Infrared [Fe II] and Dust Emissions from Supernova Remnants
aa r X i v : . [ a s t r o - ph . GA ] A p r Supernova Environmental ImpactsProceedings IAU Symposium No. 296, 2013Alak K. Ray eds. c (cid:13) Infrared [Fe II] and Dust Emissions fromSupernova Remnants
Bon-Chul Koo Department of Physics and Astronomy, Seoul National University,Seoul 151-747, KOREAemail: [email protected]
Abstract.
Supernova remnants (SNRs) are strong thermal emitters of infrared radiation. The mostprominent lines in the near-infrared spectra of SNRs are [Fe II] lines. The [Fe II] lines are fromshocked dense atomic gases, so they trace SNRs in dense environments. After briefly reviewingthe physics of the [Fe II] emission in SNR shocks, I describe the observational results whichshow that there are two groups of SNRs bright in [Fe II] emission: middle-aged SNRs interactingwith molecular clouds and young core-collapse SNRs in dense circumstellar medium. The SNRsbelonging to the former group are also bright in near-infrared H emission, indicating that bothatomic and molecular shocks are pervasive in these SNRs. The SNRs belonging to the lattergroup have relatively small radii in general, implying that most of them are likely the remnantsof SN IIL/b or SN IIn that had strong mass loss before the explosion. I also comment on the“[Fe II]-H reversal” in SNRs and on using the [Fe II]-line luminosity as an indicator of thesupernova (SN) rate in galaxies. In the mid- and far-infrared regimes, thermal dust emission isdominant. The dust in SNRs can be heated either by collisions with gas species in a hot plasmaor by radiation from a shock front. I discuss the characteristics of the infrared morphology ofthe SNRs interacting with molecular clouds and their dust heating processes. Finally, I give abrief summary of the detection of SN dust and crystalline silicate dust in SNRs. Keywords. shock waves, ISM: supernova remnant, infrared: ISM
1. Introduction
Infrared (IR) covers 3 decade logarithmic scales in wavelength, from 1 to 1000 µ m. Thisis the waveband in which we observe emission from dust, forbidden fine-structure linesfrom various metallic atoms and ions, molecular lines, and H-recombination lines. Thesediverse and unique emission features, together with their relatively small extinctions,make the IR band particularly useful for studying various physical and astrophysicalprocesses related to shocks and supernova remnants (SNRs).During the past 10 years, significant progress has been made in the IR study of SNRs asa result of space missions equipped with mid- and far-IR instruments and the developmentof wide-field IR cameras and broadband spectrometers. In this paper, I shall talk abouttwo particular spectral features often found in SNRs: (1) [Fe II] emission lines in thenear-IR (NIR) band, which is the most prominent NIR spectral feature in SNRs, and (2)dust continuum emission in mid- and far-IR spectra. For the [Fe II] lines, I briefly reviewthe basic physics, summarize observational results, and then discuss the characteristicsof [Fe II]-bright SNRs along with some related issues. For the dust emission, as there areother papers on this topic in this volume, I simply present some recent topics that arerelevant to supernovae (SNe) and SNR environments.1 Bon-Chul Koo µ m)0.000.050.100.150.20 i n t en s i t y ------ [ F e II] ------ [ F e II] [ F e II] ------ [ F e II] ------ [ F e II] ------ [ F e II] ------ [ F e II] ------ [ F e II] ------ ------ [ F e II] ------ H ------ [ F e II] ------ H ------ H ------ H I B r γ ------ H /[ F e II] µ m)0.000.050.10 i n t en s i t y ------ H e I ------ H I P a γ ------ [ F e II] ------ ------ [ F e II] ------ [ F e II] ------ ------ H I P a β ------ ------ [ F e II] ------ [ F e II] ------ [ F e II]
Figure 1.
TripleSpec spectrum of the SNR G11.2-0.3, showing numerous [Fe II] lines and otheremission lines (Courtesy of D.-S. Moon). The extinction to the source is large ( A V = 16 mag),so that the observed line intensities can be significantly different from the intrinsic ones.
