Repeated transient jets from a warped disk in the symbiotic prototype Z And: A link to the long-lasting active phase
Augustin Skopal, Taya. N. Tarasova, Marek Wolf, Pavol A. Dubovsky, Igor Kudzej
aa r X i v : . [ a s t r o - ph . S R ] M a y Draft version March 19, 2018
Preprint typeset using L A TEX style emulateapj v. 01/23/15
REPEATED TRANSIENT JETS FROM A WARPED DISK IN THE SYMBIOTIC PROTOTYPE Z AND:A LINK TO THE LONG-LASTING ACTIVE PHASE
Augustin Skopal Astronomical Institute, Slovak Academy of Sciences, 059 60 Tatransk´a Lomnica, The Slovak Republic
Taya. N. Tarasova
Scientific Research Institute, Crimean Astrophysical Observatory, 298409 Nauchny, Crimea
Marek Wolf
Astronomical Institute, Charles University Prague, CZ-180 00 Praha 8, V Holeˇsoviˇck´ach 2, The Czech Republic andPavol A. Dubovsk´y and Igor Kudzej
Vihorlat Astronomical Observatory, Mierov´a 4, SK-066 01 Humenn´e, The Slovak Republic
Draft version March 19, 2018
ABSTRACTActive phases of some symbiotic binaries survive for a long time from years to decades. The accretionprocess onto a white dwarf (WD) sustaining long-lasting activity, and sometimes leading to collimatedejection, is not well understood. We present the repeated emergence of highly collimated outflows(jets) from the symbiotic prototype Z And during its 2008 and 2009-10 outbursts and suggest their linkto the current long-lasting (from 2000) active phase. We monitored Z And with the high-resolutionspectroscopy, multicolor
U BV R C –and high-time-resolution–photometry. The well-pronounced bipolarjets were ejected again during the 2009-10 outburst together with the simultaneous emergence of therapid photometric variability (∆ m ≈ .
06 mag) on the timescale of hours, showing similar propertiesas those during the 2006 outburst. These phenomena and the measured disk-jets connection could becaused by the radiation-induced warping of the inner disk due to a significant increase of the burningWD luminosity. Ejection of transient jets by Z And around outburst maxima signals a transientaccretion at rates above the upper limit of the stable hydrogen burning on the WD surface. Theenhanced accretion through the disk warping, supplemented by the accretion from the giant’s wind,can keep a high luminosity of the WD for a long time, until depletion of the disk. In this way, thejets provide a link to long-lasting active phases of Z And.
Subject headings:
Stars: binaries: symbiotic – stars: individual: (Z And) – ISM: jets and outflows INTRODUCTION
Symbiotic stars are the widest interacting binaries withorbital periods of a few years. They consist of a red giant(RG) as the donor and a white dwarf (WD) as the accre-tor (Boyarchuk 1967; Kenyon 1986). Symbiotic stars arewell detached binary systems (M¨urset & Schmid 1999),which implies that their activity is triggered via the windmass transfer.According to the behavior of optical light curves(LC) we distinguish between the quiescent and ac-tive phases. During quiescent phases, the WD ac-cretes throughout the accretion disk formed from thegiant’s wind. This process heats up the WD to > K and increases its luminosity to ∼ − L ⊙ ,which in return ionizes a portion of the neutral windfrom the giant, giving rise to the nebular emission(Seaquist et al. 1984). Optical LCs are characterizedwith a periodic wave-like variation. No sudden bright-enings are indicated. On the other hand, activephases are characterized by a few magnitude (multi-ple) brightening(s)–outbursts–observed on the timescaleof a few months to years/decades (see, e.g., historical E-mail: [email protected]
LCs of FN Sgr, Z And, AX Per of Brandi et al. 2005;Leibowitz & Formiggini 2008, 2013) with signaturesof a mass outflow (e.g., Fern´andez-Castro et al. 1995;Esipov et al. 2000; Sokoloski et al. 2006; McKeever et al.2011). They are called ‘Z And-type’ outbursts, be-cause they were observed in the past for a proto-type of the class of symbiotic stars – Z And (Kenyon1986). In rare cases, these outbursts are followed bycollimated ejection. To date, its signature in the op-tical spectra has been recorded only for a few ob-jects: MWC 560 (e.g. Tomov et al. 1990), Hen 3-1341 (Tomov et al. 2000; Munari et al. 2005), StH α L ⊙ (M¨urset et al. 1991;Skopal 2005), is generated by the stable nuclear burningof hydrogen on the WD surface (e.g., Paczy´nski & Rudak Skopal et al. M a gn it ud e Z And
Q U I E S C E N T P H A S EOutb. Outb. Outb.A C T I V E P H A S EOutb. Outb.
UBVvis. 8 9 10 11 12 52000 53000 54000 55000 56000 57000 2002 2004 2006 2008 2010 2012 2014 2016 M a gn it ud e Julian date - 2 400 000jets jetsno jets no jets
ACT. PHASE
Fig. 1.—
Top:
UBV light curves of Z And from 1981 covering its quiescent and active phases. References for the data used are inSkopal et al. (2000). Bottom: Current active phase that started in September 2000 showing multiple outbursts. Vertical bars indicatetimes of our spectra (Table 1). The presence of bipolar jets is denoted by vertical gray belts. No jets were indicated during the major2000-03 outburst (Skopal et al. 2006) and during 2011-12 and 2014 brightenings (Fig. 3). Data are described in Sect. 2.2. InSeptember 2000, Z And started a new active phase show-ing a series of outbursts with the main optical max-ima in December 2000, July 2006, December 2009 andDecember 2011 (see Fig. 1). At/after the 2006 maxi-mum, highly collimated bipolar jets were detected forthe first time as the satellite emission components to H α and H β lines (Skopal & Pribulla 2006). Their presencewas confirmed by Burmeister & Leedj¨arv (2007) andTomov et al. (2007). Temporal development of jets andtheir possible origin were investigated by Skopal et al.(2009). The event of jets was transient, being detectedthrough the end of 2006, along the decrease of the star’sbrightness. During this period, a low amplitude irregularphotometric variation from the low stage ( . .
