Raman-scattered O vi Features in the Symbiotic Nova RR Telescopii

Symbiotic stars are interacting binary systems consisting of an evolved giant transferring mass to a hot compact object, in most cases a white dwarf (WD; e.g., Kenyon 1986). In the general case, it is believed that the accretion occurs via the stellar wind from the primary, known as Bondi–Hoyle–Lytteleton (BHL) wind accretion. For the D-type (dust) symbiotic stars, the widest binary systems composing of a Mira-type variable, several hydrodynamical simulations suggest a new mass-transfer model called wind Roche-lobe overflow (WRLOF; Mohamed & Podsiadlowski 2012) or gravitational focusing (de Val-Borro et al. 2009). Accretion through WRLOF focuses on the binary plane, leading to a higher mass-transfer rate than that of standard BHL wind accretion. This lends support to the suggestion that D-type symbiotic stars are promising candidates of type Ia supernova progenitors (e.g., Iłkiewicz et al. 2019).

RR Telescopii (RR Tel) is a dusty-type symbiotic nova comprising a Mira component and a WD. A nova-like outburst occurred in 1944 when RR Tel brightened by ∼7 mag in the visual band (Mayall 1949). In 1949, the extended atmosphere started to shrink and entered a nebular phase. Subsequently, its luminosity has been slowly fading for several decades (e.g., Mürset et al. 1991; Nussbaumer & Dumm 1997). A recent light curve from the ASAS-SN photometric database shows that the overall V magnitude is still in a trend of steady decline. It changed from 11.7 to 11.9 during the last four years (Shappee et al. 2014; Kochanek et al. 2017). Meanwhile, the effective temperature of the WD has continued to rise from T ∼ 6750 K in 1949 May (Pottasch & Varsavsky 1960) and it has exceeded >140,000 K after 1978 (e.g., Jordan et al. 1994; Hayes & Nussbaumer 1996). Based on the recent X-ray observations with XMM-Newton, González-Riestra et al. (2013) estimated its temperature T ∼ 154,000 K and luminosity L ∼ 5000 L_⊙ assuming a distance of 2.6 kpc. The orbital period of this object is still quite uncertain but it is believed to exceed several decades (Lee & Park 1999; Schmid & Schild 2002).

From time-series spectroscopic observations of RR Tel between 1950 and 1960, Thackeray & Webster (1974) reported a strong, unidentified emission band at 6825 Å having a broad width of ∼30 Å with a double-peaked profile. They noted that whenever the 6825 band appeared, the rest of the spectrum showed lines of high ionization, which implied that the origin of the band is related to high-ionized species. A second, weaker band at 7082 Å was also detected, whose profile bears a striking similarity to the 6825 feature (Allen 1979). Because of their observational properties, including abnormally broad width and the double-peaked profile, Allen (1979) proposed that both emissions come from the same atomic or molecular species.

The identification of the two bands was made by Schmid (1989), who proposed that the two broad bands at 6825 Å and 7082 Å are due to the Raman scattering of O vi λλ 1032 and 1038 doublet by atomic hydrogen H i. Synchronous far-UV and optical observations for symbiotic stars including RR Tel provided support for his identification of the Raman O vi features (Espey et al. 1995; Birriel et al. 2000). In this scattering process, an O vi 1032 photon is incident on H i in the ground state, and de-excitation into 2 s state produces an optical Raman-scattered photon with λ = 6825 Å. A similar process with an O vi 1038 photon results in an optical photon at 7082 Å. Figure 1 shows the scattering process in the symbiotic star. The heterogeneous environment in symbiotic binaries includes the nebular region photoionized by intense far-UV radiation from the hot source and the dense neutral region around the donor star, providing an excellent laboratory to study Raman scattering by H i (e.g., Nussbaumer et al. 1989; Schmid 1989).

Zoom In Zoom Out Reset image size

The inelasticity of Raman scattering requires that the wavelength of an emergent Raman photon is almost solely determined by the relative motion between the far-UV O vi emission region and the H i scattering region and negligibly dependent on the observer's line of sight. This unique property allows us to investigate the mass-transfer process with an edge-on view of the cool star. Follow-up spectropolarimetric observations revealed that the Raman O vi bands of RR Tel are strongly polarized and show complicated line structures with a polarization flip in the red wing (Schmid & Schild 1994; Harries & Howarth 1996; Schmid et al. 1999). The spectropolarimetric properties were explained by a very extended scattering region responsible for a large mass-loss rate. Another model was suggested by Lee & Park (1999), who adopted an accretion disk around a WD with an asymmetric matter distribution, resulting in the asymmetric double-peak profiles of the Raman O vi in RR Tel. Previous studies on line formation of Raman O vi features in other symbiotic stars show that their profiles provide valuable kinematic information on the accretion disk. In particular, the main peak separation of Raman O vi features indicates a representative velocity scale ∼50 km s⁻¹ of the accretion disk around the WD (e.g., Lee & Kang 2007; Heo & Lee 2015; Heo et al. 2016).

