OPEN CLUSTERS IN THE MILKY WAY OUTER DISK: NEWLY DISCOVERED AND UNSTUDIED CLUSTERS IN THE SPITZER GLIMPSE-360, CYG-X, AND SMOG SURVEYS

Because of the high incidence of bright diffuse emission in the vicinity of many clusters, the original annuli used for evaluating the cluster candidates were deemed insufficient to serve as appropriate comparison regions for the steps of filtering the cluster CMDs by extinction (Section 2.4.2) and fitting them with isochrones (Section 2.4.4). To refine the CMD comparison, we chose a new circular region with the same radius as the cluster, situated nearby but placed carefully so that the "MW disk" stellar sequences in the CMD have the same apparent color range (which includes the intrinsic color plus foreground reddening) as the corresponding sequences in the cluster CMD. This allows us to identify the cluster sequence amongst a more reliable representation of the disk stellar population.

The 2MASS CMDs for each cluster represent the coaddition of several different stellar populations, including the Galactic disk in both the fore- and background, as well as the cluster itself. In addition to these distinct stellar sequences, these regions often harbor active star-forming sites that contain YSOs and other objects with atypical NIR or MIR colors. Thus to identify and fit the candidate star cluster properly, we take measures to isolate the true CMD signal of the cluster stars.

Typical statistical field star decontamination techniques (such as those applied by, e.g., Kharchenko et al. 2005; Piatti et al. 2003; Bica & Bonatto 2011), are not reliably effective for our sample, because the MSTO overdensities of the majority of the candidate clusters occur near the completeness limit of the 2MASS photometry (H ∼ 15), where the H and (J − K_s) measurements can be dominated by photometric scatter induced by the uncertainties. This effect is even stronger in the dereddened CMDs, where the uncertainties in the MIR photometry also play a role. Some of the cluster candidates are in regions with highly variable reddening, as evidenced by the presence of small (subarcminute) absorption and emission structures in the MIR or 2MASS images. The quality of the photometry, the potential for source confusion, and the robustness of stellar detection may thus be variable across the cluster field and between the cluster and comparison region.

In lieu of statistical field star decontamination, we adopt a technique that utilizes our additional leverage from the MIR imaging, via star-by-star reddening estimates (Section 2.3). For each pair of cluster and comparison regions (Section 2.4.1), we evaluated the distribution of RJCE reddening values, both spatially and in the CMDs. Since one criterion for including a cluster in the samples presented here is a prominent cluster CMD sequence distinct from the underlying disk population(s), we were able to identify an initial range of reddening that encompassed the greatest number of cluster stars while rejecting the largest number of disk stars, which are defined as those with colors, magnitudes, and reddening values equally well-represented in both the cluster and comparison CMDs. We checked our selected reddening range by inspecting the spatial distribution of reddening in the cluster and comparison regions and confirming both that the cluster stars within the adopted reddening range clump together and that stars beyond the adopted range have similar scatter in both the cluster and comparison maps. The width of the RJCE reddening "filter" is one of the parameters optimized in the isochrone-fitting iterations (Section 2.4.4).

In Figure 2, we illustrate this process for a dusty cluster field with a large range of reddening. Panel 2(a) contains the reddening distribution for stars within the cluster radius, shaded by separation from the adopted mean RJCE reddening for this field, 〈E(J − K_s)〉_RJCE ∼ 1.2. Panel 2(b) contains the CMD for this cluster region, with the points shaded on the same scale as the distribution in panel 2(a). Note how the cluster sequence stands out from the foreground disk stars with low reddening and the background, more heavily reddened disk stars (which may also include some YSOs). Panel 2(c) shows just this isolated cluster sequence with the best-fitting isochrone, as explained in Section 2.4.4. We note that even if YSOs are inappropriately removed from a cluster field because their reddening is overestimated (Section 2.3), the remaining stellar data are better matched to the theoretical isochrones we use to derive the cluster parameters because the isochrones do not include a YSO sequence.

After applying the RJCE reddening filter (Section 2.4.2, Figure 2), we make initial estimates of the key parameters for each cluster in our sample using underlying principles of single stellar populations. We emphasize that these are initial estimates only; the final values reported in Section 3 and Tables 1 and 3 are derived from the iterative isochrone-fitting approach outlined in Section 2.4.4.

