Multiwavelength radio observations of the compact starburst in Arp - PowerPoint PPT Presentation

Multiwavelength radio observations of the compact starburst in Arp 220 Rodrigo Parra & John Conway Onsala Space Observatory O B D S M E Y R V R A A T L O A R S I U N O M T A E L CHALMERS K O N K I S S G K A Ö H 8 th EVN symposium Toru´ n, Poland, 26–29 September 2006 1

Colaborators The Arp 220 team Philip J. Diamond Jodrell Bank Hannah Thrall Jodrell Bank Colin J. Lonsdale Haystack, MIT Carol J. Lonsdale IPAC, CALTECH Harding E. Smith CASS, UCSD 2

Outline • 18 cm Global VLBI continuum observations of Arp 220 Previous results on the compact sources in Arp 220 • 6 cm Eb–Wb–Ar @ 1 Gbit s − 1 (Delay–Rate map) • Simultaneous 13, 6 and 3.6 cm VLBA • Spectra – Model fits – Speculation 3

Arp 220 in context ULIRG and OHMM 1 kpc 5" HST NICMOS HST HCS/HRT Scoville et al. (1998) Wilson et al. (2006) ⇒ 10000 km s − 1 = 28 µ asec year − 1 ) • D = 77 Mpc (1 asec = 373 pc = • L 8 − 1000 µ m = 1 . 3 × 10 12 L ⊙ ⇒ SFR = 50 − 100 M ⊙ year − 1 • ∼ 10 10 M ⊙ in the central kiloparsec ⇒ n H 2 ≈ 2 × 10 4 cm − 3 4

18 cm Global VLBI “A starburst revealed” Smith et al. (1998), Rovilos et al. (2003), Lonsdale et al. (2006) 18 cm continuum image of the nuclear region of Arp 220 observed in November 2002. The resolution is ∼ 4 mas and the map noise is ∼ 8 µ Jy beam − 1 . (Taken from EVN newsletter number 5, May 2003). See also Lonsdale et al. 2006 for a more recent and deeper image. 5

18 cm Global VLBI Main conclusions • From rate of appearence of new sources, the estimated supernova rate is ν SN = 4 ± 2 year − 1 • The SFR derived from ν SN is consistent with the SFR derived from the FIR luminosity if all detected sources are as luminous as SN1986J • 18 cm lightcurves too stable to be SNe. • Not detected at shorter wavelengths 6

6 cm Eb–Wb–Ar @ 1 Gbit s − 1 COLA Sample 11.8 11.6 seconds of DEC 11.4 11.2 100 pc 11.0 57.30 57.28 57.26 57.24 57.22 seconds of RA Overall view of the Arp 220 nuclear region. Contours are 6 cm “single baseline snapshot image” made using the Eb–Ar data. The map noise is 40 µ Jy and the contours are separated by 500 µ Jy . The resolution is ∼ 100 mas but sensitive only to features of < 1 mas (Parra et al. In prep.). Circles indicate the positions of the 49 compact sources catalogued by Lonsdale et al. (2006). 7

13, 6 and 3.6 cm VLBA Color radio image 240 220 DEC offset [mas] 200 180 160 140 −300 −350 −400 −450 −500 RA offset [mas] RGB (SCX bands respectively) composite image of the central region of the western nucleus (Parra et al. Submitted). The axes are in mas from the refer- ence position indicated in the previous figure. 18 sources detected. One source possibly resolved at 3.6 cm. 8

18 cm variability classification We classified the detections in 2 categories according to their 18 cm flux variation between 2003 November 9 (Lonsdale et al. 2006) and 2005 March 6 (Thrall et al. In prep.)  Single epoch (S)     Variable Non variable (L) Rising (R)   Ambiguous (A)   (10/18) (8/18) First seen after ∼ 2002 Known since ∼ 1994 Stable 18 cm light curves 9

Source model Core collapse supernova Before SN phase SNR phase ≈ 10 51 erg m ∗ ν α ˙ m v wind ρ ( r ) ∝ r − 2 τ CSM P ISM τ Foreground τ Foreground τ Foreground t < 0 0 < t < 5 − 10 years t > 5 − 10 years m × v wind /P ISM ) 0 . 5 m/v wind ) 1 − 4 r s ∝ ( m ejecta /n ISM ) 1 / 3 r b ∝ ( ˙ L synch ∝ ( ˙ 10

Expected radio spectrum An expanding SN/SNR shell can be modelled as a free–free obscured synchrotron source. The expected radio flux at a fixed time is given by S ν ∝ ν α exp − τ 18 ν − 2 . 1 � � (1) where τ 18 = τ CSM ( t ) + τ Foreground is the total free–free opacity at 18 cm along the line of sight and α is the synchrotron spectral index. 11

