Young Supernovae in Arp 299 and Related Starbursts

1
Young Supernovae in Arp 299 and Related Starbursts

• Jim Ulvestad
• NRAO
• 29 Jan. 2009

2

• Arp 299 program in collaboration with Susan Neff
  (NASA Goddard) and Stacy Teng (Univ. Md.)
• Thanks also to Kelsey Johnson for discussions of
  super star clusers

• Outline
• Context
• Starburst radio emission
• Merger galaxies
• Initial radio imaging of Arp 299
• Multi-epoch VLBAGBT monitoring of Arp 299
• What does the future hold?

3
Some Global Questions

• How do galaxies merge and grow?
• Relation of starbursts and AGNs?
• How does the radio/far-IR relation originate in
  starbursts?
• How do supernovae in nearby merger galaxies
  compare to supernovae in normal galaxies, and
  can we extrapolate their properties to more
  distant and powerful mergers?

4
What Can Radio Emission Reveal about
Extragalactic Starbursts?

• Optical radiation from youngest star-formation
  regions is hidden by dust
• Radio emission due to
• Complexes of dense H II regions energized by
  Super Star Clusters
• Estimate ionizing flux?massive star population
• Individual supernova remnants or young SNe
• Estimate supernova rate evolution
• Overlapping supernova remnants

5
Nearby Starbursts

• M82 (Kronberg et al. 1985 Muxlow et al. 1994)

• NGC 253 (Ulvestad Antonucci 1997)

8 mJy thermal source
25 pc
6
Results from M82, NGC 253

• Little or no source variability
• Steep spectrum sources resolve into SNRs
• Flat-spectrum sources typically H II complexes
  energized by hot stars
• N(UV)/s 1051 (D/2.5 Mpc)2 (S5 GHz/1 mJy)
• 1049 photons/s 1 O7 star
• Strongest NGC 253 thermal source is 8 mJy
• 750 O7-equivalent stars in a few parsecs
• About 20,000 solar masses or less

7
NGC 5253 (4 Mpc)
7mm VLA, 25 pc (2 arcsec) square (Turner Beck
2004)
Linear Resolution 2-4 pc NLyc ? 7 ?1052
s-1 ? Super star clusters
From KJ
8
Global Lessons from Radio SSCs

• Most recent star formation regions are bright in
  mid-IR and radio
• Radio SSC diameters are a few parsecs
• Tens to thousands of O7-equivalent stars
• Recombination linewidths and sizes indicate some
  SSCs are bound, depending on stellar mass
  function (e.g., Turner et al. 2003)
• Otherwise, overpressure would cause the SSCs
  expansion in 106 yr or less

9
Nearest MergerThe Antennae

• WFPC2, with CO overlay (Whitmore et al. 1999
  Wilson et al. 2000)

• VLA 5 GHz image (Neff Ulvestad 2000)
• Needs EVLA sensitivity and resolution

5 mJy ?30,000 O7-equivalent stars
10
Arp 220

• Luminous IR galaxy at 77 Mpc distance
• Two merging galaxy nuclei separated by 370 pc
• Star formation rates 100 Msun/yr
• Extensive radio supernova and SNR imaging by
  Smith et al. (1998), Rovilos et al. (2003, 2005),
  Lonsdale et al. (2006), Parra et al. (2007)

11
Arp 220 VLBI Images
Parra et al. 2007
12
Arp 220 Results

• 50 compact radio sources detected, initially at
  1.7 GHz, with later observations detecting 18 at
  multiple frequencies
• At least half of the 18 are likely to be Type IIn
  SNe interacting with their own stellar wind
• Others are probably young SNRs
• Four new sources over 12 months at 1.7 GHz

13
Arp 299 History
Tidal tails in Arp 299 Stars blue HI
contours

• At least one previous interaction
• 700 million years ago
• HI and stellar tidal tails,
  150 kpc in extent
• Near beginning of current interaction
• Two disks still clearly identifiable
• Nuclear separation 3.5 kpc
• Disks interacting and distorted
• Burst of star formation 6-8 million years ago (at
  the beginning of current pass)
• Should be seeing supernovae

30 kpc
J. Hibbard
HST WFPC2 (Alonso-Herrero et al. 2000)
14
Arp 299 Radio Emission

• No radio emission at optical supernova positions
• Four Strong Radio Peaks
• A and B galaxy nuclei
• C and C overlap region
• Alonso-Herrero et al. IR/opt. (2000)
• Assume starbursts Gaussian in time, 5 Myr wide,
  peak 5 Myr after start
• A 7 Myr post-peak, 0.6 SN/yr
• 700 million solar masses in young stars
• 140 solar masses/yr in star formation
• B1 5 Myr post-peak, 0.1 SN/yr
• C C 4 Myr post-peak, 0.05 SN/yr

SN2005U
Neff, Ulvestad, Teng (2004)
Red VLA 6cm Blue HST
250nm Green HST 814nm
Arp 299
15
Arp 299 Source A Properties

• Brightest radio source, 70 mJy (1.4 x 1022 W/Hz),
  steep spectrum
• Unresolved with VLA (lt 40 pc resolution)
• If supernova-powered, expect
• 1 SN about every 2 years
• Clumped, small radio sources (clusters or
  associations of clusters, confined by dense gas)
• Need higher resolution, very sensitive radio
  observations!

