Infrared Spectroscopy of the Recent Outburst in V1047 Cen (Nova Centauri 2005)

A cataclysmic variable (CV) generally consists of a compact star (the primary) in a semi-detached binary system with a donor star that transfers matter onto the primary through the inner Lagrangian point via an accretion disk (AD). In classical and dwarf nova (CN and DN, respectively) systems the primary is a white dwarf (WD), and the secondary is a late-type star (see Warner 1995).

A CN eruption is the result of a thermonuclear runaway (TNR) in the degenerate material accreted on the surface of the WD (see Bode & Evans 2012). Following the TNR, ∼10⁻⁶–10⁻⁴ M_⊙ of material, enriched in CNO, is ejected explosively at several 100–1000 km s⁻¹. The system luminosity increases by a factor of as much as ∼15 mag. The bolometric luminosity remains approximately constant during the eruption, so the visual flux declines as the effective temperature of the stellar remnant increases. Once the ejecta become optically thin to 0.1–1.0 keV photons, the source becomes a "super-soft X-ray source" (see Krautter 2012). CN eruptions are believed to repeat on timescales of ∼10⁴ yr.

Eruptions of DNe of the U Gem type are gentler affairs, and are due to an increase of the mass-flow through the AD as a consequence of a thermal instability within the AD itself; the amplitude of the outburst, which typically lasts a few days, is generally ≲5 mag, although WZ Sge stars can have outburst amplitudes up to 8 mag (Warner 1995). There is little or no material ejected, and there is no enhancement of elemental abundances.

A small number of CNe have, since their eruptions, shown DN outbursts (Livio 1987; Mróz et al. 2016, and references therein), while some DNe seem to have circumstellar shells that appear to be the result of a previous CN outburst (e.g., Shara et al. 2007). In this cyclical scheme, CN–DN–CN (Shara et al. 1986), a DN eruption is not expected to follow within a decade of the CN outburst. This is because the cooling timescale for the WD after the CN eruption is of the order of a century or more, during which high mass transfer rates from the secondary continue due to irradiation by the WD (Shara et al. 1986; Kovetz et al. 1988). Such circumstances are not conducive for a DN eruption unless special circumstances prevail (see Section 5 below). Hence the present outburst in V1047 Cen, if it is indeed a DN eruption, is a rare and exceptional event.

Here we present two epochs of near-infrared (NIR) spectroscopy of the CN V1047 Cen taken around the peak brightness of its 2019 outburst, and also present optical and NIR light curves that document the detection and evolution of the present re-brightening event.

V1047 Cen (Nova Cen 2005) was discovered by Liller (2005) on 2005 September 1.031 at m_vis ≃ 8.5, but the 2005 eruption was not well observed. It was a CN of the "Fe ii" spectroscopic type (Walter et al. 2012). The Hα line was found to have a FWHM of ∼840 km s⁻¹ on day 3 (Liller 2005), and ∼1200 km s⁻¹ on day 6.9 (Walter et al. 2012).

V1047 Cen (AT2019hik, Gaia19cfn) was found to be in outburst again by the Optical Gravitational Lensing Experiment (OGLE; Udalski et al. 2015) survey (Mróz & Udalski 2019, see Figure 1). The outburst started on 2019 April 6 and, at the time of writing (2019 mid-October, when it approached solar conjunction), had persisted for ∼160 days. The nature of the slow rise to maximum, the outburst amplitude, and the subsequent variability, rule out a recurrent nova eruption for the 2019 event in V1047 Cen (see Anupama 2008). Currently, the duration of the 2019 April outburst is of the same order as that of DN eruptions in the long-period systems, and considerably longer than those in GK Per (e.g., Szkody et al. 1985; Salazar et al. 2017). The likelihood is that we are seeing a protracted DN eruption in a CN system.

Zoom In Zoom Out Reset image size

High-resolution optical spectroscopy shortly after the 2019 April outburst was reported by Aydi et al. (2019a, 2019c). At that time the spectrum was dominated by H i (FWZI >2500 km s⁻¹) and [O iii] (FWZI > 1100 km s⁻¹), which they suggest might have arisen in material ejected in the 2005 eruption. There were also narrow He i emission lines (FWZI <300 km s⁻¹), which may have originated in the 2019 DN eruption.

