What's Behind the Elephant's Trunk? Identifying Young Stellar Objects on the Outskirts of IC 1396^*

We obtained an 87.3 ks observation of the western outskirts of IC 1396 beginning on UT 2015 December 1 (XMM-Newton observation 0762360101). This observation was designed to observe the outer part of the cluster behind the "base" of the Elephant's Trunk (the region highlighted in blue in Figure 1), specifically for purposes of identifying Class III sources. The observations used the medium-thickness optical blocking filter. We used the pipeline reduction of these data produced as part of the full reprocessing of the XMM-Newton archive for the 4XMM catalog in 2019. ⁴ Because there were no particularly bright targets in the center of the field of view, we do not incorporate data from either RGS camera. The reduction detects 152 X-ray sources.

A square degree of IC 1396 was monitored with the Wide-Field Camera (WFCAM) on the United Kingdom Infrared Telescope (Casali et al. 2007) over 21 epochs from 2014 July 18 to 2016 July 12, first presented in Meng et al. (2019). While that work looked at variability between epochs, we use here the combined photometry across all 21 epochs, which allows us to obtain deeper photometry than that provided by the Two Micron All-Sky Survey (2MASS; Skrutskie et al. 2006). Seeing ranged from 06 to 14, with average seeing of 079. We refer the reader to Meng et al. (2019) for details on the observations and reduction of this data. The pipeline-reduced data, which we use here, are available in the database WSERV9v20170222, at the WFCAM Science Archive ⁵ (Hambly et al. 2008).

We noticed in our comparison of data that the use of the star/galaxy probabilities provided by the WFCAM pipeline to eliminate sources without high probabilities of being stars (as was done in Meng et al. 2019) produced an underdensity of sources identified as stars along the path of the dust bands observed in the WISE data, without a corresponding drop in the number of sources overall along the dust band but with a drop in a number of total sources behind the cometary globule at the center of IC 1396A. This most clearly appears when comparing the ratio of objects that are classified as stars in each spatial bin to the total number of objects in each bin, shown in Figure 2. This suggests that the automated classification of star versus galaxy was affected by the dust (either due to foreground dust reddening the stars into colors more typically found in galaxies, or background dust producing a less defined point source). We therefore make no cuts based on the automated classification of each object in the UKIRT/WFCAM pipeline. We note the duplication of some sources in the overlap between regions observed at different times due to imperfect source matching; these observations typically had consistent (within uncertainties) photometry with each other at very low separations and came from different frame sets, suggesting that the sources were multiple detections of the same source. To eliminate these, we matched the UKIRT data set to itself with a search radius of 05, below the typical spatial resolution of the observations, and for each set of duplicated sources selected one source identifier as the "true" source.

Zoom In Zoom Out Reset image size

We retrieved photometry of the region observed in the UKIRT data from the early third data release from ESA's Gaia mission (Gaia EDR3; Gaia Collaboration et al. 2016, 2021), the AllWISE catalog of data from the Wide-field Infrared Survey Explorer (WISE; Cutri et al. 2021), and various pointed observations with the Spitzer Space Telescope (Werner et al. 2004), obtained via the Spitzer Heritage Archive.

We use the UKIRT/WFCAM and 2MASS data as the backbone of our target matching. We searched the Spitzer Catalog for point sources matching the coordinates of our near-IR sources with a search radius of 1'', chosen based on the typical 08 seeing of the UKIRT data and the Spitzer point-spread function, yielding 19,744 sources. We then crossmatched the near-IR data set with Gaia EDR3 and AllWISE using the CDS xMatch functionality as implemented in astroquery (Ginsburg et al. 2019), again using a search radius of 1'' for consistency with the Spitzer crossmatch.

Finally, we used the CDS xMatch service to match the XMM and UKIRT/2MASS objects, with a match radius of 2'' because the point-spread function (PSF) of XMM-Newton is considerably wider than the UKIRT and 2MASS PSFs. We examined by hand each of the matches between the XMM-Newton and UKIRT sources to determine its relevance to our work (e.g., identifying the UKIRT source most likely to correspond to the X-ray data in cases where there were two or more UKIRT sources matched to the same X-ray source). The XMM field is aligned such that some of the field falls outside the region for which we have WFCAM data; we thus do not consider the six X-ray sources that fall outside of this range.

