12. Megamasers

Rolf Kudritzki SS 2015 12. Megamasers

Maser sources in the ISM of galaxies

amplification of ISM molecular line radiation at microwave and radio wavelengths classical maser lines

OH at λ = 18cm H2O at λ = 1.35cm

Milky Way:

- masers in immediate neighborhood of young stellar objects or circumstellar envelopes of evolved stars

- isotropic luminosities up to 1 Lsun (integrated over all lines) or up to 0.1 Lsun in one line (example star formation region W49N)

1 Rolf Kudritzki SS 2015 External Galaxies:

- H2O maser in high density molecular gas located in circumnuclear disks within parsecs of AGN cores (Seyfert2, LINER, spirals, ellipticals)

- OH maser within 100 pc of nuclear starburst regions

2 4 6 - isotropic luminosities ~ 10 - 10 Lsun à 10 brighter than MW maser à Megamaser

- archetype megamaser galaxy is NGC 4258 at 7.6 Mpc

- most distant megamaser detected at red shift z = 0.66 (Barvainis & Antonucci, 2005, ApJ 628, L89)

Megamaser are potential distance indicators in the Hubble flow key publications: K.Y. Lo, 2005, Annu.Rev.Astron.Astrophys. 43, 625 Herrnstein, J.R. et al., 1999, Nature 400, 539 Humphreys, E.M.L. et al., 2013, ApJ 775, 13 2 Rolf Kudritzki SS 2015 Maser mechanism most likely process: - collisional excitation from below (symbolic level a) to upper (symbolic) level b

- spontaneous transitions from b downwards overpopulate level u relative to level l - level inversion à line absorption coefficient negative h⌫ gl ⌫ = (⌫)Blu nl nu à 4⇡ gu < 0 line emission exponentially amplified ✓ ◆ by stimulated emission

L 0 0 ⌫ dL I⌫ I⌫ = I⌫ e 0 R note: exponent is positive exponential amplification

3 © Fred Lo, 2011

Rotational Transitions of H2O (An Asymmetric Top molecule)

J K+K–

E(616)/k = 640 K

616-523: λ = 1.35 cm ν = 22.235 GHz

Riess et al., 2011 using PLR of Cepheids extragalactic distance scale New anchor galaxy of Humphreys et al., 2013 accuratedistance to 3% “ MIAPP 2014 Galaxy Maser

”

NGC 4258 NGC7.6 Mpc 4258

Rolf Kudritzki SS 2015 NGC 4258 – observations

1. spectrum at 1.35 cm of the central region

• strong H2O emissions lines

• three radial velocity systems

vrad ~ -470 km/s (systemic velocity) and

vrad ~ -470 ± 1000 km/s (blue and red shifted)

• for each systems many different components

• obvious interpretation: we look almost edge on at a circumnuclear disk which is rotating with ~1000 km/s

6 RolfHumphreys, Kudritzki SS2014, 2015 MIAPP WS NGC 4258 – spectrum at 1.35 cm Strong nuclear water maser at 22 GHz

31& vsys&±&1000&kms MIAPP,&28&May&2014& GBT&(Modjaz&et&al.&2005)&&7 RolfHumphreys, Kudritzki SS2014, 2015 MIAPP WS

2. VLBA Mapimaging structure and spectroscopy of AGN of disks NGC 4258 < 1 pc from SMBH

VLBA: Angular resolution = 200 µas (0.006 pc at ~ 7.0 Mpc) Spectral resolution < 1 kms-1 MIAPP,&28&May&2014& 8 Rolf Kudritzki SS 2015

2. VLBA imaging and spectroscopy

• individual masers resolved and vrad assigned to individual components

• objects in front of AGN have vrad ~ -470 km/s (systemic velocity) the inclination angle relative to AGN is i ~ 82°

• objects at the edges have vrad ~ -470 ± 1000 km/s (blue and red shifted)

• a rotating Keplerian disk provides a perfect fit to each individually observed components of the three radial velocity systems with an accuracy better than 1%

9 RolfHumphreys, Kudritzki SS2014, 2015 MIAPP WS This Dataset: 18 epochs

1&mas&@&7.2&Mpc&!&0.04&pc&

• Insert the time seried

1220&km&s31& 1650&km&s31& 3525&km&s31& 3280&km&s31& Argon, Greenhill, Reid, Moran & Humphreys (2007) VLBA + VLA +EFLS MIAPP,&28&May&2014& 10 RolfHumphreys, Kudritzki SS2014, 2015 MIAPP WS Red-shifted emission

MIAPP,&28&May&2014& 11 Argon et al. (2007) RolfHumphreys, Kudritzki SS2014, 2015 MIAPP WS Systemic Maser Emission

MIAPP,&28&May&2014& Argon&et&al.&(2007)& 12 RolfHumphreys, Kudritzki SS2014, 2015 MIAPP WS Blue-shifted Emission

MIAPP,&28&May&2014& Argon&et&al.&(2007)& 13 Rolf Kudritzki SS 2015 Masers probe sub-pc portion of nuclear disk 0.15&pc&