Timing analysis of the V1047 Cen light curve from the Transiting Exoplanet Survey Satellite (TESS; Ricker et al. 2014) indicated a possible 8.66 ± 0.07 hr periodicity (Aydi et al. 2019b), although this has yet to be verified. If this is confirmed to be the orbital period, it is at the higher end of CV orbital periods, and possibly suggests an evolved secondary.

The distance and reddening to V1047 Cen are poorly constrained. The Gaia survey Data Release 2 (Bailer-Jones et al. 2018; see that paper for the distance derivation, and its reliability) gives an estimated distance of 1.86 kpc, with lower and upper confidence bounds of 1.01 kpc and 4.25 kpc, respectively. This renders it difficult to assign an accurate value for the reddening using 3D extinction maps. For example, the extinction in the direction of V1047 Cen from Marshall et al. (2006) gives E(B − V) values of 0.29, 0.54, and 2.54 at the Gaia lower, most probable, and upper distance estimates. Literature values for E(B − V) give a wide range (e.g., Senziani et al. 2008, give E(B − V) in the range 1.28–1.38). Using the He ii λ4686 /λ10124 line ratio, E(B − V) = 1.1 is suggested by Di Mille et al. (2019), who also used the 5780 Å diffuse interstellar band to estimate E(B − V) = 1.

In view of the considerable uncertainty in E(B − V) we do not deredden the data here.

Spectra of V1047 Cen were obtained at the 8.1 m Gemini South Telescope using its facility spectrograph FLAMINGOS-2 (Eikenberry et al. 2004) in conditions of poor seeing and possible thin clouds. The log of the observations is given in Table 1. Data reduction was done using standard procedures for NIR data using IRAF²¹ and Figaro tasks. Because of the poor conditions under which the spectra were obtained, flux calibration of the spectra used broadband NIR photometry from SMARTS (see Figure 1). The uncertainty in the flux scaling is 20% for the June 30/July 1 spectrum, and higher for the August 1/2 spectrum, as no NIR photometry was obtained on that night. However, as we do not use absolute flux values in this study, our results and conclusions are not affected. The observed spectra are shown in Figure 2.

Zoom In Zoom Out Reset image size

Table 1.  Log of Spectroscopic Observations

Date	Grism, Coverage	IT	Standard Star
	Resolutiona	(s)
2019 Jun 30/Jul 1	R3K, 1.9–2.5 μm	1920	HIP 63036
2019 Jun 30/Jul 1	JH, 0.9–1.8 μm	240	HIP 67360
2019 Aug 1/2	R3K, 1.25–2.45 μm	1260	HIP 67360
2019 Aug 1/2	JH, 0.9–1.8 μm	240	HIP 67360

Note.

^aWith the 036 slit, for the JH grism, R ranges from ∼200 near the short wavelength edge, to a maximum of ∼1100 at 1.3–1.4 μm, dropping to ∼400 near the long wavelength edge. For the R3K grism, R varies from 1200 near 1.9 μm, increasing to 3200 near 2.2 μm, and then decreasing to ∼1200 at 2.4 μm.

Download table as:  ASCIITypeset image

The IR spectra rule out the possibility that V1047 Cen is a stellar merger event (or "Luminous Red Nova"), as suggested by Mróz & Udalski (2019). While stellar mergers have shown a similar slow rise to maximum, they also show a distinct shift to cooler spectral type, and display very specific NIR molecular features (e.g., AlO, H₂O) within 45 days of eruption. Our data show none of these characteristics. The absence of first overtone CO bands in the IR spectra (see Figure 2) also rules out the possibility that V1047 Cen might be a symbiotic system.

Both spectra show lines of H i, He i, and He ii. The C, N, and O lines that are generally seen in the NIR spectra of RNe—irrespective of whether the RN is of the T Pyx, T CrB, or U Sco sub-types (Banerjee et al. 2009, 2010; Joshi et al. 2014)—are absent in V1047 Cen, with the exception of the O i line at 1.1289 μm in August.

