aa r X i v : . [ a s t r o - ph ] S e p Astrophysical Masers and their EnvironmentsProceedings IAU Symposium No. 242, 2007J.M. Chapman & W.A. Baan, eds. c (cid:13) Arp 220 - IC 4553/4: understanding thesystem and diagnosing the ISM
Willem A. Baan
ASTRON, Oude Hoogeveensedijk 4, 7991PD Dwingeloo, The Netherlandsemail: [email protected]
Abstract.
Arp 220 is a nearby system in final stages of galaxy merger with powerful ongoingstar-formation at and surrounding the two nuclei. Arp 220 was detected in HI absorption andOH Megamaser emission and later recognized as the nearest ultra-luminous infrared galaxy alsoshowing powerful molecular and X-ray emissions. In this paper we review the available radioand mm-wave observational data of Arp 220 in order to obtain an integrated picture of the denseinterstellar medium that forms the location of the powerful star-formation at the two nuclei.
Keywords. galaxies: ISM – galaxies: nuclei – ISM: molecules – ISM: masers – galaxies: star-burst, – galaxies: OH Megamaser
1. OH MM and ULIRGs
The galaxy IC 4335, later referred to as Arp 220, was recognized early as a mergingand strongly interacting system with a blue nucleus by Arp (1966) and Nilson (1973).The detection of strong HI absorption against the radio continuum of the nuclear star-burst (Mirabel 1982) led to a search for corresponding OH absorption, which resulted inthe detection of powerful OH maser emission later classified as OH MegaMaser emission(OH MM) (Baan et al. et al. et al. et al. et al. et al. et al. et al. et al. et al.
2. Dynamics of Arp 220 and where is what?
The two nuclei of Arp 220 are separated by only 0.97” (365 pc) as determined fromtheir radio positions. They appear not yet severely deformed at this advanced stage ofinteraction and have a relatively small velocity difference. Therefore, the two mass centersare still sufficiently far away from each other and need to be one behind the other. Onlythe high-density tracer emission lines accurately trace the nuclear ISM of the two nuclei0 nderstanding Arp 220 Figure 1.
Orbital dynamics of Arp 220. The north-western nucleus is located in front of thesouth-eastern nucleus. The systemic velocities of the two equal-mass nuclei are indicated andtheir direction of motion in an orbital plane inclined by 40 o . The orbital velocity ( V orb ) is 278km s − . The combined mass of the nuclei is based on an observed stellar and gas mass. Thedirections of the observed velocity gradients of the CO and OH molecular gas in the nuclei areindicated. The MERLIN 6cm continuum map is from Baan, Cohen & Kl¨ockner (2007). having systemic velocities of 5365 km s − for the western nucleus and 5682 km s − forthe eastern nucleus (see section 6).The systemic velocities of the nuclear regions obtained from the high-density traceremission lines may be used for dynamic modelling of the system. Similar models werepresented earlier using H CO (1-0) and CO (2-1) molecular data (Baan & Haschick 1995;Downes & Solomon 1998; Scoville et al. CO emission data gives slightly different pictureof the dynamics of the system (Baan, Cohen & Kl¨ockner 2007). The eastern nucleus liesbehind the western nucleus, which would result in the observed larger column densitiesin front of the eastern nucleus (Iwasawa et al. o (Scoville et al. M ⊙ (Sakamoto et al. et al. µ m) onboardHubble ST confirm the presence of high column densities in the system (Scoville et al. cm − H). The NIR emission actuallypeaks north of each of the nuclei; there are areas where virtually no radiation can escape,even at 2.2 µ m. Reversing the symmetry of the model could place the eastern nucleus infront of the western nucleus, but would not explain the presence of OH emitting gas atthe velocity of the western nucleus in front of the eastern nucleus (see section 4).The hard X-ray (3 - 7 KeV) signature of Arp 220 as observed by Chandra shows twoseparate (extended) nuclei and there is an extended nebula at 2 keV (Iwasawa et al. α feature at 6.5 keV. The obscuration ofthe X-rays required N H = 10 − cm − (Note: N H = 10 cm − is enough for keVabsorption).
