Toward Understanding the Engines of Post AGB Outflows

Eric G. Blackman (U. Rochester)

Planetary Nebulae: <50%(?) of 1-8 Msun stars make PN (DeMarco Soker ’11)

Abell 39 (Image credit: NOAO) NGC 3242 (B. Balick)

M2-9 (NASA/STScI)

Pre-Planetary Nebulae (AGB transitions to pPNe in few 100yr):CRL2688 (egg) (R. Sahai)

M27(dumbell) ( NASA)NGC 6543 (cat’s eye) (NASA)

20% spherical

more like 100% PPN are aspherical,fraction of PPN that become PN is unclear but assumed 0(1)

32% with jets

Red Rectangle NASA/ESA

48% elliptical without jets

Rotten Egg/Calabash

Calabash/Rotten Egg

SHAPING AND POWERING PPN vs. PN

• Both PN and PPN often bipolar jets with d(Mj vj)/dt > L rad/c

• PPN are more kinematically demanding than PN: (Bujarrabal et al. 01 CO PPN outflows) :

• ~28/32 objects have fast outflows (>~50km/s) with d(Mj vj)/dt >1000 L rad/c; (for PN d(Mj vj)/dt > L rad/c)

• suggests accretion power, MHD, and, in turn, binary systems (Soker...etc)

• what fraction of PPN does high momentum population characterize? need more data

• use basic kinematic data to asssess the minimum system requirements:

• energy and timing contraints in PPN outflows and tori help (Huggins 07,12) BUT need to constrain power

• B. and Lucchini (2014) PPN accretion rates

• Tocknell et al. (2014) 4+2 post CE PN, power and timing of outflows wrt CE, (also indirect B-field strength constraints)

Lmec =12M j ,Nvj ,N

2 ≤ 12GMa

Mac

rin= 1

2Macvk

2 (rin )

vj ,N ≡Qvk (rin ); 5Q1 from standard MHD jet models

so the inequality implies : Mac ≥Q2 M j ,N

Mac ≥Q2 M j,N Q

2 M j,N

tac

Q2 M j ,obsvj ,obs

Qvkep

1tac

= 1.4 ×10−4 M

yrQ3

⎛⎝⎜

⎞⎠⎟

Ma

M

⎛⎝⎜

⎞⎠⎟

−1/2RacR

⎛⎝⎜

⎞⎠⎟

1/2M j ,obs

0.1M

⎛⎝⎜

⎞⎠⎟

vj ,obs100km/s

⎛⎝⎜

⎞⎠⎟

tac

500yr⎛⎝⎜

⎞⎠⎟

−1

M j ,obs,vj ,obs,tac from observationsvkep from assumed accretor (e.g. MS or WD)

Momentum conservation constrains :

All jets obey basic kinematic constraints:

Mac

Accretion onto primary?

• Bondi-Hoyle-Lyttleton onto companion (Huarte-Espinosa et al ‘13): NO WAY

• WRLOF (Mohamed & Podsiadlowski 12) separation, (a=20AU), dust accel. radius (Rdust =6Rp =10AU) primary Roche Lobe =8.5 AU) NO FOR MIRA-TYPE

• Red-Rectangle (Witt et al. 2009) special case of RLOF PROBABLY NOT

• RLOF (Ritter 1988) OK FOR MOST

• CE (Ricker & Tamm 12) on companion: (highly super-Eddington) OK (but...)

• accretion onto primary WD from RLOF of secondary post CE (Soker & Livio 94) ( or shredded low mass companion (e.g. Nordhaus & B. 2006; B. et 2001; Reyes-Ruiz & Lopez 1999); perhaps thick disk and super-Edd. OK

• CE (Ricker & Tamm 12) on primary OK (but...)

Which modes of accretion work for Bujarrabal PPN?