We estimate the ejection velocity from the FWZIs of the H i Pa-β and Br-γ lines (Figure 3) to be ∼2000 km s⁻¹, as reported by Aydi et al. (2019a, 2019c) for the optical H i lines, with no significant change between the two IR observations. The FWHMs of these lines are ∼1500 km s⁻¹, somewhat larger than the velocities reported by Liller (2005) and Walter et al. (2012) for Hα during the 2005 eruption. The Br-γ line recorded at our highest resolution of R ∼ 3000 clearly shows a double-peaked profile at one epoch, with a separation of ∼500 km s⁻¹. This is consistent with the signature for emission lines originating from an AD (Horne & Marsh 1986). Because the dip between the peaks is not very deep, a system at low inclination (≲30°) is indicated (Horne & Marsh 1986) if the emission comes from the AD only.

Zoom In Zoom Out Reset image size

Velocities of up to 2000 km s⁻¹ are also indicated by the He i 2.0587 μm line, although there are dramatic changes in the line profile. (Based on the close agreements of the central wavelengths of the centers of the H i lines with their laboratory wavelengths, we conclude that the radial velocity of center of mass of the V1047 Cen system is no more than several tens of km s⁻¹.) The July spectrum of this line displays a complex profile, with absorption extending from ∼0 to −2000 km s⁻¹, sharp absorption troughs at −300 and −1400 km s⁻¹, and weak emission extending from ∼0 to at least 1600 km s⁻¹. The troughs are replicated in the He i λ2.1138 μm profile (see Figure 3). In the August spectrum only a single absorption trough is present, at −1300 km s⁻¹, the wing to higher negative velocities has weakened, and the emission now is much more extensive, from −800 to +1600 km s⁻¹.

The He i triplet at 1.0833 μm is almost entirely in emission, but is more difficult to characterize quantitatively, due to the lower resolution at which it was observed and its blending on the long wavelength side with Pa-γ (1.0941 μm). In the July 1 spectrum only a single weak P Cygni absorption is present at −1600 km s⁻¹, close in velocity to the more blueshifted absorption in the He i singlet profile, and with an absorption wing extending to ∼−2000 km s⁻¹, similar to the singlet. In the August 2 spectrum absorption is completely absent, and the emission extends to ∼−2000 km s⁻¹. At both times the FWHM of the line emission is ∼1200 km s⁻¹. In contrast, Aydi et al. (2019a, 2019c) found relatively narrow (FWHM < 300 km s⁻¹) He i emission lines in an optical spectrum obtained on July 19.86 UT. We note that the (deconvolved) FWHM of the O i line at 1.1289 μm, present in the August 2 spectrum, is ∼700 km s⁻¹.

The cause of the P Cyg profile in the He i 2.0587 μm line observed on July 1 is puzzling, but its reality is beyond doubt (see Figure 3). We do not find a similar profile in the literature, though NIR spectra of DNe during outburst are rare (e.g., Howell et al. 2003). In quiescence such P Cyg profiles are not generally seen (see, e.g., the DN sample observed by Dhillon et al. 2000). It is possible that the 2019 eruption has generated two separate outflows, as witnessed in the July 1 profile of the singlet line. However, radiative transfer associated with this line is complex (see Geballe et al. 1984), and what we have observed may be a consequence of it rather than multiple outflows.

Many ultraviolet resonance lines (C iv 1549 Å, Si iv 1397 Å, N v 1240 Å) in DNe often display P Cyg profiles, implying the presence of mass loss in stellar winds (e.g., Holm et al. 1982; Warner 1995, and references therein). He ii 1640 Å and N iv 1719 Å have also been found to have P Cyg features, in RW Sex (Drew 1990), and it is possible that the He ii recombination lines (1640 Å, 4686 Å) are produced in both the disk and the wind (Warner 1995). We may thus speculate about the presence of a wind in this system, which is responsible for the P Cyg profile in He i 2.0587 μm.

