PowerPoint Presentation - Presentaci

1
Explosion, turn-off and recovery of accretion in
novae revealed by X-rays Margarita
Hernanz Institut de Ciències de lEspai
(CSIC-IEEC) - Barcelona (Spain)
2

• OUTLINE
• Classical and recurrent novae explosions
  scenarios
• Origin of X-ray emission
• Summary of X-ray observations - Theoretical
  implications
• Possibility to accelerate cosmic rays in novae
  (symbiotic recurrent) RS Oph and V407 Cyg
• Conclusions

3
White dwarfs

• Endpoints of stellar evolution (Mlt 10M?) no
  Enuc available compression until electrons
  become degenerate
• Chemical composition He, CO, ONe masses
  typical 0.6 M?, maximum MChandrasekhar (1.4M?)
• When isolated, they cool down to very low L
  (10-4.5L?)
• When in interacting binary systems, they can
  explode

4
White dwarfs in close binary systems
Symbiotic system WD Red Giant
accretion from a red giant wind
Cataclysmic variable WD Main Sequence
Roche lobe overflow
Classical nova
Recurrent nova
Hydrogen burning in degenerate conditions on top
of the white dwarf
Credit David Hardy
Prec104-105 yr Porbhr-day afew 105 km few
1010cm rate 35/yr in Galaxy
Preclt100 yrs Porb100s days a 1013-1014
cm rate 10 known in Galaxy
5

• In recurrent novae, initial mass of the WD should
  be very large (close to Chandrasekhar mass), to
  drive such frequent outbursts
• Feasible scenario of type Ia supernova
  explosions, provided that less mass is ejected
  than accreted in each explosion
• X-ray observations way to study if WD mass grows
  or diminishes after each nova explosion
• Solve the controversy between single degenerate
  and double degenerate scenarios of type Ia
  supernovae

Credit David Hardy
6
Scenario for classical novae
Mass transfer from the companion star onto the
white dwarf (cataclysmic variable) Hydrogen
burning in degenerate conditions on top of the
white dwarf Thermonuclear runaway Explosive
H-burning
Decay of short-lived radioactive nuclei in the
outer envelope (transported by convection) Envelo
pe expansion, L increase and mass ejection
7
Novae observations optical light curve
apparent luminosity (mv)
time
L increases very fast by factors greater than 104
-absolute Lmax104-5L?
8
Novae observations light curves
UV satellites Lbol(LVLUV)ct.
VUV
FH Ser 1970 - Gallagher Code 1974
V
TOT
V
UV
Lbol(LVLUV LIR) ct.
IR
Nova Cyg 1978 Stickland et al. 1981 IR emission
dust formation
9
Photosphere recedes as matter expands and becomes
transparent Supersoft X-ray emission reveals
the hot white dwarf photosphere, close to the
burning shell
10
Origin of X-ray emission (I)

• Residual steady H-burning on top of the white
  dwarf
• photospheric emission from the hot WD
• Teff (2-10)x105K (L1038erg/s)
  supersoft X-rays
• detected by ROSAT/PSPC in only 3 classical novae,
  out of 39 observed up to 10 years after
  explosion
• GQ Mus (N Mus1983), N Cyg 1992, N LMC 1995
• (Orio et al. 2001). A few more detections with
  BeppoSAX, Chandra, XMM-Newton many more with
  Swift/XRT
• duration related to H-burning turn-off time.
  Old theory tnuc100yr observations lt9 - 12
  yr typically lt 2yr
• new models L-MH,rem-Teff compatible with short
  duration of soft X-ray phase (Tuchmann Truran
  1998 Sala Hernanz, 2005) ? very small remnant
  H-mass

11
Origin of X-ray emission (II)

• Internal (external) shocks in the ejecta thermal
  plasma emission
• detected early after explosion (N Her 1991, N Pup
  1991, N Cyg 1992, N Vel 1999) internal shocks
  recurrent nova RS Oph external, V2491 Cyg 2008 ?
• Reestablished accretion emission as a CV
  (idem)
• Hard (but also soft) X-rays, depending on the
  thermal plasma T

12
Origin of X-ray emission (II, contd)

• Restablished accretion
• emission CV-like
• How and when?
• Interaction between ejecta and new accretion
  flow?
• Magnetic or non magnetic white dwarf?