3. Modelling the starburst
The power generation for the FIR luminosity of 10 . L ⊙ is dominated by star-burst activity for more than 60% in the radio and more than 90% in the near-infrared(Baan & Kl¨ockner 2006; Genzel et al. IR, M ⊙ yr − = 340 M ⊙ yr − and a supernova rate SNR= 0.2 L IR, yr − = 2.8 yr − for Arp 220 (Elson, Fall & Freeman 1989). Lonsdale et al. (2006) have observed a SNR of 4 ± − and find the rate in the western nucleus to bethree times higher than in the eastern nucleus. The mass injection rate for the nucleiof Arp 220 is estimated at 53 M ⊙ yr − . Recent evaluation of the radio data suggests atop-heavy initial mass function (IMF) or a short starburst with a duration of 3 x 10 yrwith a SFR ≈ M ⊙ yr − (Parra et al. et al.
4. Imaging the OH emission lines
Knowledge of the systemic velocities of the high-density gas at the two nuclei (seesection 6) also helps to understand the complex OH groundstate line emission signature.Two pairs of (1665 and 1667 MHz) OH emission lines are offset by 317 km s − and havea 351 km s − separation within the pair. Therefore, triple emission lines are seen: the1667 MHz West line, a convolved 1667 MHz East and 1665 MHz West line, and a 1665MHz East line. However, also emission in the 1667 MHz line at the systemic velocityof West is seen at the East location (see VLA-A and VLBI data by Baan & Haschick1987; Rovilos et al. − ,while the 1720 MHz line extends by 950 km s − , although this second outflow wing maybe less-pronounced (Ghosh & Salter 2007, personal communication).The OH emission in Arp 220 at both nuclei encompasses the high-resolution radiocontinuum structure and the high-density ISM where the star-formation is taking place. nderstanding Arp 220 Figure 2.
The OH emission at the two nuclei of Arp 220. The high-resolution EVN map of the1667 MHz emission from Rovilos et al. (2003) has been overlayed with the map of supernovae inthe two nuclei from Parra et al. (2007). Continuum contours indicate the location of the radionuclei. The superposition is a best effort. (0.1” = 35 pc)
At MERLIN resolution (0.17” = 61 pc) the 1667 MHz OH emission at the eastern nucleusis centrally peaked and has an extent of 170 pc along the velocity gradient (Rovilos et al. et al. et al. et al. et al. et al. et al. , these proceedings). The edge-on torus may serve for further amplification.Successful modelling of observed OH emission structures has been done for Mrk 273 andIIIZw 35 (Parra et al. , these proceedings; Kl¨ockner 2004; Parra et al.
5. Clumpy low-gain amplification for OH
The standard model for masers of low-gain amplification of radio continuum by fore-ground molecular material has been suggested in order to explain the extremely powerful Baanemissions in OH MM (Baan 1985, 1989; Henkel & Wilson 1990). The spatial distributionof the background radio continuum and the availability of a pumping agent will allow aforeground screen to selectively amplify the distributed continuum. Assuming a uniformscreen gives a first order estimate of the global amplification. A more realistic non-uniform and clumpy nuclear ISM will give non-uniform amplification in high-resolutiondata. Emission is found in all OH ground state transitions in Arp 220, while the known 5GHz and 6 GHz lines are all in absorption in both OH MMs and absorbers (Henkel et al. et al. CO, H O, and CH OH.Evaluation of the high-resolution data suggests a gain factor of 4 to 5 relative to thebackground ( τ = 1.4 - 1.6) for the extended components that account for the one thirdof the 1667 MHz emission (Rovilos et al. τ = 3.8 - 4.8). Clump superpositions or edge-on (shocked) shells need toincrease the optical depth by a factor up to three.Pumping scenarios for the OH molecules (Henkel et al. et al. et al. µ m FIR lines puts requirements on the FIR-SEDand requires a threshold temperature for the dust of at least 50 K, whereas Arp 220has 61 K. Typically clouds of 1 pc should have an opacity of one and a linewidth of 10km s − for line-overlap conditions. Pumping can be achieved across a density range ofn(H ) = 10 to few 10 cm − . For T ex ( OH ) approximately constant, the line ratios are afunction of optical depth and give an optical depth of order 2.0, which is consistent withobservations. Simulations of the cloud superposition and the amplification in a torus-likestructure has been done by Kl¨ockner (2004) and Parra et al. (2005).
