COOL DUST IN THE OUTER RING OF NGC 1291

Comparing the morphology of NGC 1291 over a range of wavelengths allows us to determine the spatial distribution of the various stellar populations with respect to the spatial distribution of the dust (see Buta 2011 for a comprehensive review on the uses of morphology). Figure 1 shows that the typically identified features of the galaxy in the optical are also seen at infrared wavelengths. There is a bright inner-elongated, oval-shaped bulge, an area outside this bulge with very little flux, and an outer discontinuous ring marked with dense knots of emission of varying brightnesses, which are likely regions of recent star formation. The inner lens prominent at the B band disappears in the infrared. In particular, the outer ring is faint relative to the bulge at bandpasses below 100 μm and becomes more prominent at the 160 μm and SPIRE wavelengths, consistent with findings in the proceeding by Bendo (2006). There are localized bright spots in the ring in the northwest and southeast arcs.

The resolution gains made by Herschel over Spitzer (52 for PACS versus 18'' for MIPS at 70 μm) are illustrated by the level of detail in the central portion of the galaxy seen in the PACS image. Figure 2 shows the inner 3' × 3' of both images. The PACS image has a bright central region surrounded by diffuse emission in the form of two extensions with a brighter patch on the northwest side. This is not visible in the 70 μm MIPS image, where the lower resolution smears these details.

Zoom In Zoom Out Reset image size

Given the larger beam size of the MIPS instrument and the difference in the operating temperatures of the two telescopes, it was anticipated that the 70 μm and 160 μm MIPS images would be more sensitive to diffuse emission, such as in the outer disks of galaxies, by a factor of five for dust emission at T = 15 K and by a factor of seven for dust emission at T = 20 K for a similar number of passes by each instrument (e.g., Hinz et al. 2008a). Figure 2 also shows an 11' × 11' view, where the MIPS image has a faint detection of the ring at the northwestern and southeastern hot spots that PACS does not. This may be due to the intrinsic instrument differences as suggested above or due to the methods with which the PACS data are reduced currently. Highpass filtering used during the production of the PACS maps by HIPE may remove such diffuse features. Using the Scanamorphos software (Roussel et al. 2010; Roussel 2012) after an initial process with HIPE recovers some diffuse features for KINGFISH galaxies. This software produces maps after subtracting the thermal drift and the low-frequency noise of each bolometer purely by using the available redundancy. However, the outer ring of NGC 1291 does not appear at 70 μm even when Scanamorphos is used. In this case, we find that is helpful to use both the PACS and MIPS complementary data sets for analysis. A discussion of apparent discrepancies between MIPS and PACS photometry can be found in Aniano et al. (2012).

The contribution from the young stellar population to the emission from the ring can best be seen through a combination of ultraviolet imaging, with the near-UV (NUV) thought to trace the few hundred Myr stellar population, Hα imaging, tracing the ⩽10 Myr population, and infrared images at 24 μm, representing the warm dust emission generated near star-forming regions. Figure 3 shows GALEX NUV and far-UV (FUV) images (Gil de Paz et al. 2007), an Hα continuum-subtracted image (SINGS Legacy ancillary data), and the MIPS 24 μm image (SINGS; Dale et al. 2007). NGC 1291 is not classified as an extended ultraviolet disk by Thilker et al. (2007) since the UV morphology is coincident with the optical morphology of the ring; this coincidence implies that there is a more evolved stellar component in the ring in addition to the young stars. The ring in Hα and 24 μm is very thin (∼40'' across at 24 μm) and is discontinuous, with discrete sources appearing mostly in arcs on the northwest and southeast sides of the galaxy. The UV images show many more star formation knots to the north and west of the main ring than the other wavelengths.

