TCP J18224935-2408280: a symbiotic star identified during outburst

TCP J18224935-2408280 was reported to be in outburst on 2021 May 19. Follow-up spectroscopic observations confirmed that the system was a symbiotic star. We present optical spectra obtained from the Himalayan Chandra Telescope during 2021-22. The early spectra were dominated by Balmer lines, He I lines and high ionization lines such as He II. In the later observations, Raman scattered O VI was also identified. Outburst in the system started as a disc instability, and later the signature of enhanced shell burning and expansion of photospheric radius of the white dwarf was identified. Hence we suggest this outburst is of combination nova type. The post-outburst temperature of the hot component remains above 1.5 x 10$^5$ K indicating a stable shell burning in the system for a prolonged time after the outburst. Based on our analysis of archival multiband photometric data, we find that the system contains a cool giant of M1-2 III spectral type with a temperature of $\sim$ 3600K and a radius of $\sim$ 69 R$_\odot$. The pre- and post-outburst light curve shows a periodicity of 631.25 $\pm$ 2.93 d; we consider this as the orbital period.


INTRODUCTION
Symbiotic stars are interacting wide binaries consisting of a cool giant of spectral type M (or K) as a donor star and a hot component, mostly white dwarf (WD) accreting from the giant's wind and surrounded by a circumstellar nebula (Mikołajewska 2012).Symbiotic stars manifest a wide variety of variability, from orbital motion to outbursts.Outbursts in symbiotic stars are classified into three, symbiotic novae or slow novae, symbiotic recurrent novae, and classical symbiotic outburst (Z And-type).Symbiotic novae and symbiotic recurrent novae are powered by thermonuclear runaway reactions, whereas classical symbiotic outbursts are believed to be either caused by the release of potential energy from the extra-accreted matter or due to the increased mass accretion rate followed by the expansion of hot component (Munari 2019).Classical symbiotic outbursts are commonly seen feature in symbiotic stars and typically show a 1-3 B mag brightening in the system during the outburst.
Although classical symbiotic outbursts are one of the most common features of symbiotic stars, we still have limited knowledge about the exact mechanism of these outbursts.In the literature, four different models are proposed to explain these outbursts: 1) Expansion of WD photosphere at a near constant bolometric luminosity due to an increased accretion rate that exceeds steady burning (Tutukov & Yungel'Son 1976;Iben 1982), 2) Shell flash or thermal pulse similar to nova and recurrent nova (Kenyon & Truran 1983), 3) Dwarf nova-like outburst due to accretion disc instability (Duschl 1986a,b;Mikolajewska et al. 2002), 4) Combination nova, where an outburst is initiated by disc instability following an enhanced shell burning ★ E-mail: sonith.sls@gmail.com,sonith.ls@iiap.res.in,(Sokoloski et al. 2006).In symbiotic stars, it could also be possible to see the same systems showing outbursts with different mechanisms as in AG Peg (Tomov et al. 2016) or Z And (Sokoloski et al. 2006).To understand the nature of classical symbiotic outbursts, we require spectroscopic follow-up observations of more systems.
TCP J18224935-2408280 (hereafter referred to as TCP J1822) was discovered by Tadashi Kojima, Tsumagoi, Gunma-ken, Japan, on 2021 May 19.683 UT.Discovery was reported in the 'Transient Object Followup Reports' pages 1 of the Central Bureau for Astronomical Telegrams (CBAT).It was suggested to be a symbiotic star outburst by Patrick Schmeer (Saarbrucken-Bischmisheim, Germany) after he found a 2 arcsec nearby Gaia LPV source (Gaia DR2 4089297564356878720) with an approximate orbital period of 800 d.This star is included in the Gaia DR2 catalogue of large-amplitude variables by Mowlavi et al. (2021).The spectroscopic follow-up observation by Merc et al. (2021) on 2021 June 09 showed strong emission lines of H I, He I, O [III], and He II in addition to the K5-M0 continuum.They noted that TCP J1822 is an S-type symbiotic star based on infrared colours; also, the distance and apparent magnitude of the system suggest that it contains a cool component of luminosity class III.The follow-up observations by Aydi et al. (2021) reached a similar conclusion, and in addition, they have also reported Bowen blend and relatively weak emission lines of Fe II (42,48,49 multiplets).Earlier observations by Taguchi et al. (2021) on 2021 June 07, and later observation on June 09 confirm similar nature of TCP J1822 although the identification of the O [III] line is reported weak or absent.We present optical spectroscopic observations of TCP J1822 during 2021-22 and confirm the symbiotic nature of the system and try to understand the nature of the outburst.

