Unwrapping the Mystery of Common Envelope Episodes Morgan Macleod

Unwrapping the Mystery of Common Envelope Episodes

Morgan MacLeod

Einstein Fellows Symposium Oct 18, 2016 Common envelope interactions transform binary systems

Example: formation of merging pairs of neutron stars

Pair of massive stars (>8x sun’s mass) Common envelope interactions transform binary systems

Example: formation of merging pairs of neutron stars

Pair of massive stars (>8x sun’s mass) much closer pair of neutron stars ? draws the binary closer together Common envelope interactions transform binary systems

Example: formation of merging pairs of neutron stars

Pair of massive stars (>8x sun’s mass) much closer pair of neutron stars ? draws the binary closer together

gravitational wave inspiral Orbital transformation is key in formation of compact binaries Common envelope interactions transform binary systems Example: formation of merging pairs of neutron stars Pair of massive stars (>8x sun’s mass) much closer pair of neutron stars Common Envelope Phase

Drag on surrounding Orbit stabilizes as gas tightens the orbit envelope is ejected

Evolution to contact NS

Companion Common envelope interactions transform binary systems envelope is ejected and orbit stabilizes

common envelope outcomes: tightened binary or merger

Drag on surrounding gas tightens the orbit Evolution to contact envelope is retained and binary merges Many open questions…

Challenging problem for “kitchen sink” numerical approach because of the huge range of relevant spatial and temporal scales. ?? Many open questions…

Challenging problem for “kitchen sink” numerical approach because of the huge range of relevant spatial and temporal scales.

?? … and inﬁnite combinations of binary properties! Many open questions…

Challenging problem for “kitchen sink” numerical approach because of the huge range of relevant spatial and temporal scales.

?? … and inﬁnite combinations of binary properties!

But, there are some general questions:

• What are the initial conditions of CE episodes?

• How is energy/momentum transfer between embedded object and the gas mediated?

• What is the timescale of envelope ejection?

• How do initial binary parameters map to outcomes? Two strategies for progress

• Gas dynamics of common envelope: goal: isolate and understand aspects of the physics of CE ﬂows

• Catching common envelope in action: goal: model-to-data comparison to directly constrain long-uncertain theoretical models of CE processes Gas dynamics of common envelope

In the frame of the orbiting star:

Hoyle & Lyttleton (1939), Flow is gravitationally focussed Bondi & Hoyle (1944) toward the embedded star

2GM R a ⇠ v2 1

(Edgar 2004)

…interacts with a “column” of gas with 2 Area = ⇡Ra Gas dynamics of common envelope 405

THE EFFECT OF INTERSTELLAR MATTER ON CLIMATIC VARIATION

BY F. HOYLE AND R. A. LYTTLETON

The effect of interstellar matter on climaticReceived variation 19 Apri409l 1939 by the sun's gravitational attraction, and the action of collisions in this con- densation can be shown to give the sun an effective capture radius much larger than its ordinary radius. 1. INTRODUCTION There is direc(a) Calculationt astronomica of the capture radiusl evidenc of the sune for the existence of diffuse clouds of matter Imagine the cloud to be streaming past the sun, from right to left in the figure, anind leinterstellat the velocity of anr yspace element o. f itAn relativy esectio to the sunn whe ofn atht greae tMilk distanceys Way containing a large number of bstare v. Consides usuallr the pary t showof the clous dregion that if undeflectes in dwhic by theh su nn woulo stard passs appear, and the extent of these within a distance o~ or less of its centre. It is clear that collisions will occur to thpatchee left of thse suisn becausofteen th elarg attractioe comparen of the latter dwil witl produch eth tweo opposinaveragg e apparent distance between the streamstarss o f themselveparticles and thes effec (seet of ,suc foh collisionr examples is to destro, yRussell the angula,r Dugan, and Stewart (2), p. 820). The existence of the so-called cosmical cloud in interstellar space, sharing in the general motion of the galaxy, is now well established, and observational investiga- tion shows that the obscuration referred to above occurs also on a galactic scale. Thus the diffuse obscuring clouds appear as irregularitieHow sthe in thsune genera gravitationallyl cosmical captures cloud. The dimensions of such regions are comparablinterstellare with the distancegas ands betwee how thisn might the stars, and may be very much greater. In some affectinstance solars the systempresence evolutionof such clouds is revealed by their illumination by a star or stars lying in, or near them, so that the matter then can be directly observed. In shape the clouds are very irregular; some appear like long dark lanes, while other tracts are devoid of Fig. 1. any particular form. momentum of the particles about the sun. If after collision the surviving radial component of the velocity is insufficient to enable the particles to escape, such particleSincs will eeventuall the yexistenc be swept intoe th eo sunf suc. Supposeh cloud, for examples appear, that ans to be general in the galaxy it is elemenof importanct of volume of thee clouto d consideA whose initiar l thangulae effectr momentus mtha is vo~t losecouls d be produced if a star passed this momentum through its constituent particles suffering collisions at C; then ththrouge effectiveh radiu onse a ocafn them be calculate. Thd suce frequench that the velocity oy fradiall s u cyh a toccurrence C is s for a particular star would less than the escape velocity at this distance. The element describes a hyperbola whosclearle equationy depen, with thed usua upol notationn th, mae ydistributio be written n of the clouds in space, and the intervals between these event- = s1 +woul e cos 6.d accordingly be irregular. But it is to be observed r Thate directiooncen parallethatl ttho thee initiaintervall asymptots e woulcorrespondd si nto general be of the order of the periods of time 7 8 occurring in galactiecos0+c problemsl = 0, , that is, of the order of 10 or 10 years, the average period of revolution of a star in the galaxy26- bein2 g about 2-5 x 108 years. The density of an obscuring cloud and the velocity that a star would have relative to it are known as far as orders of magnitude are concerned from astronomical considerations, and it is shown in the sequel that these clouds may have a considerable effect upon a star's radiation during the time of passage of the star through the cloud. The importance to terrestrial climate of such an effect upon the sun is at once evident, and it is to this aspect of the process that the present paper is directed, though it would seem that encounters between stars and the diffuse clouds may also have some bearing on questions of a more general astronomical nature. If any appreciable change in the sun's radiative power Gas dynamics of common envelope