6. The high-density ISM
Diagnostics of the heavily obscured nuclear ISM can also be done by interpreting molec-ular tracers to determine densities, temperatures and the molecular chemistry. Accurateestimates of global (integrated) properties can be obtained with LVG (large velocitygradient) radiation transfer studies using line emission along the energy ladder of themolecule. In addition, the chemical evolution of molecular species combined with theirradiative transfer can be obtained by modelling X-ray and Photon (UV) dominated re-gions (XDR and PDR) (Meijerink, Spaans & Israel 2006, 2007, also Spaans et al. andLoenen et al. , these proceedings).Many molecular transitions have been detected in Arp 220 and in similar nearby(U)LIRGs like NGC 6240, and Mrk 231. Single-dish spectra obtained with the IRAM30 m and SEST 15 m telescopes of some of the tracer lines in Arp 220 are presented inFigure 3 (Baan et al. − for the western and eastern nuclei (Baan et al. + (Aalto et al. et al. et al. et al. et al. nderstanding Arp 220 Figure 3.
Spectra of emission lines of high-density tracermolecules in Arp 220. From top to bottom the emissionline spectra of CO (2-1), HCO + (1-0), CS (3-2), CN (1-0),and CN (2-1). Data taken from Baan et al. (2007). Figure 4.
Simulations of the amplified OH emission fromoutflows driven by supernova remnants. The three curvesresult from different covering factors for the spherical lay-ers of molecular gas relative to the radio continuum gen-erated by the SNR. The chosen terminal velocity is asobserved in Arp 220. an extrapolation of well-resolved Galactic phenomena to integrated and unresolved nu-clear environments. High-density components may be used to discriminate between PDRand XDR excitation conditions (Baan et al. + with HCN and HNC discriminate between the gas density due todifferent critical densities of the molecules. The intensity ratio of a high-density tracermolecule with that of a lower density tracer, such as CO (1-0) that is more extended andis less affected by the star-formation activity, shows that the high-density medium variesstrongly during the lifetime of the activity (Baan et al. et al. these proceedings and Loenen, Baan & Spaans 2007).LVG simulations suggest a two-phase ISM for the combined nuclei of Arp 220 with adiffuse component with n(H ) = 2 x 10 cm − and T k = 40 - 60 K and dense componentswith n(H ) = 10 − cm − and 50 - 70 K (Greve et al. M ⊙ .The considerably higher pressure in the warm high-density medium as compared withthe low-density medium will gradually cause disruption of the medium.
7. Signature of nuclear outflows
Arp 220 is one of the OH MM galaxies showing blueshifted outflow components fortheir 1667 MHz emission lines (Baan, Haschick & Henkel 1989; Baan 2007). The ob-served outflow velocities vary with L F IR and are likely associated with the shockedshells surrounding supernova remnants. A continuing high-intensity starburst as foundin such ULIRGs would eventually result in merging the localized outflows into a large-scale blowout or nuclear superwind Heckman, Armus & Miley (1990). The OH outflowemission results from a superposition shocked SNR shells propagating into the dense ISM Baanthat amplify the embedded radio continuum. A simple simulation of a superposition ofsuperposed spherical outflowing shells amplifying a spherical continuum structure pro-vides the correct shape and signature of the outflow (see Figure 4). The observed highvelocities of the outflows suggest that J-shock chemistry is responsible for the molecularabundances. Molecules are destroyed and reassembled in J-shocks for V >
50 km s − while for V >
300 km s − this reassembly process happens at a slower rate. Using anenergy conserving scenario for the mechanical energy (see Elson, Fall & Freeman 1989),we find that: N SN R kpc V n ISM ≈
23 L
IR, , which balances the total energy of theoutflows from N SN SNRs with outflow velocity V (in units 100 km s − ) and with radiusR (in kpc) and the injected energy L IR (in units of 10 ⊙ ). Using L IR = 1.6 x 10 L ⊙ for Arp 220, an observed number of N SN ≈ − , and a typical SNR radius = 1 pc, we find that the mean ambient densityof he pre-shock material n ISM = 1.5 x 10 cm − , which is in an acceptable range. Theinterpretation of these outflows clearly requires more simulations as it provides furtherevidence of the eruptive nature of the star-forming environment.