???
16
VLBA Observations (Neff, Ulvestad, Teng 2004)

Note 1 mas 0.2 pc at Arp 299 distance of 41
Mpc
17
Arp 299 Inside Source A (2004)

• A nest of four young SNe, within 100 pc

• and
• A young supernova(?), only 2 pc from one of the
  other sources
• Tracing super-star clusters?

3 pc
Neff, Ulvestad, Teng 2004
April 2002 Feb. 2003 13cm
3.6cm
18
Supernova Factory?
Arp 299 HST / WFPC2 814nm
3 kpc

• Detected
• Supernova factory in merging galaxy pair
• Is A0 or A1 an AGN??
• A1 Optically thin, relatively stable flux
• A0 Optically thick
• Compact X-ray emission LX(A)4 x 1039 ergs/s
  (Zezas et al. 2003)
• ?L?(5 GHz, A1)1037 ergs/s
• Radio/X-ray2.5 x 10-3
• Location of X-rays?

April 2002 Feb. 2003 2.3 GHz
8.4 GHz
VLBA
3 pc
19
Why Would an AGN Matter?

• Early Universe dominated by building up galaxies
  and merging them.
• Now, internal galaxy evolution is becoming
  important, but a few mergers still are happening
• Distant mergers form many hundreds of solar
  masses into stars each year, often in galaxies
  that also have massive black holes.
• Do we have a young supernova at 2 pc from an AGN?
  (Or dual AGNs???)

12 pc
6 kpc
20
Radio/X-ray For Arp 299
Radio/X-ray ?L? (5 GHz)/L(2-10 keV) Terashima
Wilson 2003
21
Inside Source A Component Spectra
2003 May

• Components appear to be young SNe / SNR in dense
  environments
• A0 resembles young (1yr) embedded SN
• Inverted spectrum turns over 15 GHz
• No detection at 2.3 GHz
• SN ejecta has not yet broken through remnants of
  former stellar wind.
• Other components appear to be young (gt 10 yr)
  SNR
• Flat / slightly steepening spectra
• SN ejecta expanding beyond
• former stellar winds

Young SN Young SNR
SN1986J
22
VLBAGBT Monitoring

23
Inside Source A, 1st Nucleus
4 epochs combined, 2.3 GHz
24
Core of Source A
8.4 GHz
2.3 GHz
10 pc
25
Source A0 Flux Decay
Young supernova
26
Inside Source B1, 2nd Nucleus
4 epochs combined, 2.3 GHz
27
Source Spectra in Arp 299A/B

• Most compact sources are detected only at 2.3 GHz
  or 8.4 GHz
• Typical errors in a are about 0.2 (1 bin)

lt
gt
28
Another SN Appearance in B
8.4 GHz, 2005Jan02
8.4 GHz, 2005Jul12

• Within 2 mas (0.4 pc) of older, steady 2.3 GHz
  source
• From VLA archive data, 8.4 GHz flux density of B1
  increased from 7.1 to 13.6 mJy between 2004.84
  and 2006.29
• Peak power of gt1.2 x 1021 W/Hz is 2000 times
  Cas A

29
Radio Luminosity Functions
30
Arp 299 Compared to Arp 220

• Supernova rate in Arp 220 is much higher, and two
  nuclei are quite different
• Arp 299 L.F. goes deeper in power since the
  merger is twice as close
• What happens at Cas A power?

B
A
E
W
31
Individual and Integrated Properties

• SN powers and evolution imply Type II SNe
• SNR luminosity function is rather steep, with
  dN(P)/d(lnP)P-2.6 in Nucleus A
• Integrating down to Cas A power, we expect 900
  SNRs more powerful than Cas A
• Integrated power of these is 20 mJy, roughly 15
  of the total flux density
• Conclude tentatively that most of the radio
  emission is in older (background) SNRs

32
Current Summary for Arp 299

• 30 compact radio sources seen in two primary
  merger nuclei of Arp 299, above 10 x Cas A
• Four objects within 10 pc diameter may outline
  the primary super star cluster in Nucleus A
• Very wide range of radio spectra
• No SNRs or radio supernovae seen in other two
  nuclei (C and C)
• Two new supernovae in 2.5 yr, one in A with SN
  rate 0.6/yr, one in B with SN rate 0.1/yr
• Both occurred 2 pc from existing SNRs
• Powers and evolution consistent with Type II or
  IIn
• Expect to see 500-1000 SNRs above Cas A power

33
VLBI Observations of Antennae

• Nearest merger, at 20 Mpc
• 30 hours of VLBAGBT at 2.3 GHz, aiming to reach
  10x Cas A power
• Compact radio emission is predominantly outside
  the nuclei
• Many VLA sources may be single SNRs, detectable
  with VLBI