V1047 Cen also shows an intriguing Hα excess in images obtained prior to its current re-brightening. A SUPERCOSMOS (Hambly et al. 2001) Hα image of the V1047 Cen field, obtained in 2001, shows no source at the position of the nova (see Figure 4). On the other hand, a VPHAS (Drew et al. 2014) image shows that V1047 Cen had a strong Hα excess in 2013, several years after its 2005 eruption, but before the 2019 event. The Hα excess in V1047 Cen is comparable with, or even stronger than, the excess in novae with giant secondaries that show strong Hα in the SUPERCOSMOS images, as well as very strong Hα emission in their quiescent spectra (Anupama & Mikołajewska 1999). The presence of an excess may lend support to our DN interpretation, because not only DNe in quiescence, but also old CNe and their remnants, show Hα emission in their spectra (Stauffer et al. 1979; Shafter & Hessman 1988). We also note that N ii (6548 Å, 6584 Å), known to be seen in CNe, could also be contributing partially to the Hα images. Nonetheless, the strength of the Hα excess is puzzling and its evolution should be monitored.

Zoom In Zoom Out Reset image size

There is some evidence for the presence of the [Si vi] 1.9641 μm coronal line in the earlier spectrum (see Figure 2). We note that a number of the H i recombination lines have weak features on their red wings from He ii, but this is not the case for Br-δ. The ionization potential of Si v is 167 eV, so the production of Si vi requires high-energy conditions such as might occur in a shock. We have noted the likely presence of winds having different velocities; conditions for coronal emission might well occur in regions where the winds collide. We note that [Si vi] 1.9641 μm emission has previously been detected in the CV TT Ari (Ramseyer et al. 1993).

If the 2019 April event in V1047 Cen is a DN eruption in a CN system, then we are witnessing a thermal instability in the AD. In their model of the disk-instability-induced DN eruptions in GK Per, Kim et al. (1992) found that AD temperatures in the range 6000–10000 K are expected. In order for the IR emission to reflect such temperatures, dereddening by E(B − V) ≳ 1.6 is required, for both the July and August spectra. This implies either that the reddening (and hence distance) to V1047 Cen is at the higher end of the range implied by the Gaia data, or that what we are seeing is not a thermal instability event.

Only seven CNe are known to have exhibited DN outbursts within a few decades of their original CN outburst (see Vogt 1989; Honeycutt et al. 2011). The best-documented DN outbursts are those of GK Per (CN 1901), V1017 Sgr (CN 1919) and V446 Her (CN 1960), while their subsequent DN eruptions were first recorded decades later: GK Per in 1948 (Bianchini et al. 1986), V446 Her (around 1991; Honeycutt et al. 2011), and V1017 Sgr (1973; Warner 1995). While DN outbursts have been observed regularly in GK Per and V446 Her (Bianchini et al. 1986; Honeycutt et al. 2011), only two DN outbursts are known to have occurred for V1017 Sgr, though it displayed a DN eruption in 1901 prior to its CN outburst (Salazar et al. 2017). The durations of individual DN events in these CN systems also vary substantially, lasting typically about 2 months for GK Per (Bianchini et al. 1986), 10–20 days for V446 Her (Honeycutt et al. 2011), and ∼200 days for V1017 Sgr.