MBH = 1.1×106M / yr, for primary wind Mw = 8 ×10−5M / yr; wind speed vw = 10km/sprimary mass Ms = 1.5Mp secondary mass Ms = 1M

MWR = 2 ×10−5M / yr; primary mass M p = 1M; secondary Ms = 0.6M

MCE = 10−2M / yr for primary mass M p = 1M; secondary Ms = 0.6M

Accretion onto Secondary?

MRR = 2 − 5 ×10

−5M / yr

MRL (theory) ≥10

−4M / yr

MCE = 10

−2M / yr

PPN Scenario for Red Rec. (Witt et al. 09)

-Upper left: HST composite from Cohen et al. (2004). bipolar axis length ~15,000 AU.

-Main: artist (S. Lane) depiction of of basic paradigm, scale ~ 1 AU. Accretion onto secondary via Roche lobe overflow

-lower left: HD 44179 may be embedded in the central cavity of a circumbinary disk of thickness 90 AU and cavity diameter 30 AU.

-Accretion rate constraint comes from needed luminosity to maintain Lyman continuum that sustains ionization of the HII region in RR (Jura 1997) (assume 710pc distance) need 2 -5 x 10-5 Msun/yr, and gives maximum disk temp of 17,000K

-jet produces blue shift in Hα emmission, modulating primary envelope

from Witt et al. (2009,11):

HD 44179: Red Rectangle

(NASA, ESA, Van Winckel Cohen)

Carbon-rich outflows with oxygen-rich contamination.

Accretion Rate Constraints for 19 objects (18 from Bujarrabal 01 &1 from Sahai et al. 08) (B. & Lucchini 2014)

White Dwarf

Main Sequence

MWR

MCE

Med ,WD

MBH

MRR

MCE

Med ,MS

MBH

MWR

Eddington for WD

Eddington for MS

Scaling of Accretion Rates with Accretor Mass

• mass radius relation for low mass MS (Demircan & Kahraman 91):

• empirical mass radius relation for planets

• for all cases, and only weakly dependent on mass for the MS case

• correlation between jet and tori (>100AU) outflow power (Huggins 12) might be explained if torus depends on either equatorial ejection by companion and jet from accretion, or if both depend on accretion.

(Torres 2012)

m= M/Me

dLmech / dMa > 0

Ra ~ 0.99R* RMa

M

⎛⎝⎜

⎞⎠⎟

0.89

, for zero age MS and

Ra ~ 2R* RMa

M

⎛⎝⎜

⎞⎠⎟

0.75

, for terminal age MS.

Connecting PN to PPN:

vj ,N =Qvk = 1000 km/s Q2 2

⎛⎝⎜

⎞⎠⎟

MM

⎛⎝⎜

⎞⎠⎟

RR

⎛⎝⎜

⎞⎠⎟

−1

Max(vj ,obs ) =M j ,Nvj ,N

fMenv(1− e−τ (RN ) )+M j ,N

~M j ,Nvj ,NfMenv

~ 100 km/s for PPN

f ~ 0.1 (solid angle of envelope intercepted) ~M j ,Nvj ,NM j ,N

~ 1000 km/s for PN

τ (RN ) is optical depth at radius where jet speed equals naked launch speed.

• evolution of outflow from PPNe to PNe can be predicted for each accretion scenario as accretion rate evolves, but launch speed is fixed.

• Lower power but FASTER outflow speeds for PN compared to PPN if observations probe deeper in core & jet not as slowed by mass loading.

• Tocknell et al. 2014: Kinematics of 4+2 post CE, PN; three with jets that precede CE ejection; one with 2 pairs of jets that follow CE:

• Huggins 2007: PPN jets follow tori by ~250 yr in PPN

• Likely: multiple accretion scenarios with subclasses needed.