The RC is an oft-utilized standard candle for stellar population studies dating back to, e.g., Stanek et al. (1994) and Paczynski & Stanek (1998). Groenewegen (2008) used the revised Hipparcos parallaxes to revisit the calibration of the absolute magnitude of the RC in the I and K_s bands and found M_Ks = −1.54 ± 0.04, without a significant trend with either (I − K_s) or metallicity. Though other studies have identified a relationship between the RC absolute magnitude and metallicity (e.g., Girardi & Salaris 2001; Salaris & Girardi 2002), the range reported (σ ≲ 0.28 mag) is sufficiently narrow for our purposes. For our clusters with a RC, we estimate the initial distance from the difference between the intrinsic M_Ks magnitude and the mean extinction-corrected K_s magnitude of the stars in the RC.

The magnitude difference ΔK_s between the RC and the MSTO can also be used to estimate the age of a stellar population, and this approach has been used extensively for clusters with optical CMDs (e.g., Phelps et al. 1994). Beletsky et al. (2009) extended this diagnostic to the NIR by deriving empirical relationships between age and ΔK_s for 14 open clusters with metallicities in the range −0.4 < [Fe/H] < +0.4. (Though only a small sample of clusters were used, the study found a negligible trend with metallicity.)

Thus for systems with a clear RC and MSTO, we estimate the age of the cluster using Equation (2) of Beletsky et al.:

$$\begin{eqnarray*} &&\log ({\rm age}) = 8.27 + 0.35 \times \Delta K_s. \end{eqnarray*}$$

For the subset of our cluster candidates with a clear RC, we estimate an age of ∼1 gigayear (Gyr)—i.e., we find ΔK_s ∼ 2.1 mag for nearly all of the clusters with a RC.

For those clusters without a clear RC, we have no means of estimating the age and distance independent of each other; some assumption must be made for either age or distance to estimate the other parameter using a comparison to theoretical isochrones. To simplify our task, we constrain the age of our isochrone suite to the typical age of those clusters with reliable age estimates, 1 Gyr. With the age fixed, we use theoretical isochrones (Girardi et al. 2002) to estimate an initial distance using the mean absolute magnitude of the MSTO over a range of metallicities, compared to the observed magnitude, and this estimate is then iteratively refined as described in Section 2.4.4.

We make an initial estimate of the reddening, E(J − K_s), through a comparison of the mean (J − K_s) color of the MSTO and RC (if available) and the characteristic color predicted by the isochrones, without any reference to the adopted RJCE reddening range (though we certainly expect them to be consistent). To estimate the intrinsic spread in color for our 1 Gyr isochrones, we evaluate the mean color, (J − K_s)_RC and (J − K_s)_MSTO, and the spread in color (i.e., the difference between minimum and maximum color), Δ(J − K_s)_RC and Δ(J − K_s)_MSTO, for the full range of −2.0 ⩽ [Fe/H] ⩽ +0.2. We find (J − K_s)_RC = 0.76 with a spread of Δ(J − K_s)_RC = 0.06 and (J − K_s)_MSTO = 0.49 with a spread of Δ(J − K_s)_MSTO = 0.17. These ranges are on the order of the average observed scatter at the MSTO magnitudes of our sample, and we iteratively refine the estimates as described in Section 2.4.4.

As in the initial parameter estimates (Section 2.4.3), we utilize the isochrones of Girardi et al. (2002) in the 2MASS photometric system for the more advanced determination of cluster parameters. For the approach outlined in this section, we include the isochrone suite's full metallicity range (−2.0 ⩽ [Fe/H] ⩽ +0.2, Δ[Fe/H] = 0.1 dex), at a fixed age of 1 Gyr. Older isochrones (at 2, 5, 8, 10, and 13 Gyr) were available to include if indicated by the ΔH or Δ(J − K_s) separation between the MSTO and RC for an individual cluster but were not found to be necessary (Section 2.4.3). These isochrones include stage designations for the main phases of stellar evolution, and to eliminate confusion in the fitting process, the scope of the isochrones was limited to include only the main sequence, subgiant branch, red giant branch, and the location of the RC.

