Swift Observations of Novae Julian Osborne the Swift

Swift Observations of Novae Julian Osborne & the Swift nova-CV group http: //www. swift. ac. uk/nova-cv/ Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

Swift Observations of Novae Julian Osborne & the Swift nova-CV group http: //www. swift. ac. uk/nova-cv/ Nova: Thermo-nuclear runaway on an accreting white dwarf Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

Previous best nova in X-rays Nova Cyg 1992 (V 1974 Cyg) • Rosat PSPC • 18 observations • Super-soft Source lasts ~400 days Krautter et al 1996 • Duration of SSS phase generally unknown – important constraint on WD mass (Tturnoff ~ M-6. 3) Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

X-rays from novae Potential sources of X-ray emission from classical/recurrent novae: • Thermal emission from hot white dwarf • shock breakout • residual nuclear burning after ejecta dispersal Swift • High velocity shocks • internal shocks within the ejecta • shock of ejecta with shell from previous nova or planetary nebula • shock of later fast wind with earlier slower wind • Re-established accretion Swift • Compton degraded gamma-ray emission (before optical peak) • not yet seen Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

Nova X-ray light-curves Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

Swift novae stats • Swift has observed ~40 novae within 4000 days of outburst • 25 detected in X-rays • 7 novae have >100 ksec each: U Sco, KT Eri, N LMC 2009, CSS 081007, V 2491 Cyg, V 458 Vul & RS Oph • Observations start within 1 day (pre-nova for V 2491 Cyg & U Sco) Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

RS Oph: XRT light curve 0. 3 -10 ke. V light curve shows: Osborne et al 11 • Cooling hot gas emerging from red giant wind • Bright interval which lasts ~64 d • Shock breakout from wind since last outburst 21 yrs ago during bright phase Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

RS Oph: XRT light curve 0. 3 -10 ke. V light curve shows: 32 milli. Crab Osborne et al 11 • Cooling hot gas emerging from red giant wind • Bright interval which lasts ~64 d • Shock breakout from wind since last outburst 21 yrs ago during bright phase Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

RS Oph BAT detection BAT 14 -25 ke. V • RS Oph was detected by the BAT in the softest band Bode et al 2006 • Emission was likely thermal rather than compton degraded radioactivity Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

RS Oph shock development Asymmetric shock propagates into RG wind Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec) Walder et al 2008

RS Oph: XRT light curve 0. 3 -10 ke. V light curve shows: • Cooling hot gas emerging from red giant wind • Noisy onset of super-soft phase, which lasts ~64 day in total • Turnoff time → MWD~1. 35 M☼ Osborne et al 11 Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

RS Oph X-ray spectra • 1 st appearance of hot WD on day 26 • k. TBB=31 e. V on day 29, = 56 e. V on day 50. 5 • Variable neutral oxygen absorption (0. 54 ke. V) • Late flux decline consistent with temperature drop to k. T = 41 e. V at day 76. 9 Osborne et al 11 Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

RS Oph X-ray spectra • NLTE WD model atmosphere (Rauch+ 10) spectral fits (declining wind absorption with O=OSun/3, Ness et al 07) wavelength • Initial low temperature and large radius change to changes to high temperature and smaller radius • Peak Luminosity ~ LEdd • Late count rate decline due to cooling & shrinkage days Osborne et al 11 Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

A quasi-periodic modulation Day 33 -45 Day 45 -58 Period near 35 s in soft X-rays between days 3359 WD spin? Beardmore et al 2009 Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec) Nuclear burning instability?

A quasi-periodic modulation Period near 33 sec also seen by Chandra LETG on day 40 Period near 35 sec confirmed by XMM Day 33 -45 Orio et al 2007 Ness et al 2007 Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

Periods in SS novae Drake et al 2003 Nova 1999 b Aql (V 1494 Aql) • Chandra ACIS • ~42 minute period Ness et al 2003 Nova Sgr 2002 c (V 4743 Sgr) • Chandra grating • 22 minute period • Softening decline in ~2 hrs Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

Quasi-periods in SS novae • Previously seen modulation has been multiply periodic – therefore not due to WD rotation • These 20 -30 min periods have been ascribed to nonradial g-mode oscillations in the white dwarf (cf hot planetary nebula nuclei) induced by compressionopacity instability [Drake et al 2003] • RS Oph periods at ~35 sec are very much shorter • Possible origin in the ε mechanism (nuclear burning instability) for which models predict periods 4 min - 1 sec [Kawaler 1988, Sastri & Simon 1973] – Consistent with modulation (mostly) visible during SS phase Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