40&000&Rsch&

Almost perfect Keplerian rotation Miyoshi&et&al.&(1995)& to ~ 1% MIAPP,&28&May&2014& 14 Herrnstein&et&al.&(1999)& Rolf Kudritzki SS 2015

more detailed discussion of model will follow later

15 Rolf Kudritzki SS 2015

3. Time dependence and observed rotational motion

• the individual components of the vrad ~ -470 km/s system in front of the AGN show a common change with time in vrad and position dvrad • average accelleration = a =9 . 3 km/s/yr dt d⇥ • average change of impact parameter position = 31 . 5 mas/yr dt

• the ± 1000 km/s components at the edges do not show such a drift

• this can be perfectly understood as the result of Keplerian rotation

16 Rolf Kudritzki SS 2015 dv rad = a =9.3 vrad drift dt km/s/yr

Herrnstein et al., 1999 17 Rolf Kudritzki SS 2015 d⇥ position drift = 31 . 5 mas/yr dt

18 Rolf Kudritzki SS 2015

4. A simple model thin rotating Keplerian disk - AGN with mass M at center - maser at angle φ, radius R - projected angular distance of maser from center (impact parameter) is Θ - D is distance of galaxy to observer - projected linear distance of maser from the center is D⇥ cos()= D⇥ = Rcos() R 1 M 2 v = 2G rot R Keplerian velocity ✓ ◆ observed radial velocity

vrad = cos()vrot

19 Rolf Kudritzki SS 2015 5. Likelihood to detect masers at φ = 0° and 90°

for strong maser amplification one needs a long light path along which Doppler shifts of photons through differential rotational velocity are dl smaller than a thermal line width v v ⌫ = ⌫ rad ⌫ = ⌫ th 0 c  th 0 c or vrad vth 1km/s 0 lc  dl  ⇡ I = I e 0 ⌫ ⌫ ⌫ The longest coherence lengths R lc are obtained for

φ = 90° (no Doppler shifts) and

φ = 0° (smallest gradient of vrad along light path)

20 Rolf Kudritzki SS 2015 calculation of maser coherence length lc 1 M 2 vrad = cos()vrot v = 2G rot R ✓ ◆ dv dv dcos() rad = cos() rot + v dl dl rot dl dv 1 1 dR rot = v dl 2 rot R dl

dcos() d = sin() dl dl

dv d 1 dR | rad| = v sin()( )+ cos() dl rot dl 2R dl (1) ✓ ◆ 21 Rolf Kudritzki SS 2015 (R + dR)2 =(Rcos())2 +(Rsin()+dl)2

2RdR =2Rsin()dl + dl2 dl dR dl sin()+ ⇡ 2R (2) ✓ ◆

applying the law of sines and noting that the angle

between vrot and vrad is φ sin(d) sin( ⇡ + ) cos() = 2 = dl R + dr R + dr cos() d dl (3) ⇡ R

combining (2) and (3) with (1) we obtain…

22 Rolf Kudritzki SS 2015

dvrad 1 3 dl | | = v cos() sin()+ or dl R rot 2 4R ✓ ◆

2 vrad 3 4R = lccos() 4R sin()+lc (4) v 2 rot ✓ ◆

1 2 for the tangential angle φ = 0° we obtain from (4) vrad l0 =2R and we then re-write (4) as v ✓ rot ◆ 2 1 l v 2 l 1 c +3 rot sin() c =0 l v l cos() ✓ 0 ◆ ✓ rad ◆ 0 which has the solution 1 l 3 v 2 9 v 1 c = rot sin()+ rot sin2()+ l 2 v 4 v cos() (6) 0 ✓ rad ◆ s rad

23 Rolf Kudritzki SS 2015 with vrot ≈ 1000 km/s and Δvrad ≈ 1 km/s

1 2 lc 3 vrot 9 vrot 2 1 = sin()+ sin ()+ l 2 v 4 v cos() 0 ✓ rad ◆ s rad

the solution varies dramatically as soon as sin(φ) ≠ 0, as shown in the figure on the next page.

lc is different from zero only at φ = 0° and 90°. l0

For vrot ≈ 1000 km/s and Δvrad ≈ 1 km/s we have l 0 =0 . 06 R .

For R = 0.2 pc this corresponds to 3.7 1016 cm.

This explains why we see masers mostly at φ = 0° and 90°.

Rolf Kudritzki SS 2015

lc l0

25 Rolf Kudritzki SS 2015 6. Discussion of the model and distance 1 M 2 D⇥ v = 2G cos()= rot R R ✓ ◆ vrad = cos()vrot

in front of AGN with φ ≈ 90° we see the systemic components which have slightly different cos φ according to their different Θ.

D à v = v ⇥ linear behaviour in vrad vs impact parameter figure. rad R rot obviously, the components are at roughly the same radius.