13
Origin of X-ray emission (III)

• Compton degradation of ?-rays emitted by
  classical novae CAN NOT be responsible of their
  early hard X-ray emission
• Cut-off at 20 keV (photoelectric abs.)
• Fast disappearence 2days (w.r.t Tmax, i.e.,
  before visual outburst)

Gómez-Gomar,Hernanz,José,Isern, 1998, MNRAS
14
Observations Supersoft X-ray emission

• EXOSAT and ROSAT discoveries
• GQ Mus (1983) 1st detection of X-rays in a nova,
  EXOSAT (Ögelman et al. 1984). One of the longest
  supersoft X-ray phases 9 yr Ögelman et al.1993
  Shanley et al. 1995 Orio et al. 2001 Balman
  Krautter 2001
• V1974 Cyg (1992) complete light curve with
  ROSAT- rise, plateau and decline 1.5 yr
  Krautter et al. 1996, Balman et al. 1998
• N LMC 1995 ROSAT XMM-Newton 8 yrs
• ROSAT discovery Orio Greiner 1999 XMM-Newton
  obs. Orio et al. 2003

15
V1974 Cyg (1992) ROSATs soft X-ray light curve
rise until day 147 plateau 18 months BB fits
not good too large L Krautter et al. 1996, ApJ
ONe WD atmospheres MacDonald Vennes Balman et
al. 1998, ApJ
16
V1974 Cyg (1992) ROSATs soft X-ray spectra
F6x10-10erg/cm2/s kTBB21eV, kTbr0.32keV
F3.2x10-9erg/cm2/s kTBB30eV ( kTbr0.002keV)
F3x10-11erg/cm2/s kTBB20 eV,kTbr0.29keV
F3.1x10-9erg/cm2/s kTBB30eV (kTbr0.002keV)
17

• Models that best explain the supersoft X-ray
  emission of V1974 Cyg 1992 and its evolution
• WD envelope models with steady H-burning (no
  accretion)
• Mwd0.9 M?, 50 mixing with CO core (but V
  1974Cyg 1992 was a neon nova!)
• or
• Mwd1.0 M?, 25 mixing with ONe core
• in goog agreement with models of the optical and
  UV light curve (Kato Hachisu, 2006)
• Menv2x10-6 M?
• WD properties from X-ray observation of turn-off

Sala Hernanz, AA 2005
18
Observations Supersoft X-ray emission

• BeppoSAX
• V382 Vel (1999) supersoft X-ray flux not
  constant model atmosphere not a good fit
  emission lines from highly ionized nebula were
  required (Orio et al 2002)
• Chandra grating observations detected emission
  lines (Burwitz et al., 1992, Ness et al. 2005).
• Turn-off 7-9 months