2. NIR [Fe II] Emission from SNRs
NIR [Fe II] Emission Lines and J Shocks
In the NIR spectra of SNRs, [Fe II] emission lines are usually the most prominent unlessthere is heavy-element-enriched SN ejecta (Fig. 1). This contrasts with photoionized HIIregions where H recombination lines are much stronger; i.e., [Fe II] 1.257 µ m/Pa β ( ∼ × [Fe II] 1.644 µ m/Pa α )=2–8 in SNRs whereas it is 0.013 in Orion (Oliva et al. 1989;Mouri et al. 2000). Such a large difference arises because Fe atoms in photoionzed gasare in higher ionization stages and also probably because the Fe abundance in shockedgas is enhanced by dust destruction. Therefore, [Fe II] emission can be used as a tracerof fast radiative atomic shocks, although strong [Fe II] lines may be observable in sourcesionized by X-rays, e.g., in active galactic nuclei (Mouri et al. 2000).The Fe + ion has four ground terms, each of which has 3–5 closely-spaced levels to forma 16 level system (Pradhan & Nahar 2011). The energy gap between the ground level andthe excited levels is less than 1 . × K, and thus, these levels are easily excited in thepostshock cooling region. The emission lines resulting from the transitions among theselevels appear in the visible to far-infrared bands (Fig. 2 right). In the NIR JHK bands,10–20 [Fe II] lines are visible; these include the two strongest lines at 1.257 and 1.644 µ m.The ratios of these lines provide a very good density diagnostic and an accurate measureof extinction to the emitting region.[Fe II] emission lines in SNRs are mostly emitted from cooling gas behind radiativeatomic shocks. Figure 2 (top left) shows the temperature profile in the postshock coolinglayer of a radiative shock. At N H ∼ × cm − the cooling becomes important and thetemperature abruptly drops to ∼ ,
000 K. Then the temperature remains constant overan extended region, where the heating is maintained by UV radiation generated fromthe hot gas immediately behind the shock front. The corresponding profiles of H nucleiand electron densities are shown in the bottom left frame of Figure 2, together with theFe + fraction profile. Note that, since the ionization potential of the iron atom is 7.90 eV,far-UV photons from the hot shocked gas can penetrate far downstream to maintain theionization state of Fe + where H atoms are primarily neutral. Most of the [Fe II] emission,however, originates from the temperature plateau region where the ionization fraction AU 296. [Fe II] and Dust Emissions from SNRs l og T ( K ) H column density (cm -2 )0.00.20.40.60.81.01.2 ph ys i c a l pa r a m e t e r s µ m)0.00.51.01.5 i n t en s i t y Figure 2. (Top left) Temperature profile as a function of swept-up H-nuclei column density fora 150 km s − shock propagating into an ambient medium of n = 100 cm − and B = 10 µ G.[Fe II] 1.644 µ m and H β line emissivities are overplotted in an arbitrary linear scale. (Bottomleft) Profiles of H nuclei density ( n ), electron density ( n e ) and fraction of Fe in Fe + (Fe + /Fe)for the same shock. (Right) Synthesized IR spectrum of [Fe II] lines from the shock, normalizedto the [Fe II] 1.644 µ m line intensity. The calculation is done by using the Raymond code. is not too low, as shown in Figure 2. Numerical shock models covering some parameterspaces are available in Hollenbach et al. (1989), Mouri et al. (2000), and Allen et al.(2008). A grid of shock models with updated atomic parameters is in preparation by theauthor of this paper. 2.2. NIR [Fe II] Observations of SNRs
The first detection of the [Fe II] 1.644 µ m line in an SNR was reported by Seward etal. (1983) on MSH 15 −
