Insights into the evolution of symbiotic recurrent novae from radio synchrotron emission: V745 Scorpii and RS Ophiuchi
N. G. Kantharia, Prasun Dutta, Nirupam Roy, G. C. Anupama, C. H. Ishwara-Chandra, A. Chitale, T. P. Prabhu, D. P. K.Banerjee, N. M. Ashok
aa r X i v : . [ a s t r o - ph . H E ] N ov Mon. Not. R. Astron. Soc. , 1– ?? (2014) Printed 15 September 2018 (MN L A TEX style file v2.2)
Insights into the evolution of symbiotic recurrent novae from radiosynchrotron emission: V745 Scorpii and RS Ophiuchi
N. G. Kantharia ⋆ , Prasun Dutta , Nirupam Roy , G. C. Anupama ,C. H. Ishwara-Chandra , A. Chitale, T. P. Prabhu † , D. P. K.Banerjee , N. M. Ashok National Centre for Radio Astrophysics, TIFR, Pune, India Indian Institute of Science Education and Research, Bhopal, India Indian Institute of Technology, Kharagpur, India Indian Institute of Astrophysics, Bangalore, India Physical Research Laboratory, Ahmedabad, India
Accepted 2015 October 7. Received 2015 August 19; in original form 2015 May 23
ABSTRACT
We present observations at 610 MHz and 235 MHz using the Giant Metrewave RadioTelescope (GMRT) of the recurrent nova V745 Scorpii which recorded its last outburst on6 February 2014. This is the second symbiotic recurrent nova whose light curve at low fre-quencies has been followed in detail, the first being RS Ophiuchi in 2006. We fitted the 610MHz light curve by a model of synchrotron emission from an expanding shell being modifiedby radiative transfer e ff ects due to local absorbing gas consisting of a uniformly distributedand a clumpy component. Using our model parameters, we find that the emission at 235 MHzpeaked around day 35 which is consistent with our GMRT observations. The two main resultsof our study are: (1) The radio emission at a given frequency is visible sooner after the out-burst in successive outbursts of both V745 Scorpii and RS Ophiuchi. The earlier detection ofradio emission is interpreted to be caused by decreasing foreground densities. (2) The clumpymaterial, if exists, is close to the white dwarf and can be interpreted as being due to the ma-terial from the hot accretion disk. The uniform density gas is widespread and attributed to thewinds blown by the white dwarf. We present implications of these results on the evolution ofboth novae. Such studies alongwith theoretical understanding have the potential of resolvingseveral outstanding issues such as why all recurrent novae are not detectable in synchrotronradio and whether recurrent novae are progenitor systems of type 1a supernova. Key words: binaries: symbiotic – novae – radio continuum:stars
Recurrent novae are binary systems with the primary being a whitedwarf and the secondary star being a red giant or a main sequencestar. A symbiotic recurrent nova is a system comprising a red gi-ant secondary. The white dwarf accretes matter from the secondarywhich leads to a runaway thermonuclear explosion on its surface af-ter a critical mass is reached. This results in a spectacular increasein optical light by a few to 15 magnitudes.The last outburst of the recurrent nova V745 Sco, predicted bySchaefer (2010) to be in year 2013 ±
1, was recorded on 6 / Feb / ⋆ E-mail: [email protected] † E-mail: [email protected] bands ranging from the γ rays (Cheung, Jean, & Shore 2014). tolow frequency radio waves (Kantharia et al. 2014).In this paper, we describe the low radio frequency observa-tions of V745 Sco with GMRT and combine it with VLA archivaldata from the 1989 outburst to understand the evolution of the sys-tem. We also study the results on the symbiotic recurrent nova RSOphiuchi from its outbursts in 1985 and 2006. The GMRT (Swarup et al. 1991) observations of V745 Sco wereconducted either in the 610 MHz band or in the dual band which al-lows simultaneous observations in both the 235 MHz and 610 MHzbands between 7 February and 7 September 2014. Bandwidth set-tings of 32 MHz at 610 MHz and 16 MHz at 235 MHz were used.A combination of Director’s Discretionary Time and regular TAC- c (cid:13) N. G. Kantharia et. al. approved time in Observing Cycle 26 as part of the project
GalacticNovae with GMRT (GNovaG) were used for GMRT observations.The observations started with a 10 minute run on an amplitude cal-ibrator (3C286 or 3C48) followed by 3 minute and 30 minute runson the phase calibrator (J1830–360) and V745 Sco. Most of the ob-serving slots were ∼
2h in duration with ∼ .
