THE Spitzer SURVEY OF INTERSTELLAR CLOUDS IN THE GOULD BELT. VI. THE AURIGA–CALIFORNIA MOLECULAR CLOUD OBSERVED WITH IRAC AND MIPS

The majority of objects in our fields are not YSOs. The maps are contaminated by background/foreground stars and background galaxies. We have selected our YSO candidates (YSOcs) by various methods, augmenting the list where possible based on data outside the Spitzer IRAC/MIPS wavelength bands. The fundamental criteria use IRAC, MIPS, and Two Micron All Sky Survey (2MASS) data (Cutri et al. 2003) and are based on identification of infrared excess and brightness limits below which the probability of detection of external galaxies becomes high. The total number of sources is 704,045. In regions observed by both IRAC and MIPS, the YSOc selection follows that of Harvey et al. (2008). We refer to these as IRAC+MIPS YSOcs. For objects with upper limits on the MIPS 24 μm flux, we follow the method outlined by Harvey et al. (2006). We refer to these as IRAC-only YSOcs. In regions observed only by MIPS and not IRAC, we have used the formalism of Rebull et al. (2007), except we use a tighter 2MASS K_S cut of [K_S]  <  13.5. This tighter magnitude cut removed objects that were similar in color and magnitude to others that had already been eliminated. We further remove galaxies from the MIPS-only source list by including photometry from the Wide-field Infrared Survey Explorer (WISE; Wright et al. 2010). We apply color cuts suggested by Koenig et al. (2012; see their Figure 7) and require the WISE Band 2 magnitude criterion of [4.6] < 12. We refer to these as MIPS-only YSOcs. Note that the MIPS-only YSOcs were not observed with IRAC, as opposed to the IRAC-only YSOcs which were observed, but not detected, with MIPS.

Figure 6 shows the IRAC color–magnitude and color–color diagrams relevant for classifying IRAC-only sources. The different domains occupied by stars, YSOcs, and other (e.g., extragalactic) sources are shown.

Zoom In Zoom Out Reset image size

For sources in regions observed by both IRAC and MIPS, Figure 7 shows the color and magnitude boundaries used to remove sources that are likely extragalactic. This identification is done by comparing the observed fluxes and colors to results from the SWIRE extragalactic survey (Surace et al. 2004). The sources in the AMC field are compared to a control catalog from the SWIRE data set that is resampled to match our sensitivity limits and the extinction level derived for the AMC. (See Evans et al. 2007 for a complete description.)

Zoom In Zoom Out Reset image size

Finally, we vetted the YSOcs through individual inspection of the Spitzer maps (and optical images where available), and determined that 24 of the original 159 IRAC+MIPS YSOcs, 14 of the original 17 IRAC-only YSOcs, and 56 (26 based on WISE and other photometric criteria) of the original 84 MIPS-only YSOcs were unlikely to be YSOs. Henceforth we refer to the list of vetted YSOcs, totaling 166, as YSOs to distinguish them from the raw unvetted list. While we have undergone an extensive process to construct a list of sources that are very likely to be YSOs, we stress that these YSOs have not been confirmed spectroscopically. Table 3 lists the final source counts for objects in the observed fields. The IRAC and MIPS fluxes of the IRAC+MIPS and IRAC-only YSOs are listed in Table 4. The 70 μm fluxes have been listed where available. (There are fewer YSOs with fluxes at 70 μm because of the lower sensitivity and, in some cases, the bright background.) The fluxes of MIPS-only vetted YSOs are listed in Table 5 with their WISE and MIPS fluxes (and IRAC fluxes where available). In Tables 4 and 5, we have noted which YSOs are in regions of low column density ($N_{\rm {H_2}} < 5 \times 10^{21}$ cm⁻²) according to the column density maps by Harvey et al. (2013), as these are more likely to be contaminants than YSOs in regions of high column density.