Additionally, an interesting point regarding Raman O vi is that the two profiles, as usual, are not identical in a given object, even though the O vi λ λ 1032 and 1038 photons are formed in the same region. The difference in flux between the two peaks is more evident in the 7082 feature than the 6825 profile. O vi λ λ 1032 and 1038 resonance doublet lines arise from 2s_1/2 − 2p_3/2,1/2 transitions. Since the oscillator strength of the 2s_1/2 − 2p_3/2 transition is twice larger than that of the 2s_1/2 − 2p_1/2 transition, it is expected that the flux ratio F(1032)/F(1038) = 2, in the optically thin region. In accordance with the ratio of oscillator strength, O vi 1032 is twice more optically thick than O vi 1038, τ(1032) = 2τ(1038), which makes it more difficult for an O vi λ 1032 line photon to escape from the site. We, therefore, expect that the flux ratio F(1032)/F(1038) may deviate from the optically thin limit of 2 and approach unity in an optically very thick medium. Heo & Lee (2015) adopted the local variation of the flux ratio F(1032)/F(1038) in the accretion disk to account for the profile disparity of the Raman O vi features in a symbiotic nova V1016 Cyg. As demonstrated by these results, Raman O vi features can be an ideal spectroscopic tool to investigate the vicinity of the WD and further probe the mass-transfer process associated with stellar wind accretion in symbiotic stars.

Although RR Tel is a well-studied symbiotic nova, there has been a lack of optical studies for the past two decades. The most recent observation in the literature was conducted in 2000, analyzed by Young (2012). In this paper, we present and discuss our high-resolution spectroscopic observations of the Raman O vi features in RR Tel, conducted in 2016 and 2017 using the Magellan Inamori Kyocera Echelle (MIKE) spectrograph installed on the 6.5 m Magellan-Clay telescope, Chile. In Section 2, we investigate the Raman O vi features comparing the archival far-UV data taken in 2002 with the Far-Ultraviolet Spectroscopic Explorer (FUSE) satellite and the almost contemporary optical spectrum taken in 2003 with the Fibre-fed Extended Range Optical Spectrograph (FEROS) at the MPG/ESO 2.2 m telescope. A detailed profile analysis of the Raman O vi features based on the accretion disk model was performed, and the results are shown in Section 3. Our finding of the O vi λλ 3811 and 3834 doublets from our profile decomposition is described in Section 4. Discussion and summary follow in Sections 5 and 6.

2.1. MIKE Spectrum

We obtained high-resolution optical spectra of RR Tel using the MIKE spectrograph (Bernstein et al. 2003). The double echelle spectrograph MIKE covers the wavelength range of 3350–5000 Å (blue) and 4900–9500 Å (red) in its normal configuration. A slit width of 07 × 5'' was used, resulting in resolving power R ∼ 27,000 and ∼35,500 on the blue and red sides, respectively. To increase the signal-to-noise ratio (S/N), a 2 × 2 binning was applied in both the spectral and spatial directions. Observations were carried out on 2016 July 30 and 2017 July 26, for total exposure times of 2000 s and 2400 s, respectively. With these integration times, the spectra have significant S/N > 100 for strong lines suitable for profile analysis. ThAr spectra were taken for wavelength calibration.

We reduced the MIKE data using the Carnegie Python Distribution (CarPy) pipeline (Kelson et al. 2000; Kelson 2003) and NOAO.onedspec package of the Image Reduction and Analysis Facility (IRAF). ⁷ The data are overscan-corrected, bias-subtracted, extracted, wavelength calibrated, and flat-fielded with quartz lamps within the pipeline. Individual exposures were then combined using the IRAF task scombine. Due to the poor weather conditions on the night of 2016 July 30, flux calibration was not possible. For the observing night 2017 July 26, we observed the spectrophotometric standard star HR 7950 under the photometric condition. Flux calibration was performed with the IRAF tasks standard, sensfunc, and calibrate.