5h on the nova and ∼ µ Jy rms noise on the final image at 610 MHz.The data were converted from the native lta format into
FIT S format and imported into
AIPS These data were calibrated andimaged using 25 facets at 610 MHz and 49 facets at 235 MHz.The synthesized beamshapes were elliptical with typical sizes be-ing 10 ′′ × ′′ at 610 MHz and 20 ′′ × ′′ at 235 MHz with a positionangle ∼ ◦ .The results of GMRT observations of V745 Sco are listed inTable 1 and shown in Fig. 1. The first detection was on day 12(18 / / ff setin the map, Kantharia et al. (2014) missed reporting the detectionon day 12. Observations on 6 March 2014 detected a strong radiosource at both 610 MHz and 235 MHz (Table 1). No correlationbetween the varying strength of the target source and the calibratorsis observed (Table 1).We analysed the VLA archival data of V745 Sco from 1989and the results are shown in Table 2 and Fig. 1. The data on RSOphiuchi are obtained from literature – the 1.4 GHz data from the1985 outburst is from Hjellming et al. (1986) and the model fit tothe 2006 outburst is from Kantharia et al. (2007). V745 Sco is the second symbiotic recurrent nova which has beenstudied at GMRT frequencies where the emission is predominantlysynchrotron in nature. The peak radio power estimated for a dis-tance of 7 . ± . . × W Hz − and 4 × W Hz − . For compar-ison, the peak power from RS Ophiuchi at 610 MHz following itsoutburst in 2006 (Kantharia et al. 2007) was 1 . × W Hz − .Assuming equipartition of energy between the relativistic par-ticles and magnetic field, expected under minimum energy condi-tion, we estimate a magnetic field of 0.03 G and energies of ∼ Joule for the 2014 outburst of V745 Sco. An emitting shell of ra-dial extent 1 AU was assumed at a distance of 30 AU from the whitedwarf as inferred from the near-infrared (Banerjee et al. 2014) andX-ray observations (Orio et al. 2015). A magnetic field of strength0.04 G with energy in relativisitic particles being 2 . × Jouleie about 0.02% of total energy was estimated for the 1985 outburstin RS Ophiuchi (Bode & Kahn 1985). The magnetic field strengthsand particle energies match to within an order of magnitude in thetwo recurrent novae and for two consecutive outbursts.We use the parameteric model that has been presented in Eqn.1 in Weiler et al. (2002) for explaining supernova light curves. Theassumptions in the model are implicitly included and we do notexplore any possible di ff erences due to the observed bipolar na-ture of synchrotron emission from RS Ophiuchi in its 1985 outburst(Taylor et al. 1989) and its 2006 outburst (O’Brien et al. 2006). Inthe Weiler et al. (2002) model, there is a frequency-dependent de-lay in the detection of synchrotron emission due to the opacity of AIPS is produced and maintained by the National Radio Astronomy Ob-servatory, a facility of the National Science Foundation operated under co-operative agreement by Associated Universities, Inc. t−t days F l u x d e n s i t y [ m J y ] Figure 1.