Express typical ﬂow properties in terms of the

dimensionless scales of Hoyle-Lyttleton theory

26-

C s

o t

a hyperbol the

;

y n

a lose

a t h

a vo~

suc radiall

tha i

0 y

describe correspond

= t radia

e l g

escape collision 6.

velocit s

o g

writte

e example

e t

elemen

b momentu

ecos0+

th e

fo y

t survivin

, asymptot

e l

Th Then simulation results can be

sufferin

1 th

s tha

r particle =

initia angula e

suc

Suppose

distance

particle

s collisio

t notation

t r initia

? l l

e sun

thi e

enabl

afte

usua

th applied to any stellar pair:

calculate t o

whos

I t

e paralle

int

h constituen

d t

sun

velocit e

wit

, h

clou

swep

Fig

directio s

e insufficien

escap f

Th y r y e

abou o

throug

e radiu

equation g th

m e

e n o

velocit angula

tha

e whos s volum

eventuall

particle

effectiv

opposin e

les

o occu

pas e f

t l 2 y

, momentu

d th

wil

s o

th f

wil

o e thi

woul

figure

distance Length: elemen destro t n e

particle

t th

su R =2GM/v

r produc n

componen a grea

s collision

momentu

wil y

lef

9 large

h tha

con

t r d s

whe

latte

collision

thi muc

sun

h clea

right n su

i 40

f m

i s the

suc

o th

fro

o of I

radiu undeflecte

t f

o e

t e

sun

collision

radius

f variation

effec

tha centre

e s

captur t d attractio

relativ

t e

e n

th f

th pas d

o o

clou e

g t capture s

an actio

climatic

s e th

effectiv les

the

f r

o th o

becaus t

elemen d

y streamin

e an

par particle

, b e e f

e f

o matter o

y Time: f d e

r R /v

t a

Calculation

distanc

giv )

clou

(a e

lef

stream t

velocit attraction e

Conside

radius

withi

interstellar v. y

show le

b e

Imagin b

an n

ordinar

gravitationa

ca s

s effect n

The sun'

tha e

densatio th

y b Mach number: = v/c M s Gas dynamics of common envelope

Express typical ﬂow properties in terms of the

dimensionless scales of Hoyle-Lyttleton theory

26-

C s

o t

a hyperbol the

;

y n

a lose

a t h

a vo~

suc radiall

tha i

0 y

describe correspond

= t radia

e l g

escape collision 6.

