Non-planet mechanisms for Non-planet mechanisms for clearing and sculpting discsclearing and sculpting discs

• James Owen (IOA -> CITA)James Owen (IOA -> CITA)

• Barbara Ercolano (IOA -> LMU)Barbara Ercolano (IOA -> LMU)

Hope to convince you:Hope to convince you:• (Xray) photoevaporation now well (Xray) photoevaporation now well

understoodunderstood

• Significant factor in disc evolutionSignificant factor in disc evolution

• (Probably) responsible for ultimate clear out (Probably) responsible for ultimate clear out of protoplanetary discsof protoplanetary discs

• En route, produces structures with properties En route, produces structures with properties overlapping those produced by planetsoverlapping those produced by planets

Are observed structures Are observed structures (gaps, holes) pure dust (gaps, holes) pure dust

phenomena?phenomena?• Sharp edge due to radiation Sharp edge due to radiation

pressure?pressure?

• Sharp edge due to photophoresis Sharp edge due to photophoresis (Krauss et al 2007)? (Krauss et al 2007)?

Probably not….

See Dominik & Dullemond 2011

Hard to suppress small dust production

Do discs clear via viscous Do discs clear via viscous accretion?accretion?

• UV excess => do accrete at rates ~ UV excess => do accrete at rates ~ M_disc/ageM_disc/age

• Far too slow decline at late times….Far too slow decline at late times….

Phenomenological description as due to action of (pseudo) viscosity e.g for

~ R ( =>

get similarity solution:

M = M_in ( 1 + t / t )

(Lynden-Bell & Pringle 1974,

Hartmann et al 1998)

-1.5

in

~ 1/R )

Need extra gas clearing Need extra gas clearing mechanismmechanism

• Planets ?Planets ?

• Photoevaporation ?Photoevaporation ?

• Clearing by MRI driven winds? Clearing by MRI driven winds?

To clear discs within ~ 5 Myr, to make observed hole/gap structures…

Suzuki et al 2010: inner hole forms very early (0.1 A.U. @ 10^5 years,

1 A.U. @ 10^6 years) Halts accretion onto

star from very early times

also….Actually destination of `wind’ unclear (doesn’t attain escape velocity….)

Key facts for understanding Key facts for understanding Xray photoevaporationXray photoevaporation

• Temperature of Xray heated gas set byTemperature of Xray heated gas set by

• In a Parker wind, sonic transition occurs at In a Parker wind, sonic transition occurs at radius whereradius where

= isothermal , spherical

max T is 1-2 x 10^ 4 K

Results of radiation Results of radiation hydrodynamical modeling of hydrodynamical modeling of

Xray photoevaporation (Owen Xray photoevaporation (Owen et al 2010, 2011b)et al 2010, 2011b)

• R => T and (beyond 10 A.U.) fixes R => T and (beyond 10 A.U.) fixes n and hence mass fluxn and hence mass flux

sonic surface where

still holds approximately

T ~ 2000 K

T ~ 10,000 K

Result:Result:

• Proportional to L_xProportional to L_x

• Independent of M_*Independent of M_*

• Doesn’t depend on properties of Doesn’t depend on properties of underlying disc!underlying disc!

Discs with inner holesDiscs with inner holes

As vary R_hole , topology of innermost As vary R_hole , topology of innermost streamline and variation of c_s and u with streamline and variation of c_s and u with

scaled distance along streamline is scaled distance along streamline is invariantinvariant

photoevaporation rate INDEPENDENT of inner

hole size (~ 10^{-8} solar mass/yr for L_X - 10^30 erg/s)

Owen et al 2011b)

Owen et al 2010

What about other radiation What about other radiation sources?sources?

• EUV? Can’t EUV? Can’t penetrate Xray windpenetrate Xray wind

• FUV? Within 100 FUV? Within 100 A.U. only heats A.U. only heats below Xray sonic below Xray sonic surface - doesn’t surface - doesn’t change mass loss change mass loss ratesrates

(But may affect structure of subsonic region:

See Gorti & Hollenbach 2004,2008,2009)

Owen et al

2011 b)

ALSO MAY BE IMPORTANT MASS LOSS MECHANISM AT > 100 A.U.

Combining Combining photoevaporation with photoevaporation with

viscous evolution:viscous evolution:Initial

75% total lifetime 76% total lifetime 77% total lifetime

78 % total lifetime

79 % total lifetime

76 % total lifetime

80%

Etc.Four stage evolution:

Stage I

Stage II

Stage III

I Viscous dominatedI Viscous dominatedII Draining inner holeII Draining inner holeIII Outer disc clearingIII Outer disc clearingIV Thermal sweepingIV Thermal sweeping Accreting, dust free (migration), <10 AU

Empty inner hole, > 10 AU

I

II

Constant L_X

Owen et al 2011a)

NEW

Stage IV: thermal sweepingStage IV: thermal sweeping• Once Xrays penetrate a radial Once Xrays penetrate a radial

distance ~ H into disc, heated gas distance ~ H into disc, heated gas evaporates vertically in `plume evaporates vertically in `plume flow’flow’

• Residual disc clears on ~ dynamical Residual disc clears on ~ dynamical time of inner rim (~ 10s of A.U.)time of inner rim (~ 10s of A.U.)