The left and right panels of Figure 2 show the Raman-scattered O vi features at 6825 and 7082 Å. To make quantitative profile comparisons of the two Raman O vi features, we transform the observed wavelengths into the O vi Doppler factor ΔV_atomic as shown by the upper horizontal axis. First, we convert the observed wavelength to the vacuum wavelength using the refractive index of air n_air = 1.0003. By measuring the center wavelengths of the strong emission lines (e.g., He i λ7065 and λ6678, [N ii] λ6583 and λ6548, [O iii] λ4959, Hβ and Hγ), we obtain the systemic velocity ν_sys of − 55.5 km s⁻¹, and − 57.8 km s⁻¹, respectively, for the 2016 and 2017 data. We correct for the systemic velocity and calculate the corresponding far-UV wavelength of incident O vi photon using the energy conservation principle. Adopting the O vi center wavelengths of λ₁₀₃₂ = 1031.928 Å and λ₁₀₃₈ = 1037.618 Å, we finally compute the O vi parent Doppler factor ΔV_atomic. More detailed information about the transformation between the wavelength and the Doppler factor is available in Heo et al. (2016).

Zoom In Zoom Out Reset image size

We fit the Raman O vi features with two Gaussian components, from which we derive the peak separation of the Raman O vi 6825 feature ΔV = 47.5 km s⁻¹ and ΔV = 48.7 km s⁻¹ in the data of 2016 and 2017, respectively. For the Raman O vi 7082 features, the peak separations are measured to be ΔV = 45.7 km s⁻¹ and ΔV = 46.4 km s⁻¹ in 2016 and 2017, respectively. The Gaussian fitting parameters are presented in Table 3.

We overplot the MIKE data in Figure 2, whose medium gray lines represent the 2016 data and black lines show the 2017 data. The vertical axes show flux density in units of 10⁻¹⁴ erg cm⁻² s⁻¹ Å⁻¹. Since the 2016 data is not flux calibrated, we multiplied an artificial factor to match the continuum value with the 2017 data. As shown in Figure 3, the continuum around the bands is virtually flat with a constant value of 2.0 × 10⁻¹⁴ erg cm⁻² s⁻¹ Å⁻¹ due to the dust obscuration. The Mira in RR Tel is heavily obscured and so no trace of photospheric absorptions is visible (e.g., Kotnik-Karuza et al. 2006). Dust obscuration in RR Tel is also supported by its optical light curves, which do not show any pulsation of Mira, whereas the near-IR light curves show a clear periodicity (Gromadzki et al. 2009).

Zoom In Zoom Out Reset image size

To measure the total line fluxes for the Raman O vi features, we assume that the emission having the Doppler factor ΔV_atomic between − 60 km s⁻¹ and 100 km s⁻¹ is a Raman O vi feature. In these regions, the contributions of other spectral lines, including TiO or VO absorption lines from the Mira and telluric lines, are negligible. There is a telluric O₂–B band between 6870 and 6950 Å, but not affecting both Raman O vi features (Groppi & Hanson 1996). A few unknown emission lines with FWHM < 0.5 Å are also detected at the wavelength range of Raman O vi features. However, their contribution to the total line fluxes are negligible due to the relatively small intensities compared with the Raman features. Therefore we obtain the total line fluxes of the Raman O vi features F(6825) = (30.34 ± 0.11) × 10⁻¹³ erg cm⁻² s⁻¹ and F(7082) = (5.98 ± 0.08) × 10⁻¹³ erg cm⁻² s⁻¹ for the 2017 data.

The Raman line fluxes of RR Tel were studied in previous observations between 1993 and 1996 (e.g., Espey et al. 1995; Crawford et al. 1999; Schmid et al. 1999). A steady decrease was found over the period, and the same behavior is shown for the far-UV O vi 1032 and 1038 emissions. We note that the fluxes of both Raman O vi in the MIKE data have decreased by about 30% from 1996 October. The measurements are summarized in Table 1. The decrease in the line intensity is consistent with the trend of general slow fading started in 1949 after its outburst (e.g., Nussbaumer & Dumm 1997; Contini & Formiggini 1999).

Table 1. Observed Line Fluxes of the Raman O vi 1032 and 2018 Features at 6825 and 7082 Å from 1993 to 2017

Date	Raman O vi 1032	Raman O vi 1038	References
	(erg cm−2 s−1)	(erg cm−2 s−1)
1993-11-21	7.2 × 10−12	1.2 × 10−12	Schmid et al. (1999)
1995-03-18	6.7 × 10−12	1.3 × 10−12	Espey et al. (1995) a
1996-07-22	4.78 × 10−12	8.98 × 10−13	Crawford et al. (1999)
1996-10-08	4.5 × 10−12	9.1 × 10−13	Schmid et al. (1999)
2017-07-26	3.03 ± 0.01 × 10−12	5.98 ± 0.08 × 10−13	This work

Note.

^a Extinction-corrected flux.