The first step of the fitting procedure is to use the initial parameter estimates to shift each isochrone to the apparent H magnitude and (J − K_s) color of the identified cluster feature(s) in the CMD. For each star within 1.5 initial R of our cluster center, the H magnitude separation, (ΔH)_star, is computed as the distance between that star and the point on each reddened isochrone corresponding to the (J − K_s) color of the star. For each isochrone, (ΔH)_star is summed over all stars in the spatial region to find (ΔH)_isochrone, or the net H offset of that isochrone from our identified feature. We identify the "isochrone of best fit" for the assumed E(J − K_s) and (m − M) as the isochrone that has the smallest (ΔH)_isochrone. The isochrone of best fit is then overlaid on the (J − K_s, H) CMD, and a series of manual inspections are applied to iterate on the input E(J − K_s) and (m − M), as well as the applied CMD filtering parameters.

Several visual diagnostics of the cluster CMD and fitted isochrone are used to assess and iterate over five input parameters: R, (l, b), (m − M), E(J − K_s), and the range of E(J − K_s)_RJCE. First, we assess how well the fitted isochrone traces the visually identified stellar sequence and adjust E(J − K_s) and (m − M) accordingly. Second, we compare the MSTO and RC features in the CMD to those in the nearby comparison region's CMD. Strong overdensities associated with the cluster sequence confirm that the range of reddening values is reasonable and is not erroneously including or excluding regions of the CMD. Third, we adjust R and (l, b), based on the CMD-filtered spatial map, to produce the strongest MSTO and RC signal in the CMD. Once we are satisfied with our fits in all of these dimensions, we determine uncertainties on our values for E(J − K_s), (m − M), and [Fe/H] (Section 2.4.5).

We explored a variety of approaches to "weighting" the individual stars by their photometric uncertainties in the calculation of (ΔH)_isochrone. However, we found that all of these methods tend to produce poorer fits to the stellar CMD sequences, primarily because the fits are then dominated by the brightest stars, which are not necessarily the most likely cluster members. A proper weighting methodology would need to account for our stellar reddening selection, the likely type of each star (e.g., MS versus RC), and the CMD features specific to each of the clusters before it could emulate the multi-parameter visual assessment and iteration process used here.

Uncertainties on our fitting method are obtained for [Fe/H], E(J − K_s), and (m − M) using Monte-Carlo simulations of the method. These values are "measurement errors" only and do not include systematic errors in our procedure or biases in our cluster detection.

We create simulated datasets by varying the apparent J, H, and K_s magnitudes of each star in our cluster region by a Gaussian deviate, centered on the star's magnitude and color and with a width (σ) of the photometric uncertainty for each dimension. We generate 1000 realizations of each cluster dataset using this process.

The uncertainties in the fitted [Fe/H] are estimated by passing each of the simulated datasets independently through our isochrone fitting algorithm, with E(J − K_s) and (m − M) fixed at the final values from the procedure described in Section 2.4.4. This produces an output distribution of best-fit [Fe/H], from which we measure the standard deviation, add it in quadrature with the [Fe/H] grid spacing, and adopt the result as our best estimate of the measurement error on [Fe/H]. Uncertainties of our iteratively determined values of E(J − K_s) and (m − M) are estimated using a similar procedure—holding two parameters fixed while allowing the one under investigation to vary.

This procedure is performed for each of the clusters in our NIR+MIR sample, resulting in the distribution of uncertainties shown in Figure 3. For [Fe/H], the uncertainties varied from σ_[Fe/H] ∼ 0.1, meaning the same [Fe/H] was measured for all of the 1000 realizations, to σ_[Fe/H] ∼ 0.28, indicating considerably more degeneracy in the fitting procedure. The mean of the distribution of the uncertainties for all 16 clusters is σ_[Fe/H] = 0.17. Given that the isochrone grid has Δ[Fe/H] = 0.1, our method described here appears to reliably identify the best-fitting isochrones to the clusters. The uncertainty in (m − M) range from 0.26 to 1.49, with a mean of 0.71. The (m − M) grid used in estimating these uncertainties has a spacing of 0.05 mag. Finally, the E(J − K_s) uncertainties span 0.02–0.28, with a mean of 0.12 and a grid spacing of 0.01.