02 mag)increased its amplitude to ∼ .
06 mag on the timescale ofhours. Its source was identified with a large disk aroundthe WD. The authors ascribed these phenomena to adisruption of the inner parts of the disk caused by itsradiation-induced warping.In this contribution, we present an analysis of colli-mated mass outflow, which we measured again duringthe 2008 and 2009–10 outbursts. We compare the newjet event with that from the 2006 outburst to establisha more trustworthy disk-jets connection. In this way, weaim to understand better accretion process onto the WDthat keeps the current active phase of Z And for a longtime, showing outbursts with or without jets. Our ob-servations are described in Sect. 2. Their analysis andresults are introduced in Sect. 3. A discussion of the For a comparison, accretion-powered systems generate the lu-minosity of the order of 10 L ⊙ by converting the gravitationalpotential energy of the WD via the accretion disk. results and a summary are provided in Sects. 4 and 5,respectively. OBSERVATIONS AND DATA REDUCTION
Our observations of Z And during its current activephase, from 2008 to 2014, were carried out at differentobservatories.
Spectroscopy (i) At the Crimean Astrophysical Observatory (de-noted as CrAO in Table 1) the high-resolution spectrawere performed with the coud´e spectrograph of the 2.6m ZTSH telescope. The size of the ANDOR IKON-L CCD detector DZ936N was 2048 × ∼
25 000 around H α . Thelow-resolution spectra (R ∼ × SPERED code developed by S. I. Sergeevat the CrAO.(ii) At the Ondˇrejov Observatory (Ondr.), spectra ofZ And were secured with a SITe-005 800 × − per pixel). Stan-dard initial reduction of CCD spectra was carried outusing modified MIDAS and
IRAF packages. Final pro-epeated transient jets from Z And 3
TABLE 1Log of spectroscopic observations
Date JD Region Obs. ⋆ (mm/dd/yyyy) 2 45... (nm)2008 burst01/23/2008 4488.547 642-670 DDO06/11/2008 4628.844 642-670 DDO06/11/2008 4628.856 461-491 DDO06/13/2008 4630.836 642-670 DDO06/19/2008 4636.822 642-670 DDO2009-10 outburst09/26/2009 5100.535 335-740 CrAO a b b b b a b b b b Notes: ⋆ Observatory, a Nasmyth, b coud´e cessing of the data was done with the aid of the SPEFO -package software (Horn et al. 1996; ˇSkoda 1996).(iii) At the David Dunlap Observatory, University ofToronto (DDO), spectra of Z And were performed witha Jobin Yovon Horiba CCD detector (2048 ×
512 pixels of13.5 µ m size) attached to the single dispersion slit spec-trograph of the Cassegrain focus of the 1.88 m telescope.The slit width was 240 µ m corresponding to 1.5 arcsecat the focal plane. Around the H α region, the resolvingpower was ∼
12 000. Standard reduction of the spectrawas performed with the
IRAF -package software.A journal of the spectra is given in Table 1.
Photometry (iv) At the Skalnat´e Pleso and Star´a Lesn´a (pavilionG2) observatories, classical photoelectric
U BV R C mea-surements were carried out by single-channel photome-ters mounted in the Cassegrain foci of 0.6 m reflectors.Data are plotted in Fig. 1. Each value represents theaverage of the observations during a night. Correspond-ing inner uncertainties are of a few times 0.01 mag inthe B , V and R C bands, and up to 0.05 mag in the U band (see Vaˇnko et al. 2015a,b). The data were pub-lished by Skopal et al. (2012). New observations (fromNovember 2011) will be published elsewhere (Seker´aˇs etal., in preparation). To get a better coverage of the inves-tigated period, we complemented our photometry withthe BV R C CCD measurements available at the AAVSOdatabase. .(v) At the Star´a Lesn´a observatory (pavilion G1)the high-time resolution CCD photometry was per- formed during nights on 06/24/2007, 09/21/2007 and07/11/2008. The SBIG ST10 MXE
CCD camera(2184 × µ m) mounted at the New-tonian focus of the 0.5 m telescope was used (seeParimucha & Vaˇnko 2005, in detail). The starBD+47 4192 ( V = 8.99, B − V = 0.41, U − B = 0.14, V − R C = 0.19; Skopal et al. 2012, and referencestherein) was used as the standard star for both photo-electric and CCD observations.(vi) At the Astronomical Observatory on the KolonicaSaddle, the fast CCD photometry was performedduring the nights on 11/01/2009, 11/02/2009 and08/14/2011. A FLI Pro Line PL1001E
CCD camera withthe chip 1024 × µ m) was attached tothe Ritchey-Chretien telescope 300/2400 mm. The starTYC3641-00678-1 (B=9.43, V=9.05, Rc=8.83, Ic=8.62,Henden & Honeycutt 1997) was used as the comparison.All CCD frames were dark subtracted, flat-fielded, andcorrected for cosmic rays. Our high-time-resolution pho-tometry is plotted in Fig. 4.Arbitrary flux units of the high-resolution spectraaround the H α line were converted to absolute fluxeswith the aid of R C magnitudes corrected for the H α equivalent width (see Eq. (10) of Skopal 2007). Magni-tudes were converted to fluxes according to the calibra-tion of Henden & Kaitchuck (1982) and Bessel (1979).Observations were dereddened with E B − V = 0.30 andresulting parameters were scaled to a distance of 1.5 kpc(M¨urset et al. 1991; Miko lajewska & Kenyon 1996). Or-bital phase was calculated according to the ephemerisof the inferior conjunction of the cool giant (Fekel et al.2000) as JD spec . conj . = 2 450 260 . . × E. (1) ANALYSIS AND RESULTS
Photometric evolution
Figure 1 shows the
U BV
LCs of Z And covering itsquiescent and active phases from 1981. The recent activephase began in September 2000 (Skopal et al. 2000). Ourspectra (Table 1) cover three brightenings in the recentevolution of the LC. The burst in 2008 with U max ∼ . U ∼ . U max ∼ . ∼ B . .