52. After that, about a dozen Galactic and LMC SNRs havebeen observed in NIR [Fe II] lines. This number will increase with the completion of theUWIFE (UKIRT Wide-field Infrared Survey for Fe + ) project, which is an unbiased surveyof the [Fe II] 1.644 µ m line of the inner Galactic plane ( ℓ = 7 ◦ to 65 ◦ ; | b | . ◦ ) usingthe UKIRT 4-m telescope. The UWIFE project is a “cousin” of the UWISH2 project,which covers the same area in the H v = 1 → µ m (Froebrich et al.2011). Lee, Y.-H. et al. (this volume) introduce the two projects and present preliminaryresults on SNRs. In short, there are 77 SNRs in this area, and about 20%–30% of themare detected in [Fe II] and/or H lines, more than half of which are new detections.The SNRs bright in [Fe II] emission lines may be divided into two groups: (1) middle-aged SNRs interacting with dense molecular (or atomic) clouds, and (2) young SNRsinteracting with the dense circumstellar medium (CSM). Middle-aged SNRs bright in [Fe II] emission . Prototypical SNRs belonging to this cat-egory are W44, 3C391, and IC 443. All of them are interacting with molecular clouds(MCs). An indication of the MC interaction in the NIR band is the presence of H ro-vibrational emission lines that arise from slow, non-dissociative C shocks propagatinginto dense molecular gas of low-fractional ionization. Therefore, these middle-aged SNRsthat are bright in [Fe II] emission are also bright in H emission (Fig. 3).Figure 3 shows [Fe II] 1.644 µ m and H µ m line images of W44; we see that theoverall morphologies of the SNR in the two lines are similar, although the former appearsrather diffuse and confined to the SNR boundary, whereas the latter is considerablyfilamentary and fills the entire SNR (see also Reach et al. 2005). A detailed inspection Bon-Chul Koo Figure 3.
UWIFE [Fe II] 1.644 µ m (left) and UWISH2 H µ m (middle) images of W44.The scale bar corresponds to 10 ′ . The images on the right show zoomed-in views of the southernarea in the [Fe II] (top) and H (bottom) emissions, respectively. The contours of the H emissionare overlaid on the top right [Fe II] image. shows that, in some areas, there is a good spatial correlation between the two, whereasin other areas, the correlation is less significant. The detection of both [Fe II] and H emission lines is consistent with the general consensus that a molecular cloud is clumpy,being composed of dense clumps embedded in a rather diffuse interclump gas. An SNRproduced inside a MC expands into the interclump medium while engulfing dense clumps.In late stages, the SNR shock in the interclump medium becomes radiative, so that theSNR becomes surrounded by a fast-expanding ( ∼
100 km s − ) atomic shell, which isobservable in Hi emissions shouldreveal more detailed information about the structure of molecular clouds. Young SNRs bright in [Fe II] emission . There are also young SNRs bright in [Fe II]emission. Prototypical ones are Cas A, G11.2 − ∼ .
05 cm − , whereas the Kepler SNR is interactingwith a relatively dense ( > ∼ − ) medium. Young CCSNRs are interacting either withCSM or a wind bubble created in the main-sequence lifetime of their progenitors. CasA, which is Type IIb SN, for example, is interacting with a dense red supergiant (RSG)wind. The solid line is a model for Cas A from Chevalier & Oishi (2003), but assuming˙ M w = 3 × − M ⊙ yr − , v w = 15 km s − , M ej = 5 M ⊙ , and E SN = 10 ergs where ˙ M w is the wind mass-loss rate, v w is the wind speed, M ej is the ejecta mass, and E SN is theexplosion energy. Cas A follows this “Cas A-like” line as long as it continues to interactwith the dense CS wind. There are CCSNRs which fall well below the Cas A-like line, i.e.,G11.2 − < ergs) or/and dense CSM. AU 296. [Fe II] and Dust Emissions from SNRs Figure 4. (Left) Radius versus age of young SNRs. The dotted lines are models for SNRs inuniform ambient media of n = 1, 0.1, and 0.01 cm − , respectively (Truelove & McKee 1999; n = 7 ejecta model with M ej = 5 M ⊙ and E SN = 10 ergs). The solid line is for an SNR in theRSG wind case (Chevalier& Oishi 2003; see text for the parameters of the model.) Note thatthe ones marked by empty circles are pulsar wind nebulae, so they do not represent true sizesof SNRs. (Right) [Fe II] images of Cas A and G11.2-0.3 from top to bottom. The scale barscorrespond to 1 pc at the distances of the SNRs. The strong [Fe II] lines in these SNRs suggest that it is more likely because of denseCSM and that the CSM is much denser than that of Cas A. Hence, they are likely theremnants of massive SN IIL/b or SN IIn. On the other hand, there are remnants muchlarger than Cas A: MSH 15 −