19
Observations Supersoft X-ray emission

• Chandra LETGS
• V382 Vel (1999)
• Burwitz et al. 2002
• Ness et al. 2005

20
Observations Supersoft X-ray emission
Chandra and XMM-Newton (novae in
outburst) puzzling temporal behaviours grating
observations V1494 Aql (1999) - burst and
pulsations Drake et al. 2003 V4743 Sgr (2002) -
strong variability and complex spectra Ness et al
2003, Rauch, Orio, González Riestra et al., 2010
fits with NLTE WD atmospheric models ? see
Rauchs talk C1
21
Observations Supersoft X-ray emission
V4743 Sgr (2003) Temporal variability P 22
min. Ness et al. 2003, ApJ
22
Observations Supersoft X-ray emission
V4743 Sgr (2003) Non LTE model
atmospheres Rauch, Orio, González-Riestra et
al., 2010, ApJ
23
Observations Supersoft X-ray emission
XMM-Newton Monitoring campaigns of
post-outburst novae Nova LMC 1995 - Orio et al.
2003 H-burning still on in 2000 ? see Orios
talk C2, about Nova LMC 2009 Galactic novae
V5115 Sgr and V5116 Sgr 2005 Hernanz, Sala et al.
24
XMM-Newton - AO1 Cycle -Summary
Target Discovery date Date of observation Time after outburst Detection
N Sco 1997 V1141 Sco June 5 Oct. 11, 2000 1224d, 3.4yr Mar. 24, 2001 1388d, 3.8yr Sep. 7, 2001 1555d, 4.3yr NO
N Sgr 1998 V4633 Sgr March 22 Oct. 11, 2000 934d, 2.6yr Mar. 9, 2001 1083d, 3.0yr Sep. 7, 2001 1265d, 3.5yr YES but no SSS
N Oph 1998 V2487 Oph June 15 Feb. 25, 2001 986d, 2.7 yr Sep. 5, 2001 1178d, 3.2 yr Feb. 2002 1352d, 3.7yr Sept. 24, 2002 1559d, 4.3yr YES but no SSS
N Sco 1998 V1142 Sco October 21 Oct. 11, 2000 721 d, 2.0 yr Mar. 24, 2001 885 d, 2.4 yr Sep. 7, 2001 1052 d, 2.9 yr 2.6?0.3 2.2?0.4 1.2?0.2 (10-2 cts/s)
N Mus 1998 LZ Mus December 29 Dec. 28, 2000 730 d, 2.0 yr Jun. 26, 2001 910 d, 2.5 yr Dec. 26, 2001 1093 d 3.0 yr NO?

• No supersoft X-ray emission related to residual
  H-burning detected
• ? all novae had already turned-off
• 3 out of 5 were emitting thermal plasma ( BB)
  spectrum ? ejecta/accretion

25
Target Discovery date Date of observation Time after outburst Detection
N Oph 1998 V2487 Oph June 15 Mar. 24, 2007 8.8yr AO6 long exposure YES but no SSS
N Cyg 2005 V2361 Cyg February 10 May 13, 2006 - 15mo bkg Oct. 20, 2006 - 20months AO5 -- YES marginal (4.0?0.8)x10-3 cts/s
N Sgr 2005a V5115 Sgr March 28 Sep. 27, 2006 18months Apr. 4, 2009 49 months YES supersoft source YES but no SSS
N Sgr 2005b V5116 Sgr July 4 Mar. 20, 2007 20 months Mar. 13, 2009 44 months YES supersoft source YES but no SSS
N Cyg 2006 V2362 Cyg April 2 May 5, 2007 13 months affected by bkg AO6 Dec. 22, 2008 32 months YES but no SSS YES but no SSS
N Oph 2006a V2575 Oph February 9 Sep. 4, 2007 19 months AO6 NO
N Oph 2006b V2576 Oph April 6 Oct. 3, 2007 18months AO6 NO
Supersoft X-ray emission related to residual
H-burning found in 2 novae from 2005 (V5115 Sgr
V5116 Sgr)? novae had not turned-off yet
26
Nova Sgr 2005 b V5116 Sgr 610 days
post-outburst
partial eclipse by an asymmetric disk? Sala,
Hernanz, Ferri Greiner, ApJL 2008
27
Nova Sgr 2005 b V5116 Sgr 610 days
post-outburst
RGS spectra
Sala,Hernanz, Ferri, Greiner, AN (2010)
28
Nova Sgr 2005 b V5116 Sgr new obs. March 2009
1348 days post-outburst
U filter
L(3-7)x1032 erg/s (10 kpc)
Swift/XRT light light curve SSS turn-off 2 - 3
years post-outburst compatible with Hachisu
Kato (2007) prediction
29
SUMMARY of XMM-Newton campaign on Galactic novae

• 11 novae have been observed between 3 months and
  5 years after outburst (9 years)
• Only 2, V5115 Sgr 2005a and V51116 Sgr 2005b,
  were still bright in supersoft X-rays, revealing
  remaining H-nuclear burning one of them with a
  puzzling temporal behavior
• SSS phase absent means that either we missed it
  or Mejected gt Maccreted Mwd decreases after
  each nova outburst ? WD cant reach MCHANDRA and
  explode as SNIa