02 mag, comparable with that of standard stars.During the 2008 burst, we observed light variations with∆ B ∼ .
025 mag on the timescale of 1–2 hours, whereasat the maximum of the main 2009 outburst, the lightvariations increased to ∆ B ∼ .
065 mag and enlargedits timescale to 7–9 hours throughout the whole night(Fig. 4). Similar evolution in the rapid photometric vari-ability was also indicated during the previous 2006 ma-jor outburst, when the jet features were measured for thefirst time (see Fig. 3 of Skopal et al. 2009).
Evolution of the H α profile Skopal et al. F l ux [ - e r g c m - s - A - ] Radial velocity [km/s] H β F λ /10 S + Radial velocity [km/s] H α F λ /10 S + Fig. 2.—
Example of the H β and H α profile observed at themaximum of the 2008 burst. Only the S + satellite componentwas indicated at ∼
700 km s − . The gray line represents the modelcontinuum of a normal M5 giant according to Fluks et al. (1994). During the 2008 burst, the H α line showed broad wingsextended to ∼ ± − and a sharp absorption cut-ting the emission core at ∼ -50 km s − . A faint S + satel-lite component was seen at ∼
700 km s − (Fig. 3, spectrafrom January 23 to June 19, 2008). Its presence wasconfirmed by a similar feature in the H β profile (Fig. 2).During the major 2009-10 outburst, the H α profile wasalso of a P-Cyg type with broad wings (Fig. 3, spectrafrom September 26, 2009 to September 24, 2010). Thefirst satellite components were recorded on September 26,2009, prior to the light maximum, at ∼ ± − .However, at the maximum of the star’s brightness (ourspectrum from 12/02/2009) the satellite componentswere hardly recognizable. They appeared again when thebrightness of Z And began to fall (spectrum from January5, 2010). Their position with respect to the referencewavelength increased to ± (1700 − − duringJanuary to September 2010 (Table 2, Fig. 6), when thestar’s brightness gradually decreased by ∼ U , from ∼ ∼ α profile. We observed a simple emissioncore with a reduced wings and the width of the line withrespect to the 2009-10 outburst (Fig. 3). Parameters of satellite components
To determine measured parameters of the satellitecomponents, we isolated them from the whole line profileby fitting the emission line core and its extended wingswith two Gaussian functions. Then the residual satelliteemissions were compared with additional Gaussians. Us-ing their fitted parameters (the central wavelength, maxi-mum, I , and the width σ ) we determined the radial veloc-ity of satellite components, RV S , their flux F S = √ π I σ and the width F W HM S = 2 p σ . Resolution ofour spectra allowed us to estimate uncertainties in RV S to 10 −
25 km s − , in F S fluxes within 10 −
20% of the ob-served values and in
F W HM S widths within 0.2 − Physical parameters of jets
According to reasons discussed by Skopal et al. (2009),the geometry of the emitting medium that gives rise tothe satellite components can be approximated by a nar-row cone with the peak at the central object character- ized with an opening angle θ . This implies that themeasured narrow satellite components can be producedby highly collimated emitting particles – jets. There-fore, using the measured parameters of the satellite com-ponents, we can determine some physical parameters ofjets. Opening angle
Assuming that jets were launched with a constant ve-locity, v jet , perpendicularly to the disk plane, which co-incides with the orbital one (i.e. RV S = v jet cos( i )), theopening angle can be approximated as θ = 2 sin − h HW ZI S RV S tan( i ) i , (2)where i is the orbital inclination (see Skopal et al. 2009).Corresponding parameters from Table 2, orbital incli-nation i = 63 . ◦ . ◦ / − . ◦ HW ZI S = F W HM S yield the average value ofthe jet opening angle as θ = 16 . ◦ . ◦ / − . ◦ θ = 9 . ◦ . ◦ / − . ◦ −
10 outburst , (3)after its maximum. Prior to the maximum of the 2009-10 outburst, θ ∼ . ◦
2. Uncertainties were determinedas standard error of function (2) using its total differen-tial for the uncertainty in i . Influence of uncertainties in F W HM S and RV S (Sect. 3.3) can be neglected. Individ-ual values of θ are given in Table 2. Figure 6 depicts alsodata from the 2006 outburst recalculated for i = 63 . ◦ Emission measure
Assuming that (i) the jet radiation is produced by therecombination transitions in the H α line, (ii) the mediumis completely ionized, (i.e. n e = n p ≡ ¯ n jet ) and radiatesat a constant electron temperature T e , the luminosity ofjets in H α is related to the line emissivity, ε α ¯ n , by L jet (H α ) = ε α ¯ n V jet , (4)where ε α is the volume emission coefficient in H α , ¯ n jet isthe mean particle concentration and V jet is the volumeof the jets. For the optically thin medium of jets, theluminosity can be determined from the observed fluxesas L jet = 4 πd × F S . According to the definition of theemission measure, EM jet = ¯ n V jet , and assumptionsabove, we can express it as, EM jet (H α ) = 4 πd F S ε α . (5)The emission measure for one jet is of a few × cm − for ε α ( T e = 2 × K) = 1.83 × − erg cm s − (e.g.Osterbrock 1989) (Table 2). Radius
The conical shape of jets defines their volume as V jet =1 / R × ∆Ω, where ∆Ω = 2 π [1 − cos( θ / R jet ,can be expressed by means of the parameters L jet andepeated transient jets from Z And 5 Z And α F λ /10+0.5 F λ /10+0.5 F λ /10+0.5 F l ux [ - e r g c m - s - A - ] F λ /10+0.5 ? F λ /10+2 Radial velocity [km s -1 ]01/05/2010 F λ /10+2 Z And α F λ /10+1.5 F λ /10+1.5 F λ /10+1.5 F l ux [ - e r g c m - s - A - ] Radial velocity [km s -1 ]07/13/2010 Z And α F λ /10+0.5 F λ /10+0.5 F λ /10+0.5 F l ux [ - e r g c m - s - A - ] F λ /10+0.5 F λ /10+0.5 Radial velocity [km s -1 ]09/04/201307/22/2014 F λ /10+0.5 Fig. 3.—
Evolution of the H α broad wings in our spectra (Sect. 3.2). The satellite components are denoted by arrows. Their parametersare in Table 2. When the coverage of broad wings was insufficient, we compared low-resolution spectra (09/26/2009 and 05/21/2010). Thepresence or absence of bipolar jets along the current active phase is denoted in Fig. 1. Skopal et al. F l ux Radial velocity [km/s] F λ /10 F l ux F λ /10 S + S - F l ux F λ /10 S + F l ux F λ /10 F l ux H α F λ /10 S + S - -0.05 0 0.05 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 ∆ M a gn it ud e ∆ S + const. ∆ B ∆ M a gn it ud e September 21, 2007June 24, 2007 ∆ B M a gn it ud e B ∆ S + const. M a gn it ud e B ∆ S + const. M a gn it ud e Fraction of JD V ∆ S + const.