Zoom In Zoom Out Reset image size

As would be expected from its dusty surroundings (clearly visible in Figure 7(a)), this cluster has a heavily reddened main sequence with a few stars that are RGB candidates, though the RGB density is comparable to that of the nearby comparison region. Fitted with a 1 Gyr isochrone, the MSTO corresponds to a reddening of 1.20 mag in (J − K_s) and a distance of 1.69 kpc. The best-fit isochrone indicates a metal-rich cluster, with [Fe/H] ∼ 0.2, though we note that just under a third of the MS stars have local stellar densities (as determined from the IRAC catalog; Section 2.5) indicative of potential crowding effects on the 2MASS photometry.

This cluster also sits in a region of heavy reddening, with diffuse emission visible at MIR wavelengths (Figure 8); however, the contrast between the cluster and comparison regions suggests more localized extinction, with very few stars in the comparison region within the reddening range of the cluster. We note the very clear RC feature here, which tightly constrains the reddening and distance to E(J − K_s) = 1.27 and D = 1.36 kpc, with the best-fitting 1 Gyr isochrone having [Fe/H] ∼ 0.2.

Both a dense MSTO and RC-like feature characterize this object (Figure 9), even within a narrow reddening range of ΔE(J − K_s)_RJCE = 0.14. The RC is only ∼1 mag brighter than the MSTO, suggesting an age possibly younger than our 1 Gyr nominal isochrone. Our suite of models does not extend to younger ages, however, and using the 1 Gyr isochrone, we derive E(J − K_s) = 0.60, D = 3.50 kpc, and [Fe/H] ∼ 0.2. The large scatter in the RC cannot be explained by crowding effects, as all of the stars are comfortably isolated (Section 2.5), but we do note that nearly a third of the MS stars are potentially affected.

Though without a particularly strong overdensity in the MIR source map, this cluster has significantly overdense MS and RC features in the NIR CMD, in contrast to the comparison region (Figure 10). We note the presence of diffuse light in the cluster vicinity, which may be impacting the NIR photometry to produce the increased scatter in the RC feature. Despite this, the presence of such a feature constrains the cluster to E(J − K_s) = 0.60, D = 2.91 kpc, and [Fe/H] ∼ 0.2.

This cluster is visible in the source maps as a clear overdensity (Figure 11), and comparison of our candidate positions to the Dias et al. catalog yielded a potential match in Czernik 45 (Czernik 1966). However, the offset of 575 is larger than our original or best-fit cluster radii (34 and 4', respectively), Czernik's original size (15), and the calculated size from the only identified follow-up study to Czernik 45 (125; Tadross 2009). Furthermore, the 2MASS CMD presented by Tadross (2009) bears little resemblance to the CMD at our candidate position, including a complete lack of the RC stars we see at H ∼ 11.75. Given the coordinate, size, and CMD discrepancies, there does not appear to be sufficient evidence to identify this stellar overdensity as Czernik 45. Some scatter remains on the red side of the MS even after applying the reddening filter, but the MS and RC of this cluster correspond most closely to a stellar population with E(J − K_s) = 0.50, D = 3.04 kpc, and [Fe/H] ∼ 0.2.

The overdensity in the MIR images and source maps (Figure 12) is weaker here than in many of the other clusters, but the CMD demonstrates a much stronger MS than the comparison region, along with ∼4 possible RC stars. We note that there is considerable scatter at (J − K_s) ∼ 1.0, but this is present in both the cluster and comparison CMD. The MS and RC of this cluster correspond to E(J − K_s) = 0.45, D = 3.11 kpc, and [Fe/H] ∼ 0.0.

This object is included in our sample due to its significant source count overdensity and MSTO (Figure 13). However, this MSTO is located very close to the limits of the 2MASS photometry, and the residual scatter in the RC/RGB part of the CMD prevents us from placing reliable constraints on the metallicity. We are able to derive both a reddening and a distance: E(J − K_s) = 0.40 and D = 4.81 kpc.

This object, shown in Figure 14(a), appears to have a significant amount of diffuse emission in the MIR image, which also appears (though much weaker) in the DSS-Red and 2MASS images. It is also coincident with a 6 cm emission radio source (GB6 B2010+3635; Gregory et al. 1996) and an IRAS source (IRAS 20103+3633). Odenwald (1989) noted the presence of several YSOs near the IRAS source, but we could not find any identification of a stellar cluster at this position.