Zoom In Zoom Out Reset image size

In a detailed study of H ii regions in early-type-ringed galaxies by Crocker et al. (1996), NGC 1291 is considered to be the most unusual galaxy of their sample in Hα emission. They describe a diffuse and spiral-like Hα distribution in the inner region of the galaxy, which appears uncorrelated with the bar and lens features visible in optical images, with differing shape and orientation. The central emission also takes on this spiral-like appearance at 24 μm. Additionally, the Hα emission near the galaxy center is described by Crocker et al. (1996) as asymmetric, and weak Hα arms extend beyond the inner lens. These regions are filled with diffuse patterns of ionized gas filaments to a radius of ∼5 kpc, which is analogous to a pattern seen in the later type galaxy M 31 (Ciardullo et al. 1988) and M 81 (Devereux et al. 1995), and assumed to be caused by the ejection of the interstellar medium via a galactic wind. The diffuse and unusual Hα emission and the H i observations, showing an outer ring with a pronounced central hole (van Driel et al. 1988), have led to the supposition that there is little to no star formation in the central bulge (Hogg et al. 2001). However, we now see that the inner portion of the galaxy appears to be brighter compared to the ring at both Hα and 24 μm, making it unlikely that there is no star formation occurring centrally.

Global star formation rate estimates for NGC 1291 differ slightly with method. The rates are log SFR(M_☉ yr⁻¹) =−0.5  ±  0.4 (Hα; Kennicutt et al. 2008), −0.45  ±  0.04 (FUV+TIR; Skibba et al. 2011), and −0.07  ±  0.05 (Hα + 24 μm; Calzetti et al. 2010). While these star formation rates are within the known range for such galaxies (Kannappan et al. 2009; Skibba et al. 2011; D. A. Crocker et al., in preparation), the H i gas mass of NGC 1291 is higher than expected from its morphological type and is not coincident with any central star formation.

To understand the star formation rates specifically in the ring, we calculate rates using the GALEX UV data for this portion of the galaxy only. The FUV and NUV total magnitudes, within a circle of radius 65, are found to be 13.5 and 14.4, respectively, in agreement with the GALEX Atlas of Nearby Galaxies (Gil de Paz et al. 2007). For the aperture between 3' and 65 that contains only the ring, the NUV magnitude is 15.1, which translates to a log star formation rate (M_☉ yr⁻¹) = −0.52, following Donovan et al. (2009), who use a relation from Kennicutt (1998) for a similar early-type galaxy with an ultraviolet ring. Given the GALEX resolution and the distance to the galaxy and following Kennicutt (1989), we then calculate an average star formation rate density for the ring of 6.9 × 10⁻⁵ M_☉ yr⁻¹ kpc⁻², with no corrections for reddening or dust extinction. Foreground extinction is estimated to be small here, only A_U = 0.07 mag at U band (Schlegel et al. 1998). Even if the extinction is two to three times larger than this estimate in the UV, the revised log star formation rate (M_☉ yr⁻¹) is only −1.44, for a star formation rate density of 8.3 × 10⁻⁵ M_☉ yr⁻¹ kpc⁻². This density is quite low for an outer ring (though similar to those derived for ESO 381-47 and NGC 404; Donovan et al. 2009; Thilker et al. 2010) and may be due to the fact that the UV emission is discontinuous across the large area of the ring (i.e., between 3' and 65).

Summing the sky-subtracted fluxes in the inner galaxy and in the ring area in each band and taking their ratio yields a simple measurement of the change in brightness of the central portion of the galaxy relative to the brightness of the ring. Figure 4 shows these ratios from 24 to 500 μm, where the errors are calculated from photometric uncertainties. We see that the ratio of the flux in the ring aperture to the flux in the central circle increases with increasing wavelength, indicating that there is a corresponding decrease in temperature of the dust in the ring compared to the inner galaxy.