velocit s

o g

writte

e example

e t

elemen

b momentu

ecos0+

th e

fo y

t survivin

, asymptot

e l

Th Then simulation results can be

sufferin

1 th

s tha

r particle =

initia angula e

suc

Suppose

distance

particle

s collisio

t notation

t r initia

? l l

e sun

thi e

enabl

afte

usua

th applied to any stellar pair:

calculate t o

whos

I t

e paralle

int

h constituen

d t

sun

velocit e

wit

, h

clou

swep

Fig

directio s

e insufficien

escap f

Th y r y e

abou o

throug

e radiu

equation g th

m e

e n o

velocit angula

tha

e whos s volum

eventuall

particle

effectiv

opposin e

les

o occu

pas e f

t l 2 y

, momentu

d th

wil

s o

th f

wil

o e thi

woul

figure

distance Length: elemen destro t n e

particle

t th

su R =2GM/v

r produc n

componen a grea

s collision

momentu

wil y

lef

9 large

h tha

con

t r d s

whe

latte

collision

thi muc

sun

h clea

right n su

i 40

f m

i s the

suc

o th

fro

o of I

radiu undeflecte

t f

o e

t e

sun

collision

radius

f variation

effec

tha centre

e s

captur t d attractio

relativ

t e

e n

th f

th pas d

o o

clou e

g t capture s

an actio

climatic

s e th

effectiv les

the

f r

o th o

becaus t

elemen d

y streamin

e an

par particle

, b e e f

e f

o matter o

y Time: f d e

r R /v

t a

Calculation

distanc

giv )

clou

(a e

lef

stream t

velocit attraction e

Conside

radius

withi

interstellar v. y

show le

b e

Imagin b

an n

ordinar

gravitationa

ca s

s effect n

The sun'

tha e

densatio th

y b Mach number: = v/c M s

Density gradient: ✏⇢ = Ra/H⇢ Flow parameters for common envelope inspiral

Relations between ﬂow parameters:

The stellar envelope is in HSE with the same gravity that drives the orbital motion!

~g Flow parameters for common envelope inspiral

Relations between ﬂow parameters:

The stellar envelope is in HSE with the same gravity that drives the orbital motion!

Envelope structure vs Density gradient gas compressibility

relative velocity: fraction of Keplerian

Mass ratio 1 2q 2 4 s ✏ = f ⇢ (1 + q)2 M k ✓ ◆

Mach Num Common Envelope Wind Tunnel

3D (AMR) calculation in FLASH Cartesian geometry: “unwrapped” star

stellar limb ~g inject ﬂow with polytropic (HSE) proﬁle

speciﬁed by (✏ , ,q) PM + absorbing central BC ⇢ M

stellar core

-y boundary enforces HSE Common Envelope Wind Tunnel

3D (AMR) calculation in FLASH Cartesian geometry: “unwrapped” star

stellar limb ~g inject ﬂow with polytropic (HSE) proﬁle

speciﬁed by (✏ , ,q) PM + absorbing central BC ⇢ M

stellar core

-y boundary enforces HSE Common Envelope Wind Tunnel = s =5/3 Two strategies for progress

• Gas dynamics of common envelope: goal: isolate and understand aspects of the physics of CE ﬂows

• Catching common envelope in action: goal: model-to-data comparison to directly constrain long-uncertain theoretical models of CE processes Catching common envelope in action

Another way to make progress in this ﬁeld is through direct comparison to observable transients: Catching common envelope in action

Another way to make progress in this ﬁeld is through direct comparison to observable transients:

Luminous Red Novae:

V1309 Sco

Bright ﬂare

Orb. period decreases Catching common envelope in action Luminous Red Novae? Catching common envelope in action Luminous Red Novae? Catching common envelope in action Luminous Red Novae? (Kasliwal 2011) Catching common envelope in action Andromeda galaxy… Catching common envelope in action Andromeda galaxy… Catching common envelope in action (M31 LRN2015) M31 LRN 2015

(Kurtenkov+ 2015) Outburst in Andromeda galaxy in Jan 2015 Catching common envelope in action (M31 LRN2015) M31 LRN 2015

The Astrophysical Journal Letters, 805:L18 (6pp), 2015 June 1 Williams et al.

(Williams+ 2015)

Figure 3. F814W HST (negative) image of the location of M31LRN 2015 taken 2004 August 16.5. The 1σ and 3σ errors on the calculated position of the quiescent system are represented by green circles. The red crosses represent the positions of the nearby sources in the HST data.

