FQ Circini: An Ordinary Nova with a High-mass B1 V(n)(e) Companion Whose Decretion Disk Transfers Mass to the White Dwarf via Roche-Lobe Overflow

TL;DR

FQ Cir is identified as a fast He/N nova in a close binary where a high-mass Be star (approximately $M_{\rm comp}\approx13\,M_{\odot}$) with a decretion disk transfers gas via Roche-lobe overflow into the white-dwarf’s accretion disk. The white dwarf is massive ($M_{\rm WD}\approx1.25\,M_{\odot}$) and likely of ONe composition, ruling out a Type Ia supernova progenitor; the system defines a new class of high-mass cataclysmic variables (HMCVs) and reveals a disk-to-disk accretion mechanism. The companion’s rapid rotation and decretion disk, truncated by the Roche lobe, feed the WD-driven nova, representing a novel accretion topology in interacting binaries. The study combines archival photometry, multi-wavelength SEDs, and spectroscopy to derive robust constraints on masses, radii, temperatures, and the orbital geometry, and discusses the evolutionary path and fate, including potential accretion-induced collapse. Collectively, FQ Cir establishes a prototype HMCV and highlights a new evolutionary channel for mass transfer in compact binaries with massive donors.

Abstract

FQ Cir was an ordinary fast He/N classical nova, peaking at $V$=10.9. The pre-eruption and post-eruption counterpart was at $V$=14.0, making the smallest known classical nova amplitude of 3.1 mag. The nova light and the counterpart coincide to 0.034 arc-seconds, and the counterpart is a rare hot/blue emission-line star with flickering, so the identification of the quiescent nova is certain. The counterpart is a weak Be main sequence star, B1 V(n)(e). A coherent photometric period appears in all four {\it TESS} Sectors and in the AAVSO post-eruption light curve, as ellipsoidal modulation with orbital period 2.041738 days. The companion must have been spun-up to a fast rotation, and like all Be stars, a decretion disk is exuded. With the constraints of the blackbody radius and the main sequence, the companion mass is 13.0$^{+0.2}_{-0.5}$ $M_{\odot}$, with radius 6.2$\pm$0.2 $R_{\odot}$. This is the discovery of a cataclysmic variable with a high-mass companion, a new class that we call `High Mass Cataclysmic Variables'. The white dwarf mass is 1.25$\pm$0.10 $M_{\odot}$ and must have an accretion disk that supplies fuel for the nova eruption. FQ Cir represents a new mode of accretion in interacting binaries, with Roche lobe overflow from the decretion disk feeding mass into the usual accretion disk around the white dwarf, for disk-to-disk accretion. From the mass budget of the binary, the primary star must have its initial mass $>$7.6 $M_{\odot}$, forming an ONe white dwarf, so FQ Cir can never become a Type Ia supernova.

Figures (9)