34
EVLA/SKA Possibilities

• Need much better sensitivity to get near Cas A
  power, or to observe many more modest starbursts
• EVLA is much more sensitive, but 70 mas
  resolution at 32 GHz is inadequate
• Need 10 times the resolution?lots of collecting
  area on few hundred km baselines
• However, the much greater sensitivity would lead
  to more confusion in densest clusters
• SKA needs lots of long(ish) baselines, and
  frequencies above a few GHz

35
VLBI Possibilities

• 4 Gbps data rate will increase sensitivity by 4x
  in 2011, so a 30-hr VLBAGBT observation will
  reach 4x Cas A power
• Detection of 200 individual SNRs in Arp 299
• True luminosity function, and locations of super
  star clusters in the nuclei, will be achievable
• Trade off larger EVN dishes vs. frequency
  flexibility
• Yet more sensitivity could give luminosity
  functions in individual clusters, to compare SNR
  evolution in different environments

36
Extra Slides

• Following

37
Inside Source A Component Lightcurves
Normal Type II SN (on same time axis as below)
We observe A0 on falling part of light curve at
8 GHz
Confined Type II SN
SN1986 J

• A0 behaves like an expanding, Type II SN.
• Flux falling at 8 GHz
• Still no detection at 2.3 GHz
• Other sourcelets behave like young, confined SNR.
• Do not vary over 1 year

Bietenholz et al. 2002
38
Super Star Clusters in Radio/mid-IR
Beck, Turner, Gorjian 2001
Left mid-IR image Right 2-cm radio contours
overlaid (from Kobulnicky Johnson 1999)

• Can account for most of the hosts radio and
  mid-infrared emission
• Poor correspondence with optical peaks

39
The Problem with the Antennae

• At 20 Mpc, 1 arcsec100 pc. Need resolution
  better than 10 pc, but then brightness
  sensitivity is a problem.
• Need EVLA-1 and EVLA2 for sensitivity and
  resolution

6 cm ? 2cm
40
Inside Source A Component Spectra

• Components appear to be young SNe / SNR in dense
  environments
• A0 resembles young (1yr) embedded SN
• Inverted spectrum turns over 15 GHz
• No detection at 2.3 GHz
• SN ejecta has not yet broken through remnants of
  former stellar wind.
• Other components appear to be young (gt 10 yr)
  SNR
• Flat / slightly steepening spectra
• SN ejecta expanding beyond
• former stellar winds

Young SN Young SNR
SN1986J
41
Observational strategyIf we want to understand
cluster formation, its not a bad idea to observe
them while they are forming.
Problem Once clusters are visible to HST, they
have already emerged from their birth material
From Kelsey Johnson
42
Typical Radio Supernovae
In a Dense Environment

• Majority are Type II
• M gt 8 Msun progenitor
• Massive pre-explosion wind
• Light Curves
• Type II have slow radio turn-on (20- 100d) /
  turn off (few years)
• Turn-on first at shorter wavelengths
• Spectral evolution with time
• Turn-on (flux rising) spectrum peaks at shorter
  wavelengths
• Long wavelengths self-absorbed, thermal emission
  dominant
• Later, radio spectrum is steep
• Optically thin, non-thermal synchrotron emission
  dominates

M 20-30 Msun Wind gt 10-4 Msun / yr Turn-on
few years / turnoff gt 20yrs
Spectrum flattens but does not become non-thermal
(steep)
43
Arp 299 Inside the Starburst
Arp 299 CO (cold gas)

• Extremely dense gas and dust in both nuclei
• Prograde retrograde encounter
• Gas driven into nuclei, not thrown out into
  tails
• High star formation rate (SFR)
• Nuclei forming 30 (East) and 15 (West)
  Msun/ year, in massive stars
• SFR led Weedman and collaborators to coin the
  term starburst (1983).
• Four known SNe since 1990E
• Very high extinction in nuclei
• Optical observations cannot penetrate
• Require radio observations to see into the galaxy
  centers

Aalto
Arp 299 Ha (Ionized Hydrogen)
E
W
Hibbard
44
Arp 299 Science Summary
HST WFPC2

• Detected
• Supernova factory in merging galaxy pair
• Imaged, with the only telescope that can
• See through the dust
• AND
• Provide adequate resolution to separate
  individual supernovae
• Is A1 an AGN??
• Optically thin spectrum and relatively stable
  flux.
• LX(A)1.3 x 1040 ergs/s (Zezas et al. 2003)
• ?L?(5 GHz, A1)1037 ergs/s

10 LY
1 HST/PC2 pixel
2.3 GHz
8.4 GHz
45
NGC 3256

• Colorsoptical DSS
• ContoursHI (Hibbard et al.)

• RedULXs (Lira et al. 2002)
• Radio (Neff, Ulvestad, Campion 2003)

46
NGC 3256 The Nuclei

• 8.4 GHz and 15 GHz VLA images of two nuclei (with
  X-ray positions indicated)

• 1.5-10 keV contours (Chandra archive)

47
Radio/X-ray Ratio as a Diagnostic
Radio/X-ray ?L? (5 GHz)/L(2-10 keV) Terashima
Wilson 2003
48
NGC 3256 Northern Nucleus

• BUT, 300 km/s velocity gradient across 40 pc
• 108 solar masses enclosed

• Strong diffuse component to X-rays