Fig. 4.— H α broad wings (left panels) and short-term variability in the optical continuum (right panels) at different stages of Z And activ-ity. During the 2006 and 2009-10 major outbursts, when the satellite S − and S + components emerged, a slow higher-amplitude photometricvariation developed (Sect. 3.1). Data from 2006 and 2007 were published by Skopal et al. (2009). Fluxes are in 10 − erg cm − s − ˚A − . epeated transient jets from Z And 7 TABLE 2Parameters of Gaussian fits to the satellite emission components S − and S + : radial velocity RV S ( km s − ), flux F S (10 − erg cm − s − ) and F W HM S ( km s − ). Derived parameters are: opening angle θ ( ◦ ) and emission measure of bothjets, EM jet (10 cm − ) Date RV S F S F W HM S θ EM jet mm/dd/yyyy S − S + S − S + S − S + S − S + ⋆ -1157 110510/01/2009 † † ⋆ -1670 174005/25/2010 † Notes: ⋆ low-resolution spectrum, † spectrum does not cover the S − component ∆Ω as a function of ¯ n jet , R jet = (cid:18) L jet (H α ) ε α ( H, T e )∆Ω ¯ n − (cid:19) / . (6)For average values of F S and θ of one jet measured dur-ing 2010 (Table 2, Eq. (3)), the average jet radius can beexpressed as a function of ¯ n jet aslog( R jet / AU) = 6 . −
23 log(¯ n jet ) . (7) Mass-loss rate
According to the mass continuity equation, the geo-metrical and kinematics parameters of the jets allow usto determine the mass-loss rate through jets as˙ M jet = ∆Ω R µm H ¯ n jet v jet , (8)where µ is the mean molecular weight and m H is the massof the hydrogen atom. For the average value of the jetvelocity during 2010, v jet = RV S /cos ( i ) = 3900 km s − (Table 2) and the angle of the jet nozzle of 9. ◦
4, we obtainlog( ˙ M jet /M ⊙ yr − ) = − . −
13 log(¯ n jet ) (9)and/or as a function of the radius R jet aslog( ˙ M jet /M ⊙ yr − ) = − . R jet / AU) (10)for one jet. Determination of ˙ M jet thus requires a knowl-edge of ¯ n jet or R jet (see Skopal et al. 2009, for details). Comparison with the 2006 outburst
To date, bipolar jets in the form of satellite compo-nents to the H α (H β ) line were identified in the spec-trum of Z And only during the 2006 and 2009-10 out-bursts (Fig. 1). Our monitoring of the latter showedthat the spectroscopic and photometric characteristics of both outbursts are very similar. We summarize themas follows:(i) The jets appeared around their optical maxima, andwere gradually vanishing along the decrease of the star’sbrightness. Thus the event of jets was transient (Figs. 6and 7).(ii) In both cases, large variation in RV S were observedat the beginning of their emergence, around optical max-ima, and then settled on a constant level (see Fig. 2 ofTomov et al. (2007), Fig. 6 of Skopal et al. (2009) andFig. 6 here).(iii) During both outbursts, rapid light variation on thetimescale of hours within ∆ m ∼ .
