The near-infrared window is favorably located within the electromagnetic spectrum to possess a lambda-bandpass that is short enough to be sensitive to the photospheric emission of stars (late-type cool stars, including luminous giants), while long enough to penetrate foreground dust. It is used to study the stellar and kinematic properties of spheroidal/elliptical galaxies, and the internal structure (gas, dust and stars) of disk galaxies. Here we consider spiral galaxies, in particular the special case of disks whose orientation is (mostly) facing us. Although a detailed galaxy-by-galaxy comparison is beyond the scope of this paper, we can identify some qualitative trends and generalize the near-infrared properties of modestly inclined disk galaxies.
The near-infrared light efficiently penetrates dust (AK ~ 10% AV) associated with the interstellar medium of galaxies, revealing enduring internal structures that have emerged over their history. We observe (LGA Atlas, Figures 1-5) both flocculent types (lacking spiral symmetry, chaotic or piece-meal arm structures; e.g., M31, NGC 2841, NGC 7793) and beautiful grand design types (continuous, two-arm symmetry; e.g., M83, M100, NGC 4254). Both types may have ringed arms and large-scale bar configurations (e.g., NGC 1365, NGC 2903, NGC 4274). Distorted arms, one-armed spirals, warps and other asymmetries, are also observed, typically associated with tidal interactions (see e.g., NGC 1097, NGC 2442, NGC 3627; see also M51 below). For most disk galaxies the disk/spiral structure is both symmetric and modest in complexity. The disk stellar distribution is exponential and symmetric disk (e.g., Figure 7), representing the mass density of the disk. The arms are typically smooth and broad in comparison to the visual (see also Rix & Rieke, 1993; Seigar & James 2002.) We observe m=1 (most common) or 2 (less common) modes for most spirals, which suggests that higher-order density waves are rare as seen in the infrared (but consider the peculiar NGC 908, which might have multiple arms). Note that the 2MASS images may not be sensitive enough to detect multiple arm structures for the more distant galaxies; nevertheless, our results are consistent with the more sensitive infrared observations of Block & Puerari (1999). Moreover, as Block et al (1994) point out, there is theoretical reason to believe that m=1 modes should dominate the internal structure of spirals, given that the Inner Lindblad Resonance (Linden-Bell & Kalnajs 1972) disrupts the higher-order density waves, suppressing the formation of arms beyond the first mode.
It is striking that the collective morphologies, density wave patterns and arm classifications (Elmegreen & Elmegreen 1987) we observe in the near-infrared are often so different from that observed at visual wavelengths. In addition to NGC 253 (Figure 6), a prime example is M63 (NGC 5055). It is visually classified as a flocculent, with Elmegreen arm class 3, but the near-infrared images reveal a continuous symmetric spiral (indicative of an arm class > 5, grand design). In fact Block & Puerari (1999) measured low order Fourier modes in M63, showing that the spiral modes were similar to that of NGC 5861 (arm class 12, grand arm spiral). Another example is NC 3521, classified as a flocculent type (Elmegreen arm class 3), whereas the near-infrared mosaic reveals beautiful two-arm symmetry (grand design). Discrepancies between the visual and infrared imaging must be due to singular or combined effects of extinction and population differences. The emerging paradigm is that of dual co-existing spiral structure¨Population I elements (young, massive stars, including gas and dust) luminous at visual wavelengths and Population II elements (stars, evolved giants) luminous at infrared wavelengths. The two components may be strongly coupled, in which case the visual and infrared morphology should be similar, or strongly decoupled, in which case the morphologies may look nothing alike. It is the latter case that seems to be the most prevalent for spirals of the 2MASS LGA, although this is only a qualitative finding. At present, decoupling between the underlying "backbone" and the newly forming stars would suggest that their evolution is following separate tracks, and thus demanding study beyond the visual wavelengths. There is of course a third scenario: partial coupling between the disk components. Consider the intriguing case of the Whirlpool Galaxy.