Sets in when column density at inner rim is

~ 0.5 g/cm^2

Remove few -> 10 Jupiter masses of gas

Thermal sweeping limits Thermal sweeping limits lifetime of non-accreting lifetime of non-accreting

hole stage (stage III)hole stage (stage III)• Fraction of lifetime spent with hole (stage II Fraction of lifetime spent with hole (stage II

+ III) ~ 10%+ III) ~ 10%

• Fraction of lifetime spent with `transparent Fraction of lifetime spent with `transparent accreting’ hole (stage II) ~ 5%accreting’ hole (stage II) ~ 5%

• Fraction of lifetime spent with non-Fraction of lifetime spent with non-accreting hole (stage III) ~ 5% accreting hole (stage III) ~ 5%

Which inner hole sources Which inner hole sources could be due to could be due to

photoevaporation?photoevaporation?• Systems evolve Systems evolve

as inner holes as inner holes draindrain

• Initial Mdot Initial Mdot depends on L_Xdepends on L_X

• Initial radius Initial radius depends on M_*depends on M_*

Cyan = Brown et al 09, blue = Cieza et al 10,

Black open - Ercolano et al 09, Black filled =

Espaillat et al 08,09, red= Kim et al 09, magenta -

Merin et al 10, green = Najita et al 10

These are upper limits

Around half (those in shaded region)

These can’t

Owen et al 2011b)

Evidence for Xray Evidence for Xray photoevaporationphotoevaporation

Both Xray and EUV photoevaporation Both Xray and EUV photoevaporation explain line profiles of NeII 12.8 explain line profiles of NeII 12.8 mm

Only Xray photoevaporation explains low Only Xray photoevaporation explains low velocity (~ 5 km/s) component of OI 6300 velocity (~ 5 km/s) component of OI 6300 in T Tauri stars (cf EUV models: Font et al in T Tauri stars (cf EUV models: Font et al

2004 )2004 )

Xray:Ercolano & Owen 2010

EUV: Alexander 2008

Cf observed profiles for TW Hydra, Pascucci & Sterzik 2009

…….but note lack of blueshifted OI 6300 in TW Hyda ……Pascucci et al 2011

Evidence for Xray Evidence for Xray photoevaporation?photoevaporation?

• High L_X stars lose discs earlier - impliesHigh L_X stars lose discs earlier - impliesWTTs should have higher L_X on average.WTTs should have higher L_X on average.

Well known observational correlation. Well known observational correlation. Usually argued that Xrays Usually argued that Xrays suppressed/absorbed by accretion: suppressed/absorbed by accretion: perhaps instead accretion suppressed perhaps instead accretion suppressed by Xrays…….by Xrays…….

Preibisch et al 2005,

Greogory et al 2007

Population synthesis: Owen et al 2011a

Dependence on Dependence on metallicity:metallicity:

Photoevaporation more efficient at low Z: Photoevaporation more efficient at low Z: lower dust extinction => Xrays heat to higher lower dust extinction => Xrays heat to higher

columncolumnEXPECT SHORTER DISC LIFETIMES AT LOW Z

A DISCRIMINANT FOR DISC CLEARING: PLANETS V. PHOTOEVAPORATION?

Disc lifetime **increases*Disc lifetime **increases*strongly with decreasing Z strongly with decreasing Z if it’s instead set by time if it’s instead set by time

required for planet required for planet formationformation

• A possible observational discriminant?A possible observational discriminant?

Z

Z

-5/2

-11/2

(see Ercolano & Clarke 2009)

• Recent Recent claim of claim of shorter shorter disc disc lifetimes in lifetimes in lower Z lower Z environmeenvironmentnt

…further studies at low Z may hold the key to discriminating between photoevaporation and planet formation

CONCLUSIONSCONCLUSIONSXrays can drive photoevaporative winds of Xrays can drive photoevaporative winds of

10^-8 M_sun/yr at upper end of XLF: like 10^-8 M_sun/yr at upper end of XLF: like EUV winds, these produce a RAPID clearing EUV winds, these produce a RAPID clearing

phase but the Xray wind cuts in at much phase but the Xray wind cuts in at much higher accretion rate.higher accretion rate.

Produce small holes at range of accretion Produce small holes at range of accretion rates but no accreting holes beyond ~ 20 rates but no accreting holes beyond ~ 20 A.U.; expect accreting and non-accreting A.U.; expect accreting and non-accreting

holes to have similar frequency.holes to have similar frequency. Xray photoevaporation ==> line diagnostics ([Ne II] 12.8 m and [OI] 6300 lines )

Xray photoevaporation: shorter disc lifetimes at low Z (opposite to clearing by planet formation)

Explaining higher L_x in general in discless stars (shorter lifetimes)

Predicts shorter disc lifetimes at low Z (opposite to planets!)