Photometry
For understanding the behaviour of TCP J1822 before and during the outburst, we have obtained V and g band photometric data from ASAS-SN sky survey (Shappee et al. 2014;Kochanek et al. 2017), covering the period JD 2457461.83 to JD 2460138.2 (2016March 14 -2023 July 12), G, G  and G  band magnitudes from Gaia DR3 (Gaia Collaboration et al. 2022), covering the period JD 2456913.47 to JD 2457506.75 (2014September 12 to 2016 April 28).Gaia and ASAS-SN light curves are shown in Fig. 1.

Spectroscopy
Low-resolution optical spectra of TCP J1822 were obtained from Himalayan Faint Object Spectrograph Camera (HFOSC) mounted on Himalayan Chandra Telescope (HCT) situated at Indian Astronomical Observatory, Hanle.Observations were carried out between 2021 June 10 and 2021 September 19, using grism 7, having a wavelength range of 3500 to 8000 Å with a resolution of R ∼ 1300 and grism 8, having a wavelength range of 5200 to 9000 with a resolution of R ∼ 2200.A majority of these observations were carried out in the Target of Opportunity (ToO) mode.The details of the observations are given in Table 1.The data reduction was carried out using a pipeline based on python using pyraf modules, following the standard procedure using different tasks in the Image Reduction and Analysis Facility (IRAF2 ).Wavelength calibration was carried out for grism 7 and grism 8 using FeAr and FeNe arc lamp spectra, respectively.Feige 110 and Feige 66 were used as standard stars.On the nights in which observation was carried out in the ToO mode, spectrophotometric standards observed on the nearest night were used for correcting the instrumental response.The response corrected spectra in two grism were scaled to a weighted mean and combined to give the final spectrum.ASAS-SN g-band photometric light curve was used for calibrating the spectra to the absolute flux scale.

Optical light curve and periodicity
The ASAS-SN g-band light curve (Fig. 1) shows that TCP J1822 started brightening on 2021 May 16 and peaked at around 13.5 mag with an increase of 2.2 mag from its quiescent state.Such a 2-3 mag brightening is often seen in Z-And type symbiotic outbursts.The triangular-shaped outburst peak is similar to the light curve of Z-And outburst of 2000 (Sokoloski et al. 2006), where it is suggested that a disc instability event causes an initial brightening in the light 2459822.171800 + 1200 3800-9000 curve.There followed a decrease of 0.5 mag in the next ten days and a re-brightening to a second maximum, with a broader peak.This follow-up event appears to be related to the nuclear burning on the surface of the WD.After 2021 June 5, the g-band magnitude started to decline again, and TCP J1822 returned to its photometric quiescent state about a year after outburst.
The pre-and post-outburst light curves of TCP J1822 show wavelike variations.Using the Lomb-Scargle periodogram (LSP) (Lomb 1976;Scargle 1982), we have obtained periods of 598.95, 618, 598.95 and 609 d corresponding to the highest peaks in the Gaia G, G  , G  and ASAS-SN V bands, respectively.A similar analysis using the ASAS-SN g band removing magnitudes during the outburst and shifting the post-outburst quiescent magnitudes to the pre-outburst level gives a period of 629.75 d for the highest peak.Furthermore, we estimated the period using multiband data after applying appropriate magnitude shifts and combining them, resulting in the highest LSP peak at 631.69 d, with a false alarm probability of <0.01 per cent.These periodograms are shown in Fig. 2. Other peaks obtained from the Gaia data are due to the sampling effect (see appendix A).Additionally, we verified our result for the multiband data using LombScargleMultiband function implemented in astropy (Astropy Collaboration et al. 2018), which gives the same result.Using this period as an initial guess, we have fitted a sinusoidal curve and estimated period, light curve minima and associated errors.Based on the above analysis, we have obtained an ephemeris of TCP J1822 given by    = 2457541.69± 5.78 + 631.25 ± 2.93 × (1) We attribute this periodicity of 631.25 ± 2.93 d to the orbital period.Using the period and phase based on the ephemeris, the G  , G and G  data points were fitted using a sinusoidal function by varying the amplitude of the sinusoid (see Fig. 3).The resultant amplitudes are approximately 0.36, 0.2 and 0.15 mag in the G  , G and G  bands, respectively.The larger amplitude at the shorter wavelength is indicative of the irradiation of the red giant by the hot WD.This suggests that there was quiescent burning on the suface of the WD in pre-outburst quiescence.Photometry over the next few years will help in refining this value.Future multiband observations would help to delineate the effects of ellipsoidal modulation (which should be prominent in the  band) and irradiation by the hot component ( band).