52, G292.2 − > ergs) could be another possibility. Figure 4 suggeststhat all four [Fe II]-bright young CCSNRs mentioned at the beginning of this paragraphare SN IIL/b candidates interacting with dense CSM. (For W49B, however, the bipolarType Ib/c SN origin has been suggested. See Lopez et al. 2013 and references therein.)The [Fe II] emission in these young SNRs originate from both shocked CSM andshocked SN ejecta. In Cas A, it is well known that there are two types of knots detectedin the visible waveband: quasi-stationary flocculi (QSFs), which are dense CS knotsmoving at a few hundred kilometers per second, and fast moving knots (FMKs), whichare metal-rich SN ejecta knots moving at several thousand kilometers per second. Lee,Y.-H. et al. (this volume) show that there are also fast-moving [Fe II] knots that lackother metallic lines, which could be pure Fe ejecta synthesized in the innermost SN region.In G11.2 − ∼ ,
000 km s − , whichsuggests that they might be SN ejecta (Moon et al. 2009). Again, their spectra do notshow metallic lines other than Fe. The bright filament in the southeast of G11.2 − Some Issues[Fe II] - H reversal . Since the early days of NIR observations of SNRs, it has beenknown that there are SNRs with H filaments lying beyond [Fe II] filaments, i.e., furtherout from the SNR center, which is not easily explained by shock models (Graham etal. 1991; Oliva, Moorwood, & Danziger 1990; Burton & Spyromilio 1993). We now havemore sources showing similar patterns, e.g., G11.2-0.3, W49B, and 3C396 (Koo et al.2007; Keohane et al. 2007; Lee et al. 2009). W44 in Figure 3 is another example. Hence,we need an explanation for this “[Fe II]-H reversal”; i.e., we need to know what theexciting mechanisms of the H emission is and how they excite the H gas beyond theSNR. Some proposed mechanisms are fluorescent UV excitation, X-ray heating, magneticprecursors, and reflected shocks, but high-resolution NIR spectroscopic studies are neededto address the issue. [Fe II] luminosity as an SN rate indicator . Can the [Fe II] 1.257 or 1.644 µ m lumi-nosities be used as an indicator of galactic SN rates? Several groups have addressedthis for starburst galaxies and have derived a conversion factor of SN rate =0.03–0.1 L [ F eII ] / L ⊙ yr − where L [ F eII ] is the [Fe II] 1.644 µ m luminosity (Morel et al. 2002;Alonso-Herrero et al 2003; Rosenberg et al. 2012). In the case of nearby starburst galax-ies M82 and NGC 253, however, 70%–80% of the [Fe II] emission is known to be diffuseemission of unknown origin, although it is speculated to be related to the SN activitytherein (Greenhouse et al. 1997; Alonso-Herrero et al. 2003). In the Galaxy, the [Fe II]-bright SNRs represent 20%-30% of the known ∼
300 SNRs, which occupy a small fractionof the entire set of SNRs present in the Galaxy. It is clear that we need to have betterunderstanding of the population of [Fe II]-bright SNRs and also the origin of the diffuse[Fe II] emission to obtain a more reliable relation between L [ F eII ] and the SN rate ingalaxies.
3. Dust IR Emission from SNRs
Dust Heating in SNRs Interacting with MCs
Dust in SNRs can be heated either collisionally or radiatively. In SNRs with fast, non-radiative shocks, dust grains are heated by collisions with gas particles, mainly electrons,in a hot plasma behind the shocks (e.g., Dwek et al. 2008). Many SNRs have mid- andfar-IR morphology almost identical to that of X-ray, which suggests that the IR emissionin these SNRs is primarily from collisionally-heated dust grains.