This object (Figure 18) does have a coordinate match in the Dias catalog as "IC 1848," with an offset of 14, but it is unclear from the rest of the literature that a detection of the same open stellar cluster has been made here. Many papers refer to IC 1848 as an H ii region or star forming region alone (e.g., Chauhan et al. 2011; Koenig et al. 2012). Hillwig et al. (2006) refers to IC 1848 as an open cluster, dominated by the star system HD 17505 at the center, and find a distance of 1.9 kpc. This system is within 07 of our observed overdensity center, though Hillwig et al. (2006) identify IC 1848 with OCL 364 (Ruprecht et al. 1981), which lies ∼485 away. The cluster radius given in the Dias catalog, 9', appears better-matched to the rim of diffuse emission surrounding our much smaller (∼4') grouping of stars, and we are not able to identify the source of the catalog's distance (2 kpc) to determine whether this is a photometric stellar distance or a kinematical distance of the nebula. The presence of the bright HD 17505 system in the center makes the stellar cluster nearby very difficult to discern in shorter wavelength images, such as 2MASS and the DSS, but in the Spitzer image (Figure 18(a)), a large number of fainter, densely packed stars are clearly visible.

FSR 0442 (Figure 27) was identified in the same location in the MIR source images but with a larger radius (7') than that published in its discovery (4') in the automated cluster search by Froebrich et al. (2007). Based on the RJCE extinction estimates and the MIR source maps, we isolated a strong MS and fit a 1 Gyr isochrone to the CMD features with D = 1.84 kpc, E(J − K_s) = 0.70, and [Fe/H] ∼ −0.1.

A cluster of stars was identified at (l, b) = (114617, 0222) in the center of a large region of gaseous emission (Figure 28(a)). We identified this cluster as BDS 2003 46, cataloged by Bica et al. (2003), who used 2MASS images to look for new clusters around the central positions of optical and radio nebulae. They identified an "Infrared Group" at the same position as our star cluster, measured the size and ellipticity of the group, and recorded a kinematical distance using the nebula in which the cluster is embedded (Sh2-165; 1.6 kpc). From our isochrone-fitting procedure, we derive E(J − K_s) = 0.95, D = 3.76 kpc, and [Fe/H] ∼ −0.1.

The distance is ∼2 kpc greater than that of Bica et al. (2003), though no distance uncertainties were reported in that study. However, the method cited (Downes et al. 1980) reports distance uncertainties on the order of 1–2 kpc, so the difference may not be that highly significant. (This assumes that the cluster is truly coincident with the gaseous nebula.) Kinematic distances in the vicinity of outer disk spiral arms are often offset from photometric distances due to non-circular gas motion associated with the arms. Generally the kinematical distances in this part of the Galaxy, which includes the Perseus spiral arm, exceed the photometric ones by a kpc or more, particularly on the near side of the arm (where a nebula at 1.6 kpc would lie), though exceptions are known where the kinematic distance underestimates that obtained by other means (Humphreys 1976; Moisés et al. 2011).

A properly reddened isochrone at D = 1.6 kpc placed in the CMD in Figure 28(c) has a MSTO ∼0.4 mag brighter than could be attributed to the stellar data (excluding the four bright stars near (J − K_s) ∼ 1.5), so this kinematical distance appears genuinely inconsistent with the stellar cluster. However, the vertical scatter in Figure 28(c) suggests that the isochrone fit may not be sufficient to determine this object's parameters unambiguously—if the brighter blue stars are genuine cluster members, then perhaps a smaller distance is more appropriate. But treating those stars as cluster members produces a considerably worse overall match to the isochrone, so in Table 3, we report the parameters derived for the fit shown in Figure 28(c). However, we note that the bright diffuse emission in this region may be causing problems with the NIR and MIR photometry, despite our photometric quality restrictions.

We initially identified a stellar overdensity at (l, b) = (120067, 1037) with R = 44 in the MIR images but optimized the cluster sequence shown in Figure 29 by adjusting the coordinates to (120067, 1050) with R = 50. This sequence corresponds to a population with D = 2.80 kpc, E(J − K_s) = 0.48, and [Fe/H] ∼ −0.2. This position was matched to that of the cluster FSR 0494 (Froebrich et al. 2007), which was originally identified with a much smaller radius (R = 08).

