Spectroscopy of the 2015 Outburst of AG Pegasi

Symbiotic stars are interacting binaries with long orbital periods (hundreds of days to many dozens of years) typically consisting of a white dwarf (WD) accreting mass via the gravitation capture of the wind from a red giant; this wind forms a surrounding nebula partially ionized by the hot WD producing spectral emission from the far-UV to the optical (Murset & Schmid 1999; Mikołajewska 2007). This quiescent scenario can be interrupted by either classical (also referred to as Z And-type) outbursts producing 1–3 magnitudes of optical brightening and lasting months to a few years; or those labeled thermonuclear powered which cause greater brightening lasting weeks to months (fast or symbiotic recurrent novae), or decades (slow or symbiotic novae).

This note focuses on the second peak of the 2015 classical outburst of symbiotic star AG Pegasi, which consists of a 0.6 M_⊙ WD and an M3 III red giant, with an orbital period of 818 days (Skopal et al. 2017). The classical outburst, commencing 2015 June, followed a return to quiescence from the slowest nova event recorded (summarized in Ramsey et al. 2016). Lasting little over a year, the 2015 event resulted in two abrupt optical brightening events of approximately 1.5 and 1 magnitudes respectively (Mistry & Steele 2020).

Presented are WD parameter derivations, line flux measurements, and nebular emission models derived from spectroscopic analysis of data from the 2.0 m Liverpool Telescope LOTUS spectrograph (Steele et al. 2016) covering the second brightening event.

Wavelength calibrated LOTUS spectra from 2015 September 4th to 2016 January 11th (corresponding to the outburst period) and one from 2016 April (post-outburst) were flux calibrated in a two-stage process. Correction for atmospheric absorption was performed using a spectrum of G191-B2B (Oke 1990) from 2015 November 5th that was scaled according to the Sloan r' band magnitude (Mistry & Steele 2020) for each night. The resultant flux and wavelength calibrated spectra were then dereddened according to the model of Fitzpatrick (1999).

Models of the nebular contribution to the spectrum (right panel of Figure 1) are constructed based on assumptions from Skopal (2005): the observed spectrum, F λ, is a superposition of the WD [F_h (λ)], giant [F_g (λ)], and nebular [F_neb(λ)] flux densities. Therefore

$$\begin{eqnarray*}\begin{array}{rcl}F\lambda & = & {F}_{h}\lambda +{F}_{\mathrm{neb}}\lambda +{F}_{g}\lambda \\ & \approx & {F}_{\mathrm{neb}}\lambda +{F}_{g}(\lambda )\end{array}\end{eqnarray*}$$

where F_h is approximately negligible in the spectral range, and F_g (λ) is assumed constant in time. Model giant surface flux densities are from Kurucz (1993). Based on the fit to molecular bands at the red end of our spectra, the chosen stellar model corresponds to an effective temperature and surface gravity of T_eff = 3500 K and $\mathrm{log}\,g=0.99,$ with solar metallicity.

Zoom In Zoom Out Reset image size

WD temperature, T_h , calculations (left panel of Figure 1) are based on Equation (8) of Iijima (1981) using the line fluxes of H β and He ii (4686) emission features. These recombination lines are generated by ionizations within the nebula caused by the blackbody approximated WD flux, so that:

$$\begin{eqnarray*}&&{T}_{h}=\left[19.38\ {\left(\mathrm{match}\displaystyle \frac{2.22{F}_{4686}}{4.16{F}_{{\rm{H}}\beta }}\right)}^{\tfrac{1}{2}}+5.13\right]\times {10}^{4}\end{eqnarray*}$$

WD luminosities, L_h , (left panel of Figure 1) are derived through their proportionality to the hydrogen ionizing photon flux, f(T_h ). Since f(T_h ) is balanced by the rate of nebular case B recombinations, α_B, from Skopal et al. (2017) we have:

$$\begin{eqnarray*}&&{L}_{h}={\alpha }_{B}\ \mathrm{EM}\displaystyle \frac{\sigma {T}_{h}^{4}}{f\left({T}_{h}\right)}\end{eqnarray*}$$

where the emission measure, $\mathrm{EM}={\int }_{V}{n}_{+}{n}_{e}{dV}=4\pi {d}^{2}\left[{F}_{{neb},U}\lambda /{\epsilon }_{U}\right]$, is the weighted average of the U band nebula flux, F_neb,U λ either side of the Balmer Jump. WD radius, R_h , (left panel of Figure 1) is then calculated using the Stefan–Boltzmann equation.

T_h evolution follows changes in the light curve (LC), with values of 144,000–173,000 (±5000) K. The rebrightening peak occurs with T_h  = 166,000 (±5000) K. Comparable figures (140,000–184,000 K) were produced by Skopal et al. (2017), Tomov et al. (2016), and Ramsey et al. (2016) using the same method.

Use of nebular flux in deriving the EM helps to produce a L_h evolution matching LC changes, where L_h  = 4000–14,300 (±600–2000) L_⊙. Similar figures (5000–18,000 L_⊙) were found by Skopal et al. (2017) using the same method, who explained that values 4–8 times lower, produced by Tomov et al. (2016) when using line fluxes for EM calculations, are due to a greater opacity in spectral lines than in the nebular continuum.

Radial changes follow those of the LC, with values of R_h  = 0.1–0.15 ( ± 0.01) R_⊙ comparable with those of Skopal et al. (2017), but a factor of 10 greater than those of Tomov et al. (2016) due to the reasoning above.

Hβ and He ii (4686 Å) line fluxes show a rise and fall coinciding with the rebrightening event and the subsequent decline. Peak values of 78 and 43 × 10⁻¹¹  ergs s⁻¹ cm⁻², respectively, occur ∼20–40 days after the rebrightening peak. These lines are the strongest emission features in the spectral range.

The changing nebular flux is a key feature of the spectra, rising and falling with the light curve (right panel of Figure 1; for first peak see Skopal et al. 2017) This is indicative of WD activity via the reprocessing of high energy photons by the nebula. The WD influence on nebular flux is strongest in the near-UV to B passbands; at redder wavelengths, this influence is greatest during LC peaks. The gradual return to quiescence (spectrum 4 onwards) allows the giant flux to have a larger influence on the spectral flux distribution.

Overall, the features described are consistent with our current understanding of classical outbursts—caused by an increase in the rate of shell burning under non-degenerate conditions, driving an expansion of the WD pseudo-photosphere and an increase in luminosity.

LT is operated on La Palma island by LJMU in the Spanish Observatorio del Roque de los Muchachos of the Instituto de Astrofisica de Canarias with financial support from the UK Science and Technology Facilities Council.

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