Download table as:  ASCIITypeset image

2.2. FUSE Spectrum

RR Tel was observed on 2002 June 14, with the FUSE satellite as part of the Cycle 3 program C141 (data ID: C1410102000; PI: P.R. Young). The FUSE spectrum of RR Tel was taken for an exposure time of 128 s through the low-resolution aperture (LWRS: 30 × 30') using TTAG (Photon Address) mode. The SiC2 channel covers a wavelength range of 917–1104 Å at a spectral resolution of 12,000–20,000 (Moos et al. 2000; Sahnow et al. 2000). The data were reprocessed and archived at MAST with the CalFUSE pipeline version 3.2.1. The extracted data is binned in wavelength with a bin size of 0.013 Å (Dixon et al. 2007). Considering that the observation was made under the low-resolution mode, we binned the data by a spectral resolution of 0.362 Å.

In Figure 4, we present a zoom-in of the FUSE spectrum centered on the two strongest emission lines, O vi 1032 and 1038. The light gray line represents the raw data, whereas the black line shows the binned spectrum. The O vi emissions have line fluxes of F(1032) = (2.01 ± 0.56) × 10⁻¹⁰ erg cm⁻² s⁻¹ and F(1038) = (1.18 ± 0.32) × 10⁻¹⁰ erg cm⁻² s⁻¹, respectively, yielding the flux ratio of F(1032)/F(1038) = 1.70. There has been a steady decrease in line fluxes of both O vi emissions from F(1032) = 2.56 × 10⁻¹⁰ erg cm⁻² s⁻¹ and F(1038) = 1.54 × 10⁻¹⁰ erg cm⁻² s⁻¹ in 1993 (Schmid et al. 1999)—a drop of 22%. We summarize the observed values of O vi lines and compare them with previous measurements from other research works in Table. 2. We note that the flux ratio F(1032)/F(1038) has remained almost constant at ∼1.7 throughout this period.

Zoom In Zoom Out Reset image size

Table 2. Observed Line Flux of O vi λλ 1032 and 1038 Resonance Doublet and Their Flux Ratio from 1993 to 2002

Date	O vi 1032	O vi 1038	F(1032)/F(1038)	References
	(erg cm−2 s−1)	(erg cm−2 s−1)
1993-09-16	2.56 × 10−10	1.54 × 10−10	1.70	Schmid et al. (1999)
1995-03-12	2.32 × 10−10	1.43 × 10−10	1.64	Birriel et al. (2000)
2002-06-14	2.01 ± 0.56 × 10−10	1.18 ± 0.32 × 10−10	1.70	This work

Download table as:  ASCIITypeset image

2.3. FEROS Spectrum

2.4. Profile Variation of Raman O vi

We introduce our line formation modeling via a Monte Carlo approach to compute the flux and the profiles of Raman-scattered O vi features. The Monte Carlo simulation starts with a generation of a far-UV O vi photon at a random place in an accretion disk. The O vi photon enters the H i region where it is Rayleigh scattered. It escapes the region if the photon has an optical depth τ larger than τ_∞, the scattering optical depth to an observer at infinity, or becomes an optical photon through Raman scattering. A full description of the simulations, including the escape condition and the scattering process, can be found in Lee & Lee (1997).

It is worth noting once again that the line profiles of Raman-scattered features are almost independent of the observer's line of sight and are almost solely determined by the relative motion between the far-UV O vi emission region and the H i scattering region. Because the H i region around the Mira commands an almost perfect edge-on view of the accretion around the WD, the profile analysis of Raman-scattered features provides crucial information, including physical extent, density distribution, and kinematics of the O vi 1032 and 1038 emission region around the WD (Lee & Kang 2007; Heo et al. 2016; Lee et al. 2019).

3.1. Stellar Wind and the STB Ionization Front

First of all, to model the density distribution of H i around the Mira, it is assumed that the stellar wind from the Mira follows a beta law, where the radial wind velocity v_r ( r ) is given by

$$\begin{eqnarray}&&{v}_{r}{({\boldsymbol{r}})={v}_{\infty }(1-{R}_{* }/r)}^{\beta }.\end{eqnarray} \tag{ 1 }$$

Here, R_* is the launching site of the wind and v_∞ is the wind terminal velocity (e.g., Lamers & Cassinelli 1999). We choose the parameter β = 1 for simplicity in our Monte Carlo simulations.

The slow stellar wind around the Mira is illuminated by intense far-UV radiation from the hot component, and hence some parts facing the WD are photoionized. Seaquist et al. (1984) presented their photoionization calculation to find the ionization front, known as the STB model. In the STB model, the ionization front in the stellar wind region is determined by the balance of photoionization rate by the H-ionizing luminosity from the hot component and recombination rate set by the mass-loss rate.