We compare our final YSO source list to those found for LkHα 101 in Gutermuth et al. (2009). All 103 YSOs in Gutermuth et al. (2009) are identified as sources in our catalog with positions that are within a couple tenths of an arcsecond agreement. Where this work and Gutermuth et al. (2009) provide fluxes, they agree at the shorter IRAC bands (IRAC1-3) typically within 0.05–0.1 mag. At IRAC4 and MIPS1, the agreement is typically within 0.2 mag. These differences are what one might expect for fitting a point-spread function (used here) versus aperture fluxes (used by Gutermuth et al. 2009) at wavelengths where there is substantial diffuse emission. (Recall that we have incorporated their data set into our own.) Therefore no previously identified sources have been missed in this study, and our measurements agree well with those of Gutermuth et al. (2009). Note, however, that the different classification methods used in this work and by Gutermuth et al. (2009) each yield a different total number of YSOs in this region; we have identified 42 YSOs whereas Gutermuth et al. (2009) identified 103. Our total breaks down into 7 YSOs identified here that were not identified by Gutermuth et al. (2009) and 35 YSOs shared between the two lists. (The c2d pipeline identified 47 YSOcs that were listed as YSOs by Gutermuth et al. (2009), but 12 were removed during the vetting process.) The major source of this discrepancy is that we require 4 (or 5) band photometry with a signal-to-noise ratio (S/N) ⩾3 in IRAC (and MIPS 24 μm bands) to identify YSO candidates. Such criteria are especially difficult to satisfy in the region of bright and diffuse emission around LkHα 101. Therefore, our results do not contradict those in Gutermuth et al. (2009); rather we believe that the stringent criteria used here have excluded some YSOs. We keep these criteria for consistency with other c2d and Spitzer GB survey observations and analyses, but note the limitations in such a bright region.

The diffuse emission problem is isolated to the immediate vicinity of LkHα 101. To demonstrate this point, in Figure 8 we have plotted the location of all the sources having an SED consistent with being a reddened stellar photosphere and an associated dust component, which do not have S/N ⩾3 at all IRAC bands. The SEDs of these sources are classified as "star+dust" in our catalog. Of the 56 YSOs listed by Gutermuth et al. (2009) that were not identified as YSOs in this work, the majority of them (34 of 56) have a "star+dust" SED. There is a total of 465 "star+dust" sources without robust four-band IRAC fluxes in the AMC field. These sources are relatively evenly distributed throughout the field, with the exception of a striking over-density at the center of LkHα 101 compared to other IRAC regions. Therefore, we believe this over-density is an effect of the difficulty in getting detections with S/N ⩾3 across four bands in the bright LkHα 101 region and not that there are significantly fewer YSOs than suggested by Gutermuth et al. (2009).

Zoom In Zoom Out Reset image size

Harvey et al. (2013) identified 60 YSOs in the AMC with Herschel/PACS, 49 of which are also identified in this work. Four of these Spitzer-identified YSOs are members of pairs of YSOs that are blended in the Herschel images. Herschel is more sensitive to the rising- and flat-spectrum sources, i.e., most (76%) of the other 45 Spitzer-identified YSOs that are also detected in the Herschel maps are Class I/F objects and the remaining 24% are Class IIs.

The YSOs are classified according to the slope of their SED in the infrared (see Evans et al. 2009 for a description). The spectral index, α, is given by

$$\begin{equation} \alpha \equiv {{d \log (\lambda S(\lambda))} \over {d \log (\lambda)}} \end{equation} \tag{ 1 }$$

and determined by fitting the photometry between 2 μm and 24 μm. The distribution of α values is shown in Figure 9 along with the relative number of YSOs in each SED class. The majority of YSOs identified in the cloud are Class II objects (55%). The percentage of sources in each SED class for the AMC is strikingly similar to that of Perseus (23%, 11%, 58%, and 8% for Class Is, Fs, IIs, and IIIs, respectively; Evans et al. 2009).

Zoom In Zoom Out Reset image size

Table 6 lists the breakdown of Class Is, Fs, and IIs for the AMC and other clouds in the GB and c2d surveys to estimate their relative ages. We did not include Class IIIs in this analysis since this population is typically incomplete in Spitzer surveys (e.g., see discussions in Harvey et al. 2008; Evans et al. 2009; Gutermuth et al. 2009) due to their weak IR excess. This simplifies the comparison to other clouds where the completeness limits may vary. We compared the ratio of Class Is and Fs to Class IIs, N_{I + F}/N_II, for the different cloud populations in other GB and c2d surveys which use the same classification scheme. We also include YSOs in the OMC identified with Spitzer by Megeath et al. (2012); since they use a different classification scheme, however, we have re-calculated the α values for their sample. The Class I/F lifetime is relatively short compared to the Class II lifetime, and therefore a higher ratio indicates a younger population (see discussion in Evans et al. 2009). The high number of Class Is and Fs suggests that the AMC is relatively young compared to other clouds.