30
Observations Supersoft X-ray emission

• Swift/XRT
• Ness et al. 2007, Osborne (todays talk, C1)
• The largest sample. Example two extreme cases
• V723 Cas (1995) L and Teff not well determined
  (BB) Ness et al. 2008 Still SSS 12 yrs after
  outburst. New XMM-Newton observations in 2010,
  still active
• V2491 Cyg (2008) duration SS phase 10 days
• Also observed with XMM-Newton and Suzaku Ness et
  al. 2011, Takei et al. 2011

31
Observations Supersoft X-ray emission

• V723 Cas (1985)
• Swift observations in 2007 Ness et al. 2008,
  MNRAS
• not turned-off yet
• XMM-Newton obs. in 2010 still on

32
Observations Supersoft X-ray emission
V2491 Cyg (2010) Page et al., 2010, MNRAS
(Other interest of this nova later)
33
Observations Supersoft X-ray emission
V2491 Cyg (2010) Ness et al. 2011
34
SUMMARY of Swift/XRT campaign
From Julian Osborne see talk in C1
35
2007-2008
Novae in M31 XMM-Newton Chandra monitoring d
and line of sight absorption known Henze et al.
2011, AA
2008-2009
? see talks by Henze C1 Pietsch C2 this afternoon
36
Observations of novae where H has turned off
Recovery of accretion and/or ejecta emission
37
Nova Oph 1998 V2487 Oph - 4.3 yrs post explosion

• Identification of three Fe Ka emission lines
  neutral Fe 6.4 keV He-like Fe 6.68 keV
  H-like Fe 6.97 keV
• If Thigh (10-20) keV, He-like and H-like lines
  well reproduced only 6.4 keV fluorescent line
  added
• If complex absorption -partial covering
  absorber- low (ISM) high NH ? Thigh(10-20) keV

Fluorescent Fe Ka line at 6.4 keV reveals
reflection on cold matter (disk and/or WD)
accretion
38
Nova Oph 1998 V2487 Oph 4.3 yrs post explosion
d10 kpc

• LBB 50 LTOT0.2-10 keV - f(emitting
  surface/wd surface)10-4 (hot spots)
• Luminosity, spectral shape ..? Intermediate
  polar? need Pspinvs. Porb

39
N Oph 1998 V2487 Oph Mar. 24, 2007 8.8yr post
outburst

• Spectral model similar to previous observations
• No clear periodicities in X-rays, but hint of
  orbital period 6.5 hrs
• Optical observations seem to confirm the orbital
  P

40
V2487 Oph (1998) 1st nova seen in X-rays before
its explosion (ROSAT)
Positional correlation with a source previously
discovered by ROSAT (RASS) in 1990 suggests that
the host of the nova explosion had been seen in
X-rays before the outburst (Hernanz Sala 2002,
Science)? new case V2491 Cyg (2008b) previous
ROSAT, XMM and SWIFT detections (Ibarra et al.
2009, AA)
41
Nova Oph 1998 V2487 Oph Hard X-rays

• Detection with INTEGRAL/IBIS survey in the
  20-100 keV band (Barlow et al. 2006, MNRAS)
  kT25 keV flux compatible with our XMM-Newton
  results, but the IBIS spectrum has low S/N.
• Hints for large MWD from the optical light curve
  (Hachisu Kato, 2002, ApJ)
• also large MWD from large Thigh deduced from
  X-ray spectra but Thigh not well constrained
• The recent nova V2491 Cyg (2008b) has also
  been detected in hard X-rays with Suzaku (Takei
  et al. 2009)

42
Observations wih Suzaku and XMM-Newton V2491
Cyg (2008) prompt and short duration hard
X-rays Takei et al. 2009 and 2011
43
Nova Oph 1998V2487 Oph - Recurrent Nova