8. Formaldehyde emission
Formaldehyde emission has been verified in Arp 220 as well as two other OH MM, andan H O MM using the Arecibo and Effelsberg telescopes (Baan, Haschick & Uglesich1993; Araya, Baan & Hofner 2004). At low resolution the formaldehyde in Arp 220 peaksat the western nucleus and has weak emission towards the eastern nucleus and the con-necting structure (Baan & Haschick 1995). Recent MERLIN data shows an elongatedemission region of 80 pc at the western nucleus (resolution 0.04” = 15 pc) centered at5325 km s − , which is 40 km s − below the systemic velocity of the western nucleus(Baan, Cohen & Kl¨ockner 2007). The velocity gradient of H CO aligns with the weakgradient of the OH emission.The diagnostics of this emission is still limited by the understanding of the pump-ing mechanism. Because the formaldehyde emission at lower resolution mimics the 2.2 µ m emission, the IR radiation field could play a role in the pumping. A first pump-ing model proposed for the excitation of the extragalactic formaldehyde was based onpumping by the non-thermal radio continuum, which gives small (few percent) opticaldepths for densities of 10 − cm − (Baan, G¨usten & Haschick 1986). While this modelcould operate in extragalactic environments, it poses a problem for small scale Galac-tic environments. Evaluation of collisional and other pumping schemes for Galactic andextragalactic environments shows that optical depths at the relevant density range arestill relatively small and may not yet satisfy maser requirements under Galactic con-ditions (see Araya, Baan & Hofner 2004; Baan, Haschick & Uglesich 1993; Araya et al.
9. Conclusions
Radio observational tools contribute significantly to the understanding of the nuclearmedium of Arp 220. Arp 220 is dynamically advanced such that the two nuclei have devel-oped a common gas/stellar structure in advance of the merger. The current dynamicalpicture can be determined on the basis of the observed velocities of the two nuclearISM components. Placing the eastern nucleus, with the highest obscuration, behind the nderstanding Arp 220 et al. to10 cm − . The low-density medium with 10 − cm − is traced by the extended diffuseand clumpy CO components. The CO (2-1) transition traces the upper end of this rangein nuclear regions enveloping the highest density medium. The high-density constituentshave densities in the range 10 to 10 cm − depending on the critical density of themolecules. LVG simulations suggest that the temperature of the low-density tracer emis-sions ranges from 40-60 K, while those of the higher density components suggest 50-70 K.This overpressure of the high-density components results from radiative and mechanicalfeedback of the star-formation process. The integrated dust temperature from the FIRSED is about T d = 40 K, while the local conditions in Arp 220 with T d >
60 K providethe required pumping photons for the OH emission.Both nuclear regions harbor a powerful and rapidly evolving starburst at the pointdestroying its own ISM. The star-formation rate is on the order of 340 M ⊙ yr − andoccurs predominantly in the western nucleus. The ISM in these nuclei is still mostly intactbut there is evidence of radiative and mechanical feedback at work. The overpressureof the high-density molecular constituents and the observed OH outflows will in timeredistribute and homogenize the nuclear ISM. Radiative feedback also results in a PDRdominated molecular chemistry. The effects of feedback on the molecular medium andthe star-formation process would lead to a top-heavy (truncated) IMF and lead to theearly termination of the star-formation. There is no evidence for the presence of an AGNin the nuclear regions and any associated XDR conditions. In addition, the observedemission due to X-ray binaries is too weak for the level of star-formation activity. Thenuclei of Arp 220 have become the prototype of an extreme nuclear environment withdominant PDR conditions.The FIR-pumped OH emission extends beyond the nuclear activity regions and showsdiffuse structures as well as clumpy and shell-like components with densities of 10 − cm − (see Fig. 2 Rovilos et al. CO can be collisionallyor radiatively pumped for densities