Light curve of V745 Scorpii: 610 MHz and 235 MHz from the2014 outburst using GMRT and 1.4 GHz from the previous outburst in 1989using VLA archival data. t is 6 February 2014. the foreground thermal gas. The radio emission rises as opacitydecreases with peak emission at unity opacity and then declinesas the emitting shell expands. The model includes opacity due toseveral di ff erent components. We only included the local opacitiesdue to the uniform and clumpy parts of the circumbinary materialwhich well explained the light curves from the 2006 outburst in RSOphiuchi (Kantharia et al. 2007). This model was fitted to the lightcurves at 610 MHz from the 2014 outburst (Fig. 2) and at 1.4 GHzfrom the 1989 outburst in V745 Sco and to the 1.4 GHz data fromthe 1985 outburst in RS Ophiuchi (Fig. 2). The model outputs arelisted in Table 3. Using the model parameters listed in Table 3, we have plotted thevariation in the optical depth with time (see Eqn 1 and Fig 3 top).The temporal variation in the optical depth due to uniformly dis-tributed gas, τ uniform and the clumpy gas, τ clumpy is (Weiler et al.2002): τ uniform [ τ clumpy ] = K [K ] (cid:18) ν (cid:19) − . t − t ! δ uniform [ δ clumpy ] (1)K , K , δ are determined from the fits to the light curve (see Table3). The variation in both optical depths at three frequencies 235MHz, 610 MHz and 1.4 GHz is shown in Fig. 3. As seen in thefigure, τ clumpy drops to one within five days following the outburstat all bands. τ uniform shows a slower decline – is one around day10 at 1.4 GHz, day 18 at 610 MHz and day 35 at 235 MHz whichwould roughly correspond to the peak emission at those bands. Ourobservations at 235 MHz are consistent with this model. Thus theturnon of the synchrotron emission at the low radio frequencies isdetermined primarily by the optical depth of the uniform densitygas in the 2014 outburst of V745 Sco. Due to the absence of datapoints leading to the peak of the 1989 outburst (see Table 3), allthe model parameters are not well constrained and do not allow astudy of the opacity variation. The temporal flux variation is similarin both epochs. The model fit predicts the peak at 1.4 GHz in 1989to have been around day 18 as compared to day 10 in 2014. Thissuggests evolution in the environment of the nova. We also did a similar study of opacity variation in RS Ophiuchi us-ing our model fit to the 1.4 GHz data from the 1985 outburst taken c (cid:13) , 1– ?? ecurrent nova evolution from radio synchrotron emission Table 1.
GMRT results on V745 Sco at 610 and 235 MHz. t is 6 February 2014. The flux density of the background source (bcksrc) is estimated from theimage whereas the flux density of the phase calibrator J1830-360 is obtained from GETJY.t–t Date Flux density at 610 MHz Flux density at 235 MHz α V745Sco σ bcksrc J1830-360 V745Sco σ bcksrc J1830-360days 2014 mJy / beam mJy / beam mJy Jy mJy / beam mJy / beam mJy Jy3 9 Feb < .
55 - 16.26(0.16) 18.1(0.17) - - - -12 18 Feb 1.07 0.14 16.41(0.14) 16.87(0.05) - - - -27 05 Mar 6.9 0.13 17.04(0.13) 18.07(0.11) - - - -28 06 Mar 6.8 0.17 15.3(0.11) 16.63(0.11) 4.3 0.9 30.9(0.9) 29.44(0.3) 0.4935 13 Mar 5.89 0.18 14.76(0.1) 16.27(0.16) 4.7 1.1 31.2(1.2) 30.1(0.4) 0.2440 18 Mar 5.11 0.13 15.91(0.13) 17.39(0.22) < . > . − . < > − .
671 18 Apr 2.52 0.19 18.08(0.19) 19.2(0.4) - - -86 03 May 2.06 0.07 17.59(0.07) 18.54(0.08) - - - -101 18 May 1.17 0.21 13.42(0.22) 16.44(0.94) < . > − . < . t−t days F l u x d e n s i t y [ m J y ] V745 Sco 610MHz t−t days F l u x d e n s i t y [ m J y ] RS Oph 1420MHz
Figure 2. (left) The best model fit (solid line) to the light curve of V745 Sco at 610 MHz from its outburst in 2014. (right) The best model fit (solid line) to thelight curve of RS Ophiuchi at 1.4 GHz from its outburst in 1985. Data points are taken from Hjellming et al. (1986).
Table 2.
Analysis of VLA archival data (AH 383, AH 389) at 1.4 GHz onV745 Sco following the outburst in 1989. t is 30 July 1989.t-t date S Referencedays mJy33 01 Sep 1989 5.7 Hjellming (1989)49 17 Sep 1989 3.3(0.48) VLA archives56 24 Sep 1989 3.17(0.22) VLA archives76 07 Oct 1989 1.96(0.36) VLA archives88 19 Oct 1989 1.28(0.18) VLA archives194 02 Feb 1990 < .