• Figure 1: Light curve for FQ Cir 1894--2025. The $V$ magnitudes represent the brightness only of the B-star, because the WD and disks are greatly fainter than the hot star. This light curve includes magnitudes from DASCH 1-year averages (blue diamonds), ASAS 0.1-year averages (green circles), APASS magnitudes (black squares), ASAS-SN 0.05-year averages (orange plus signs), AAVSO 0.05-year averages (magenta dashes), and TESS 1-day averages (red triangles). The two red lines show the two most prominent trends that last longer than one decade. The six labeled flares show the cases of well-resolved and significant flares, variously month-long and decade-long. The 2022 nova eruption is indicated with the red up-arrow. One of the points from this light curve is that the 14th-mag star has long-term trends and flickering on all time scales, with this being impossible for all normal B-stars, yet such is seen in all Be stars. Such a rare star is incredibly unlikely if the 14th-mag star is a random foreground star. Another point is that the star is not systematically brightening at an accelerating rate, so this indicates that the high-mass-ratio binary is not accreting by ordinary Roche lobe overflow from a companion filling its Roche lobe (because such would require runaway accretion).
• Figure 2: TESS light curve for FQ Cir. These three panels plot the TESS normalized flux for Sector 12 (top panel), Sectors 38 and 39 (middle panel), and Sector 65 (bottom panel). These are smoothed with a simple running average. The dominant feature is that of $\sim$5% flickering on timescales of hours, days, and weeks. Such behavior is never seen for normal B-stars with no emission lines. Such behavior is always seen for Be stars. That a unique hot-blue star happens to be within 0.034 arc-seconds of the nova position is highly improbable -- unless the star is the nova counterpart. Further, the TESS light curves show a highly significant photometric periodicity (1.020869$\pm$0.000018 days) superposed on the flickering. The thin vertical lines indicate the times of minimum light according to the best fit ephemeris, with epoch BJD 2459740.9710$\pm$0.0065. In general, the times of local minima are close to the lines, while the local maxima are between the vertical lines, matching the ephemeris. Nevertheless, often there is little apparent connection with the ephemeris. The periodicity is highly significant, see the Fourier transform in Figure 3, so the ubiquitous flickering is often hiding the sinewave modulation. The 1.02 day photometric modulation is caused by ordinary ellipsoidal modulation, for which the orbital period is twice the photometric period, so $P$=2.04 days. See the average folded light curve in Figure 4. This periodicity is coherent and stable over 6.3 years, and the only such clock is the orbit.
• Figure 3: Fourier transform of the TESS light curve. The primary point of this figure is that a highly significant Fourier peak (at a frequency of 0.98 cycles/day, with photometric period 1.02 days) appears centered on the nova pixel. This figure is shown so that it can be seen that the Fourier peak is isolated and high above the local background noise level, demonstrating that the peak is highly significant. This high significance is also proven by the chi-square fit to the light curve (c.f. Figure 4). This same peak is independently visible with the same strength and period in each of TESS sectors 12, 38, 39, and 65. No significant Fourier peak is seen for the double-period of 2.04 days (for a frequency of 0.49 cycles/day). The Fourier peak could arise from an orbital period of 1.02 days with irradiation and hotspot effects dominating over the ellipsoidal effect, or from an orbital period of 2.04 days with the ellipsoidal modulation dominating over any irradiation and hotspot effects. With the hot B-star swamping the irradiation and hotspot effects, the ellipsoidal mechanism must be dominating, so the orbital period is actually $P$=2.04 days. This plot also shows another significant peak for a period of 2.26 days, also centered on the nova pixel, with the folded light curve showing the characteristic shape of a Cepheid, so this is just a random foreground variable star near to the nova. For frequencies lower than 0.4 cycles/day, the background noise starts rising due to the chaotic flickering on longer time scales.
• Figure 4: Folded light curve for the orbital periodicity. Here, the TESS magnitude is the flux converted to magnitude with the average flux corresponding to 14.00 mag. The light curves for individual cycles have their baselines varying from 13.92 to 14.08 because the counterpart is flickering up and down. But when the 50 cycles are averaged together, a blatant sinewave periodicity is seen. The photometric period is 1.020869 days for a simple sinewave, but the orbital period is twice that at $P$=2.041738 days for a double sinewave. For FQ Cir, the ellipsoidal mechanism makes for two identical photometric minima each orbit, at the times of the inferior and superior conjunctions. This figure plots the phase-averaged folded light curves in bins of 0.04, with the individual points having error bars of $\pm$0.002 mag. The phase is calculated with the orbital period of 2.041738 days with a zero-phase for an epoch of minimum light of BJD 2459740.9710. Each point is plotted twice, once with phase 0.00--1.00, and a second time with a phase 1.00 larger.
• Figure 5: Spectral energy distribution for the quiescent counterpart of FQ Cir. The flux (in units of Jansky) is plotted versus the central photon frequency, $\nu$. These data are from the infrared with WISE (burnt-red diamonds), the near-infrared with 2MASS (red squares), the optical with APASS (green circles), and the ultraviolet with Galex (the blue square). The data sets do not follow any one smooth curve due to the known variability and the data sets being taken in different years. Nevertheless, there is a clear Rayleigh-Jeans slope throughout the infrared, and there is a turnover towards the ultraviolet. The Rayleigh-Jeans slope shows that the SED is dominated by a blackbody, and the blue-turnover shows that the surface temperature is 20,000$\pm$3000 K. Further, we see no evidence of any accretion disk or dust shell or decretion disk, and we do not expect any because the Be-star dominates over any plausible emission.