Figure 5. (Left) ATCA 20-cm image of Kes 17 with overlaid XMM 0.2–12 keV X-ray (thin)and AKARI 65 µ m (thick) contours. (See also Lee et al. 2011.) (Right) VLA 20-cm image of IC443 with overlaid ROSAT 0.2–2.4 keV X-ray (thin) and AKARI 90 µ m (thick) contours. AU 296. [Fe II] and Dust Emissions from SNRs α photons, heat the dust in the cooling layer, and, subsequently,the infrared radiation from these hot dust grains heats the dust at larger column densities(e.g., Hollenbach et al. 1979). The far-IR bright regions in MC-interacting SNRs (Fig. 5)are probably where the radiation field is strong and the ambient density is high. Andersenet al. (2011) carried out a systematic study of the dust emission from MC-interactingSNRs found in the GLIMPSE survey, and derived dust temperatures of 29–66 K from theSpitzer MIPS spectral energy distribution (60–90 µ m). There could be a dust componentat a lower temperature, however, because the SNRs interacting with MCs are bright inthe far-IR waveband beyond the MIPS coverage (e.g., see Lee et al. 2011). A systematicstudy of MC-interacting SNRs, including the AKARI and Herschel far-IR data, will beuseful to understand the heating mechanisms and also the processing of dust grains inthese SNRs. 3.2. Star Dust in SNRsSN dust in young CCSNRs . The dense, metal-rich, cooling SN ejecta can effectively pro-vide an environment for dust to condense. In the high-redshift galaxies, where low-massstars do not have enough time to evolve to AGB stars, SNe could be the main contribu-tors of dust, depending on the dust yield (Dwek & Cherchneff 2011). Theoretical studieshave shown that as much as 1 M ⊙ of different dust species can form in SN IIP withmassive H envelopes, whereas in SN IIL/b or SN Ia with little or no H envelopes only alimited amount of dust could form (Nozawa et al. 2010; Nozawa et al. 2011).In observational studies, however, only a very small amount of dust in SNe has beendetected, i.e., < ∼ − M ⊙ . (see Gall et al. 2011 and references therein). It is only towardthe LMC SN 1987A and some young Galactic SNRs where a significant amount of SNdust has been detected: In 1987A, Matsuura et al. (2011) reported detection of 0.4–0.7 M ⊙ of dust, whereas, in the Galaxy, 0.1–0.2 M ⊙ of dust has been detected in CasA and the Crab nebula which are SNIIb and SN IIP(?), respectively. In another SN IIPcandidate, G54.1+0.3, a dust ring of 0.58–0.86 M ⊙ has been detected around its pulsarwind nebula, but the nature of the ring is not yet conclusively identified (Koo 2012). Crystalline silicate dust in MSH 15 − . Essentially all dust grains in the ISM areamorphous. Crystalline silicate dust grains have been found mainly in evolved stars andyoung stellar objects, indicating that they form in situ in circumstellar disks and/or out-flows of these objects (Henning 2010). In this regard, the detection of crystalline silicatesin the SNR MSH 15 −
52 is interesting (Koo et al. 2011).As we mentioned in § −
52, is a young ( ∼ ,
000 yr) SNR probably ex-panding inside a bubble, suggesting progenitor SN type of Ib/c (see Fig. 4). The remnanthas a central pulsar, and there is an O star (Muzzio 10) and a bright MIR source (IRAS15099 − ′′ and 31 ′′ (or 0.35 pc and 0.60 pcat 4 kpc) to north, respectively. IRAS 15099 − − −
52 appears to be the first case in which crystalline silicates have been observedto be associated with a SNR.
Acknowledgements
I wish to thank Lee, Y.-H., Jeong, I.-G, and Moon, D.-S. for their help with figures. Myresearch is supported by Basic Science Research Program through the National ResearchFoundation of Korea (NRF) funded by the Ministry of Education, Science and Technology(NRF-2011-0007223).