• Previous outburst in 1900 June 20, discovered in
  the Harvard College Observatory archival
  photograph collection Pagnotta and Schaefer, IAUC
  8951, 200 2009 AJ)
• recurrent nova - P98 yrs
• MWD very close to MCHANDRA ? relevance for the
  SNIa scenario
• challenge for theory to get recurrent
  nova explosions with such short time scales
• X-ray emission CV-like ? RN scenario
• The recent nova V2491 Cyg (2008b) has also
  been claimed to be recurrent. It was also a very
  fast nova, expected to be massive, very luminous
  in X-rays (Ibarra et al. 2009, AA), and detected
  in very hard X-rays (Takei et al. 2009)

44

• Models of recurrent novae TNR on accreting WDs
• Search combinations of initial conditions leading
  to short recurrence periods
• Prec ?Macc / (dM/dt) 98 yrs (21 years for RS
  Oph)
• ?Macc required accreted mass on top of the WD
  to power the outburst through a TNR
• Mwdini? Accretion rate? Lwdini?
• Accretion rate related to mass loss from the
  red giant wind
• effective dM/dt onto the WD 2x10-7 - 10-8 M?/yr

45
Accreted masses to reach H-ignition conditions
critical accreted mass does not depend only on
Mwd Mwd very close to MCHANDRA
Lini
Hernanz José 2008
46
Recurrence Periods
V2487 Oph 1998 Prec98 yr
Lini

RS Oph Prec21 yr
.
M2 10-7 M?/yr L10-2 L?

Hernanz José 2008
47
CONCLUSIONS (recovery of accretion)

• X-rays are crucial to study the recovery of
  accretion in post-outburst novae type of CV,
  mass of the WD
• Magnetic WD challenge for accretion
  traditionally assumed to occur through a normal
  accretion disk in a non magnetic WD. But some
  cases of novae in magnetic CVs are known V1500
  Cyg (1975), V4633 Sgr (1998) asynchronous polar
  as a consequence of the nova outburst (Lipkin
  Leibowitz, 2008), V2487 Oph (1998), V2491 Cyg
  (2008)
• ? see Pietschs talk C2 M31N 2007-12b, an IP?
• Massive WD if Thigh(plasma) is large and/or the
  nova is recurrent. Novae as scenarios for type Ia
  supernovae
• but very ad-hoc conditions are required to
  obtain a recurrent nova (Precurrence lt 100
  years)
• but XMM spectra (V2487 Oph) looks CV-like ? RN
  scenario

48
An interesting case the recurrent nova RS Oph,
which erupted in 2006 Previous eruption in 1985
Prec 21 yrs Short recurrence period ? large
MWD close to MChandra (deduced from models) ?
possible SNIa scenario (but should be CO
WD!) Porb456d RG companion symbiotic
recurrent nova Detected as a very variable SSS
by Swift/XRT (Bode et al. 2006), XMM-Newton
(Nelson, Orio et al. 2008, Ness et al. 2009)
49
Supersoft X-ray light curve of the recurrent nova
RS Oph (Swift observations, Bode et al. 2006)
Mwd1.35 M? Menv 4x10-6 M?
Kato Hachisu, 2007
50
RS Oph in quiescence observed with XMM-Newton
Nelson, Mukai, Orio, et al., 2011 observations
in quiescence, 537 and 744 after outburst ?
accretion rate theoretical In previous
eruptions very faint X-ray source in quiescence,
hard to reconcile with large accretion rates
needed to explain frequent (every 20 yrs)
outbursts
51
RS Oph grating observations with XMM-Newton and
Chandra Ness et al, Drake et al. 2009 See as
well observations of U Sco (another recurrent
nova, eclipsing, not of the symbiotic nova type
Ness talk C2
52

• RS Oph
• Prec 21 yr last eruptions in 1985, 2006
• MWD gt 1.35 M? (deduced from models not
  measured)
• Mejec (3-4)x10-6 M?
• not all accreted mass is ejected (deduced from
  models)?
• MWD increases M close to MCHANDRA ? SNIa
  scenario
• Interaction between nova ejecta and red giant
  wind expanding shock wave sweeps through the red
  giant wind (mini SN remnant)
• Detected from radio to X-ray wavelengths
• X-rays reveal interaction between ejecta and
  wind (hard) and hot white dwarf surface with
  remaining nuclear burning (soft)
• Acceleration of particles (Tatischeff Hernanz
  2007, ApJL)