In the STB geometry, the ionization front is specified by the ionization parameter X, which is given by

$$\begin{eqnarray}&&X=\displaystyle \frac{4\pi {{aL}}_{H}}{{\alpha }_{B}}{\left(\displaystyle \frac{{m}_{H}{v}_{\infty }}{\dot{M}}\right)}^{2}.\end{eqnarray} \tag{ 2 }$$

Here, a is the binary separation, L_H is the H-ionizing photon number luminosity, α_B is the case B recombination coefficient for H i, and m_H is the proton mass. In this work, we adopt the values of a = 56 au and L_H = 7 × 10⁴⁷ s⁻¹ that were used in previous works (González-Riestra et al. 2013; Hinkle et al. 2013).

3.2. An Asymmetric O vi Accretion Disk

3.3. Monte Carlo Results

To reproduce the peak separation of the Raman O vi features, the Keplerian velocity is set to >35 km s⁻¹. It corresponds to the physical size of <0.8 au when we adopt M_WD = 0.65 M_⊙ as the mass of the WD (González-Riestra et al. 2013). The asymmetry of the matter distribution in the accretion disk is given as a function of the azimuthal angle θ measured from the line toward the Mira. The best fit is made using the following density function,

$$\begin{eqnarray}&&n{(\theta )={n}_{0}(1+\theta \times (2\pi -\theta )}^{6}/8500).\end{eqnarray} \tag{ 3 }$$

In Figure 5, we present the schematic top view of the O vi disk with the corresponding O vi emissivity (left). In the right panel, the solid line represents the O vi density profile n(θ)/n₀ as a function of θ. The O vi emissivity is dominant in the RER (0 < θ < π), whose peak is at around θ = 0.3π.

Zoom In Zoom Out Reset image size

Considering the disparate profiles of the two Raman O vi, we use the flux ratio F(1032)/F(1038), varying from 1.5 to 2, as a function of θ

$$\begin{eqnarray}&&F{(1032)/F(1038)=2-\theta \times (2\pi -\theta )}^{4}/1500,\end{eqnarray} \tag{ 4 }$$

which is schematically illustrated in the middle panel and represented by the dotted line in the right panel of Figure 5.

Adopting the asymmetric O vi accretion disk model supplemented by the locally varying F(1032)/F(1038), Monte Carlo simulations were carried out under various parameters, a mass-loss rate $\dot{M}$ ranging from 1 × 10⁻⁶ to 1 × 10⁻⁵ M_⊙ yr⁻¹ with a step width of 1 × 10⁻⁶ M_⊙ yr⁻¹ and terminal velocities of the Mira wind v_∞ with 10, 15, 20, and 25 km s⁻¹. These parameter sets give the ionization parameter X from 0.2 to 80.

We find that the double-peak structure of Raman O vi features in RR Tel is diluted if the stellar wind velocity exceeds v_∞ > 20 km s⁻¹ or the ionization parameter X < 1.25. With the maximum velocity of v_∞ = 20 km s⁻¹, we, therefore, constrain $\dot{M}\lt 8\times {10}^{-6}\,{M}_{\odot }\,{\mathrm{yr}}^{-1}$, which gives X ∼ 1.25. The corresponding ionization front is presented in Figure 1. The best-fit result is obtained with v_∞ ∼ 20 km s⁻¹ and $\dot{M}\sim 2\times {10}^{-6}\,{M}_{\odot }\,{\mathrm{yr}}^{-1}$. In Figure 6, we present our best-fit profiles for the Raman O vi features. The black lines show the MIKE data, while the light gray lines represent the results of our simulations.

Zoom In Zoom Out Reset image size

Raman scattering efficiency, η, is defined as the number ratio N(FUV)/N(Raman) of the incident and Raman-scattered photons. The conversion rate for Raman-scattered O vi features was derived from direct measurement of the O vi far-UV emissions in far-UV spectra and O vi Raman features in optical spectra (Schmid et al. 1999; Birriel et al. 2000). Those authors suggested that N(6825)/N(1032) ∼ 6% and N(7082)/N(1038) ∼ 2%, respectively. From our Monte Carlo simulations, we obtained 10.5% for the λ1032 → λ6825 conversion and 4.3% for λ1038 → λ7082 conversion, respectively.