Pre-main sequence stars in Pre-main sequence stars in the Xraythe Xray

• Properties well characterisedProperties well characterised

• Correlation with stellar mass (but broad spread at Correlation with stellar mass (but broad spread at given mass)given mass)

• NOT CAUSED BY ACCRETION (some anti-correlation NOT CAUSED BY ACCRETION (some anti-correlation between L_X and strength of disc diagnostics)between L_X and strength of disc diagnostics)

• Previous studies of Xray-disc interaction focused Previous studies of Xray-disc interaction focused on non-thermal ionisation by hard Xrays (=> on non-thermal ionisation by hard Xrays (=> implications for MRI)implications for MRI)

Preibisch et al 2005

See also Albacete-Colombo et al 2007

Neuhauser et al 1995, Flaccomio et al 2003

Synthetic X-EUV spectrum Synthetic X-EUV spectrum of T Tauri starof T Tauri star From Ercolano et al 2009a

EUV = 10^41 /s (cf Alexander et al 2005) L_X= 2 x 10^{30} erg/s

Generated by plasma code of Kashyap

& Drake 2000 from Chandra Xray

emission measures of T Tauri stars by

Maggio et al

2007 and a low T (~ 10^4 K)

component based on emission

measures of RS CVn binaries (Sanz

Forcada et al 2002)

Coronal emission doesn’t Coronal emission doesn’t necessarily reach discnecessarily reach disc

• Effect on spectrum of different Effect on spectrum of different screening columns close to starscreening columns close to star

EUV screened out by N_H =10^20/cm^2 Xrays penetrate to N_H = 10^22/cm^2

Use strong correlation Use strong correlation between between

T in X ray heated gas and T in X ray heated gas and ionisation parameterionisation parameter

= F_X/n

(defined locally)

HOW TO SIMULATE HYDRODYNAMICALLY:

ZEUS 2D SIMULATION: T IN XRAY ZEUS 2D SIMULATION: T IN XRAY HEATED REGION SET BY HEATED REGION SET BY

Owen, Ercolano, Clarke & Alexander 2009

Mass loss rate = 1.5 x 10^{-8} M_/yr

Use MOCASSIN on converged Use MOCASSIN on converged flow structure to check flow structure to check

temperature parametrisationtemperature parametrisation

Gas T set

by dust T

Base of Xray heated region

Xray

heated region

~few 1000 K

WHEN COMBINE EITHER WHEN COMBINE EITHER EUV OR X-EUV WIND EUV OR X-EUV WIND

PROFILES WITH VISCOUS PROFILES WITH VISCOUS EVOLUTIONEVOLUTION

3 stage evolution:3 stage evolution:a)a) Normal viscous evolutionNormal viscous evolution

b)b) Creation of inner holeCreation of inner holec)c) Evaporation of outer discEvaporation of outer disc

Similarities between EUV Similarities between EUV & &

X-EUV modelsX-EUV models

Differences:Differences:

3 stage evolution

Gap opens at few A.U.

Inner hole and outer disc clearing are both fast cf first (viscous) stage

Gap opens at much higher accretion rate in X-EUV case:

Higher accretion rate on star during inner hole draining

10 x higher outer disc mass when gap opens

Observational diagnostics Observational diagnostics of X-EUV photoevaporationof X-EUV photoevaporation

Also agrees with lack of observed Also agrees with lack of observed blueshift in inner hole source GM Aur blueshift in inner hole source GM Aur

(optically thin so line symmetric)(optically thin so line symmetric)

Ercolano & Owen 2010

Are observed inner holes due to Are observed inner holes due to photoevaporation?photoevaporation?

Can explain many inner hole Can explain many inner hole sources but not very large holes sources but not very large holes

with high accretion rate with high accretion rate

Tracks produced by models with

different L_X_

Owen et al in prep.

Herschel and ALMA will improve

statistics

• Correlation between accretion rate Correlation between accretion rate andand

L_X in T Tauri starsL_X in T Tauri stars

Red = no wind

Black = wind for

L_X =2 .10^30

Blue = low L_X

(wind x 0.1)

Broad spread in L_X at given mass

(Preibisch et al 2005):

Note phase of photoevaporation starved

accretion prior to rapid

Decline : see Drake et al 2009

The observational The observational situationsituation

L_X cf average for

stars of that mass

But not always…..But not always…..

Inner hole in image

No clue from SEDIRS 48

(Geers et al 2007)

Evidence for Evidence for photoevaporation:photoevaporation:

• Best evidence to date from line Best evidence to date from line profiles of [Ne II] 12.8 micron profiles of [Ne II] 12.8 micron emission - suggests outflow at ~ 10 emission - suggests outflow at ~ 10 km/s (=> ionised gas)km/s (=> ionised gas)

Alexander 2008

Resulting Xray mass loss Resulting Xray mass loss ratesrates

10 x higher than EUV10 x higher than EUV

Mass loss peaks at around 10-20 A.U.

(cf EUV, peaks here)