53
RS Oph
A supernova remnant-like, but faster and dimmer
free expansion phase days
54
RS Oph (2006 eruption) blast wave evolution

• IR (Das06,
• Evans07)
• X-rays RXTE
• Swift (Sokoloski06, Bode06)

• 2 caveats
• Why shock cooling started at 6 days, when Ts was
  108K and radiative cooling was not important?
  Particle acceleration - CRs
• Why vshock (X-rays) lt v (IR) (for test
  particle strong shock) underestimates vshock
  when particle accel. is efficient, because Ts is
  lower (particle ecape and softer EOS)

55
RS Oph (2006 eruption)

• Non-linear diffusive shock acceleration model of
  Berezhko
• Ellison (1999)
• accelerated proton spectrum and post shock
  temperatures as a funtion of ?inj - the fraction
  of shocked protons injected into the acceleration
  process
• Tatischeff Hernanz, ApJL 2007

56
RS Oph (2006) cosmic-ray modified shock

• Good agreement with X-ray measurements of Tshock
  for moderate CR accel. efficiency ?inj10-4 and
  Alfvén wave heating of the precursor
• Energy loss rate due to particle escape
• 100 times larger than Lbol of postshock plasma ?
  energy loss via accelerated particle escape much
  more efficient than radiative losses to cool the
  shock

57
RS Oph (2006) predicted gamma-ray emission

• p0 production from eCR and (dM/dt)RG
• IC contribution from non thermal synchrotron L
  (Kantharia et al.07, radio 1.4 GHz), Lsyn5x1033
  td-1.3 erg/s, and ejecta L, LejLEdd 2x1038
  erg/s LIC Lsynx Urad/(B2/8p) Lsyn
• p0 production dominates
• RS Oph would have been detected by Fermi!

58
V407 Cyg

• Detected by Fermi/LAT 2 days after outburst
  Cheung et al. 2008, Science
• Main differences wrt RS Oph
• Not a standard recurrent novae no regular
  eruptions before 2010 (Munari 2010)
• Porb 43 yr (456d in RS Oph) ? a 15 AU (10
  times larger than for RS Oph) ? shock wave needs
  7 days to reach da, so it propagates through
  the RG wind perturbed by the orbital motion in
  V407 Cyg (RS Oph, free exp. unperturbed wind at
  1d)
• Lsyn needed to compute IC not available from
  early radio observations
• Preliminary estimation gamma-ray flux from p0
  production

59
SUMMARY (of GeV emission from symbiotic RNe)

• Recurrent novae in symbiotic binaries are
  expected to accelerate particles and emit VHE
  gamma-rays detectable with Fermi, because of the
  shock wave propagation in the dense red giant
  wind
• RS Oph would have been detected by Fermi
• V407 Cyg, detected by Fermi, did not behave as
  RS Oph, regarding X-ray and radio emission. So
  computing IC contribution is difficult.
• Other similar systems exist in the Galaxy
  eventually 1-2 novae with RG companion per yr are
  expected (but not necessarily detected in the
  optical)

60
Summary

• Variety of behaviours of post-outburst novae
  still need more observations
• Grating spectra are very rich, but still lack of
  emission models (e.g. WD expanding atmospheres)
  to interpret them. Blackbodies give wrong L and
  Teff emission is often a mixture of
  photospheric and ejecta components
• WD Mass and envelope chemical composition
  (mainly H content) determine duration of SS X-ray
  phase
• Duration of SS X-ray phase observed indicates
  in general Menv lt Macc-Meject from hydro models ?
  mass loss (wind and/or others?)
• Recurrent novae very short duration of SS
  phase compatible with small Menv. Challenging for
  theory narrow parameter range Mwd extremely
  large accretion rate large.
• ? main caveat for RNe as SNIa scenario not CO
  WDs but ONe