75 VLA archives (3 σ ) from Hjellming et al. (1986) and the Kantharia et al. (2007) fit pa-rameters for the 2006 outburst (Fig. 3). Interestingly, the variationin opacities due to uniform density gas and clumpy gas is compa-rable following the outburst in 2006 whereas the 1985 fit shows arapid fall in τ clumpy as noted for V745 Sco in 2014. The di ff erencesin the 1985 and 2006 outbursts of RS Ophiuchi are (1) The turn-onin 2006 at a given frequency occurs at an earlier date following theoutburst as compared to the 1985 outburst. (2) The turn-on day in1985 is primarily determined by the uniform density gas whereas in 2006, it is determined by both the uniform density and clumpygas. (3) The τ uniform in 1985 falls to unity around day 32 at 1.4 GHzand day 85 at 235 MHz whereas in 2006, it is unity around day 2 at1.4 GHz and around day 8 at 235 MHz. We infer the net e ff ect to bereduction in the ambient ionized gas densities in 2006 as comparedto 1985. Kantharia et al. (2007) had arrived at a similar result usingthe 325 MHz data from 2006 when it was detected on day 38 andusing the result from 1985 that no emission was detected upto day66 (Spoelstra et al. 1987) and inferred that the absorbing densitiesin 2006 were about 30% of those in 1985. The model fitted to the1985 1.4 GHz data on RS Ophiuchi is consistent with this conclu-sion and suggests that the emission at 325 MHz would have beenvisible ∼ day 66 in 1985. Our main results from the synchrotron light curve fitting for boththe novae are: (1) earlier turn-on and peaks observed at a given fre-quency following an outburst for successive outbursts, (2) clumpymaterial distributed closer to the system and uniformly distributedgas being more widespread. In Table 3, we have listed the emis-sion measure (EM) of the foreground thermal gas when it would be c (cid:13) , 1– ?? N. G. Kantharia et. al.
Table 3.
Model outputs. The spectrum is assumed to have a spectral index of α = − .
5. Emission measure of absorbing gas when the optical depth is one. Theelectron temperature is assumed as 10 K. N is the number of detections, t peak , S peak , t refer to the day on which peak emission occurred, peak emissionstrength and the day on which the emission was 1% of peak emission. β is the temporal decay index, K , δ uniform , K , δ clumpy are the constants from Eq. 1.Nova Outburst ν N t peak S peak t β K δ uniform K δ clumpy χ EM peak epoch days mJy days cm − pcV745 Sco 2014 610 MHz 14 23 7.1 9.5 − . − . − . − . − . − . × RS Ophiuchi 2006
610 MHz 14 29 52.9 6 − .
24 0.14 − .
29 0.53 − .
14 1.5 10 − . − . − . × Parameters from the multi-frequency light curve fits in Kantharia et al. (2007). The 1985 light curve data at 1.4 GHz is taken from Hjellming et al. (1986) t−t days -6 -5 -4 -3 -2 -1 o p t i c a l d e p t h τ uniform 235MHzclumpy 235MHzuniform 610MHzclumpy 610MHzuniform 1420MHzclumpy 1420MHz t−t days -4 -3 -2 -1 o p t i c a l d e p t h τ uniform 235MHzclumpy 235MHzuniform 610MHzclumpy 610MHzuniform 1420MHzclumpy 1420MHz t−t days -4 -3 -2 -1 o p t i c a l d e p t h τ uniform 235MHzclumpy 235MHzuniform 610MHzclumpy 610MHzuniform 1420MHzclumpy 1420MHz Figure 3.