Distance and reddening
From the Gaia EDR3 parallaxes (Gaia     2021) method.We have estimated a visual extinction of  v ∼ 1.48 in this direction and for this distance using the 3D map of interstellar dust reddening published by Green et al. (2019).We calculated  v values using the same procedure for the upper and lower bounds of the distance error, and the results are identical.The  v value derived from the Bailer-Jones et al. ( 2021) approach in the direction of the object does not show significant differences beyond 5 kpc.The map by Schlafly & Finkbeiner (2011) indicates that the visual extinction in the direction of TCP J1822 is,  v = 1.73.Goodness-of-fit of the astrometric model is -0.16 for Gaia EDR3.A lower than 3 is considered to be a good fit3 .However, Bailer-Jones et al. ( 2021) use a probabilistic approach for estimating distances, which relies on priors constructed based on single stars within our Galaxy.It can cause considerable uncertainties for binaries like symbiotic stars.In this work, we use the distance calculated using EDR3 parallaxes for calculating  v value and distance prior given for the SED fit (see section 3.3).Considering that the  v value is not showing much difference above 5 kpc distance, we are fixing our  v value to a conservative 1.48 in all the calculations done in this paper.Our spectral type estimation of the cool giant in the system will be hotter by 1-2 spectral sub-types if we take into account a larger  v value from Schlafly & Finkbeiner (2011).Since our best-fitting spectral type from SED matches well with the quiescent TCP J1822 spectrum above 6000 Å, where the contribution from giant dominates, we find the conservative  v value we have adopted is suited for subsequent calculations.Reddening corrections were done using the extinction law of Fitzpatrick (1999).

Spectral energy distribution of cool component in TCP J1822
The Virtually no excess over the stellar continuum is seen in the WISE W3 and W4 bands, indicating that this is an S-type symbiotic.The fit is quite reasonable, given that the data represent different epochs.For running ARIADNE, we have used the temperature prior based on the Gaia temperature estimate and upper limit.The distance prior is taken from the distance estimate (see section 3.2), setting the highest error as the upper limit.The  v value has been kept fixed at 1.48.We have set a uniform prior for log g from 0 to 4 based on our initial fitting, which results in a radius estimation in the giant star regime.We have used the default prior for radius and metallicity.The best fit gives a temperature of 3614.It should be noted that the temperature, radius and log g are more accurately constrained, whereas [Fe/H] is indicative, and the obtained SED represents only the mean spectrum of the red giant in TCP J1822.Since no UV photometry is available, we are unable to probe the nature of the hot component or its orbital variations in a similar manner.