References
Allen, M. G., Groves, B. A., Dept, M. A., Sutherland, R. S., & Kewley, L. J. 2008,
ApJS , 178,20Alonso-Herrero, A., Rieke, G. H., Rieke, M. J., & Kelly, D. M. 2003,
ApJ , 125, 1210Andersen, M., Rho, J., Reach, W. T., Hewitt, J. W., & Bernard, J. P. 2011,
ApJ , 742, 7Burton, M., & Spyromilio, J. 1993,
Proceedings of the Astronomical Society of Australia , 10, 327Chevalier, R. A. 1999,
ApJ , 511, 798Chevalier, R. A. 2005,
ApJ , 619, 839Chevalier, R. A., & Oishi, J. 2003,
ApJL , 593, L23Dwek, E., Arendt, R. G., Bouchet, P., et al. 2008,
ApJ , 676, 1029Dwek, E., & Cherchneff, I. 2011,
ApJ , 727, 63Froebrich, D., Davis, C. J., Ioannidis, G., et al. 2011,
MNRAS , 413, 480Gall, C., Hjorth, J., & Andersen, A. C. 2011,
A&ARv , 19, 43Graham, J. R., Wright, G. S., Hester, J. J., & Longmore, A. J. 1991, AJ , 101, 175Greenhouse, M. A., Satyapal, S., Woodward, C. E., et al. 1997, ApJ , 476, 105Henning, T. 2010,
ARAA , 48, 21Hollenbach, D. J., Chernoff, D. F., & McKee, C. F. 1989,
Infrared Spectroscopy in Astronomy ,290, 245Hollenbach, D., & McKee, C. F. 1979,
ApJS , 41, 555Keohane, J. W., Reach, W. T., Rho, J., & Jarrett, T. H. 2007,
ApJ , 654, 938Koo, B.-C. 2012, Publication of Korean Astronomical Society, 27, 225Koo, B.-C. & Heiles, C. 1995,
ApJ , 442, 679Koo, B.-C., McKee, C. F., Suh, K.-W., et al. 2011,
ApJ , 732, 6Koo, B.-C., Moon, D.-S., Lee, H.-G., Lee, J.-J., & Matthews, K. 2007,
ApJ , 657, 308Lee, H.-G., Moon, D.-S., Koo, B.-C., et al. 2011,
ApJ , 740, 31Lee, H.-G., Moon, D.-S., Koo, B.-C., Lee, J.-J., & Matthews, K. 2009,
ApJ , 691, 1042Lopez, L. A., Ramirez-Ruiz, E., Castro, D., & Pearson, S. 2013,
ApJ , 764, 50Matsuura, M., Dwek, E., Meixner, M., et al. 2011,
Science , 333, 1258McKee, C. F., Hollenbach, D. J., Seab, G. C., & Tielens, A. G. G. M. 1987,
ApJ , 318, 674Moon, D.-S., Koo, B.-C., Lee, H.-G., et al. 2009,
ApJ , 703, L81Morel, T., Doyon, R., & St-Louis, N. 2002,
MNRAS , 329, 398Mouri, H., Kawara, K., & Taniguchi, Y. 2000,
ApJ , 528, 186Nozawa, T., Kozasa, T., Tominaga, N., et al. 2010,
ApJ , 713, 356Nozawa, T., Maeda, K., Kozasa, T., et al. 2011,
ApJ , 736, 45Oliva, E., Moorwood, A. F. M., & Danziger, I. J. 1989,
A&A , 214, 307Oliva, E., Moorwood, A. F. M., & Danziger, I. J. 1990,
A&A , 240, 453Oliva, E., Moorwood, A. F. M., Drapatz, S., Lutz, D., & Sturm, E. 1999,
A&A , 343, 943Pradhan, A. K. & Nahar, S. N. 2011,
Atomic Astrophysics and Spectroscopy
Cambridge Uni-versity Press: Cambrdige and New YorkReach, W. T., Rho, J., & Jarrett, T. H. 2005,
ApJ , 618, 297Rosenberg, M. J. F., van der Werf, P. P., & Israel, F. P. 2012,
A&A , 540, A116Seward, F. D., Harnden, F. R., Jr., Murdin, P., & Clark, D. H. 1983,
ApJ , 267, 698Truelove, J. K., & McKee, C. F. 1999,
ApJS , 120, 299, 120, 299