D The average slope b can be measured from this b = vrot R (7) figure

26 RolfHumphreys, Kudritzki SS2014, 2015 MIAPP WS Masers probe sub-pc portion of nuclear disk 0.15&pc&

40&000&Rsch&

Almost perfect Keplerian rotation Miyoshi&et&al.&(1995)& to ~ 1% MIAPP,&28&May&2014& 27 Herrnstein&et&al.&(1999)& Rolf Kudritzki SS 2015

for these components we can also use the observed drifts in vrad and Θ dv d d rad = a = v cos()= v sin() dt rot dt rot dt the sin(φ) factor explains why no drift at φ = 0°. vrot For φ ~ 90°, sin(φ) ~ 1 and with (t)=!t ! = we obtain 2 R vrot a = R (8) R the spatial drift is the change of ⇥ = cos ( ) with time D

d⇥ R d R vrot = sin ( ) = sin ( ) and for φ ~ 90°, sin(φ) ~ 1 dt D dt D R

d⇥ vrot = (9). The measurements (7), (8), (9) already dt D yield R, D, vrot from the systemic components.

28 Rolf Kudritzki SS 2015 In addition, we can use the

vrad shifted components at φ ~ 0° and 180° . We start again with D⇥ v = cos()v cos()= rad rot R 1 M 2 v = 2G rot R ✓ ◆ 1 cos() using = we obtain R D⇥ 1 2 3 M 1 2 this is the fit curve in the vrad = cos() 2G 1 (10) D ⇥ 2 vrad vs impact parameter ✓ ◆ figure. all vrad shifted maser components are on one curve à same cos φ for all. Following section 5. we adopt φ = 0° and 180° .

(10) yields M/D which can be used for distance together with (7), (8), (9).

Rolf Kudritzki SS 2015

fit residuals smaller than 1%!!

Humphreys, 2014, MIAPP WS 30 Rolf Kudritzki SS 2015 first detailed distance determination by

Herrnstein, J.R. et al., 1999, Nature 400, 539 gave D = 7.2 ± 0.5 Mpc

Improved data with longer time base line and refined fitting algorithms assuming a warped confocal elliptical disk and differential precession by

Humphreys, E.M.L. et al., 2013, ApJ 775, 13 resulted in

D = 7.6 ± 0.17 ± 0.15 Mpc

fitting systematic errors

This is a 3% accuracy and the most accurate distance to a galaxy outside the Local Group.

7 The mass of the AGN black hole is (4.0 ± 0.09) 10 Msun . 31 RolfHumphreys, Kudritzki SS2014, 2015 MIAPP WS

Search for Megamasers in the Hubble Flow

2. Megamaser Cosmology Project (MCP)

&NRAO&Key&Project&(PI:&Braatz)#to& determine&H0&by&measuring&geometric& distances&with&an&accuracy&of&~10%&to& ~10&galaxies&in&the&Hubble&Flow&&

Braatz,&Condon,&ConstanUn,&Greene,&Hao,&Henkel,&Impellizzeri,&Kuo,&Lo,&Reid,&et&al.&

MIAPP,&28&May&2014& 32 RolfHumphreys, Kudritzki SS2014, 2015 MIAPP WS

Goal: Direct H0 Measurement

• Water masers in AGN have been detected well into the Hubble Flow – distances out to ~200 Mpc

• Direct estimation of H0 • Systematic uncertainties in individual disk models not likely to be correlated -0.5 – Uncertainty in H0 scales as σH0/H0≈N (σD/D) • Broad distribution on sky needed to reduce impact of large scale flows – Even after correction from models of cosmic velocity field

Braatz & Gugliucci 2008; Reid et al. 2009, 2013 RolfHumphreys, Kudritzki SS2014, 2015 MIAPP WS UGC&3789& Reid&et&al.&(2013)&

MIAPP,&28&May&2014& 35 RolfHumphreys, Kudritzki SS2014, 2015 MIAPP WS

UGC&3789& Reid&et&al.&(2013)&

D=&49.6&±&5.1&Mpc& 31 31& H0&=&68.9&±&7.1&kms Mpc 10%&

37 RolfHumphreys, Kudritzki SS2014, 2015 MIAPP WS NGC 6264 Kuo et al. (2013)

D=144±19&Mpc& 31 31& H0=68±9&kms Mpc &&&&&&&&&&&&&13%& &

MIAPP,&28&May&2014& 38 © Fred Lo, 2011 MCP: Angular Diameter Distance Scale NGC 4258 Mrk 1419 NGC 2273 J0437+2456 IC 2560 ESO 558-G009 NGC 6323 NGC 6264 UGC 3789 NGC 5765b NGC 1194 Largest Structures

Cepheids Distance Ladder

0 Mpc 50 Mpc 100 Mpc 150 Mpc

One geometric method covers all scales out to the Hubble ﬂow and the largest structures

39 RolfHumphreys, Kudritzki SS2014, 2015 MIAPP WS Challenge of Distant Maser Galaxies

NGC&4258&

NGC&6323&

Henkel et al. (2012) MIAPP,&28&May&2014& 40 Rolf Kudritzki SS 2015