There are 47 X-ray sources in the XMM-Newton observation without UKIRT/WFCAM counterparts. While most of these are likely extragalactic in origin, several are among the brightest X-ray sources, suggesting that they are Galactic in origin. To check if these X-ray sources had near-IR counterparts too bright for UKIRT/WFCAM, we also crossmatched the XMM-Newton data to 2MASS. However, we found that all XMM/2MASS matches in the UKIRT field also had UKIRT matches, suggesting that UKIRT saturation was not the cause of the lack of a near-IR counterpart.

To compare our sample to other studies of this region, we also crossmatched X-ray sources from Chandra/ACIS-I images of the cometary globule IC 1396A at the tip of the elephant's trunk (Getman et al. 2012) to the UKIRT data, again using the CDS/Xmatch functionality. For convenience, we refer to this field as IC 1396A for the remainder of the paper. While this field does not cover the core of the Tr 37 cluster (which presumably formed out of the IC 1396 molecular cloud), it is more interior to the cluster than the IC 1396-West field. We do not examine X-ray data from the cluster core (Mercer et al. 2009) because these Chandra data have diminished sensitivity due to the presence of the Chandra High-Energy Transmission Grating Spectrometer (HETG-S) and because the UKIRT field does not fully overlap with the cluster core. Considering just the IC 1396A field and the IC 1396-West fields provides near-complete coverage of the Elephant's Trunk from tip to where it meets the larger cloud, as depicted in Figure 1. We used the crossmatches to 2MASS identified by Getman et al. (2012) for those X-ray sources that did not have a UKIRT counterpart.

We estimated the distance to the cluster using Gaia EDR3 parallaxes to cluster members that have previously been identified via a variety of methods (Contreras et al. 2002; Sicilia-Aguilar et al. 2005b, 2006a, 2013; Reach et al. 2004; Morales-Calderón et al. 2009; Barentsen et al. 2011; Sicilia-Aguilar et al. 2006b; Mercer et al. 2009; Getman et al. 2012), following the methodology of Sicilia-Aguilar et al. (2019). As with Sicilia-Aguilar et al. (2019), we selected only those sources that had matching radii <05, relative parallax uncertainties with a ratio of parallax uncertainty to parallax σしぐま_ωおめが /ωおめが < 0.1, and proper-motion uncertainties below 2 mas yr⁻¹, to ensure that the Gaia source was the same as the source in the existing catalog. We also required that for those objects that had both Gaia DR2 and EDR3 parallax measurements, the estimated distance from EDR3 is consistent with its estimated distance from Gaia DR2; this ensures that any differences in our estimate relative to Sicilia-Aguilar et al. (2019) are due to incorporating sources with new or improved-quality parallax measurements, rather than potential mismatches with Gaia EDR3. We estimated distances following the prescription of Bailer-Jones (2015), using a characteristic length scale of l = 1.35 kpc in the prior based on the direction of observation (Bailer-Jones et al. 2018).

We found an average distance to the cluster of ${931}_{-116}^{+159}$ pc, where the uncertainties indicate the 5th and 95th percentiles. This measurement is consistent with both the Sicilia-Aguilar et al. (2019) estimate of ${945}_{-73}^{+90}$ pc ⁶ and the previous value of 870 pc from Contreras et al. (2002). We note that our work here specifically does not incorporate additional information beyond a literature identification of cluster membership and a parallax measurement; future work that incorporates more information from Gaia (e.g., proper-motion measurements) will likely produce a more precise result. However, this distance estimate will be sufficient for our purposes: excluding objects with distances that are inconsistent with cluster membership from our sample.

A total of 213 sources with Spitzer photometry have αあるふぁ_S > −1.6 from Spitzer. Of these, 174 have αあるふぁ_S values that correspond to Class II YSOs, while 25 have slopes corresponding to Class FS and 14 have slopes corresponding to Class I. We list these in Table 2. One of these sources is also detected with XMM-Newton.

Three hundred forty-five sources with mid-IR data from Spitzer or WISE have SED slopes corresponding to a bare photosphere (αあるふぁ ≤ −1.6) and Gaia-parallax-based distances that place them at the correct distance for cluster membership, while an additional 491 have SED slopes corresponding to a bare photosphere and no parallax measurement. Of these, 15 also have X-ray detections with XMM-Newton, making them Class III YSO candidates. We list these in Table 2. The others require additional observational data, such as deeper X-ray observations or the detection of lithium in their spectra, to determine their youth.