Evolution of the optical spectra
We obtained three spectra of TCP J1822 in 2021, during the declining phase of the outburst and four in 2022, after the outburst has almost subsided.Significant changes can be seen in the emission lines and the continuum over this period (see Fig. 6).The very blue continuum along with strong Balmer and Paschen lines in the spectrum of 2021 June 10 are suggestive of a prominent, hot accretion disc.As the photometry data also indicates, the beginning of the outburst was due to an accretion event.The circumstellar nebula, illuminated by the luminous WD, also contributed to the emission lines.As the outburst progressed, the blue continuum weakened and the contribution from the red giant in the form of visibility of some TiO bands became 4000 4500 5000 5500 6000 6500 7000 7500 8000 8500 9000 9500 Wavelength    C1.The evolution of emission lines of TCP J1822 gives an insight into the nature of the outburst and the hot component present in the system.All the emission lines fluxes, including Balmer lines, and H I lines, showed an increasing trend in the first (2021 June 10) to the second (2021 July 14) observation and later declined.The O I 7774 Å line is significantly fainter than O I 8444 Å.This may be due to the fluorescence of Ly photons, where Ly photons at 1025.72 Å pump the O I ground state resonance line at 1025.77Å and later downward cascade to produce 11287Å, 8444Å, and 1304Å lines in emission (Bowen 1947;Kastner & Bhatia 1995).
In the quiescent phase spectra (obtained on and after 2022 March 08) all the emission line strengths were reduced, except Raman scattered O VI line and O I 8444 Å. Faint lines of [Ca VII] 5618 Å, [Fe VII] 5721 Å, [Ca V] 6086 Å or [Fe VII] 6087 Å appeared and later reduced their strength in the subsequent observations.High excitation lines emerging in the quiescent spectra indicate that the expanding pseudo-photosphere becomes optically thin (see section 3.5); hence the nebular region is exposed to the heated WD.

Raman scattered O VI line
Raman scattered O VI line is a unique feature that can only be present in a system with neutral hydrogen regions and a hot component which could ionize oxygen to the fifth ionized state (Nussbaumer et al. 1989).Hence this feature alone can confirm the symbiotic nature of a system.In our optical spectrum obtained on 2021 June 10 near the g-band maximum, this line was almost invisible.However, it strengthened significantly in the second observation on 2021 July 14 and remained at a similar strength throughout the later observations (see Fig. 8).Shape of the line was broad on 2021 July 14, but narrowed subsequently.Variations of the Raman scattered O VI line correlated to the optical light curves were also seen in other outbursting symbiotics such as V426 Sge (fig.7

Bowen feature
The Bowen feature near 4640 Å is seen in all our spectra.This feature was also reported in earlier spectra obtained by Aydi et al. (2021); Taguchi et al. (2021).The strength of the feature reduced as the outburst declined.The Bowen feature is produced when X-rays emitted from a compact source interact with nearby gaseous matter.This indirectly indicates that the system produces X-rays during the outburst --type (Luna et al. 2013) -and is consistent with the presence of Raman scattered O VI line in systems showing -type X-ray emission (Akras et al. 2019).The Bowen feature is also shown by other symbiotic systems like RR Tel and AG Peg (Eriksson et al. 2005).These two stars are also reported to show -type and type X-ray emission, respectively (Luna et al. 2013).The presence of an enhanced blue continuum during the outburst also indicates the possibility of the -type X-ray component in the system, which originates from the inner layer of accretion disc (Luna et al. 2013).Another similar symbiotic with an active accretion disc is MWC 560 (fig.G1 of Lucy et al. 2020).

Nature of the hot component
The lower limit of the temperature of the hot component ( h ) in a symbiotic star can be estimated using the empirical relation Murset & Nussbaumer (1994).This is based on the highest observed ionization potential ( max ) of an emission line seen in the optical spectrum.Using this, we determine  h ≳ 114 000 K from the presence of Raman scattered O VI band at 6825 Å in the spectra of TCP J1822, corresponding to the highest ionization potential  O +5 ∼ 114 eV.
The hot component is best studied using x-ray and uv observations.In the absence of those, emission lines in the optical are a good proxy for understanding its nature.Considering the hot source to be a blackbody, its temperature and luminosity can be calculated based on H, He I and He II lines assuming case B recombination.We have used the relation (2) derived by Iĳima (1981) The luminosity of the hot component was calculated using equation (8) of Kenyon et al. (1991) and equation ( 6) given in Mikolajewska et al. (1997).Both results match within 25 per cent, and the average value of these estimates is given in Table 2.The luminosity estimate using equation ( 7) of Mikolajewska et al. (1997), which is based on H flux, gives a value nearly half of the above.This is not unexpected given that Mikolajewska et al. (1997) noted these equations have a factor of ∼2 accuracy.A similar effect was reported in the case of Hen 3-860 by Merc et al. (2022), where they have shown H lines having an absorption component seen in high-resolution observations, and hence the flux is getting underestimated.However, we do not see any absorption feature in our low-resolution spectra of TCP J1822.
The blackbody assumption also allows determination of the radius, which is given in Table 2. From Fig. 9 it is evident that radius of the hot component showed an increasing trend during the outburst decline.There is an enhancement in the blue wings of H early during the outburst; the line width is also broader (see Fig. 10).The radius suddenly dropped when TCP J1822 reached the quiescence phase (last four observations).The increase in radius was due to the physical expansion of the photosphere caused by excess burning on the surface of the WD.As the photosphere expanded, the temperature dropped.During the quiescence phase, the expanded shell became optically thin, and hence radius showed a sudden drop, which means we started seeing closer to the WD again.
H wing profiles presented in Fig. 10 are obtained by subtracting the local continuum using fit_continuum function in Specutils (Earl et al. 2023).We see H wings as broad as ∼3500 km/s in the blue region and ∼3000 km/s in the red region for the first three spectra taken during the outburst.The H wings in the blue region are stronger than those in the red region.Line broadening has been reported in past outbursts in AG Peg (fig. 3 in Tomov et al. 2016) andV426 Sge (fig. 3 in Skopal et al. 2020).However, in the case of AG Peg and V426 Sge, velocities of H wing profiles are lower (≤1500 km/s) compared to what we observe in TCP J1822.These broadenings are due to an increased outflow during the outburst.