of 10 − cm − . The SNR-driven shells seen in the OHemission are not yet well-constrained but ambient densities of 10 cm − are consistent.The diffuse and the high-brightness OH components cover regions with sizes of 1 to 100pc, and result from a distributed clumpy molecular medium contained in two edge-onthick tori with a radius of about 60 pc that surround the nuclear ISM at each nucleus. Thehighest amplifying gains result from cloud superpositions and edge-on shells (also seen asoutflows) in the nuclear ISM. The OH and H CO emission structures trace intermediate-density components surrounding the highest density ISM structures co-located with thesites of star-formation. The special conditions leading to OH and H CO MM activitymay occur during a significant part of the duration of the starburst.The high-resolution VLBI data on the OH and H CO emissions and the data for high-density molecular tracers provide a complementary picture of the nuclear ISM and theongoing SB activity in Arp 220. While the emission of the high-density tracer moleculespresents the integrated properties of the nuclear ISM and is mostly unresolved at thetwo nuclei, the high-resolution data on the OH and H CO emissions provides sufficientstructural detail of the medium surrounding the sites of star formation. The interpretationof integrated properties depends on our ability to extrapolate from the understandingof Galactic environments to the extreme case of unresolved nuclear environments instarburst nuclei. Baan
References
Aalto, S., Spaans, M., Wiedner, M.C. & H¨uttemeister, S. 2007,
A&A , 464, 193Araya, E., Baan, W.A. & Hofner, P. 2004,
ApJS , 154, 541Araya, E., et al. 2007,
ApJ , submittedArp, H. 1966,
ApJS , 14, 1Baan, W.A. 1985,
Nature , 315, 26Baan, W.A. 1989,
ApJ , 338, 804Baan, W.A. 2007,
New Astron. Rev. , 51, 149Baan, W.A., Cohen, R.J. & Kl¨ockner, H.-R. 2007, in preparationBaan, W.A. & Haschick, A. D. 1987,
ApJ , 318, 139Baan, W.A. & Haschick, A.D. 1995,
ApJ , 454, 745Baan, W.A., Haschick, A.D. & Henkel, C. 1989,
ApJ , 346, 680Baan, W.A., Haschick, A.D. & Uglesich, R. 1993,
ApJ , 415, 140Baan, W.A., Henkel, C., Loenen, A.F., Baudry, A. & Wiklind, T. 2007,
A&A , in the pressBaan, W.A., G¨usten, R. & Haschick, A.D. 1986,Baan, W. A. & Kl¨ockner, H.-R. 2006,
A&A , 449, 559Baan, W.A., Wood, P. A. D. & Haschick, A.D. 1982,
ApJL , 260, L49Downes, D. & Solomon, P. 1998,
ApJ , 507, 615Elson, R.A.W., Fall, S.M. & Freeman, K.C. 1989,
ApJ , 336, 734Genzel, R. et al. 1998,
ApJ , 498, 579Greve,T., Papadopoulos, P., Gao, Y. & Radford, S. 2007,
ApJ , submittedGraci´a-Carpio, J., Garc´ıa-Burillo, S., Planesas, P. & Colina, L. 2006,
ApJ , 640, L135Heckman, T., Armus, L. & Miley, G. 1990,
ApJS , 74, 833Henkel, C. & Wilson, T.L. 1990,
A&A , 229, 431Henkel, C., G¨usten, R. & Baan, W. A. 1987,
A&A , 185, 14Impellizzeri, C.M.V., Roy, A.L., Henkel, C. et al.
Astron. Nachr. , 326, 544Iwasawa, K., Sanders, D. B., Evans, A.S., et al.
MNRAS , 357, 565Kl¨ockner, H.-R. 2004, PhD thesis, University of GroningenLockett, P. & Elitzur, M. 2007, in preparationLoenen, A.F., Baan, W.A. & Spaans, M. 2007, in preparationLonsdale, C.J., Diamond, P.J., Thrall, H., Smith, H.E. & Lonsdale, C.J. 2006,
ApJ , 647, 185Meijerink, R., Spaans, M. & Israel, F. 2006,
ApJL , 650, L103Meijerink, R., Spaans, M. & Israel, F. 2007,
A&A , 461, 793Mirabel, F.I. 1982,
ApJ , 260, 75Nilson, P. 1973,
Nova Acta Regiae Soc. Sci. Upsaliensis , Ser. V, 0Parra, R., Conway, J.E., Elitzur, M. & Pihlstr¨om, Y.M. 2005,
A&A , 443, 383Parra, R., Conway, J.E., Diamond, P.J., et al.
ApJ , 659, 314Randell, J., Field, D., Jones, K.N., Yates, J.A. & Gray, M.D. 1995,
A&A , 300, 659Rovilos, E., Diamond, P.J., Lonsdale, C.J., et al.
MNRAS , 342, 373Rovilos, E., Diamond, P. J., Lonsdale, C.J. et al.
MNRAS , 359, 827Sakamoto, K., Scoville, N. Z., Yun, M.S., et al.
ApJ , 514, 68Sanders, D.B., Soifer, B.T., Elias, J.H. et al.
ApJ , 325, 74Scoville, N.Z., Evans, A.S., Thompson, R., et al.
ApJS , 119, 991Scoville, N.Z., Yun, M.S., & Bryant, P.M. 1997,
ApJ , 484, 702Solomon, P., Downes, D., Radford, S.J.E. & Barrett, J.W. 1997,
ApJ , 478, 144Wiedner, M.C., Wilson, C.D., Harrison, A., et al.
ApJ , 581, 229Wiedner, M., Aalto, S., Volgenau, N.H. et al.et al.