Zoom In Zoom Out Reset image size

Similarly, Figure 5 shows the spectral energy distribution (SED) of the whole galaxy at far-infrared wavelengths, along with the SED points for the inner portion of the galaxy only (within 3') and for the ring only (between 3' and 65). The SED for the ring shows that this area generates cooler emission than the central region of the galaxy. The central portion of the galaxy contributes more to the total galactic emission from warm dust at 24 μm (∼60%) but is quickly overtaken by the cool ring starting at 160 μm, as already suggested by the results of Figure 4. To calculate the temperatures of the emitting dust in various portions of the galaxy, we fit a blackbody with a frequency-dependent emissivity with an exponent of 1.5 to both the ring and inner galaxy SEDs from 70 to 500 μm, following Engelbracht et al. (2010) and Skibba et al. (2011). Trends are unaffected if emissivity values of 1 or 2 are used, although the temperatures systematically increase or decrease, respectively. The uncertainties in the temperatures are computed via a Monte Carlo simulation in which 10,000 trials are performed. The photometric measurements are allowed to vary in a normal distribution with a standard deviation indicated by the photometric uncertainty. The cool dust temperature of the ring is T_ring = 19.5 ± 0.3 K and for the inner portion of the galaxy it is T_inner = 25.7 ± 0.7 K. This gives T_inner/T_ring  ∼  1.3.

Zoom In Zoom Out Reset image size

Both of these dust temperature values are typical of cool dust temperatures in nearby galaxies (e.g., Hinz et al. 2006, 2008b). The values for NGC 1291 are slightly different from those found by Engelbracht et al. (2010), likely due to the difference in apertures used, but consistent with their overall finding that earlier type spirals have higher central temperatures by 20% to 50% over their late-type counterparts. The temperature ratio is closer to that found for strongly barred galaxies compared to those derived in weakly barred galaxies, though the bar classification for this galaxy changes slightly between optical and near-infrared analyses, from SB to SAB (Buta et al. 2010).

The dust masses for NGC 1291, as a whole and for its components, can be calculated using the mid-infrared photometry from IRAC, PACS, MIPS, and SPIRE and the models of Draine & Li (2001, 2007) and constructed as suggested in Draine et al. (2007), using a mixture of amorphous silicate grains and carbonaceous grains with a distribution of grain sizes. All images are convolved to the MIPS 160 μm resolution before modeling (Aniano et al. 2011). Details of the use of these models specifically with Herschel photometry are described extensively in Aniano et al. (2012) for two other spiral galaxies from KINGFISH. The results of such fitting for NGC 1291 are summarized here, with the black solid line in Figure 5 representing the final fit to the data for the whole galaxy. The apertures, dust masses (M_dust), aperture areas, mean starlight heating intensities (〈U〉), fluxes radiated by dust corresponding to photodissociation regions (f_PDR), and the fractions of grain mass contributed by polycyclic aromatic hydrocarbons (q_PAH) are listed in Table 1. The total dust mass for the galaxy, calculated from a sum of the modeling for each individual galaxy pixel, is M_dust = 2.8 × 10⁷ ± 3.5 × 10⁶ M_☉, with a total dust luminosity of L_dust = 2.4 × 10⁹ ± 2.3 × 10⁷L_☉. Dust masses calculated for annular apertures with radii in intervals of 1' out to 6' show that the dust in the ring dominates the total dust mass of NGC 1291. The average starlight heating intensity is calculated to be the largest in the inner portion of the galaxy. This 〈U〉 can be related to the characteristic dust temperature, T_dust, by the approximate relation T_dust = 18K × 〈U〉^1/6, where 〈U〉 is in units of the starlight energy density in the solar neighborhood (Aniano et al. 2012).

First, we explore the idea that all galaxies have similar dust temperature gradients to NGC 1291. Engelbracht et al. (2010) showed that the central areas of early-type spiral galaxies generally have enhanced dust heating of their cool dust components compared to their disks. They find that, on average, the cool dust temperature of the central component is 15% ± 3% hotter than the disk. Similarly, Pohlen et al. (2010) showed that SPIRE surface brightness ratios seem to decrease with radius in spirals, implying that the dust in the outer regions is colder than dust in the centers of galaxies. Galametz et al. (2012) study 11 galaxies in the KINGFISH sample, including NGC 1291, fitting Spitzer and Herschel data SEDs with two modified blackbodies and creating spatially resolved maps of their dust properties. They also see systematic drops in dust temperature with radius for disk galaxies, also on the order of 10–15 K from inner to outer galaxy.