Nature of the outburst
The optical eruption observed in TCP J1822 shows an amplitude of around 2.5 mag in the ASAS-SN g-band and is similar to Z And-type outburst seen in classical symbiotic stars.They show a brightening of 1-3 mag with time scales from months to years (e.g.Z And, CI Cyg, and AG Dra).In addition, spectroscopic observations of TCP J1822 after the optical maximum show that forbidden lines (e.g.Kenyon et al. 1991;AG Dra -Mikolajewska et al. 1995;LIN 9 -Miszalski et al. 2014).The multi-peak light curve of TCP J1822, with a sharp rise during the outburst, resembles that of Z And, which showed a combination nova outburst in 2000 (Sokoloski et al. 2006).Dominance of the blue continuum, strong Balmer and Paschen lines in the early outburst, and the nature of the light curve indicate that some sort of disc instability was responsible.This probably deposited additional matter on the already burning WD, causing the second peak in the light curve (see Fig. 7).Dwarf nova-like disc instability as a triggering mechanism for Z And outburst is also examined in the theoretical model by Bollimpalli et al. (2018).It is estimated that a high accretion rate of the order of 10 -6 M ⊙ yr -1 is required for such a scenario to be feasible in a symbiotic star like Z And.Bollimpalli et al. (2018) suggest that such an enhancement in mass transfer could be attributed to the magnetic activity on the surface of the giant as suggested by Leibowitz & Formiggini (2008).In this scenario, the increased mass transfer could act as a trigger mechanism for enhanced shell burning.The continuum observed during the outburst of TCP J1822 is derived from multiple components, including the nebula, accretion disc, and WD.Understanding the individual contributions of each component requires rigorous modelling, which is beyond the scope of this paper.
The presence of high ionization lines like He II 4686 and Ramanscattered 6825 from outburst through near-quiescence indicates that the WD continued to burn matter on its surface.This could give rise to detectable soft X-rays.X-ray data would be needed to understand the relative contributions of steady nuclear burning and accretion in the system.The strength of the Raman scattered O VI line remains high even after nearly a year after the outburst declined, indicating that enough material reached the surface of the hot component to maintain the shell burning for a prolonged time.From the ASAS-SN g band light curve (Fig. 1), it is seen that the post-outburst magnitude is brighter than the pre-outburst magnitudes, which further confirms our finding.
After returning to quiescence, TCP J1822 exhibits a temperature of above 10 5 K, luminosity of order 10 3 L ⊙ , which is typical for the hot component in quiescently burning symbiotic stars (fig.4 in Mikołajewska 2003, andMunari 2019).which unambiguously confirm the symbiotic nature of the system.