06 mag developed fromirregular . RV S . For the 2009-10outburst, the period with jets lasted for ≈
10 months andsatellite components settled on ± (1700 − − ,whereas for the 2006 outburst the relevant quantitieswere ≈ ∼ ± − , respectively(Figs. 6 and 7). DISCUSSION
Here we discuss basic conditions of the jet’s ejectionby the nuclearly powered WD in the symbiotic binaryZ And, and its possible consequences for the accretionprocess during Z And-type outbursts of symbiotic stars.Skopal et al. (2009) interpreted the ejection of jets dur-ing the 2006 outburst as a result of disruption of the innerparts of the disk due to the radiation-induced warpingcaused by a significant increase of the WD luminosityat the outburst maximum. Close similarity of both theoutbursts with jets (Sect. 3.5) suggests their common na-ture.However, there is no theoretical application of theradiation-induced warping of disks around WDs in sym-biotic binaries. Therefore, we first briefly introduce mainresults of its application to selected types of objects. Skopal et al. F l ux [ - e r g c m - s - A - ] Wavelength [A] [ O III] H α H β H γ H δ H e II H e I H e I H e I Z And F l ux [ - e r g c m - s - A - ] + c on s t . Heliocentric radial velocity [km/s] H α phase date Z And
Fig. 5.—
Left: Optical spectrum from the 2009-10 outburst (magenta line) and its model SED (solid black line). It is given bysuperposition of radiation from the warm pseudophotosphere (blue line), the cool giant (orange line) and the nebula (green line). Right:Absorption component in the P-Cyg profile of the H α line during this outburst indicates expansion of the warm pseudophotosphere. On the radiation-induced warping of disks
Petterson (1977) found that the effect caused bythe pressure of radiation from the center of an X-ray source on the disk structure is very important.Iping & Petterson (1990) confirmed that radiation forcescan make rings of the disk precess in either direction,and change their inclination angle. In this way, they ex-plained the inclination of the disk in Her X-1. Usinga simple analytic approach, Pringle (1996) showed thateven an initially flat disk is unstable to warping, becausethe surface of a warped disk is illuminated by a central ra-diation source in a non-uniform manner. Livio & Pringle(1996) showed that also the accretion disk around thecentral stars of PNe can become unstable to a radiation-induced self-warping. Livio & Pringle (1997) demon-strated that this effect is accompanied by a wobblingmotion in both the inclination and azimuthal directions.They used the term ”wobble” to describe rather er-ratic motions. Pringle (1997) investigated the effectof radiation-induced warping on accretion disks aroundmassive black holes in active galactic nuclei. He calcu-lated that the axis of the jets can be severely misalignedfrom the normal to the outer disk. Southwell et al.(1997) derived radiation-induced precession of the diskwith a timescale of the order of months for a strong su-persoft X-ray source CAL 83. Wijers & Pringle (1999)explored the effect of self-induced warping of disks in dif-ferent types of X-ray binaries. They found that at highluminosities the inner disk can tilt through more than90 ◦ , which may explain the torque reversals in systemssuch as 4U 1626-67. For symbiotic stars, a disruption of the disk has beenconsidered for the accretion-powered symbiotic systemCH Cyg by Sokoloski & Kenyon (2003). They inter-preted the change of the fastest variations (timescale ofminutes) into smooth, hour-timescale variations by dis-ruption of the inner disk in association with the mass Her X-1 is an intermediate-mass X-ray binary containing anaccreting neutron star with jets (van den Eijnden et al. 2018).
4U 1626-67 is an ultra-compact X-ray binary bearing a neu-tron star with the orbital period of 42 min (Beri et al. 2018, andreferences therein). ejection event. For the nuclearly powered symbiotic starZ And, Skopal et al. (2009) suggested that the jets ejec-tion during its 2006 outburst could be also triggered bythe radiation-induced disk warping.In the following section, we will show how the basicconditions for the disk warping and jet ejection are sup-ported by observations for the Z And outbursts with jets.
Indication of radiation-induced warping of the diskduring Z And outbursts with jets
According to Sect. 4.1 the necessary ingredients fordisk warping and jets ejection are: (i) The presence of adisk, (ii) emergence of a strong central radiation source,and (iii) an observational response of the wobbling mo-tions of the disk. The corresponding critical results con-firming these terms for the case of Z And are describedas follows.(i) The presence of a disk-like formation around theWD during active phases of symbiotic stars was provenby modeling their UV to near-IR SED (see Skopal2005). The disk-like structure is indicated by the two-temperature-type of the hot component spectrum. Thecooler spectrum is produced by a warm stellar pseu-dophotosphere radiating at 1 − × K, whereas thehotter one is represented by the highly ionized emissionlines and a strong nebular continuum. The former is notcapable of producing the observed nebular emission andthus the latter signals the presence of a strong ionizingsource ( & K) in the system, which is not seen di-rectly by the outer observer. This type of the SED canbe explained by a disk-like structure of the hot compo-nent viewed under a high inclination angle. Then theflared outer rim of the disk (which is the warm pseu-dophotosphere indicated by model SED) occults the cen-tral ionizing source in the line of sight, while the nebulaabove/below the disk can easily be ionized (see Fig. 27of Skopal 2005).During the 2009-10 outburst of Z And, creation ofsuch a disk around the burning WD was documentedby models of SED from September 2009 to November2010 (see Fig. 8 of Tarasova & Skopal 2012). Here,Fig. 5 shows example of a representative model SEDof the optical spectrum from July 5, 2010. The two-epeated transient jets from Z And 9 O p e n i ng a ng l e θ [ ° ] Julian date - 2 400 000 θ (S - ) θ (S + ) 500 1000 1500 2000 R a d i a l v e l o c iti e s o f j e t s [ k m s - ] |RV S - |RV S + B m a gn it ud e Z And
Fig. 6.—
Evolution of RV S (middle) and θ (bottom) along the B -LC (top) of the 2006, 2008 and 2009-10 outbursts. Maxima of2006 and 2009-10 events are denoted by vertical bars. Uncertaintiesin RV S and magnitudes are within the size of points, and those in θ are described in Sect. 3.4.1. Data are from Burmeister & Leedj¨arv(2007), Tomov et al. (2007), Skopal et al. (2009) and our Table 2. temperature-type of the hot component spectrum con-sists of a stellar radiation produced by a warm pseu-dophotosphere (the cooler component) and a strong neb-ular emission (the hotter component). The former radi-ates at ∼ ∼ L ⊙ (the blue line in the figure), while the latter composesof the nebular continuum with a high emission measure EM ∼ . × cm − (the green line) and emissionlines of highly ionized elements (e.g. H I , He I , He II ,[O III ]). The warm pseudophotosphere generates the fluxof hydrogen ionizing photons L H ∼ × s − , which isnot capable of producing the observed EM that requires L H = α B (H , T e ) × EM ∼ × s − for the total hy-drogen recombination coefficient α B (H , T e ) = 1 . × − cm s − (e.g., Nussbaumer & Vogel 1987). This impliesthat the warm pseudophotosphere cannot be a sphere,but a disk, as described above. Independently, the pres-ence of jets indicates the presence of a disk, because jetsrequire an accretion disk (e.g. Livio 1997).(ii) The emergence of a strong central radiation source is connected with outbursts, when a significant increaseof the central source luminosity is indicated. For ex-ample, the emission measure of 1 . × cm − requiresthe luminosity of the ionizing source to be of the or-der of 10 erg s − (see Skopal et al. 2017, in detail).Also a high EM ∼ . × cm − , as indicated aroundthe maximum of the 2006 outburst (see Sect. 3.1. ofSkopal et al. 2009), suggests similarly high luminosity.For the 2000-03 outburst, Sokoloski et al. (2006) esti-mated the hot star luminosity to 10 L ⊙ , lasting approx-imately for one year. Thus the high luminosity of thecentral source during outbursts can illuminate the disk,which can become unstable to warping.(iii) Observational response of the radiation-inducedwarping of the disk can be associated with the higher-amplitude photometric variability (∆ m ∼ .