Finally, we also compared the number of YSOs per square degree in the AMC (11.5 deg²) ¹⁹ to that in the OMC (14 deg²). The OMC is forming vastly larger amounts of stars. It has 237 YSOs per deg² whereas the AMC only has 13 YSOs per deg², a factor of about 20 fewer. Even if we only compare the number of YSOs in the OMC with four-band photometry (as this was the source of the discrepancy between the total number of YSOs around LkHα 101 identified in this work and by Gutermuth et al. 2009, who use a similar identification method to Megeath et al. 2012), this still suggests that there is at least a factor of 15 more YSOs in the OMC than in the AMC. Despite the differences in identification methods used for the OMC and for the AMC, it is clear that the OMC is forming far more stars than the AMC is. The YSOs in the OMC are also concentrated much more strongly than in the AMC, despite both clouds having comparable sizes and masses. We note that Lada et al. (2009) attribute the difference between the amount of star formation to the different amounts of material at high A_V /column density.

The first order SED parameter α is used as a primary diagnostic of the excess and circumstellar environment and to separate the YSOs into different "classes" (Section 3.2). Once we have a model of the stellar source, however, we are able to characterize the circumstellar dust better. For each source we determined the values of α_excess and λ_turnoff defined by Cieza et al. (2007) and Harvey et al. (2007) and used in many works since. λ_turnoff is the longest measured wavelength before an excess >80% of the stellar model is observed. If no excess >80% is observed, than λ_turnoff is set to 24 μm. α_excess is the slope of the SED at wavelengths longward of λ_turnoff. α_excess is not calculated for YSOs with λ_turnoff =24 μm  as there are not enough data points to determine the slope of the excess. These parameters provide a better characterization of the excess since α can include varying contributions from the stellar and dust components.

Figure 13 shows the distribution of α_excess and λ_turnoff for the Class IIs and Class IIIs. Class II and Class III YSOs with long λ_turnoff and positive α_excess  (YSOs 2, 24, 58, 64, 74, 102, 108, 113, 115, and 133 in the 8 μm bin and YSOs 145, 150, 162, and 165 in the 12 μm bin of Figure 13) are good classical transition disk candidates; the lack of near-IR excess but large mid-IR excess is a sign of a deficit of material close to the star within a substantial disk. Cieza et al. (2012) have recently done a study on the transition disks in the AMC, Perseus, and Taurus and identify six transition disk candidates in the AMC, three of which are also in our list of candidates (YSOs 58, 102 and 115). Of their remaining candidates, two were debris-like disks (YSOs 11 and 54) and the other was not identified in our YSO list. The larger distribution of α_excess for sources with longer λ_turnoff is consistent with distributions found for disk populations in other clouds (e.g., Cieza et al. 2007; Alcalá et al. 2008; Harvey et al. 2008; Merín et al. 2008).

Zoom In Zoom Out Reset image size

Figure 14 shows the ratio of the disk luminosities to stellar luminosities for the Class II and Class III sources. The disk luminosity is the integral of the observed excesses. (The excess at a given wavelength is calculated by subtracting the flux of the stellar model at that wavelength from the observed flux). The distribution of L_disk/L_star for Class II and III sources in the AMC is similar to that found for other c2d and GB surveys with Spitzer (Serpens: Harvey et al. 2007, IC 5146: Harvey et al. 2008, Chameleon II: Alcalá et al. 2008, Lupus: Merín et al. 2008, and the Cepheus Flare: Kirk et al. 2009). We find the Class III sources in the regions typically occupied by sources with passive disks and debris disks (e.g., 0.02 < L_disk/L_star < 0.08 for passive disks; Kenyon & Hartmann 1987). The low disk luminosity may be attributable to the lack of mid-IR excess at IRAC wavelengths in these sources' SEDs.

Zoom In Zoom Out Reset image size

It is possible that some of the Class III sources identified here are field giants. Oliveira et al. (2009) followed up on 150 Spitzer identified YSOs in Serpens and obtained 78 optical spectra with sufficient S/N. They showed that there were at least 20 giant contaminants in this list, 18 of which were identified as Class III sources. The more scattered spatial distribution of Class IIIs throughout the AMC is consistent with this idea that they are contaminants. Additionally, five of our Class III objects (YSOs 11, 141, 144, 148, 164) have very high luminosities (>100  L_☉). Four of these objects (YSOs 141, 144, 148, 164), as well as YSO 149 which is not of particularly high luminosity, are quite removed from the areas of high extinction toward the AMC (see Figure 15 in the following section) and are in regions of low column density ($N_{\rm {H_2}}< 5 \times 10^{21}$ cm⁻²; see Section 3.1).