CONCLUSIONS
(ii).We probed the nature of the cool component in the system using multiband SED and found that the system contains an M1-2 III spectral-type star having a temperature of ∼ 3600K, radius of ∼ 69 R ⊙ and luminosity of ∼ 700 L ⊙ .
(iii).TCP J1822 shows a combination nova type outburst where the outburst begins as accretion disc instability during the first peak of the light curve and then enhances the shell burning in the system, which is correlated with the radius increase of WD photosphere.
(iv).The pre-and post-outburst light curve of TCP J1822 shows a 631.25 ± 2.93 day periodic variation, which most probably originates from the orbital motion of the system.
(v).The post-outburst temperature of the hot component remains above 1.5x105K, indicating a stable shell burning in the system for a prolonged time after the outburst.The strength of Raman scattered O VI band and elevated post-outburst ASAS-SN g band magnitude compared to pre-outburst also confirms the same.These findings collectively suggest an enhanced mass transfer during the outburst.

Figure 1 .
Figure 1.Light curve of TCP J1822 using the Gaia G, G  , G  magnitudes and ASAS-SN g and V band magnitudes.Periodic behaviour of TCP J1822 is evident in the pre-and post-outburst light curves.The dashed lines show the best fitting sinusoidal curve based on the ephemeris provided in equation (1).

Figure 2 .
Figure 2. Comparison of Lomb-Scargle periodograms for TCP J1822 obtained using Gaia and ASAS-SN data, including combined multiband data.LSP give dominant peaks at 598.95, 618, 598.95, 609, 629.75, and 631.69 d in G, G  , G  , ASAS-SN V, ASAS-SN g, and multiband data, respectively.The dashed line indicates a 631.69-day period value.See the text for details.

Figure 3 .
Figure 3. Gaia G, G  and G  band light curves of TCP J1822.The G, G  and G  bands are represented by green, blue and red points, respectively.The dashed lines show the sinusoidal function fitted based on the ephemeris for Gaia G, G  and G  bands.See the text for details.

Figure 4 .Figure 5 .
Figure 4. Spectral energy distribution of the TCP J1822 obtained from various bands.Blue points represent the bands used for obtaining the best fit to the SED, as mentioned in section 3.3.WISE W3 and W4 filter magnitudes are over-plotted as orange points.No infrared excess is seen over the stellar continuum.

Figure 6 .
Figure6.The de-reddened optical low-resolution spectra of symbiotic star TCP J1822 at various epochs of its recent outburst.For clarity, each spectrum has been shifted vertically by the indicated amount.
of Skopal et al. 2020) and AG Peg (fig.2ofTomov et al. 2016).However, unlike V426 Sge and AG Peg, in TCP J1822, after the initial rise, Raman scattered O VI line strength remains almost the same even after the light curve declines.It did not show any decreasing trend while the g-band light curve declined.The continued line strength for a prolonged period indicates that some extra mass reached the hot component to sustain the shell burning.

Figure 7 .
Figure 7. (Top) ASAS-SN g-band light curve of the outburst of TCP J1822.Epochs of our spectroscopic observations are marked with red arrows.Evolution of the temperature of the hot component and emission line fluxes are shown in the subsequent plots.

Figure 8 .
Figure 8. Evolution of Raman scattered O VI band during the TCP J1822 outburst.In the first epoch observation (June 10 2021), g-band light curve was at its peak (see Fig. 7) but the Raman scattered O VI band was absent in the optical spectrum.

Figure 9 .
Figure 9. HR diagram showing evolution of the hot component in TCP J1822 from 2021 (orange points) to 2022 (blue points) during the current outburst.

Figure 10 .
Figure 10.Broadening of H line during outburst of TCP J1822.The H line is plotted after subtracting the local continuum.Enhanced bluer wing of H line is an indication of outflow while the outburst happened in the system.
(i).The optical spectrum of TCP J1822 shows Balmer series lines, O I, He I, and high excitation lines such as He II, O[III], Raman scattered O VI and TiO band heads from the cool component

Figure A1 .
Figure A1.Lomb-scargle periodogram generated using simulated Gaia G magnitudes and observed Gaia G magnitudes are shown in the figure.Simulated data points are created on the same observed epochs to check the sampling effect.We assumed a sinusoidal variation in the G band light curve and used the same period we obtained from the observed G magnitudes, 598.95 d.

Table 1 .
Observational log for spectroscopic data obtained for TCP J1822.

Table 2 .
The de-reddened absolute fluxes of H, He II 4686 Å, He I 4471 Å, He I 5876 Å, and O VI 6825 Å line, together with the estimated luminosity, temperature and radius of the TCP hot component.