06 mag) onthe timescale of hours that emerges exclusively dur-ing major outbursts, during which the bipolar jets arelaunched (Fig. 4). According to models SED, this type ofvariability is produced by the warm pseudophotosphere,i.e., the outer rim of the disk, because its contributiondominates the
B, V passbands (see Fig. 5 and point (i)above). In this way, we directly indicate the responseof the disk warping in the form of the short-term varia-tions in the continuum. As the dynamical timescale ofthe disk is comparable with the timescale of the smoothlight variations (see Appendix A), this can be caused bya variable projection of the disk surface into the line ofsight. We thus observe wobbling of the outer parts ofthe disk – in agreement with the general view that thewarped disk starts to wobble or precess (Livio & Pringle1997, Sect. 4.1). Alternatively, the smooth light varia-tions could be partly caused by the brightness variationof the outer disk due to reprocessing flickering light fromits inner warping part (see Appendix B).
Jets’ radial velocities from the warping disk
Complex variations in RV S around the maxima of both2006 and 2009-10 outbursts (Sect. 3.5) could also be a di-rect result of the radiation-induced disk warping. Assum-ing jets to be perpendicular to the inner disk, their initialchaotic wobbling motions can produce shifts in RV S dueto the different inclination of a jet to the line of sight, asconsidered for PNe by Livio & Pringle (1997). One canalso imagine that the twisted inner disk can cause thataxis of both jet nozzles not to be parallel, which causesthe observed jet’s asymmetry (see Fig. 6 of Skopal et al.2009).However, after a short time of ∼ RV S settled at around ± − and ± (1700 – 1800) km s − for 2006 and 2009-10 outburst, re-spectively. During the small 2008 burst, the only presentS + component was located at ∼
700 km s − (Table 2).This suggests that the wobbling motions of the diskceased.Diversity of the final RV S during each outburst couldbe caused by different conditions for the radiation-drivenwarping during these events. According to Pringle(1996), the radiation-driven warping of the disk can oc-cur at radii, which depend on the central star massand luminosity, accretion rate, and viscosities in thedisk. During the Z And-type outbursts, additional en-ergy is liberated by thermonuclear burning on the WD0 Skopal et al.surface due to a transient high accretion rate (see, e.g.,Skopal et al. 2017, for AG Peg). Depending on its quan-tity during different outbursts with different properties ofthe disk, the warping can occur at different radii. Quali-tatively, the material liberated at different heights abovethe WD then converts different amounts of the gravita-tional potential energy into the kinetic energy at its ac-cretion and thus can be expelled at different velocities inthe form of jets. The energy power of these outbursts, asgiven by their brightness maximum and duration, seemsto be consistent with this view (see Fig. 1). Disk-jets connection in Z And
Figure 7 shows the evolution of the measured flux ofjets and the effective radius of the warm pseudopho-tosphere, R effWD , along the 2006 and 2009-10 outbursts.Note that the disk radius R D ∝ R effWD (e.g., Eq. (11)of Skopal et al. 2011). Effective radii were derived frommodeling the SED during the 2009-10 outburst (see Ta-ble 5 of Tarasova & Skopal 2012). From the 2006 out-burst, there is only one estimate of R effWD available aroundthe optical maximum (Skopal et al. 2009). Gradual de-crease of the disk radius and the flux of jets along thedecline of the star’s brightness implies a gradual dilutionof the disk.Shrinking of the disk radius is also indicated by the ra-dial velocities of the absorption component in the H α profile that decelerate along the decline of the star’sbrightness (bottom panel of Fig. 7). This effect can be aresult of expansion of the outer parts of the disk leadingto a shrinkage of the optically thick part of the warmpseudophotosphere. Therefore, the optically thick windin the H α line at/above the warm pseudophotosphere willalso originate at smaller radii, where it is driven outsidewith a smaller velocity.Below we describe how the disk–jets connection can beunderstood in the context of the accretion process ontothe WD during Z And outbursts with jets. On the accretion during Z And outbursts with jets
Significant increase of the WD luminosity around theoptical maximum, the indication of the disk warping(Sects. 4.2 and 4.3), and the presence of jets is a resultof a significant increase of the accretion rate through thedisk, so that an excess of angular momentum of the ac-creting material has to be removed via jets. A fraction ofthe disk material, accreted onto the WD, enhances nu-clear shell burning on its surface. As a consequence, thehigh radiation output makes the surrounding disk un-stable to warping and enhances the stellar wind duringoutbursts (see Sect. 4.5 for some details). The rest of theaccreting material is liberated in the form of jets, whosekinetic energy comes from their accretion energy. Thisleads to a gradual removal of the disk material from itsinner parts onto the WD and into the jets. As a result,the gradual decrease of fueling the WD and the release ofthe accretion energy lead to a gradual decline of the star’sbrightness, shrinking of the disk radius and vanishing ofthe jets as documented by observations in Fig. 7.In a nutshell, expanding of the outer parts of the diskand sucking its inner parts by the luminous WD via theradiation-induced warping lead to a gradual dilution ofthe disk. This gives reasons for the observed disk-jets E ff ec ti v e r a d i u s [ s o l a r un it s ] R WDeff (2006)R