For our best-fit results, we obtain the total line flux F(6825) = 3.0 × 10⁻¹² erg cm⁻² s⁻¹ and F(7082) = 0.7 × 10⁻¹² erg cm⁻² s⁻¹, respectively. Adopting a distance of 2.6 kpc (Whitelock 1988; González-Riestra et al. 2013), the corresponding number luminosities of Raman-scattered photons emitted per unit time are N(6825) = 5.6 × 10⁴⁵ s⁻¹ and N(7082) = 1.3 × 10⁴⁵ s⁻¹. Taking the conversion efficiency for O vi, we deduce the far-UV number flux densities N(1032) = 5.3 × 10⁴⁶ s⁻¹ and N(1038) = 3.1 × 10⁴⁶ s⁻¹. The corresponding line fluxes are obtained F(1032) = 1.3 × 10⁻⁹ erg cm⁻² s⁻¹ and F(1038) = 7.3 × 10⁻¹⁰ erg cm⁻² s⁻¹, yielding the flux ratio F(1032)/F(1038) of 1.8. This result does not deviate much from the observed value of 1.7, as shown in Section 2.2. In Table 4, we summarize the intrinsic far-UV O vi parameters including, center wavelength, Raman scattering efficiency η, calculated number flux density N_exp, expected line flux F_exp, and the observed flux F_tobs in FUSE data.

Table 4. Observed far-UV O vi λλ 1032, 1038 Line Fluxes (F_obs) and Expected Values (N_exp, F_exp) from the Best-fit Monte Carlo Result

Line	λ	η	Nexp	Fexp	Fobs
	(Å)		(s−1)	(erg cm−2 s−1)	(erg cm−2 s−1)
O vi 1032	1031.928	0.105	5.3 × 1046	1.3 × 10−9	2.01 × 10−10
O vi 1038	1037.618	0.043	3.1 × 1046	7.3 × 10−10	1.18 × 10−10

Download table as:  ASCIITypeset image

The O vi λλ 3811 and 3834 lines arise from transitions between the level $1{s}^{2}\,3p\,{(}^{2}{P}_{3/2,1/2}^{0})$ and $1{s}^{2}\,3s\,{(}^{2}{S}_{1/2}^{0})$, respectively. We present a partial Grotrian diagram for O vi in Figure 7. O vi λ 3811 appears in the optical line list of RR Tel by McKenna et al. (1997), who obtained a high-resolution spectrum in a wavelength range from 3430 to 9320 Å with a spectral resolution of R ∼ 25,000. In their spectrum, they found an emission line feature at 3811.35 Å with Δv ∼ 72 km s⁻¹ and suggested its identification as a blend of O iii 3810.96 and O vi 3811.36. Crawford et al. (1999) extended the line list using data with a higher spectral resolution of R ∼ 50,000. They proposed that the O vi 3834 line could contribute to the emission at 3833.96 Å, which is detected with Δv ∼ 74 km s⁻¹ and blended with He ii 3833.80 and He i 3833.57. Young (2012) performed an extensive study of O vi recombination lines in RR Tel, from which 19 lines, including O vi 3811 and 3834, were identified. They suggested that a highly ionized region, producing the lines by recombination onto O vii, is situated in the inner structure nearer to the hot WD than the O vi emitting region.

Zoom In Zoom Out Reset image size

In this work, we perform a detailed profile decomposition of the emission features at 3811 and 3834 Å with our MIKE data and confirm the identification of the O vi λλ 3811, 3834 doublet. The first step to isolate the O vi doublet from the blended profiles is to infer the profile width of the O vi doublet. It is expected that the profile width of the O vi doublet is comparable to that of N v λλ 4603, 4619 and C iv λλ 5801, 5812 doublets, which have similarly high ionization potential and isoelectronic configurations with O vi. With a single Gaussian function, the N v doublet lines are located at 4602.85 Å and 4619.14 Å with FWHM = 48 km s⁻¹ and 50 km s⁻¹, respectively. The C iv doublets are found at 5800.22 Å and 5810.90 Å with FWHM = 43 km s⁻¹ and 42 km s⁻¹, respectively. Our fitting parameters with a single Gaussian function for the N v and C iv doublets are summarized in Table 5.

Our MIKE data shows that the 3811 emission feature has a single-peak profile with FWHM ∼ 62 km s⁻¹. To decompose it into O vi 3811 and O iii 3810, the profile width of O iii 3810 is derived from the adjacent O iii emission line at 3791 Å, whose profile width is measured to be ∼29 km s⁻¹. Within the expected FWHM of 44 km s⁻¹ for O vi 3811 and 29 km s⁻¹ for O iii 3810, we find that O vi 3811 is centered at 3810.74 Å with a line flux F(3811) = 1.5 × 10⁻¹³ergcm⁻² s⁻¹. The fitting result is shown in the left panel of Figure 8.