Employing the conversionsPre-outburst from Sirianni et al. (2005) and sourceregion, the most likelyin source of the Hα emission is the using the distance and reddening toward M31LRN 2015 progenitor system itself, either directly from the progenitor or a in M31 (see Section 2) we estimateHST the absoluteimaging magnitude companion, or from a period of mass loss. of the progenitor candidate to be MV = −1.50± 0.23, MI = − 2.55± 0.13 with (V − I)0 =± 1.05 0.15, consistent with a 5. DISCUSSION red giant. M31LRN 2015 is a luminous, red transient in M31, The Local Group Galaxies Survey (LGGS; Massey characterized by weakening Hα emission on an increasingly et al. 2006) data reveal a nearby source, J004208.06 red continuum, with a number of absorption lines emerging +405501.7, at V 22.33 0.09, (B V ) 0.75 0.17, =± − =± after peak, including Na I D and Ba II, and TiO bands appearing (V − R)=± 0.62 0.10, and (R − I)=± 0.65 0.04. While it fi at later times. Considering the spectroscopic and photometric would rst appear that the progenitor is brighter in the Massey evolution of the outburst we conclude that this object is a LRN et al. (2006) catalog, an inspection of the images reveals this in M31, a conclusion that Kurtenkov et al. (2015b) reached source is in fact a blend of several of the stars visible in independently from their data. We now compare the properties Figure 3, including the two brightest. A source is clearly visible of this system to those of other proposed LRNe and ILRTs. at the position of M31LRN 2015 in the LGGS narrowband Hα The light curves of both LRNe and ILRTs are characterized data (taken 2002 September 11), but nothing is seen at this by slow evolution (compared to most novae) and significant (Kurtenkov+ 2015) position in either the [O III] or [S II] data. To remove any reddening as they fade. The light curves of LRNe V1309 Sco contribution from continuum emission in these narrowband (Mason et al. 2010) and V838 Mon (Munari et al. 2002b) images we subtracted the scaled LGGS R-band image from show the brightness quickly falling in the bluer filters, but both the Hα and [S II] images, and the scaled V-band image plateauing in the red filters. The light curve of V4332 Sagittarii from the [O III] image. The scaling was calculated by (another LRN; Martini et al. 1999) also shows this object to be Outburst in Andromeda galaxy in Jan 2015determining a linear fit between the broadband and narrowband reddening about two weeks after discovery, although there is data using SExtractor (v2.19.5; Bertin & Arnouts 1996) no color information for the earlier periods. The only LRN that photometry of ∼20,000 objects in each image. A clear source shows clear evidence of a significant multiple-peak light curve was seen in the continuum-subtracted Hα data, indicating a structure is V838 Mon. ILRTs are more luminous at peak than strong Hα excess in the spectrum of the pre-outburst M31LRN the LRNe discussed above and they also show different 2015. Azimlu et al. (2011), who also used the LGGS data, list photometric evolution. The first ILRT discovered, M85 it as an H II region (#1527 in their paper), with an estimated OT2006-1, shows a long plateau phase (∼70 days), with luminosity of 6 × 1034 erg s−1. If indeed this is a star-forming relatively little color evolution until later times (Kulkarni H II region, the luminosity implies a star formation rate of et al. 2007). ILRT PTF 10fqs showed a plateau phase −7 −1 ∼5 × 10 Me yr (using the conversion from Kennicutt 1998). (∼40 days), before rapidly reddening (Kasliwal et al. 2011), A SN remnant origin for this emission seems unlikely, due to and NGC 300 OT2008 also shows no major optical color the lack of [S II] emission and the small size, which implies an evolution until later times (Bond et al. 2009). Our observations age of less than a few hundred years. If this is not an H II show that M31LRN 2015 clearly resembles LRNe, rather than

4 Catching common envelope in action (M31 LRN2015) M31 LRN 2015

The Astrophysical Journal Letters, 805:L18 (6pp), 2015 June 1 Williams et al.

(Williams+ 2015)