The Whirlpool Galaxy (M51ab, NGC 5194/95) is one of the most studied objects in the sky by both professional and amateur astronomers alike (over 800 references in NED alone), owing to its proximity (8.7 Mpc), angular extent on the sky (~12 arcmin diameter in the visual), double galaxy interaction, and its beautiful grand design spiral. UV and visual images reveal the long symmetric arms of M51a (NGC 5194), illuminated by innumerable giant molecular clouds and massive star formation regions tightly confined to the arms (e.g., Rand & Kulkarni 1990). M51a is classified as a grand design Sbc, with evidence of a Seyfert 2 nucleus (Heckman 1980; Larken et al. 1998). The disk structure of M51a is a classic case of star formation induced by the passage of spiral density waves (Shu et al. 1972; Lord & Young 1990; Lord & Kenney 1991; Knapen et al. 1992; Rand 1993; Seigar & James 2002). Its northern companion, M51b (NGC 5195), is classified as a barred S0, with a LINER nucleus. And indeed the striking difference between the pair of galaxies is one of its main attractions.
In the near-infrared (Figure 1, panel 31; see also Wright et al 1991; Rix 1993) the spiral arms of M51a are clearly delineated, spiraling outward from the nucleus (the arms originate from a point at the resolution limit of 2MASS, ~2 arcsec, or 84 pc encompassing the nucleus) and stretching northward into the disk of its companion, M51b. Giant HII region complexes are resolved in the images, where the 2µm "knots" are probably dominated by Hydrogen recombination line emission associated with high-mass star formation sites. There is also significant near-infrared light between the arms, representing the disk or "backbone" of M51a, tracing the mass density of the system. Note the slightly distorted (eastern) outer-arm of M51a, probably associated with the M51b tidal interaction (see also Sweitzer 1977; Block et al 1994). The companion, M51b, is roughly half the size of its southern counterpart. It exhibits a prominent bar (extending north-south) and ringed spiral arm, in addition to significant isotropic light (hence, the S0 or lenticular morphology) and diffuse ("plume"-like) light that extends well beyond the disk. The companion is indeed peculiar.
To summarize, the near-infrared images show significant inter-arm light (presumably the Population II "backbone" disk), but also well established spiral arm structure. Although the spiral arm structure is much more modest in comparison to what is observed in the UV or visual, the qualitative appearance of the M51 system in the near-infrared is not unlike that of the visual wavelengths. This suggests that the Population I elements (discrete regions defining the spiral arms) and the Population II elements (inter-arm light and the smooth, broad light delimiting the spiral arms) are coupled to some degree. The spiral density wave is evidently long-lasting and influencing the distribution (and possibly evolution) of the underlying Population II stars. The population coupling is consistent with enhancement of the star formation efficiency in the extended near-infrared arms and the narrow UV/visual arms (see also Schweizer 1976; Seigar & James 2002). To broaden our perspective, we now consider the mid-infrared emission that tracks the hot dust associated with star formation (see also Sauvage et al 1996; Block et al 1997 for 11 µm mid-infrared images of M51). We have combined our near-infrared mosaic of M51 with an ISOCAM 6.8 µm mid-infrared mosaic to illustrate the vivid dichotomy between the stellar populations in M51 and between the pair of galaxies themselves.
In Figure 27 we show an RGB color combination of the 2MASS J (1.2 µm) and Ks (2.2 µm) M51 mosaics, with the ISOCAM LW2 (5.5- 8.5 µm) mid-infrared mosaic (ISOCAM is the mid-infrared imager of the Infrared Space Observatory; the data and reductions are kindly provided by Dr. Alessandra Contursi). Here the J-band is painted blue, the Ks-band painted green and the LW2-band painted red. The resolution of the ISOCAM data is ~7 to 8 arcsec (with 3 arcsec pixels), to which the 2MASS data has been gaussian smoothed to match. We retain the ISO spacecraft orientation as it observed M51, roughly 54 degrees east of north, to minimize degradation of the space-born image. The LW2 band is sensitive to emission associated with "Aromatic Features," the so-called Unidentified Infrared Bands (centered at 6.2 and 7.7 µm), commonly thought to be due to polycyclic aromatic hydrocarbon (PAH) molecules or tiny carbonaceous grains (see Leger & Puget 1984; Desert et al 1990; Schutte et al 1998.) The Aromatic Features are very strong in normal spiral galaxies, largely arising from photo-dissociation regions (Hollenbach & Tielens 1997; Helou et al. 2000), and totally dominating the mid-infrared light shortward of 12 µm (see also Lu 1998; Roussel et al. 2001).