HR = (0. 6 -2. 0 ke. V)/(0. 3 -0. 6 ke. V) Extreme early variability 100 sec hardness and count rate values compared to expected hardness for k. T=52 e. V blackbody and ISM plus variable neutral absorption GK Per 18 Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

Rapid variability in SS novae Flux variation unexplained Brief flux increase Drake et al 2003 Nova 1999 b Aql (V 1494 Aql) Does not look like variable absorption • Chandra ACIS • X-ray burst lasts ~14 mins Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

Variable absorption? General trend of hardening at times of low flux is consistent with neutral absorption model Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

Variable absorption? General trend of hardening at times of low flux is consistent with neutral absorption model Counter-examples exist, due to ionization of oxygen? - Reduction of variability due to blob expansion/ionization? Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

Variable absorption? Ratio of bright phase spectrum to faint phase shows complexity of the spectral change with intensity This is not simple absorption Its probably not absorption at all Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

RS Oph Summary • • • Recurrent nova in 455 day binary Initially very hard X-ray source due to wind shock Super-soft phase lasted ~64 days, short duration suggests high mass WD SS phase started with v large variability. X– ray spectral behaviour not simple, UV unaffected 35 sec qpo present at some times during SS phase – WD instability? Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

RS Oph Summary • Early decline of SSS phase strongly suggests a very high mass WD (early fuel exhaustion due to ignition at low accreted mass) • From SSS: M_burnt ~ 3 x 10 -7 M_sun • To reach ignition: M_acc ~ 4 x 10 -6 M_sun • From shock: M_ej ~ 10 -7 -10 -6 M_sun • Nova cycle mass budget: accreted mass >4 x ejected mass → future SN • but uncertainties large Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

V 2491 Cyg Page et al 2010 V 2491 Cyg was observed – and detected - pre-outburst as part of the BAT survey follow-up. This X-ray source may have been the counterpart of the BAT source. Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

V 2491 Cyg Numbers in the top right corner indicate the day after outburst. Short-term X-ray variability is seen, though there is no strict periodicity. Similar (unexplained) variability has been seen in other SS novae. Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

V 2491 Cyg The UV (1928 Å) evolution is very different from that seen in X-rays, showing a fading trend beyond at least 20 days after the outburst. Earlier observations were saturated. The UV curve has inflections at ~40 and ~57 days, at the times of the X -ray peak and the end of the rapid decline. Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

V 2491 Cyg Blackbody fit parameters while the super-soft source was visible Absolute values probably not reliable A rising temperature and shrinking radius are seen between days 30 - 57 Where is the expected constant luminosity nuclear burning phase? Grey points show luminosity upper limits from linking X-ray to UV. Page et al 2010 Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

V 2491 Cyg Mc. Donald CO LTE model fits Show same trends of rising temperature and shrinking radius as BB fits Page et al 2010 Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

V 2491 Cyg Mc. Donald LTE models are worse spectral fits than BB Page et al, subm Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

V 2491 Cyg 0. 3 -10 ke. V Page et al 2010 Flickering is visible after day ~57. If this is the recovery of accretion due to viscous expansion of a disk, then disk was removed to 2 x 1010 cm. With Porb=0. 96 days, Rdisk ~ 3 x 1010 cm. Inner disk only destroyed by nova Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

V 2491 Cyg Page et al 2010 Estimates of the time to reach ignition pressure for V 2491 Cyg derived from the observed pre-nova luminosities of Ibarra et al (2009). Curves are from the Nauenberg (1972) WD M-R relationship (points from Althaus et al 2005), ranges reflect distance and accretion efficiency uncertainties. Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

V 2491 Cyg It is highly unusual to see a nova in X-rays pre-outburst. The only others which have been detected pre/inter-outburst have actually been recurrent novae (e. g. RS Oph and V 2487 Oph). A high inter-outburst accretion rate is expected if a nova is to be observably recurrent. It seems possible that V 2491 Cyg will be a recurrent nova. Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

KT Eri Nova Eri 2009 m. V, peak = 5. 4 14 Nov 2009 Early SS X-ray behaviour shows large-scale variability like RS Oph Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

KT Eri Clear spectral variations of a very super-soft spectrum Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