We note that there is a density peak corresponding to the Froebrich et al. object within our cluster radius (Figure 29(b)).

An embedded cluster was identified in the MIR source maps at (l, b) = (138041, 1515) with a radius of R = 364 (Figure 30). We matched this detection to SAI 24, discovered by Glushkova et al. (2010), and fit a 1 Gyr isochrone with D = 1.57 kpc, E(J − K_s) = 0.95, and [Fe/H] ∼ −0.2 to the strong MS feature.

Despite the seemingly very good fit (Figure 30(c)), we find a puzzling discrepancy between the RJCE reddening estimates for the cluster stars and the NIR reddening needed to fit the isochrone. The comparison between these two reddening estimates for the entire set of fitted objects is explored in Section 4 and Figure 6, in which SAI 24 is a clear outlier. However, the explanation invoked for the other minor discrepancies—variations in the MIR extinction law—is insufficient for SAI 24. Assuming a constant NIR extinction law (i.e., A_J/A_Ks and A_H/A_Ks constant), the discrepancy in SAI 24's reddening values implies a MIR extinction ratio A_[4.5μ]/A_Ks of ∼0.9 ± 0.14, almost a factor of two higher (and beyond the uncertainties) than has been measured in even the most "extreme" cloud cores (Román-Zúñiga et al. 2007; Flaherty et al. 2007). Allowing for the NIR extinction law variations observed in both diffuse and dense regions (e.g., throughout the diffuse interstellar medium and toward the Galactic center; Nishiyama et al. 2009; Zasowski et al. 2009; Stead & Hoare 2009) only worsens the problem.

Reconciling the reddening values within the uncertainties is barely possible if the stellar locus observed at (J − K_s) ∼ 1–1.2 is a very blue (metal-poor) RGB instead of a MS, but beyond the reddening offset, there is no evidence in support of this theory, and numerous lines of evidence against it (such as the shape of the CMD stellar locus, the location of the cluster within a dust cavity, and the massive size implied by such a well-populated RGB). Spectroscopic follow-up to determine the stellar properties of the cluster members will shed valuable light on this puzzle.

In the MIR images, a cluster was identified at (l, b) = (150661, −0608) with R = 41 (Figure 31) and matched to FSR 0665 (Froebrich et al. 2007). We isolate a highly reddened sequence of stars absent from the comparison region and fit a 1 Gyr isochrone with D = 1.47 kpc, E(J − K_s) = 1.20, and [Fe/H] ∼ −0.1.

Like SAI 24 (Section 3.3.4), FSR 0665 has a discrepancy between its reddening as estimated from the RJCE method and that estimated from the fitted NIR isochrone, which cannot be explained by reasonable variations in the extinction law (nor does it lie in a visibly dusty region likely to have extreme variations). However, the CMD at this position (Figure 31(c)) inspires less confidence in the best-fit reddening, which is reflected in the relatively large reddening uncertainty of 0.10 mag. In fitting this cluster, given the absence of a RC, we were strongly guided by the overdensity of brighter stars not apparent in the comparison CMD (i.e., the stars with (J − K_s) ∼ 1.4 and H ∼ 12.5) that lie to the red side of the color distribution of cluster candidate members. Emphasis on this brighter overdensity is consistent with the fitting procedure applied to all of the clusters, and we report the best-fitting values as our best estimates for this cluster; however, we acknowledge that the high level of scatter in the CMD makes the fit less secure, and further observations (spectroscopy or deeper photometry) of this cluster may well confirm the membership of some of the bluer stars and thus revise the cluster reddening value downward to be more consistent with the RJCE-derived stellar reddening values.

A cluster was identified in MIR images at (l, b) = 168292, 1440) with an estimated radius of R = 398 (Figure 32). This detection matches well the published coordinates of Czernik 20 (Czernik 1966).

However, we noticed some differences between our NIR CMD (Figure 32) and those published in the literature. We resolve the differences by both carefully inspecting the literature and exploring the environs of the candidate cluster.