In Figure 27 we see the mid-infrared light (as painted in red) tightly confined to the spiral arms of M51a and coincident with the HII complexes traced in the visible regime. Spherical blobs, knots, filaments and shells are some of the discrete sub-structures that line the spiral arms. In comparison, the near-infrared light is primarily seen between the arms (as painted in blue-green), or smoothly tracing the spiraling arms, or confined to the nucleus (with remarkable symmetry about the nucleus.) The mid-infrared arms wrap inward until they form a bright nuclear ring, virtually independent from the near-infrared light. The ring, and relative lack of mid-infrared light in the nucleus of M51a, may be indicative of disruption of the spiral density wave, or disruption of the PAH molecules from the UV radiation field associated with the buried AGN (Seyfert 2 nuclear line emission.) The near-infrared arm light is much broader, smoother and lower surface brightness than the mid-infrared light. The Population II stars are not only organized into arms by the density wave, but broadly coincident with the newly formed Population I elements¨coupling the mass density "backbone" with the large scale shock-triggered star formation. The degree of coupling between the old and new stellar populations in M51 is probably unusual given the exceptionally strong spiral density wave coursing through its disk, but partial couplings are also seen in other prominent late-type disk galaxies (e.g., M83; see also Lord & Kenney 1991; Seigar and James 2002).
The contrast between M51a (NGC 5194) and its companion, M51b (NGC 5195) is even more spectacular. The companion's disk is nearly devoid of mid-infrared light (except for the nucleus itself; see Boulade et al. 1996), while the near-infrared reveals the inner structure: diffuse disk light, symmetric "backbone", bar extending the length of the disk, circum-nuclear ring and nucleus. It is peculiar that M51b should possess such warm far-infrared colors and the detection of CO (Sage 1990), while lacking mid-infrared emission beyond the nucleus. The modest nuclear starburst is evidently driving most of the infrared activity in M51b. For the disk/bar of M51b, we have an example of total decoupling between the Population I elements (albeit, mostly non-existent in this early-type spiral) and the Population II bar and disk "backbone". Note also that the companion does not appear to be disrupted or distorted by the tidal interaction with the much larger M51a, consistent with the findings of Block et al. (1994); but see also Boulade et al. 1996 for a contrasting view. To summarize, the near-infrared images of M51 reveal both diffuse and spiral internal structure, demonstrating the linkage between the density wave-influenced stellar populations with the symmetric underlying "backbone" populations. The mid-infrared emission is tightly confined to the spiral arms of M51a, and is only diminutively present in its companion M51b, tracing the sites of massive star formation. The M51 system may represent a case in which the Population I and II disk components are physically coupled for M51a, and completely decoupled for M51b.
Some final notes regarding morphology and internal structure. Although the statistical trends are compelling, it is still quite remarkable how insensitive the NIR is to morphological type for any given galaxy that is inspected. The reduced parameters of the NIR alone do not provide enough information to discern one Hubble T-type from the next (but in a gross sense, they can delineate early-types from late-type disks). This underscores the decoupling between the Population I elements (dictating the Hubble T-type classification) and the Population II elements (dominating the 2µm light). The best hope for robust galaxy classification seems to be some complementary, multi-wavelength parameter combination (e.g., Figure 27), optimizing/exploiting the strengths of each "mask" or view into the internal structure of galaxies. Finally, it is worth noting that there is another promising method toward galaxy classification that 2MASS is well suited for: 3-dimensional pattern matching. This method requires (in addition to sophisticated software) a very large and uniform training set to cover the phase space that the Hubble types inhabit. This is the ultimate strength of the 2MASS Extended Source Catalog--large all-sky census of the local Universe.