precise, eclipsing benchmark for the interplay of gravitational radiation and mass transfer in the shortest-period AM CVn regime and a southern-sky verification target for LISA.

Abstract

We report the discovery of ATLAS J101342.5-451646.8 (hereafter ATLAS J1013-4516), an eclipsing, mass-transferring AM Canum Venaticorum binary with an 8.56-minute orbital period, identified via periodic variability in light curves from the Asteroid Terrestrial-impact Last Alert System survey of Gaia white dwarf candidates. Follow-up spectroscopy with the Large Lenslet Array Magellan Spectrograph reveals a helium-dominated accretion disk, while high-speed photometry with ULTRACAM shows pronounced primary and secondary eclipses. We construct a decade-long timing baseline using data from ATLAS, Gaia, ULTRACAM on the Gran Telescopio Canarias, and the proto-Lightspeed instrument on the Magellan Clay telescope. From this baseline, we measure an orbital period derivative of dP/dt = -1.60 +/- 0.07 x 10^-12 seconds per second. Interpreted in the context of stable mass transfer, the magnitude and sign of the period derivative indicate that the orbital evolution is governed by the interplay between gravitational-wave-driven angular momentum losses and mass transfer, directly probing the donor star's structural response to mass loss. Assuming angular momentum loss dominated by gravitational wave emission, we constrain the component masses and infer the characteristic gravitational wave strain of the system for future space-based observatories such as the Laser Interferometer Space Antenna. We predict a characteristic strain corresponding to a four-year signal-to-noise ratio greater than approximately 20, establishing ATLAS J1013-4516 as a strong prospective source for probing long-term orbital evolution in the mass-transferring regime.

Figures (10)

• Figure 1: Phase-folded light curves of ATLAS J1013$-$4516 from synoptic surveys, folded on the timing ephemeris listed in Table \ref{['tab:inferred']} with phase zero defined at mid-eclipse. Cyan ($c$, 420–650 nm) and orange ($o$, 560–820 nm) points show phase-binned ATLAS photometry, overlaid with Gaia epoch photometry. All data are binned using inverse-variance--weighted averaging and repeated over three orbital cycles for visualization. The ATLAS light curves are linearly rescaled to match the interquartile range and median level of the phase-binned Gaia photometry. The resulting $\sim$50% eclipse-like attenuation and coherent periodic modulation motivated high-speed follow-up.
• Figure 2: Coadded, continuum-subtracted Magellan/LLAMAS spectra of ATLAS J1013$-$4516 obtained on 2025 December 14. The displayed wavelength range is restricted to the highest signal-to-noise, non--sky-dominated region of the LLAMAS coverage. Right panels show phase-resolved He ii line profiles, constructed by binning the spectra into four equal orbital phase bins and arbitrarily offset vertically for clarity. Both the phase-binned line profiles and the coadded spectrum have been smoothed with a Gaussian kernel for visualization.
• Figure 3: ULTRACAM light curves of ATLAS J1013$-$4516, folded on the orbital period. The $u_s$, $g_s$, $r_s$, and $i_s$ light curves shown here are constructed from multiple nights of observations using different filter triplets ($u_sg_sr_s$ and $u_sg_si_s$ configurations). Central wavelengths are 3520, 4729, 6196, and 7708 Å for $u_s$, $g_s$, $r_s$, $i_s$, respectively. Each panel includes insets (right columns) that zooms in on the primary and secondary eclipses. Gray and purple shaded regions mark the primary and secondary eclipse windows, respectively. The insets highlight the wavelength dependence of the eclipse depths: the primary eclipse depends towards shorter wavelengths, reflecting the dominance of a hot, blue accretion component, while the secondary eclipse becomes more pronounced at longer wavelengths as the cooler, red-emitting donor contributes more strongly. Fluxes are normalized by the out-of-eclipse median.
• Figure 4: Phase-folded proto-Lightspeed g'-band light curve of the AM CVn system ATLAS J1013$-$4516 (top; green points), repeated over three orbital cycles for visualization. The top panel shows binned photometry (green) and the best-fitting lcurve model overplotted (black). The bottom panel shows the corresponding additive model components, including the accretion disk (blue), donor star (red), accretor (brown), hot-spot contributions (orange), and an asymmetry term (purple). The sum of these components reproduces the total model light curve shown in the top panel.
• Figure 5: Comparison of ULTRACAM light curves of ATLAS J1013$-$4516 with short-period, disk-accreting AM CVn binaries ZTF J0546+3843 and ZTF J1858+2024. Each row shows simultaneous multi-band photometry phase-folded on the orbital period. The dashed line marks the phase of maximum $u$ flux. All three systems display deep disk eclipses, but the phase and amplitude of the blue hump differ, reflecting variations in hot-spot geometry and stream-impact location within the disk.