However, it is not obvious that NGC 1291 should follow this pattern and that outer rings should be dominated by cool dust emission. UV rings in S0 and early-type spiral galaxies are often associated with sites of recent star formation, sometimes marked by clumpy H ii regions, and can even be the only sites of star formation in some galaxies (e.g., Donovan et al. 2009; Thilker et al. 2010). A recent study of five nearby barred S0 galaxies with rings by Marino et al. (2011) showed that these outer rings account for a majority (up to 70%) of the flux at UV wavelengths, indicating their young stellar population. Buta et al. (2010) specifically discuss the morphology of NGC 1291 based on the B band and 3.6 μm IRAC images (Sheth et al. 2010), saying that the outer ring is "where most of the recent star formation is taking place" and is prominent in the B band (Figure 1). At 3.6 μm, however, the ring does not stand out and is instead a broad ellipse in a rounder diffuse background (Figure 1 and their Figure 7). Therefore, we might expect that, if the 24 μm emission is associated with the warm (T ∼ 50 K) dust heated by the young stellar population (seen at UV and B band to be prominent), then the warm dust at 24 μm should be brighter in the ring as well, yet the wavelengths associated with cool (T ∼ 20 K) dust are brighter relative to the flux from the rest of the galaxy.

To study the dust temperature gradient in NGC 1291 in more detail, Figure 6 displays circles of radii 2', 3', 4', 5', and 6' overlaid on the SPIRE 250 μm image, along with the estimated temperature of the dust within each annular aperture, using the same blackbody fitting technique as before. (These annuli match the dust mass annuli calculations found in Table 1. Unlike Table 1, the data for Figure 6 are not convolved to the 160 μm resolution.) We remind the reader that such temperatures are representative, characterizing large grains that dominate the emission at wavelengths greater than 70 μm, and that, in fact, such large annuli contain a wide range of temperatures, including small grains undergoing temperature fluctuations. Keeping this in mind, if all disks, with or without rings, have similar dust temperature gradients due to the gradually increasing distance between star formation regions with radius, then we should be able to identify the same drop in temperature for spirals other than NGC 1291. We choose, from the KINGFISH sample, NGC 628 (SAc; d = 7.3 Mpc), which does not contain any rings but is face-on, is at a similar distance to NGC 1291, and has a similar angular size and total infrared luminosity to NGC 1291. We show in Figure 6 the dust temperatures for this galaxy for the same annular apertures as NGC 1291. For this galaxy, we see a gradual decline in temperature with some scatter. We see neither the dramatic dip in temperature at 3' that is seen for NGC 1291 nor do we see such a large spread in temperatures, with NGC 628 having a temperature difference between inner and outer areas of 4.8 K versus 8.2 K for NGC 1291. Thus, it appears that the dust temperature from the ring is cooler than would be expected from a simple temperature decline with radius for a normal (non-ringed) spiral galaxy. Comparisons of spatially resolved dust temperatures with a larger number of KINGFISH galaxies are shown in M. Galametz et al. (2012, in preparation).

Zoom In Zoom Out Reset image size

The next possibility is that non-stellar sources such as an AGN could heat the dust in the inner portion of the galaxy, where we find a warmer dust temperature compared to the outer ring. This is suggested by the fact that outer rings are found to be much more frequent in galaxies with Seyfert nuclei (Hunt & Malkan 1999). NGC 1291 is an X-ray source, with emission detected within 25 (Bregman et al. 1995), which appears to be consistent with two energy components with their origin in a stellar component and an extended hot gas component (Hogg et al. 2001). Irwin et al. (2002) found an excess of soft emission similar to that seen in several other low-luminosity AGNs. Moustakas et al. (2010) classify NGC 1291 as an "AGN" type galaxy based on optical spectra of the nuclear region, although they note that one or more emission line(s) failed their S/N  >  2 requirement for such a classification. Recent analysis of X-ray spectra of the nuclear source also indicates a low-luminosity AGN with moderate obscuration (Luo et al. 2012), which could be a source of dust heating.