Zoom In Zoom Out Reset image size

We employed a minimum spanning free (MST) analysis, the details of which are described in Masiunas et al. (2012), on the Class I, F, and II sources in the AMC to identify the densest regions of YSOs and the largest groups. (Class III sources are omitted to avoid the risk of including field giants, see, for example, Section 4.3). This analysis connects YSOs by the minimum distance to the next YSO to form a "branch" (Cartwright & Whitworth 2004). Figure 16 shows the cumulative distribution function (CDF) of the branch lengths between YSOs. This is used to determine the MST critical branch length, L_crit, that defines the transition between the branch lengths in the denser regions to the branch lengths in the sparser regions (Gutermuth et al. 2009). Therefore L_crit is based on relative over densities of objects. We measure an L_crit of 210'' for the AMC. Group memberships are defined by members which are all connected by branches of lengths less than L_crit. Figure 17 shows that we have extracted four groups with 10 or more members (marked by colored convex hulls) and three groups with 5–9 members (marked with magenta circles). Table 9 lists the properties of these groups. The position of the group is given by its geometric center. The group's effective radius, R_eff, defines the radius of a circle with the same area as the convex hull containing the group members. The maximum radial distance to a member from the median position gives R_circ, therefore a circle with this radius would contain all group members. Finally, the elongation of the group is determined by comparing R_circ to R_eff and represented by the aspect ratio, R_circ ²/R_eff ². The MST analysis on the full cloud recovers the clustering surrounding LkHα 101. The cluster subtends a larger area than that measured in Gutermuth et al. (2009) confirming their claim that there was star formation extended beyond their field of view. The star formation is mostly extended along the north–south direction of the cluster and therefore we measure a more elongated group than measured by Gutermuth et al. (2009). This is still the largest group in the AMC in terms of area and the number of members.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

As discussed in Section 3.1, our analysis is likely to have underestimated the number of YSOs in the region around LkHα 101. To check the consistency of our analysis with Gutermuth et al. (2009), we ran the MST analysis on both YSO lists within the Gutermuth et al. (2009) area of four-channel IRAC coverage. This leaves us with 41 of the YSOs presented here and 102 of those presented in Gutermuth et al. (2009). (There is one bright YSO in Gutermuth et al. 2009 that lies just outside their four-channel IRAC coverage to the south. It was only observed at IRAC1 and IRAC3.) We get an L_crit of 120'' for our cropped list of YSOs and an L_crit of 73'' for the cropped Gutermuth et al. (2009) YSO list. (Note that running the analysis on the cropped field, which is dense compared to the rest of the cloud, yields a smaller L_crit than when the analysis is run on the whole cloud. This is expected as L_crit is based on over densities, as discussed above.) The ratio of the L_crit values for the two YSO lists (73/120 = 0.61) agrees with our expectation that it should scale with the square-root of the density, and hence the cropped YSO count ($\sqrt{102/41} = 0.63$). Therefore we report that the derived properties are consistent with those measured by Gutermuth et al. (2009). (Differences are expected as shown by Gutermuth et al. 2009 with their comparisons among several shared regions.) However, the missing YSOs at the center of the cluster complicate any further comparison with their results.

We find 76% (113 of the 149) of the Class Is, Class Fs, and Class IIs are found in groups. Rather than compare the class fractions, given by N_{I + F}/N_II in Table 9, we directly compare the underlying distribution of α to determine whether the distribution of YSOs within groups is consistent with the whole cloud. We get the same result for each group: a K-S test on the α distribution of the group and the α distribution of the whole cloud shows that we cannot reject the hypothesis that they are drawn from the same sample (p-values > 0.13). (We also did a K-S test for each group with the extended population and found the same result.)

Similarly, we compared the properties of disks within groups and those not in groups by performing a K-S test on the distributions of disk luminosities (p-value of 0.08), α_excess (p-value of 0.9), and λ_turnoff (p-value of 0.9) and find no evidence that the two populations are drawn from different parent populations.