Increasing Activity in T CrB Suggests Nova Eruption Is Impending

T Coronae Borealis (T CrB) is the nearest symbiotic recurrent nova (RN); a binary system in which a white dwarf (WD) accretes from its red giant companion through a disk. Twice in the last two centuries, the accreted material ignited on the surface of the WD via runaway thermonuclear fusion reactions and produced a nova eruption. In T CrB, outbursts were recorded in 1866 and 1946, suggesting a recurrence time of ∼80 yr. The distance as determined from Gaia data is ${806}_{-31}^{+33}$ pc (Bailer-Jones et al. 2018). As in the other symbiotic RNe RS Oph, V745 Sco, and V3890 Sgr, the WD mass is required to be well above 1 M_⊙ to generate repeated nova eruptions in less than a century, and so T CrB is a candidate supernova Ia progenitor (e.g., Shahbaz et al. 1997; Fekel et al. 2000). Being the nearest known RN, we expect the next eruption of T CrB—predicted to happen between 2023.6 ± 1 (Schaefer 2019) and approximately 2026—to shine strongly from γ-rays to radio wavelengths, providing detailed information about the mass, structure, energetics, and perhaps driving mechanism of the outflow.

The nova recurrence time is inversely related to the WD mass and the accretion rate. In the case of a system with the parameters of T CrB, M_WD of 1.2–1.37 M_⊙ (Belczynski & Mikolajewska 1998; Stanishev et al. 2004) and recurrence time of about 80 yr, theoretical models by Prialnik & Kovetz (1995) predict that an accumulated mass of 10⁻⁶ M_⊙ will be needed for a thermonuclear runaway (TNR). Over 80 yr that implies an average accretion rate of between 10⁻⁸ M_⊙ yr⁻¹ and 10⁻⁷ M_⊙ yr⁻¹.

But measuring the accretion rate in symbiotic binaries is a difficult task. Although in cataclysmic variables the optical brightness can be used as a proxy for the accretion rate, in symbiotic stars, where the contribution to the optical broadband light of the nebulae and the red giant are not negligible, optical photometry does not provide an actual measurement of the accretion rate.

In the case of T CrB, Selvelli et al. (1992) and Stanishev et al. (2004) estimated the rate of accretion onto the WD (${\dot{M}}_{\mathrm{WD}}$) from spectra obtained with the International Ultraviolet Explorer satellite between 1978 until 1990. Including periods of both high and low ultraviolet (UV) flux between 1978 and 1990, they found an average accretion rate of 9.6 × 10⁻⁹ (d/806 pc)² M_⊙ yr⁻¹. By modeling the spectral energy distribution in the UV region, Stanishev et al. (2004) obtained a high-state accretion rate (from 1980 to 1988) of 1.1 × 10⁻⁸ (d/806 pc)² M_⊙ yr⁻¹ and a low-state accretion rate (from 1978 to 1980 and 1988 to 1990) of 1.5 × 10⁻⁹ (d/806 pc)² M_⊙ yr⁻¹. It should be noted that the high state identified by Stanishev et al. (2004) did not reach B-magnitudes as bright as the ones discussed here (see Section 3).

Additional constraints on $\dot{M}$ onto the WD in T CrB come from the fact that the boundary layer is typically optically thin, making it a hard X-ray source in quiescence, with the strength of the hard X-ray emission directly related to $\dot{M}$ (Luna et al. 2008; Kennea et al. 2009). To date, this is a unique feature among all known symbiotic RNe. Suzaku observations in 2006 allowed us to measure ${\dot{M}}_{q}$ = 0.7 × 10⁻⁹ (d/806 pc)² M_⊙ yr⁻¹ (Luna et al. 2019), where ${\dot{M}}_{q}$ is the accretion rate measured between nova eruptions. Clearly both UV and X-ray measurements indicate that ${\dot{M}}_{q}$ was low since 1979 until 2006 when compared with the predicted average accretion rate necessary to trigger a nova outburst every 80 yr.

We base the findings described below on publicly available data from two archives—the Digital Access to a Sky Century @Harvard (DASCH; Grindlay et al. 2009) project to digitize the Harvard Astronomical Photographic Plate collection, which provides a photometric database with a baseline of about 100 yr; and B-band observations from the archive of the American Association of Variable Star Observers (AAVSO). Details about the DASCH photometric pipeline can be found in Tang et al. (2013). The DASCH database contains unflagged observations of T CrB from 1901 April through 1989 May with non-uniform sampling.