The third possibility is that the bar in NGC 1291, seen at the various wavelengths in Figure 1, funnels gas into the inner region of the galaxy, replenishing the supply necessary for star formation. This compact region of star formation could then generate warmer dust temperatures than in the outer ring (see also Engelbracht et al. 2010). If gas is being transported inward in this manner, the observational evidence would likely be in the form of "hot spots" at the ends of the bar—features that are brighter than the rest of the bar formed by shock heating of the gas as it is drawn to the center of the galaxy, causing a "pileup" near the inner Lindblad resonance (Combes & Gerin 1985; Shlosman et al. 1989). This is not seen in the H i data, which has a hole at the center (van Driel et al. 1988), or in the other wavelengths that are available. However, we cannot rule out this possibility, as such hot spots are thought to be small, very faint features. Hunt & Malkan (1999) argue that rings alone could be signs of inward transport of material, essentially a second "pileup" of dust and gas in the outer resonance of the galaxy; the star formation in the ring prominent in the UV, presumably indicating stellar ages of ∼100 Myr, could be a residual signature of such a buildup. We note that transport of material via the bar may be related to the AGN hypothesis in the previous section, in that AGNs as well as star formation can be fueled via inflow.

The last explanation for the change in cool dust temperature between the inner galaxy and outer ring is that the ring is composed mainly of an old population of stars with some current star formation. In this case, the dust in the ring was produced by generations of stars in the past and is heated now mainly by the old (∼10 Gyr) stellar population, shown best by the 3.6 μm image (Buta et al. 2010 and Figure 1) where the ring is a large, diffuse entity spanning much more physical space than seen at far-infrared wavelengths. Such "old population rings" have been suggested by comparisons between B- and I-band images of nearby galaxies. Buta (1995) showed a typical galaxy of this type (IC 1438) which has a nuclear ring, an oval inner ring, and an outer ring that is much brighter in the I passband than in the B. The explanation given for this difference in brightness was that the outer ring component formed first and left behind a stellar remnant as the other two inner features formed. An old outer ring would be unusual in that the dynamical timescales for forming rings or arms are much shorter at smaller disk radii (e.g., Freeman et al. 2010).

A remnant of this type could be the source of dust heating for the emission we see at far-infrared wavelengths. In addition, there would also be a contribution to the dust heating from ionizing ultraviolet photons, which have escaped from the existing (but weaker, in terms of contribution) star-forming regions. This combination of dust heating sources has been well modeled by, e.g., Misiriotis et al. (2001) and Popescu et al. (2002, 2011), and proposed to explain observations of other nearby galaxies (Hinz et al. 2004; Tabatabaei et al. 2007; Bendo et al. 2010). (Though see Groves et al. 2012 for a case where the old stellar population heats dust to warm temperatures.) Inspection of the B–I image (R. J. Buta 2010, private communication) shows such a color difference and will be investigated further in R. J. Buta et al. (2012, in preparation). Simulations by Freeman et al. (2010) addressing Hoag's object, an outer ring galaxy with no inner morphological features, show that an outer ring can be formed via perturbations from a bar which then dissipates slowly over time. In the environment of these instabilities, the majority of the simulated gas particles fall to the center, while the remaining particles create an outer ring, with the region between the two essentially empty. If this is the manner in which NGC 1291 is forming its ring, then the implication is that the ring is unlikely to be older than the inner regions.

Noll et al. (2009) use SED-fitting techniques to study the star formation history of NGC 1291 and posit two exponential starburst events for the galaxy, one of which occurred 10 Gyr ago and the second of which occurred 200 Myr ago. This analysis was performed for the entire galaxy. A similar study for the inner portion and the ring separately would help disentangle their stellar histories.