• absence of observed binary or the presently observed binary need not be the source of the presently observed jets; e.g. shredded planet

About magnetic fields• Magnetic fields are the “drive belt” not the “motor”: energy source is

accretion (diff. rotation and turbulence); can drive dynamos

• Dynamically important fields are thus consequences of binary interactions as the interactions supply the free energy

• Dynamo theory is subtle because it involves turbulence. “Applied” dynamo theory is analagous to Shakura-Sunyaev accretion disk theory; field estimates derived with transport coefficients parameterized

• MHD angular momentum transport in accretion disks leads to both local and large scale contributions. Jets are evidence for large scale angular mom. transf. relative strength and role of large scale fields in disks is open question

• PN/PPN context: Magnetic bomb vs. magnetic tower vs. magneto-centrifugal jets

• influence and the self-consistent origin are usually considered separately

LPoynt ~ 12BpBφvk (rin )rin

2 12Macvk

2 (rin ) = 6 ×1036 erg/s

(BpBφ )*1/2 ~ 7000 G

Mac

10−4M / yr⎛⎝⎜

⎞⎠⎟

1/2M*

M

⎛⎝⎜

⎞⎠⎟

1/2R*

R

⎛⎝⎜

⎞⎠⎟

−5/2

Shaping from imposed toroidal field, wind velocity, rotation, and (Garcia-Segura et al. 99,05):

Mag. Field collimates flow

Radiation wind driving reduced at equator by rotational density enhancement

Isolated AGB Dynamos (Pascoli 97, Blackman et al. 01): Transient Dynamo: No re-seeding of differential rotation

Quasi Steady Dynamo: -Re-seeding of differential rotation via convection + rotation -possibly sufficient if Poynting flux is trapped

Common Envelope Dynamo (Regos & Tout ,1995 Nordhaus, et al 07)

-Initial rotation profile is only source of differential rotation energy.-resulting transient dynamo is insufficient without strong ang. mom mixing (Soker & Zoabi 2002; Nordhaus, et al 07)

In-spiral of the secondary spins up envelope induces transient dynamos: sufficient to unbind envelope, produce ansae, cycles: Poynting flux drains in <100yr

Accretion Disk Dynamo or Flux Advection (secondary or primary) (e.g. Blackman, et al. 01; Nordhaus & Blackman 07, B 12)

-accretion provides free energy needed to amplify field sufficient to power long term outflows (PPN PN)-Note: possible strong field WD outcome (Nordhaus et al.10)

B-field Origin and Outflow Scenarios

NOT SO GOOD

BETTER

BEST, Maybe Combined with

CED

r = rc

r = r - Lc

r = r + Lc 1

! - layer

" - layer

Con

vect

ion

Zon

e

z^

x^

y^

Convective Zone

Shear Zone

Core

(e.g. Blackman et al 06; Nordhaus, et al 07)

Geometry for Transient Interface Dynamo

Meridional Slice of Progenitor Geometry

Convection + Differential rotation induced by the in-spiraling secondary can power a dynamo.

L1 1011 cm

L 4.6 ×1010 cm

Common Envelope Dynamo:

¡¡

M

T

Brown Dwarf Companion0.02M! Thermally Dominated Model

The insets represent the time evolution from 0 to 0.2 yrs.

Heat resulting from turbulent dissipation is responsible for unbinding the envelope after ~ 20 yrs.

The resulting outflow is expected to be spherical or quasi-spherical.

Common Envelope Dynamo:

¡¡

M

0.05M! Brown Dwarf CompanionMagnetically Powered Outflow

Poynting flux is responsible for unbinding the envelope within ~ 50 yrs.

Insets represent time evolution from 0 to 0.2 yrs.