WDeff (2009-10) 0 2 4 6 F j e t s [ - e r g s - c m - ] F jets (2006)F jets (2009-10) 9 10 11 12 -100 0 100 200 300 B m a gn it ud e R a d i a l v e l o c it y [ k m / s ] Julian date - 2 400 000 RV obs RV obs - RV orb Fig. 7.—
Disk-jets connection during the 2006 and 2009-10 out-bursts. The top panel compares their B -LCs shifted to their max-ima (vertical bar). The second panel shows evolution of averagefluxes of jets, F jets = ( F S + + F S − ) / R effWD , of the warm pseudopho-tosphere, and the bottom panel displays radial velocities of theabsorption component in the H α profile (crosses are values cor-rected for the orbital motion of the hot component according toelements of Fekel et al. (2000)). epeated transient jets from Z And 11connection and explains why the launching of jets is tran-sient. On jets and Z And-type outbursts
During quiescent phases of symbiotic stars, no jets havebeen indicated, although their WDs accrete through-out the accretion disk and, in most cases, at highrates of 10 − − − M ⊙ yr − to power their high lu-minosities by stable hydrogen burning on the WD sur-face (e.g., Paczy´nski & ˙Zytkow 1978; Hachisu et al. 1996;Shen & Bildsten 2007). To date, jets were indicatedonly during outbursts (see Sect. 1), during which a largeoptically thick disk is created around the WD (Skopal2005). The presence of jets thus confirms the presenceof the disk and constrains accretion at a high rate, al-though enhanced mass-outflow from the hot componentis directly indicated (see below).For the nuclearly powered symbiotics, the appearanceof transient jets signals transient accretion at rates abovethe limit of the stable hydrogen burning on the WD sur-face. According to theoretical predictions, at these ratesthe luminosity of the burning WD can increase to theEddington limit (e.g., Fig. 2 of Shen & Bildsten 2007),and the mass-outflow from the WD in the form of windenhances (Hachisu et al. 1996). Both characteristics arewell supported by observations. (i) The high luminos-ity of the WD, burning hydrogen above the stable limit,was recently evidenced for AG Peg during its 2015 Z And-type outburst (Skopal et al. 2017). For Z And, the WD’sluminosity of the order of ∼ erg s − was determinedfor the 2000-03, 2006 and 2009-10 outbursts (see thepoint (ii) of Sect. 4.2). (ii) Enhanced mass-outflow viathe wind is directly indicated by broadening of emis-sion lines during outbursts (e.g. Fern´andez-Castro et al.1995). Modeling the broad H α wings, Skopal (2006)found an increase of the mass-loss rate via the wind fromhot components during outbursts to & − M ⊙ yr − ,i.e., factor of &
10 higher with respect to values fromquiescent phases.Appearance of jets during Z And-type outbursts thusmanifests their nature by nuclear burning of hydrogenon the WD surface at rates above the upper limit of thestable burning. The highly energetic events we observeduring these outbursts (the luminosity close to the Ed-dington value and the wind outflow at & − M ⊙ yr − )require a relevant source of material that effectively fu-els the WD. Indication of a large neutral disk at/aroundthe orbital plane during Z And-type outbursts was de-scribed in the point (i) of Sect. 4.2. According toCarikov´a & Skopal (2012) such a disk can be formed dur-ing outbursts by compression of the enhanced wind to-ward the equatorial plane due to rotation of the WD.This means that the disk consists of the material pre-viously accreted onto the WD from the giant. Due tothe thermal warping, this material is reaccreted again,which prolongs the period with a high luminosity. Thecontinuous flow of material from the giant’s wind helpsto refill in the disk until the next possible instability willcause another (out)burst. This process can repeat up todepletion of the disk formed during outbursts. Then anaccretion disk will be created from the giant’s wind, andthe system enter a quiescent phase.According to Leibowitz & Formiggini (2008), the main active phases of Z And last approximately of 15–20 yearsand appear quasi-periodically with a separation of theircenters by ≈
20 years (see their Fig. 1). Assuming thisevolution also for the current active phase, it should ceasearound 2020. SUMMARY
We continued monitoring of the prototypical symbioticstar Z And after its major 2006 outburst, during whichtwo-sided jets in the optical spectrum were indicated forthe first time. We used the high-resolution spectroscopyaround H α , multicolor U BV R C photometry, and high-time-resolution photometry to search for the reappear-ance of jets during the following outbursts of the currentactive phase (Fig. 1). Our findings may be summarizedas follows.1. The bipolar jets reappeared during the major 2009-10 outburst, as indicated by well-pronounced S − and S + satellite components to the H α line. Duringthe smaller 2008 burst, only a single S + componentwas present (Sect. 3.2, Figs. 2 and 3).2. The evolution of jets during both the 2006 and the2009-10 outbursts was similar. The jets appearedaround the outburst maxima, weakened along thedecline of the star’s brightness, and ceased after ≈ ≈
10 months during 2006 and 2009-10 out-bursts. A large variation in jets radial velocitieswas measured at their emergence, but after 1–3 months they settled on a constant level of ∼± ∼ ± − , respectively (Sect. 3.5,Fig. 6).3. During both outbursts, a smooth light variationwithin ∆ B ∼ .