A total of 288 sources have αあるふぁ > −1.6 from AllWISE. We checked each of these sources that did not also have Spitzer data to determine whether the source was likely to be blended, contaminated, or confused, based on whether it was matched to multiple UKIRT sources using the 1'' search radius. We excluded any UKIRT sources for which the same AllWISE source was matched to multiple UKIRT sources. This leaves 153 sources with αあるふぁ_W > −1.6 that we believe to be uncontaminated. Of these, 72 have SED slopes corresponding to Class I; 26 have SED slopes corresponding to Class FS; and 55 have SED slopes corresponding to Class II. Two of these sources (one Class FS YSO, and one Class II YSO) are also detected with XMM-Newton. These are listed in Table 2.

We expect that there will be some differences between detections with WISE and Spitzer, due to differences in the WISE and Spitzer PSFs, observation strategies (Spitzer provides deeper observations at the expense of spatial coverage), and data reduction (e.g., background subtraction).

To check whether we could reasonably combine estimates of αあるふぁ from Spitzer and AllWISE, we considered the 105 sources in IC 1396-West that are detected with both telescopes. Of these, 99 have αあるふぁ in each telescope that corresponds to the same YSO class. Two Class II sources in Spitzer have Class III slopes in WISE, and one system has a Class II slope in Spitzer but a Class FS slope in AllWISE. Most notably, three sources with αあるふぁ_S corresponding to Class III have αあるふぁ_W corresponding to disk-hosting classes—i.e., three Class III sources are misclassified as disk hosts using WISE data. If we assume similar statistics hold for the rest of the field, ${95.2}_{-2.5}^{+1.7} \%$ of YSO classifications in the AllWISE-only sample would be classified the same way by Spitzer, indicating that we can reliably compare results from each telescope despite their differences.

Fifteen UKIRT/WFCAM sources with mid-IR matches and αあるふぁ < −1.6 are detected with XMM-Newton, as are three UKIRT/WFCAM sources with mid-IR matches and αあるふぁ > −1.6. These are listed in Table 2. This yields a disk fraction for X-ray-detected cluster members of ${17}_{-7}^{+10} \%$, comparable within uncertainties to both the ${29}_{-3}^{+4} \%$ claimed by Getman et al. (2012) for the IC 1396A field between IC 1396-West and the core, and to the ${20}_{-7}^{+10} \%$ disk fraction for X-ray-detected cluster members in the core of the Tr 37 cluster by Mercer et al. (2009). The uncertainties on these disk fractions were determined using the Wilson score interval for uncertainty in binomial proportions (Wilson 1927) with z = 1. In summary, between disk-hosting sources that we detect in the archival data and the Class III sources detected with XMM-Newton, we identify 381 new YSOs in IC 1396 with mid-IR data.

Many factors limit our ability to effectively identify all cluster members in the region where IC 1396-West overlaps with the UKIRT/WFCAM data. Higher-mass stars are intrinsically brighter than their lower-mass counterparts, making them more detectable in X-rays. This is compounded by the suppression of X-ray emission due to accretion in disk hosts, as higher-mass stars dissipate their disks more quickly than their lower-mass counterparts (e.g., Carpenter et al. 2006). An X-ray-selected sample is thus biased toward higher-mass stars, which are more likely to be diskless, resulting in a disk fraction lower than the true disk fraction for the region.

Simply identifying additional disk hosts provides another observational bias. The SED of a disk host is fairly conspicuous even in faint mid-IR data, such that we can detect a disk around a low-mass star just as readily as around a high-mass star. Similarly, a disk indicates youth just as readily as X-ray emission, so we can readily identify disk-hosting low-mass stars without X-ray emission. However, only including sources with either an X-ray detection or a disk biases our search heavily in favor of finding a disk fraction larger than the true disk fraction, as this would include the low-mass disk hosts while excluding the low-mass non-disk hosts.