Czernik 20 was first identified by Czernik (1966) in the images of the Palomar Sky Atlas. The later studies by Babu (1989) and Sujatha et al. (2006) determined cluster parameters via isochrone-fitting, but both authors expressed concern that the original cluster identified by Czernik (1966) was actually a redetection of NGC 1857 (Cuffey & Shapley 1937). NGC 1857 is offset from Czernik's original published position by ∼8', and both Babu (1989) and Sujatha et al. (2006) selected the stars at this offset position in their analyses instead of using the published coordinates of Czernik 20. However, our examination of the Spitzer images identified a cluster at Czernik's (1966) original coordinates, prior to any knowledge of the Czernik object. Furthermore, the NIR CMDs at the two positions of Czernik 20 and NGC 1857 are quite different from each other, but both show distinct cluster-like stellar loci (Figures 32(c) and (d)). The independent detection of Czernik 20, combined with the obvious differences in the CMDs at the two positions, indicate that there are in fact two distinct star clusters close together on the sky. See Figure 4 for a comparison of the positions and sizes identified by Czernik (1966), by Sujatha et al. (2006), and independently in this work.

Zoom In Zoom Out Reset image size

To augment the existing confusion over the location of Czernik 20, an isochrone with Sujatha et al.'s (2006) cluster parameters (which were ostensibly fit to the stars at the location of NGC 1857) is actually significantly better matched to the stars at Czernik's original coordinates for Czernik 20. After applying our reddening filter and isochrone-fitting procedure, we find comparable reddening but a smaller distance (D = 3.10 kpc and E(J − K_s) = 0.35), with [Fe/H] ∼ 0.0.¹²

With regard to NGC 1857, we identified a distance and reddening cataloged in Battinelli & Capuzzo-Dolcetta (1991, D = 1.9 kpc and E(J − K) = 0.2); however, when comparing isochrones of these parameters with the CMD of NGC 1857's position, we observed a substantial mismatch. The reddening later reported by Colegrove et al. (1994) for this cluster, E(J − K) = 0.36, only increases the mismatch. An isochrone with the distance, reddening, and metallicity from Battinelli & Capuzzo-Dolcetta (1991) appears to be closely matched to the Galactic disk stars, not to the clear cluster MS we observe. It is this locus that we fit in Figure 33(c), with D = 1.40 kpc, E(J − K_s) = 0.07, and [Fe/H] ∼ 0.0. We note that while our distance is somewhat similar to that given in Battinelli & Capuzzo-Dolcetta (1991), the discrepancy in reddening (coupled with the very narrow color spread in the MS in Figure 33(c)) rules out the possibility that we are fitting the same stellar sequence. This apparent disagreement emphasizes the benefit of exploring open clusters with multi-wavelength IR datasets. Namely, the increased sensitivity of the (J − K_s) color to stellar temperature, and the decreased sensitivity of MIR colors to interstellar dust, can break degeneracies between derived parameters based on intrinsic color and reddening, adding considerable leverage to understanding these low density stellar clusters.

The visual cluster search identified a stellar grouping at (l, b) = (173046, −0160) with an estimated radius of R = 44 (Figure 34), spatially coincident with FSR 0777 (Froebrich et al. 2007). We note the presence of bright emission nearby, coincident with a peak in the source counts map (Figure 34(b)), but careful inspection of the image itself does not indicate the presence of a genuine stellar cluster associated with this emission. We find a best-fit 1 Gyr isochrone with [Fe/H] ∼ 0.0 and E(J − K_s) = 0.55 at a distance D = 3.11 kpc. Despite being removed from the visible emission region, the cluster still has a large E(J − K_s) and a large ΔE(J − K_s) of 0.49 mag. We note that there is some (J − K_s) scatter in the cluster CMD not present in the comparison region (Figures 34(c) and (d)) that cannot be reduced by further restricting the E(J − K_s) range. Some of these differences could be due to the rather large angular separation of the comparison region, which was specifically chosen to avoid the widespread MIR emission, but we suspect much of the scatter in Figure 34(c) could be image artifacts or spurious detections associated with the nearby nebulosity.

We identified a stellar overdensity at (l, b) = (20079, 0625) with R = 27 (Figure 35) and matched it to FSR 0979 (Froebrich et al. 2007). The overdensity highlighted in the NIR CMD is not particularly strong (Figure 35(a)), but the MS is more densely populated, corresponding to a cluster at a distance D = 2.90 kpc, E(J − K_s) = 0.50, and [Fe/H] ∼ 0.2.