Variable at 18 cm Young SNe? 1000 1000 1000 W55 W56 E24 500 500 500 (S) (S) (S) 200 200 200 µ Jy µ Jy µ Jy 100 100 100 50 50 50 20 20 20 1000 1000 W11 W12 1 2 4 6 8 10 1 2 4 6 8 10 1 2 4 6 8 10 500 500 GHz GHz GHz (R) (R) 200 200 µ Jy µ Jy Radio spectra for the 10 100 100 50 50 detections classified as 20 20 variable at 18 cm. Filled 1000 1000 W25 W34 1 2 4 6 8 10 500 500 GHz (R) (R) circles are from our 200 200 µ Jy µ Jy 100 100 multiwavelength obser- 50 50 vations from Jan 2006. 20 20 1000 1000 1000 W15 1 2 4 6 8 10 E10 1 2 4 6 8 10 E14 Open circles are 18 cm 500 500 500 GHz GHz (A) (A) (A) 200 200 200 from Mar 2006 and Di- µ Jy µ Jy µ Jy 100 100 100 50 50 50 amonds are from Nov 20 20 20 2003. Errorbars are ± σ 1 2 4 6 8 10 1 2 4 6 8 10 1 2 4 6 8 10 GHz GHz GHz and upper limits are 5 σ 12

Non variable 18 cm SNRs? 1000 1000 1000 500 500 500 200 200 200 µ Jy µ Jy µ Jy 100 100 100 50 50 50 20 20 20 W10 W17 W42 1000 1000 1000 1 2 4 6 8 10 1 2 4 6 8 10 1 2 4 6 8 10 500 500 500 GHz GHz GHz 200 200 200 µ Jy µ Jy µ Jy 100 100 100 50 50 50 20 20 20 W8 W18 W39 1000 1000 1 2 4 6 8 10 1 2 4 6 8 10 1 2 4 6 8 10 500 500 GHz GHz GHz 200 200 µ Jy µ Jy 100 100 50 50 20 20 W30 W33 1 2 4 6 8 10 1 2 4 6 8 10 GHz GHz Radio spectra for the 8 detections classified as non variable at 18 cm. These sources present stable 18 cm light curves since Smith et al.(1998). 13

3.6 cm flux vs τ 18 Decouple intrinsic synchrotron strength from foreground absorption. 1500 1978K 1986J 2 3.6 cm flux versus 18 cm opac- 3 1 ity. Stars indicate the long lived 1000 S 3.6 [ µ Jy ] stable sources (class L). Red 3 lines indicate the loci traced by 4 well studied Type II RSNe from 4 Weiler et al. (2002) with their 500 5 fluxes scaled to the distance of 5 6 1979C 1 6 Arp 220. 7 7 2 8 3 9 10 × 1980K 10 4 5 6 7 8 This is evidence that younger 1 0 sources have higher τ 18 which 0.1 1 10 τ 18 is consistent with RSN models. 14

Σ − D relation for SNRs Huang et al. (1994) Flux–D relation for SNRs. The stars indicate sources of class 1000 86J L. Arrows indicate upper limits W42 in size (unresolved). From the 79C correlation we estimate diame- 100 ters between 0.2 and 0.4 pc. S 3.6 [ µ Jy ] 41.95+575 44.01+596 Crosses are SNRs in M82. Red 43.31+592 lines are the tracks traced by 10 well studied Type II SNe from 80K Weiler et al. (2002). These 1.0 tracks are expected to join the correlation at a diameter set by the equilibrium between the 0.1 stellar wind pressure and the 0.1 1 10 D [ pc ] ISM pressure. 15

Σ − D relation for SNRs Derived ISM properties m = 10 − 4 M ⊙ year − 1 , Using typical values of ˙ v wind = 10 km s − 1 and m ejecta = 5 M ⊙ , combined with the typical radius of ∼ 0.1 pc for the remnant size predicted by the Σ − D we obtain P ISM > 4 × 10 7 K cm − 3 (Dopita et al. 2005, estimate P ISM ≈ 10 7 by modelling the pan-spectral energy distribution) and n ISM > 3 × 10 4 cm − 3 (Scoville et al. 1997, estimate n ISM = 1 . 4 × 10 4 cm − 3 using CO(1-0) observations.) 16

Resolved source W42 Hypernova remnant? 3.6 cm uniformly weighted map of source W42. Contours are − 25, 25, 50, 75 and 95% of the peak flux. 4 The deconvolved diameter is 2.3 mas (0.86 pc). W42 ⇒ v exp < 40000 km s − 1 = 2 The integrated flux at 3.6 cm is 574 µ Jy and the mas 0 fitted spectral index is α = − 0 . 24 ⇒ W total (min) = 2 × 10 50 ergs = −2 As usual, we need more data to confirm this result. We have new data from May 2006 plus new high −4 4 2 0 −2 −4 frequency Global VLBI scheduled for November. mas 17

Main Conclusions For more, see Parra et al. submitted • First detection of compact radio sources in Arp 220 at λ < 18 cm. Previous non-detections at 6 cm due to large angular distance to calibrator used. • We find evidence that younger sources have higher foreground absorption consistent with SN models. • The bright 18 cm sources from Smith et al. (1998) in addition to show little time variability at 18 cm have significantly lower foreground absorption consistent with SNR models. • The previous two points imply that the compact sources in Arp 220 are a mixed population of supernovae and supernova remnants. • One source possibly resolved at 3.6 cm ( D = 0 . 86 pc ). Minimum energy calculations suggest initial kinetic energy in the order of 10 51 ergs. 18

THE END

M82 1 kpc 100 pc 18 cm EVN HST ACS/WFC Pedlar et al. (1999) • D = 3 . 7 Mpc (1 asec = 18 pc) 20