Zoom In Zoom Out Reset image size

As shown in the right panel of Figure 8, the blended emission feature at 3834 Å has a complicated profile consisting of strong Balmer H9 λ 3835, two peaks and a blue shoulder, implying that the 3834 blend comprises at least four emission lines. For decomposition, we first analyze Balmer H9 3835 and He ii 3833 lines by taking H8 3888 and He ii 3887, and H7 3969 and He ii 3967 as reference lines. The Balmer line profiles are well fitted using three emission components: a central main part, broad wings, and a red bump. The Balmer line fluxes are measured to be F(H7) = 2.5 × 10⁻¹² ergcm⁻² s⁻¹, F(H8) = 1.9 × 10⁻¹² ergcm⁻² s⁻¹, and F(H9) = 9.3 × 10⁻¹³ ergcm⁻² s⁻¹. The flux ratio F(H7): F(H8): F(H9) = 1: 0.78: 0.37 deviates from the case B recombination, which yields a declining trend 1: 0.66: 0.46.

He ii 3887 and 3967 lines are fitted with a single Gaussian function with FWHM = 0.54 Å. Considering the flux ratio of Balmer lines and He ii emissions, the blue shoulder located at around 3833 Å is plausibly identified with He II emission so that we use the term "He ii 3833." The multi-Gaussian fitting results for the Balmer and He ii emission lines are presented in Figure 9 and Table 6.

Zoom In Zoom Out Reset image size

We find one narrow peak with a small flux at 3833.29 Å. Using a single Gaussian, the FWHM is measured to be 20 km s⁻¹, which is smaller than the O vi 3811 line (44 km s⁻¹). In this regard, we suggest that the narrow peak at 3833.29 Å is mainly contributed by the He i 3833 line.

Taking these three emission lines into account, we finally isolate O vi 3834 with a Gaussian function having the same FWHM (44 km s⁻¹) as that of O vi 3811. We locate the line center at 3833.67 Å and measure a flux of F(3834) = 8.9 × 10⁻¹⁴ ergcm⁻² s⁻¹, resulting in F(3811)/F(3834) ∼ 1.7.

The results obtained from our simulations are based on simple kinematics with a Keplerian disk with an asymmetric matter distribution. Heo et al. (2016) proposed that the density asymmetry of the accretion disk is responsible for the profile disparity in the Raman O vi features. The density distribution yielding the best-fit profiles is depicted in Figure 5, where we note that the highest density on the entering side is 3.5 times that of the most sparse region on the opposite side. In reality, the high contrast of the density profile can be made by the elliptical shape of the accretion disk or clumpy, irregular structures in the flow, which should be general features because of the shock front, accretion wake, or the unstable behavior in the wind accretion process. A number of numerical works of wind accretion in symbiotic stars produced a strong spiral shock and eccentric stream flow, supporting the local density enhancement (e.g., de Val-Borro et al. 2009, 2017). However, the 3D spatial information, including its elongation, scale height, and radial motion, etc., is mixed into the Raman O vi profile. To disentangle the properties and geometry of the accretion flow, further modeling combining 3D high-resolution SPH simulations for wind accretion is needed. In Figure 6, our best-fit Monte Carlo profile is insufficient to reproduce the red wing part of the Raman 6825 feature at ΔV ∼ 70 km s⁻¹ only with our accretion disk model. We point out that the red wing part can be explained with an additional component moving away from the binary orbital plane (e.g., Heo et al. 2016; Lee et al. 2019).

We also note that the blue peak of the Raman 7082 is far below the theoretical range. Taking into account the ratio of the Raman scattering cross-section for the O vi 1032 and 1038 photons σ_Ram(1032)/σ_Ram(1038) = 3.04 and their oscillation strength τ(1032)/τ(1038) = 2, the acceptable range of the F(6825)/F(7082) should be <6.08. However, the MIKE data show that the flux ratio of the blue peaks f_blue(6825)/f_blue(7082) is 9.7 in 2017. One possible explanation for the relative suppression of the Raman feature was suggested by Schmid et al. (1999), who proposed the diminution is due to the presence of molecular TiO transitions in the atmosphere and wind of the Mira component. As discussed in Section 2.1, however, the Mira of RR Tel is not visible in the optical wavelength. Hence its absorption lines can not be the reason to explain the suppression of the blue portion of the Raman 7082. Further investigation is needed to interpret the discrepancy between the theory and the observation.

In comparison with the FEROS spectrum taken in 2003, we note that the profiles of Raman O vi features have changed, their red peaks in the MIKE data being stronger. The variation of the Raman O vi profiles implies that the density distribution of the accretion disk has been more asymmetric in the last two decades. The increasing asymmetry of the matter distribution could be related to the dust obscuration episode that occurred between 1996 and 2000, which is associated with the enhancement of the mass-loss rate (Kotnik-Karuza et al. 2006; Jurkic & Kotnik-Karuza 2012).