In this context the 2019 DN outburst in V1047 Cen has two notable characteristics. The duration of the outburst is relatively long, and as of early November 2019 the decline phase has yet to commence. Further, the DN event occurred only 14 yr after the CN eruption, the shortest recorded gap between CN and DN eruptions. Explaining this behavior within the framework of the "hibernation" paradigm (Shara et al. 1986) may be challenging. In that theory for cyclic nova evolution, the mass accretion rate $\dot{M}$ remains high $(\dot{M}\sim {10}^{-8}\,{M}_{\odot }\,{\mathrm{yr}}^{-1})$ on a timescale of a few centuries after the CN eruption due to the irradiation of the secondary star by the still-hot WD (Shara et al. 1986; Kovetz et al. 1988). Eventually $\dot{M}$ declines, taking the AD through the DN phase and into hibernation, from which the CV reawakens after several millennia for the next CN eruption. The theory is supported by the detection of ancient CN remnants around the DNe Z Cam (Shara et al. 2007), AT Cnc (Shara et al. 2012), and also by DNe outbursts preceding the main CN eruption of Nova Cen 2009 (Mróz et al. 2016). A DN eruption is not expected to occur during the high $\dot{M}$ phase that lasts for a few centuries after the CN outburst, as DN outbursts are believed to result from thermal instabilities associated with hydrogen ionization in the AD occurring during a regime of low mass accretion rate ($\dot{M}\sim {10}^{-10}$ M_⊙ yr⁻¹; Osaki 1974; Cannizzo 1993). Thus, a special set of circumstances is needed to account for the early DN outburst in V1047 Cen. In the case of GK Per, DN eruptions are not surprising (Schreiber et al. 2000) as it has a large AD (Kim et al. 1992), which is consistent with its 2 day orbital period. Thus, the critical mass transfer rate ${\dot{M}}_{\mathrm{crit}}$ below which outbursts are possible is large (note that ${\dot{M}}_{\mathrm{crit}}$ scales as (R₁₀)^2.6, where R₁₀ is the outer disk radius in units of 10¹⁰ cm; Cannizzo 1993). Thus DN outbursts are possible in GK Per even at high accretion rates (Kim et al. 1992). It is possible that similar circumstances are responsible for the DN outburst of V1047 Cen but a robust determination of the orbital period is needed to verify this.

In summary, we have presented NIR spectra of the 2019 outburst of V1047 Cen. The presence of mostly H and He i lines in the spectra, the presence of double-peaked line profiles, and the nature of the light curve all suggest that a DN outburst has occurred. Such an event is rare, especially coming so soon after the CN outburst, and challenging to explain. The present and future outbursts of V1047 Cen could provide an opportunity to probe the CN to DN transition phenomenon and also test the generic mechanisms that are responsible for DN outbursts.

This Letter is largely based on observations obtained for program GS-2019A-DD-111 of the Gemini Observatory, which is operated by the Association of Universities for Research in Astronomy, Inc., under a cooperative agreement with the NSF on behalf of the Gemini partnership: the National Science Foundation (United States), National Research Council (Canada), CONICYT (Chile), Ministerio de Ciencia, Tecnología e Innovación Productiva (Argentina), Ministério da Ciência, Tecnologia e Inovação (Brazil), and Korea Astronomy and Space Science Institute (Republic of Korea).

The OGLE project has received funding from the National Science Centre, Poland, grant MAESTRO 2014/14/A/ST9/00121 to AU.

This work has made use of data from the European Space Agency (ESA) mission Gaia processed by the Gaia Data Processing and Analysis Consortium (DPAC), funding for which has been provided by national institutions, in particular the institutions participating in the Gaia Multilateral Agreement.

Based on data products from observations made with ESO Telescopes at the La Silla Paranal Observatory under programme ID 177.D-3023, as part of the VST Photometric Hα Survey of the Southern Galactic Plane and Bulge (VPHAS+, www.vphas.eu).

Nova research at Stony Brook is supported by a grant from the NSF. F.M.W. acknowledges support from Stony Brook University for continued participation in the SMARTS consortium. D.P.K.B. is supported by a CSIR Emeritus Scientist grant-in-aid hosted by the Physical Research Laboratory, Ahmedabad. C.E.W. acknowledges partial support from a USRA/NASA SOFIA contract. R.D.G. is supported by NASA and the United States Air Force. K.L.P. acknowledges support from the UK Space Agency. S.S. acknowledges support from a NASA Theory grant to ASU. D.A.H.B. thanks the National Research Foundation for research support.

We acknowledge with thanks the variable star observations from the AAVSO International Database contributed by observers worldwide and used in this research.

Facilities: Gemini South (FLAMINGOS-2) - , OGLE - , Gaia - , ESO - , SMARTS - , AAVSO. -

Software: IRAF (Tody 1986, 1993); Figaro (Shortridge et al. 2014).