KT Eri Spectrum varies on both low and high energy sides Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

KT Eri Ratio: day 83. 5 over day 82. 0 Energy (e. V) Ratio of bright over faint spectra shows variation is mostly Oxygen line emission Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

KT Eri No correlation of X-ray with optical flux Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

KT Eri Is there a ~35 sec modulation in KT Eri, just like in RS Oph? Yes, there is! (ATEL 2423, Beardmore et al) NB: No such modulation is seen in the AGN Mkn 421, which also had ~150 XRT c/s (ie <0. 4%). Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

KT Eri XRT Chandra XRT 06 Feb 2010 = day 83. 6 Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

KT Eri Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

KT Eri Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

KT Eri Summary • SS phase onset around day 60 shows large scale variability like RS Oph • Spectrum is very soft (even for a SSS), k. T_BB = 16 e. V • Large spectral variability does not look like simple temperature or absorption variations • X-ray and optical flux unrelated • Variable strength ~35 sec qpo during SS phase, period & amplitude not a strong function of SSS intensity • This is the second SS nova to show ~35 sec qpo, favouring a WD instability origin over a rotation-based mechanism Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

U Sco – the fastest nova U Sco is the only eclipsing recurrent nova. Porb = 1. 2 day Outbursts regularly every 10. 3 yrs, next was predicted at 2009. 3± 1. 0 Behaviour in 1945 Previous outburst with template from 7 outbursts (Schaefer 2010) Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

U Sco Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

U Sco Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

U Sco Osborne et al ATEL 2442 Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

U Sco Days: 12. 1 -13. 8, 14. 3 -16. 2, 16. 3 -18. 5, 19. 2 -20. 6 Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

U Sco 1 mag. Quiescent eclipse (Schaefer ar. Xiv: 0912. 4426) Optical eclipses only seen after day 14 Early optical eclipses are broad and shallow Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

U Sco • BB spectral fits show shrinking photosphere revealing higher temperatures • Luminosity peak ~ Eddington • Radius is cold WD value by day ~27 Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

U Sco Summary • U Sco is the fastest declining nova known, suggesting a very high mass • It is the only eclipsing RN, offering a unique opportunity to dissect the progress of the outburst • The SS phase onset between day 8 -12 coincides with the optical plateau, although this is not generally true of novae • Broad and shallow eclipses are seen in the UV & optical from day 14 only. • Lack of strong X-ray eclipses appears to imply scattering • Very strong optical emission line wings (‘horns’) will constrain the ejecta geometry • The UV spectrum is very rich, giving great hope of a radial deconstruction of the properties of the evolving nova ejecta Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

V 407 Cyg Summary Fermi detection of Nova V 407 Cyg at <100 Me. V Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec) Abdo et al 2010 Science

V 407 Cyg X-ray flux X-ray HR UV flux Fermi detection during initial hard phase Short-lived bright X-ray phase ends at time of UV faster decline Shore et al 2011 Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

V 407 Cyg Complex X-ray spectrum: Enhanced abundances & 2 temperatures required for initially highly absorbed cooling optically thin emission (N>250, O~35, Ne<3, Na>20, Mg<2. 3) Shore et al 2011 BB-like soft component seen before day 60 Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

V 407 Cyg • First nova of V 407 Cyg • Secondary is Mira-type with strong wind • RS Oph-like nova evolution of X-rays and optical line profiles • Enhanced abundances seen in X-rays indicative of nuclear explosion • High ionization lines eventually became highly asymmetric • Indicates explosion off-centre in Mira wind • Mira has longest P~750 d and does not fill Roche lobe • Measured shock velocity of ~2500 km/s consistent with Fermi detection at ~3 days • Orbital period likely >30 years He II CV CV Shore et al 2011 Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)

Summary • The scheduling flexibility and multi-wavelength capability of Swift make it the best rapid-response nova observing facility • Extensive datasets are revealing novel behaviours • The short SS phase durations measured suggest high M WD →SN 1 a? • • Blackbody fits show similar trends to static NLTE model atmosphere fits, but lack of outflow in model limits inference from SSS spectral fits • Periodic & aperiodic X-ray variability still to be understood. Not shown: Orbital modulation likely in some cases. X-ray & UV/opt are correlated/anticorrelated/unrelated Swift Novae – Julian Osborne – MAXI 1 st year meeting – Shibuya, Japan (30 Nov – 2 Dec)