The evolution of the optical depth due to the uniform density andclumpy medium in the nova system for (top) 2014 outburst of V745 Sco,(centre) 2006 outburst of RS Oph using parameters given in Kantharia et al.(2007), (bottom) 1985 outburst of RS Ophiuchi. Clumpy medium showssimilar behaviour for the outbursts in 1985 and 2006 whereas the uniformdensity medium shows evolution. The large optical depths are not physicaland are shown only to indicate total absorption. The region of interest isclose to where opacity is one as indicated by the black horizontal line. transparent to the emission at 610 MHz and 1.4 GHz. The earlierpeaks with successive outbursts would indicate reducing EM of theabsorbing gas.Theoretical studies indicate that when the accretion rate ex-ceeds some critical limit (few times 10 − M ⊙ yr − ; Hachisu & Kato2001), then the envelope on the white dwarf can expand to the sizeof a red giant (Nomoto 1982). This can then lead to the formationof a common envelope (e.g. Nomoto, Nariai, & Sugimoto 1979)which can trigger a spiral-in of the binary and a double degeneratesystem (Iben & Tutukov 1984). Hachisu, Kato, & Nomoto (1996)found that another outcome of the larger accretion rate would befast optically thick winds ( ∼ − , > − M ⊙ yr − ) blown bythe white dwarf which they refer to as accretion winds. The accre-tion winds would stabilise the system and the binary can continueto evolve as a single degenerate system (Hachisu, Kato, & Nomoto1996). The accretion rates for V745 Sco and RS Ophiuchi were es-timated to be 2 × − M ⊙ yr − (using a recurrence timescale of 25years) and 1 . × − M ⊙ yr − (Hachisu & Kato 2001). Combiningour results with the theoretical arguments, we infer the following:1. The onset of synchrotron radio emission is delayed by the opti-cally thick winds blown by the white dwarf which constitutes theuniform density component. The clumpy component rapidly getstransparent. We suggest this to be due to the material from the ac-cretion disk close to the white dwarf which is blown o ff in eachoutburst (Hachisu & Kato 2001).2. The earlier turnon of the radio emission with successive out-bursts would then indicate the reduction in emission measure ofthe accretion winds. This could imply a reduced accretion rate( < few times 10 − M ⊙ yr − ) on the white dwarf. The estimatedaccretion rate on both the novae is about 10 − M ⊙ yr − . If indeedthe accretion rate has dropped causing the winds to stop – eitherthe two stars can spiral-in leading to a double degenerate systemor if the white dwarf is massive enough, it can explode as a type1a supernova. If the latter is not the case, then there are reasons tobelieve that the single degenerate system might not evolve into atype 1a supernova.3. Alternately since the mass of the white dwarfs in both systems isbelieved to be close to the Chandrasekhar limit, the critical accre-tion rate limit is larger at 10 − M ⊙ yr − (Nomoto 1982). Since theestimated accretion rates are lower than this larger critical limit, itwould cause the winds to stop as the white dwarf grows in mass,leading to a transparent ambience at radio wavelengths. From ourresults on V745 Sco and RS Oph over two outbursts, it can be sur-mised that the synchrotron emission at 610 MHz in the next out-burst from V745 Sco should be detected before day 9.5 and fromRS Oph before day 6. If the accretion rate has reduced, then itwould lengthen the period between two outbursts – however if onlythe accretion winds have stopped, this should have no e ff ect on the c (cid:13) , 1– ?? ecurrent nova evolution from radio synchrotron emission outburst frequency.4. The electron energy spectrum set up by the shock in two distinctoutbursts appear similar for the two systems studied here. Multifre-quency radio synchrotron data is required for further study which isfeasible in future outbursts in the fast-evolving recurrent nova sys-tems provided time allocation is made faster on major radio tele-scopes. In this paper we have presented our observations, at 