06 mag on the timescale of hoursdeveloped from rapid, . L ⊙ and drives the windat & − M ⊙ yr − .6. The large disk created at the beginning of theoutbursts ( R D ∼ − R ⊙ , Fig. 7) consists ofthe material originally accreted onto the WD fromthe giant (Carikov´a & Skopal 2012). It representsa reservoir of mass for fueling the burning WD,because its inner part can be reaccreted via theradiation-induced warping. In this way, a high lu-minosity of the burning WD can be sustained for a2 Skopal et al.longer time, until depletion of the disk (Sect. 4.5).The jets thus signal the presence of the reaccre-tion process, providing a link to long-lasting activephases of Z And.Observations of transient jets along the decline fromthe optical maximum of some Z And-type outbursts(to-date observed for Hen 3-1341 (Munari et al. 2005),Z And (e.g. Burmeister & Leedj¨arv 2007) and BF Cyg(Skopal et al. 2013)) represent new challenges for thetheoretical modeling of the radiation-induced warping ofdisks formed during outbursts of symbiotic stars. Thisshould provide us with a better understanding of the ac-cretion process during nuclearly powered eruptions onthe surface of WDs with prolonged stages of high lumi-nosity and signatures of collimated mass ejection.The authors thank the anonymous referee for criti-cal, but encouraging, comments on the original version of the manuscript. Theodor Pribulla is thanked for ac-quisition of spectra at the David Dunlap Observatory.Miloslav Tlamicha, Tereza Krejˇcov´a, Pavel Chadima,Lenka Kotkov´a, Luk´aˇc Pilarˇc´ık, Jan Sloup, KateˇrinaHoˇnkov´a, Jakub Juryˇsek, Ludˇek ˇRezba and Jan Fuchs,are thanked for their assistance in the acquisition of thespectra at the Ondˇrejov observatory. We also acknowl-edge the variable-star observations from the AAVSO In-ternational Database contributed by observers worldwideand used in this research. This work was supportedby the Czech Science Foundation, grants P209/10/0715and GA15-02112S, by the Slovak Research and Develop-ment Agency under the contract No. APVV-15-0458,by the Slovak Academy of Sciences grant VEGA No.2/0008/17 and by the realization of the project ITMSNo. 26220120029, based on the supporting operationalResearch and development program financed from theEuropean Regional Development Fund. APPENDIX
DYNAMICAL TIMESCALE OF THE DISK
Here we compare the timescale of the observed light variability (Sect. 3.1) with the dynamical time of the disk withradius R D around the WD with the mass of M WD , t dyn = 0 . × (cid:18) R D R ⊙ (cid:19) / (cid:18) M WD M ⊙ (cid:19) − / hr . (A1)A representative value of R D = 10 R ⊙ (see Fig. 7) and M WD = 1 M ⊙ corresponds to t dyn ∼
14 hrs. If we consider R D ∼ R ⊙ for the outer part of the disk, then t dyn ∼ . R D . . R ⊙ . DIFFUSION TIMESCALE OF THE DISK
A contribution to the smooth light variation could also come from the inner warping part of the disk, whoseflickering light is hidden, but can be reprocessed in the outer disk. The timescale for this process is given by thediffusion timescale, t dif = R /c l, (B1)where R out is the radius of the reprocessing medium, c is the light speed, and l is the mean-free path of a photon.For the opacity κ λ and density ρ of the medium, l = 1 /κ λ ρ . If we can assume that ρ and κ λ are constant along thepath of the photon, then the corresponding optical depth τ λ = κ λ ρR out = R out /l , and the diffusion timescale can beexpressed as R out τ λ /c , or t dif = 6 . × − R out R ⊙ τ λ hr . (B2)If the thickness R out ∼ R ⊙ and its total optical depth is as large as 1–10 , the diffusion timescale t dif = 12 s–1.3 daysis sufficiently short to observe the inner disk light variation at its outer rim (which is the warm pseudophotosphere),well within the timescale of the jet ejection period ( ≈ REFERENCESBeri, A., Paul, B., & Dewangan, G. C. 2018, MNRAS, 475, 999Bessel, M. S. 1979, PASP, 91, 589Boyarchuk, A. A. 1967, Soviet Astronomy, 11, 8Burmeister, M., & Leedj¨arv, L. 2007, A&A, 461, L5Brandi, E., Miko lajewska, J., Quiroga, C. et al. 2005, A&A, 440,239Carikov´a, Z., & Skopal, A. 2012, A&A, 548, A21Esipov, V. F., Kolotilov, E. A., Miko lajewska, J., et al. 2000,Astronomy Letters, 26, 162Fekel, C. F., Hinkle, K. H., Joyce, R., & Skrutskie, M. F. 2000,AJ, 120, 3255Fern´andez-Castro, T., Gonz´alez-Riestra, R., Cassatella, A.,Taylor, A. R., & Seaquist E. R. 1995, ApJ, 442, 366 Fluks, M. A., Plez, B., The, P. S., de Winter, D., Westerlund, B.E., & Steenman, H. C. 1994, A&AS, 105, 311Hachisu, I., Kato, M., & Nomoto, K. 1996, ApJ, 470, L97Henden, A. A., & Kaitchuck, R. H. 1982,
AstronomicalPhotometry , (New York: Van Nostrand Reinhold Company), 50Henden, A. A., & Honeycutt, R. K. 1997, PASP, 109, 441Horn, J., Kub´at, J., Harmanec, P., et al. 1996, A&A, 309, 521Iping, R. C., & Petterson, J. A. 1990, A&A, 239, 221Kenyon, S. J. 1986,
The symbiotic stars , CUP, New YorkLeibowitz, E. M., & Formiggini, L. 2008, MNRAS, 385, 445Leibowitz, E. M., & Formiggini, L. 2013, ApJ, 146:117Livio, M. 1997, in ASP Conf. Ser. 121, Accretion Phenomena andRelated Outflows, ed. D. T. Wickramashinge , G. V. Bicknell,& L. Ferrario, (San Francisco: ASP), 845 epeated transient jets from Z And 13
Livio, M., Pringle, J. E. 1996, ApJ, 465, L55Livio, M., Pringle, J. E. 1997, ApJ, 486, 835McKeever, J., Lutz, J., Wallerstein, G., Munari, U., & Siviero, A.2011, PASP, 123, 1062Miko lajewska J., & Kenyon, S. J. 1996, AJ, 112, 1659Munari, U., Tomov, T., Yudin, B. F., et al. 2001, A&A, 369, L1Munari, U., Siviero, A., & Henden, A. 2005, MNRAS, 360, 1257M¨urset, U., Nussbaumer, H., Schmid, H. M., & Vogel, M. 1991,A&A, 248, 458M¨urset, U., & Schmid, H. M. 1999, A&AS, 137, 473Nussbaumer, H., & Vogel, M. 1987, A&A, 182, 51Nussbaumer, H., & Vogel, M. 1989, A&A, 213, 137Osterbrock, D. E. 1989,