3.1. Optical High State Leading up to the 1946 Nova Eruption

By querying the DASCH and searching for photometric observations of T CrB previous to the 1946 eruption, we found clear evidence in the B-band light curve for an optical bright state that started in 1938 and lasted about 7 yr, until about 1945, about one year before the nova eruption, and then continued after the nova event. Figure 1 shows the DASCH light curve (blue dots). This light curve confirms the phenomenon mentioned in an abstract by Schaefer (2014), who described finding such a high state associated with both the 1866 and 1946 nova eruptions. However, the DASCH data revealed that the high state, which continued after the eruption for several years, had been missed in the previous studies. The visual AAVSO data suggested that a minor precursor event occurred before each eruption that lasted approximately 1 yr. But our examination of the AAVSO light curve revealed that its coverage around the time of the two novae was too sparse to reveal the full high state that is evident in the DASCH data. The DASCH light curve also shows that after ∼7 yr of "super-active" state, T CrB faded and almost reached the pre-"super-active" brightness for about 1 yr, after which time the nova eruption occurred.

Zoom In Zoom Out Reset image size

3.2. Similarity to the Current High State

The historical light curves show that T CrB experiences two active states between nova eruptions: one is the so-called "super-active" state (Munari et al. 2016), which occurs between ∼8 and ∼1 yr before the nova eruption and another one, less noticeable, which starts about 200 days after the nova eruption and lasts for about 8 yr. Although previously noted in an abstract by Schaefer (2014), to our knowledge, the DASCH data show this clearly for the first time.

We propose that both high states play a significant role in the further development toward the nova eruption. Outside these states, the quiescent accretion rate typically seems to be about 10⁻⁹ M_⊙ yr⁻¹; at this level it would not be possible to reach the required ignition mass of a few 10⁻⁶ M_⊙ in approximately 80 yr. If the accretion rate during the high states is about 5 × 10⁻⁹ M_⊙ yr⁻¹ to 5 × 10⁻⁸ M_⊙ yr⁻¹, then an order of magnitude estimation yields that about 10⁻⁶ M_⊙ could be accreted onto the WD in 20 yr, providing a large fraction of the ignition mass. We emphasize, however, that the nature of the post-eruption high state is unknown, and thus its connection to a period of increased accretion rate is only speculative.

By comparing both "super-active" states, we see that the current state reached the peak earlier than expected if the recurrence time is exactly 80 yr (we note, however, that most RNe do not recur precisely periodically; Schaefer 2010). Moreover, the plateau of the current "super-active" state is about half a magnitude fainter than the 1938–1945 state, and thus the fading to the quiescent, pre-eruption state, could have already started. Although the AAVSO light curve does not show a significant decline in the optical brightness, both X-ray softness ratio and Hα flux indicate a secular decline in the accretion rate since 2016 (G. J. M. Luna et al. 2020, in preparation). We predict that T CrB is within 3–6 yr of its next TNR.

In cataclysmic variables, a single-disk instability dwarf-nova outburst cannot cause the WD to accumulate enough mass to significantly fuel a nova eruption. For example, Cannizzo (1993) modeled the disk instability outburst in SS Cyg and found that about 6 × 10⁻¹⁰ M_⊙ yr⁻¹ is stored in the disk during quiescence, which is then all or partially dumped into the WD during a long-lasting (about 50 days long) outburst. The aforementioned theoretical models by Prialnik & Kovetz (1995) show that the ignition mass for a nova eruption in a 1.1 M_⊙ WD (as in SS Cyg) is on average 10⁻⁶ M_⊙. On the other hand, in symbiotics, where a large and unstable accretion disk can store a significant amount of mass (about 10⁻⁶ M_⊙; Wynn 2008), the WD has the potential to assemble enough mass after a disk instability to trigger a nova eruption.

Our results thus suggest a scenario in which the WDs in symbiotic recurrent novae accumulate most of the ignition mass during sporadic high states. Hints of such episodes were already noticed by Nelson et al. (2011) in their analysis of quiescent X-ray data of the symbiotic RN RS Oph. Because of their faintness when not experiencing nova eruptions, the data on the other symbiotic recurrent novae (V3890 Sgr, V745 Sco, and perhaps V2487 Oph) are too scarce for us to place constraints on low-amplitude high states like the one for T CrB we report here. With the ignition mass on the WDs in symbiotic RNe most likely to be reached during an accretion high state, states such as the one that T CrB is currently in then become indicators of an impending nova eruption.

We acknowledge with thanks the variable star observations from the AAVSO International Database contributed by observers worldwide and used in this research. The DASCH project at Harvard is grateful for partial support from NSF grants AST-0407380, AST-0909073, and AST-1313370. G.J.M.L. is a member of the CIC-CONICET (Argentina) and acknowledge support from grants ANPCYT-PICT 0901/2017 and CONICET-NSF International Cooperation Grant 2016. N.P.M.K. acknowledges support from the UK Space Agency. J.L.S. acknowledges support from NSF award AST-1616646.