610 and 235MHz using the GMRT, of the recurrent nova system V745 Sco fol-lowing its outburst in 2014. The parametric model including opac-ities due to clumpy and uniform media in Weiler et al. (2002) ex-plains the light curves of V745 Sco and RS Ophiuchi. We concludethe following from our study:(1) The radio synchrotron emission is visible sooner after the out-burst, with each outburst. In V745 Sco, the 610 MHz emissionpeaked ∼ day 23 in 2014 and ∼ day 18 at 1.4 GHz in the 1989 out-burst. Our model fit predicts that the 1.4 GHz emission would havepeaked ∼ day 10 in 2014. In RS Ophiuchi, the radio synchrotronemission at 1.4 GHz turned on on day 20.5 in 1985 whereas thefirst detection in 2006 was on day 4.7 (Eyres et al. 2009).(2) The circumbinary material in the recurrent nova with a red giantcompanion is evolving with time. Clumpy material lies closer to thesystem compared to the extent of the uniform medium. This mate-rial could be due to the accretion disk of the white dwarf whichis destroyed with each outburst. The uniform density componentis caused by the hot optically thick winds blowing from the whitedwarf. The earlier visibility could indicate that the winds are ar-rested due to the accretion rate falling below some critical rate fora given white dwarf mass. This could lead to multiple evolution-ary scenarios which need to be investigated further. Interestingly,Williams (2013) also required a medium with clumpy and uniformcomponents to explain optical and X-ray data. Well-sampled mul-tifrequency data during the rise of the light curve to peak are neces-sary to estimate the e ff ect of the uniform and clumpy components.(3) All recurrent nova systems at all wavebands in quiescent (e.g.Anupama & Mikołajewska 1999) and outburst phases need to bestudied. Novae are an important Galactic system suited to the studyof shock interaction with the ambient medium and its evolutionover short timescales and multiple epochs. ACKNOWLEDGEMENTS
We thank the reviewer, A. R. Taylor for a helpful review. Wethank the sta ff of the GMRT that made these observations possi-ble. GMRT is run by NCRA of the Tata Institute of FundamentalResearch. We thank the Centre Director, NCRA for granting DDTtime. We thank the AAVSO for all their valuable work. NGK thanksPrasad Subramanian for discussions and Dave Green for commentson the manuscript. PD acknowledges that this work is partially sup-ported by the DST INSPIRE Faculty Fellowship award [IFA-13 PH54] and performed at IISER, Bhopal. REFERENCES
Anupama G. C., Mikołajewska J., 1999, A&A, 344, 177Banerjee D. P. K., Joshi V., Venkataraman V., Ashok N. M., Mar-ion G. H., Hsiao E. Y., Raj A., 2014, ApJ, 785, L11 Bode M. F., Kahn F. D., 1985, MNRAS, 217, 205Cheung C. C., Jean P., Shore S. N., 2014, ATel, 5879Duerbeck H. W., 1989, Msngr, 58, 34Eyres S. P. S., et al., 2009, MNRAS, 395, 1533Hachisu I., Kato M., Nomoto K., 1996, ApJ, 470, L97Hachisu I., Kato M., 2001, ApJ, 558, 323Hjellming R. M., van Gorkom J. H., Taylor A. R., Sequist E. R.,Padin S., Davis R. J., Bode M. F., 1986, ApJ, 305, L71Hjellming R. M., 1989, IAUC, 4853, 2Iben I., Jr., Tutukov A. V., 1984, ApJ, 284, 719Kantharia N. G., Roy N., Anupama G. C., Banerjee D. P. K.,Ashok N. M., Dutta P., Prabhu T. P., Johri A., 2014, ATel 5962Kantharia N. G., Anupama G. C., Prabhu T. P., Ramya S., BodeM. F., Eyres S. P. S., O’Brien T. J., 2007, ApJ, 667, L171Nomoto K., Nariai K., Sugimoto D., 1979, PASJ, 31, 287Nomoto K., 1982, ApJ, 253, 798O’Brien T. J., et al., 2006, Natur, 442, 279Orio M., Rana V., Page K. L., Sokoloski J., Harrison F., 2015,MNRAS, 448, L35Schaefer B. E., 2010, ApJS, 187, 275Spoelstra T. A. T., Taylor A. R., Pooley G. G., Evans A., AlbinsonJ. S., 1987, MNRAS, 224, 791Swarup G., Ananthakrishnan S., Kapahi V. K., Rao A. P., Subrah-manya C. R., Kulkarni V. K., 1991, CuSc, 60, 95Taylor A. R., Davis R. J., Porcas R. W., Bode M. F., 1989, MN-RAS, 237, 81Weiler K. W., Panagia N., Montes M. J., Sramek R. A., 2002,ARA&A, 40, 387Williams R., 2013, AJ, 146, 55 c (cid:13) , 1–, 1–