Employing the conversionsPre-outburst from Sirianni et al. (2005) and sourceregion, the most likelyin source of the Hα emission is the using the distance and reddening toward M31LRN 2015 progenitor system itself, either directly from the progenitor or a in M31 (see Section 2) we estimateHST the absoluteimaging magnitude companion, or from a period of mass loss. of the progenitor candidate to be MV = −1.50± 0.23, MI = − 2.55± 0.13 with (V − I)0 =± 1.05 0.15, consistent with a 5. DISCUSSION red giant. M31LRN 2015 is a luminous, red transient in M31, The Local Group Galaxies Survey (LGGS; Massey characterized by weakening Hα emission on an increasingly et al. 2006) data reveal a nearby source, J004208.06 red continuum, with a number of absorption lines emerging +405501.7, at V 22.33 0.09, (B V ) 0.75 0.17, =± − =± after peak, including Na I D and Ba II, and TiO bands appearing (V − R)=± 0.62 0.10, and (R − I)=± 0.65 0.04. While it fi at later times. Considering the spectroscopic and photometric would rst appear that the progenitor is brighterBinary in the Massey Systemevolution of the outburst we conclude that this object is a LRN et al. (2006) catalog, an inspection of the images reveals this in M31, a conclusion that Kurtenkov et al. (2015b) reached source is in fact a blend of several of the stars visible in independently from their data. We now compare the properties Figure 3, including the two brightest. A source is clearly visible of this system to those of other proposed LRNe and ILRTs. at the position of M31LRN 2015 in the LGGS narrowband Hα The light curves of both LRNe and ILRTs are characterized data (taken 2002 September 11), but nothing is seen at this by slow evolution (compared to most novae) and significant (Kurtenkov+ 2015) position in either the [O III] or [S II] data. To remove any reddening as they fade. The light curves of LRNe V1309 Sco contribution from continuum emission in these narrowband (Mason et al. 2010) and V838 Mon (Munari et al. 2002b) images we subtracted the scaled LGGS R-band image from show the brightness quickly falling in the bluer filters, but both the Hα and [S II] images,Transient and the scaled V-band imageOutburstplateauing in the red filters. The light curve of V4332 Sagittarii from the [O III] image. The scaling was calculated by (another LRN; Martini et al. 1999) also shows this object to be Outburst in Andromeda galaxy in Jan 2015determining a linear fit between the broadband and narrowband reddening about two weeks after discovery, although there is data using SExtractor (v2.19.5; Bertin & Arnouts 1996) no color information for the earlier periods. The only LRN that photometry of ∼20,000 objects in each image. A clear source shows clear evidence of a significant multiple-peak light curve was seen in the continuum-subtracted Hα data, indicating a structure is V838 Mon. ILRTs are more luminous at peak than strong Hα excess in the spectrum of the pre-outburst M31LRN the LRNe discussed above and they also show different 2015. Azimlu et al. (2011), who also used the LGGS data, list photometric evolution. The first ILRT discovered, M85 it as an H II region (#1527 in their paper), with an estimated OT2006-1, shows a long plateau phase (∼70 days), with luminosity of 6 × 1034 erg s−1. If indeed this is a star-forming relatively little color evolution until later times (Kulkarni H II region, the luminosity implies a star formation rate of et al. 2007). ILRT PTF 10fqs showed a plateau phase −7 −1 ∼5 × 10 Me yr (using the conversion from Kennicutt 1998). (∼40 days), before rapidly reddening (Kasliwal et al. 2011), A SN remnant origin for this emission seems unlikely, due to and NGC 300 OT2008 also shows no major optical color the lack of [S II] emission and the small size, which implies an evolution until later times (Bond et al. 2009). Our observations age of less than a few hundred years. If this is not an H II show that M31LRN 2015 clearly resembles LRNe, rather than

4 Taken together: a binary merger outburst Taken together: a binary merger outburst

sub-giant primary star growing to point of interaction

M1 4 5M ; R1 30R ⇡ ⇡ Taken together: a binary merger outburst

sub-giant primary star growing to point of interaction

M1 4 5M ; R1 30R ⇡ ⇡

transient rise time similar to orbital period t t peak ⇠ orb Taken together: a binary merger outburst

sub-giant primary star growing to point of interaction

M1 4 5M ; R1 30R ⇡ ⇡

transient rise time similar to orbital period t t peak ⇠ orb

fast ejecta relative to escape velocity

vej >vesc Taken together: a binary merger outburst

sub-giant primary star merger transient: growing to point of interaction 46 Erad 10 erg M1 4 5M ; R1 30R ⇠ ⇡ ⇡ 1 v 400 km s ej ⇠ t 8d transient rise time similar peak ⇠ to orbital period 2 mej 10 M t t ⇠ peak ⇠ orb small fraction of system mass ejected dynamically fast ejecta relative to escape velocity

vej >vesc What can we learn about common envelope?

• Stars sometimes swallow other stars: this can make new (tighter) binaries, or cause a complete merger.

• We still have NO IDEA how the details work!! What can we learn about common envelope?

• Stars sometimes swallow other stars: this can make new (tighter) binaries, or cause a complete merger.

• We still have NO IDEA how the details work!!

• Catching these events in action constrains the properties of the onset of common envelope. What can we learn about common envelope?

• Stars sometimes swallow other stars: this can make new (tighter) binaries, or cause a complete merger.

• We still have NO IDEA how the details work!!

• Catching these events in action constrains the properties of the onset of common envelope.

• As we start to discover binaries merging through the emission of gravitational waves, it’s extremely important to understand the assembly of these close systems through common envelope phases.