Resultant poloidal outflow is expected, with weaker equatorial outflow (e.g. Matt et al. 2006)

Magnetic Burst from wind-up of Dipole Field (Matt et al. 2006 ) Initially:

va / vesc = 1vrot / vesc = 0.1χ ≡ (vrotva )

1/2 / vesc = 0.3

Magnetic Tower

e.g. Lynden-Bell 96,03

Huarte-Espinosa et al. 2012

Magneto-centrifugal Launch

Blandford Payne (82); Pellitier & Pudritz (92)

-low density magnetically dominated “cavity” grows with time -toroidal magnetic pressure pushes flow ahead of the rising tower-Field || to flow on axis, changing to toroidal field at edge. -in principle, could remain magnetically dominated out to observable scales, so flow is still slaved to magnetic field there-field primarily toroidal at large r

-magnetically dominated at base (and thus resembles magnetic tower there) base but becomes flow dominated at observable scales -flow is centrifugally launched onto field lines from disk, then pressure from toroidal magnetic pressure accelerates flow -field at large scales is slaved to flow on observed scales: e.g. if jet has nonuniform outflow speed, field can be stretched along jet direction, so field could be less dominantly toroidal than for magnetic tower

Connections to Strong Field WD

• 10% of Isolated WD, have 1-1000 MG; 90% have B < 0.1 MG

• within 20pc 109 known WD; 21 with nondegenearte companion but SDSS detected 149 HMFWD none with such detached companion (Silvestri et al. 2007+): HMFWD do not occur with detached companions .

• in strong field AM Her type polars, they, have companion linked by field, and accretion, but no disk presently.

• evidence overall is that binary interactions play a role in HMFWD (either in polar type binaries or isolated) the isolated HMFWD likely had past binary interaction with companion gone, and the Low MFWD in detached binary have not yet had such interaction. (More data on age and spin needed for HMFWD.)

• CE dynamo alone has trouble accounting for isolated HMFWD: hard to get field onto WD

• But isolated HMFWD might be the result of accretion disk dynamo, after tidal shredding of secondary: super-Eddington radiatively dominated advection dominated accretion disk with dynamo (Nordhaus et al. 2010)

• note also Tout at al. (2008)

(Nordhaus et al. 2010)

(BpBφ )*

1/2 ~ 300 MGMac

10−4M / yr⎛⎝⎜

⎞⎠⎟

1/2M*

M

⎛⎝⎜

⎞⎠⎟

1/2R*

109 cm⎛⎝⎜

⎞⎠⎟−5/2

Measured AGB Magnetic Fields

AGB stars: e.g. Masers (Herpin et al 06)

OH shell of NML Cygni (Etoka

Diamond 04) Miras (Kemball Diamond 97,09)

Syncrohtron: (Perez-Sanchez et al.

2013); 7000AU, mG field, collimates jet in post AGB star IRAS 15445−5449

Leone et al. 2014: No > kG fields on CSPN; but 655G on NGC4361 generally large error bars Herschel, VLT, spectropolarimetry

distinguish launch vs. propagation regions when interpreting observed fields

=

Evolved AGB star W43: (“water fountain” source) precessing jet in a post-AGB star; B-fields 85mG at ~500 AU; strong enough to collimate (Vlemmings et al. 06,10)

Vlemmings (2010)

Conclusions/Open Issues• Bujarrabal PPN outflows are more kinematically demanding than those of PN but

both point to need for magnetically mediated jets; need better statistics on what fraction of PPN are of Bujarrabal (high power) type

• Constraining accretion rates from theory, simuations, and observations is powerful for constraining paradigms; moreover, jet magnetic fields are likely slaved to the free energy from differential rotation

• while CE dynamos and magnetic bombs might explain jet ansae, accretion is needed for longer term steady MHD jets

• Are PN are evolved stages of PPN? can all kinematic constraints be fit with single accretion scenario?

• Likely multiple classes of engines leading to subdivided classes of PN and PPN (accretion on secondary vs. accretion on primary, etc.)

• absence of observed binary does not preculde influence of a previously shredded companion (e.g. planets, perhaps making PN appropriately named)

• Clues from isolated high magnetic field white dwarfs?

• accretion, B-field origin, ouflow generation, asymptotic propagation are still often studied separately; still have to make choices in studying pieces of the full problem

END