We also found the O vi doublet at 3811 and 3834 Å, which are blended with other emission lines. Our profile decomposition leads us to investigate whether the O vi λλ 3811, 3834 doublet has a single Gaussian profile with a width of ∼44 km s⁻¹, corresponding to a physical size of ∼0.5 au. It is worth bearing in mind that the emission region of the O vi λλ 1032 and 1038 doublet has a representative scale of <0.8 au. This result is consistent with the idea that the O vi recombination lines are produced in the inner part closer to the WD than the O vi resonance lines proposed by Young (2012). Photoionization modeling is used to obtain the ionization structure of oxygen in the nebula. We conduct a simple exercise for a spherical geometry with a uniform density assuming physical parameters of hot WD, L_H ∼ 5000 L_⊙ and T_eff ∼ 154,000 K, adopted in our work. The exercise gives a hydrogen volume density n(H) ∼ 10^9.6 cm⁻³. Our estimation appears to be consistent with 3D hydrodynamical simulations of focused wind accretion in symbiotic stars by de Val-Borro et al. (2017, see their Figure 8). Their model with a = 8 au, $\dot{M}\sim {10}^{-6}\,{M}_{\odot }\,{\mathrm{yr}}^{-1}$, and v_wind ∼ 20 km s⁻¹ produced the accretion stream around WD exceeding the local density >10⁻¹⁶ g cm⁻³. If we adopt the solar abundance, this value is translated into the hydrogen volume density of n(H) > 10⁹ cm⁻³. It should be noted that a direct comparison between our work and theirs is not feasible considering the different parameter sets adopted in this work.

The identification of the O vi λ λ 3811 and 3834 doublet strongly supports the presence of the EUV doublet at 150.09 Å and 150.13 Å, originating from 3p → 2s transitions, as shown in Figure 7. Because of a small separation of ΔV ∼ 80 km s⁻¹ between two wavelengths, we expect a blended emission at 150.1 Å with a spectrograph having a spectral resolution R < 10,000. O vii is one of the highest observed stages of ionization in the RR Tel spectrum, and thus can provide a reliable diagnostic of high-temperature plasma. With future EUV observations, our analysis for the O vi and O vii groups could reveal important information on the ionization state and nebular physics in RR Tel.

RR Tel is a symbiotic nova, consisting of a Mira variable and a hot WD. It is known as one of the objects having broad emission features at 6825 and 7082 Å, which originate from Raman scattering of O vi 1032 and 1038 by neutral hydrogen atoms. In this work, we present optical high-resolution spectra of RR Tel obtained with MIKE and FEROS and the far-UV FUSE spectrum. We have performed Monte Carlo simulations for the profiles of the Raman O vi features, assuming a Keplerian disk. It allows us to map the O vi disk and estimates the representative scale of <0.8 au. The best-fit profiles are obtained with a mass-loss rate of the Mira $\dot{M}\sim 2\times {10}^{-6}\,{{\rm{M}}}_{\odot }\,{\mathrm{yr}}^{-1}$ and a wind terminal velocity v_∞ ∼ 20 km s⁻¹. Comparison with FEROS data provides evidence of a change in the density asymmetry of the O vi disk, which can be associated with the mass-transfer history of RR Tel in the last two decades. We also identify O vi λ λ 3811 and 3834 doublet from detailed profile decomposition, from which the physical scale of the O vii region is estimated to be ∼0.5 au. It is emphasized again that Raman profiles provide essential information regarding the matter distribution and kinematics of the H i and O vi components leading to the advancement of our knowledge on the stellar wind accretion in symbiotic stars. Combining 3D hydrodynamical modeling might improve our overall understanding of the WRLOF mode and the accretion phenomena in D-type symbiotic systems.

We would like to thank the anonymous referee for constructive comments that greatly improved this paper. We are also grateful to Young-Min Lee for valuable discussions. This research was supported by the Korea Astronomy and Space Science Institute under the R&D program (Project #2018-1-860-00) supervised by the Ministry of Science, ICT, and Future Planning. J.E.H. is supported through a Gemini Science Fellowship by the international Gemini Observatory, a program of NSF's NOIRLab, which is managed by the Association of Universities for Research in Astronomy (AURA) under a cooperative agreement with the National Science Foundation, on behalf of the Gemini partnership of Argentina, Brazil, Canada, Chile, the Republic of Korea, and the United States of America. J.E.H. was also supported from Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (#2019R1A6A3A03033787). R.A. acknowledges financial support from DIDULS Regular PR#1953853 by Universidad de La Serena. This paper includes data gathered with the 6.5 m Magellan Telescopes located at Las Campanas Observatory, Chile. This work is based on observations collected at the European Southern Observatory under ESO program ID 60.A-9120(B).

Facilities: Magellan:Clay (MIKE), Max Planck:2.2 m (FEROS), FUSE.

Software: IRAF (Tody 1986, 1993), CarPy (Kelson et al. 2000; Kelson 2003).