Additionally, the effects of the surrounding nebula could lead to misidentification of an embedded cluster member as a background extragalactic source. We can illustrate this by considering the case of a hypothetical single 1.0 M_⊙ YSO at 1000 pc (within the expected range for cluster distance, but further toward the back of the cluster), at varying levels of extinction (and thus varying degrees of interstellar absorption). We use the colors of a 1.0 M_⊙ pre-main-sequence star from Pecaut & Mamajek (2013) and assume an unabsorbed flux for a 1.0 M_⊙ weak-lined T Tauri star from the L_X −M relationship of Telleschi et al. (2007), following an APEC model with $\mathrm{log}(T)=7.05$. We assume for illustrative purposes the relationship between column density N_H and A_V from Güver & Özel (2009): N_H = (2.21( ± 0.09) × 10²¹)A_V . In Table 3, we present five cases for this source: no foreground material, the typical cluster extinction per Sicilia-Aguilar et al. (2005b), and A_V = 5, 10, and 20, corresponding to varying levels to which the source is still embedded in its natal envelope. As these demonstrate, a YSO could be faint enough to go undetected in V-band observations, and on the borderline of detection in the UKIRT/WFCAM K band, while still producing detectable X-rays. This could account for some portion of detected X-ray sources without a near-IR counterpart.

More subtly, relying on Spitzer data leaves us only with data in the specific pointings available. While it has more spatial coverage, AllWISE is also magnitude limited. Heavily extincted sources may thus go undetected in IR, while their X-rays are detected as hard sources without an IR counterpart and interpreted as a background AGN. The wide PSF of AllWISE (∼6'' at W2; Wright et al. 2010) also means that inevitably several potential good disk candidates are removed from consideration due to confusion and contamination (see Section 3.2).

Astrometry, independent of X-ray and IR emission, could potentially be used to identify cluster members—ideally, we would be able to map the bulk motion of known cluster members and then algorithmically find objects with similar motions. However, Gaia EDR3 is only complete to a magnitude limit G = 17 (Gaia Collaboration et al. 2021), with a poorly defined faint limit beyond that; if we conservatively assume incompleteness below G = 17, an extinction A_G ≃ 2, and a distance to the cluster of 931 pc, Gaia data will be incomplete for cluster members with M_G ≳ 5.2. This would bias the sample both toward targets at the front of the cluster (i.e., at distances < 931 pc) due to their higher apparent brightness and minimized extinction from the nebula, and toward intrinsically brighter (i.e., high-mass) sources.

While the issues all combine to provide an incomplete picture of the cluster in this region, it is worth noting that past surveys should suffer from similar issues for similar reasons. We thus conclude that our search with X-rays is broadly comparable to past X-ray-based searches for cluster members in IC 1396. For the purposes of estimating the disk fraction of cluster members, we focus on the X-ray-detected cluster members, as this is the surest way to ensure that we are comparing similar subpopulations of the cluster. We acknowledge that this disk fraction estimate is at best a lower limit on the true disk fraction of the cluster, given the number of low-mass stars we do not consider here; however, it is the best estimate available with the data we have in hand. While we do not incorporate the disk-hosting YSOs we identify that are not detected in X-rays into our disk fraction estimate, we present their identification here so that a more accurate disk fraction can be computed as more complete data sets become available.

Only 18 of the 152 X-ray sources in IC 1396-West are identified clearly as disk-hosting or non-disk-hosting cluster members. While 6 of the remaining 134 are not included because they do not overlap with the UKIRT/WFCAM observations, it behooves us to consider the remaining 128 sources.

Forty-seven XMM sources in the UKIRT field have neither a UKIRT nor 2MASS counterpart within 2''. We expect that these are likely either extragalactic sources, or cluster sources embedded deep enough within the nebula that they are undetected in UKIRT due to extinction. For the full XMM-Newton/EPIC field of view, we would expect ∼78 extragalactic background sources with fluxes >2.5 × 10⁻¹⁴ erg s⁻¹ cm⁻² (Moretti et al. 2003); our number of 47 unmatched sources in the UKIRT/XMM overlap region (more than half of the full field of view) is consistent with that number. We do not detect as many sources that can be interpreted as extragalactic background sources as were found in the IC 1396A field (Getman et al. 2012); we expect that this is due to the apparent higher column density of the nebula in IC 1396-West compared to the IC 1396A field (Figure 1), the higher background level for XMM-Newton compared to Chandra, the larger PSF for XMM-Newton resulting in faint background sources going unresolved, and the lower sensitivity threshold we find for the XMM-Newton observations compared to the Chandra observations (Section 4.4).

Fifty-five XMM sources have UKIRT counterparts that either do not have matches to Spitzer or AllWISE sources or do not have a valid slope solution. Of these, 16 have Gaia-parallax measurements that put them outside the cluster, while 2 have parallaxes that put them within the cluster and 38 have no measured parallaxes. These 40 sources (two with in-cluster parallaxes, 38 without measured parallaxes) are likely cluster members, bringing the total number of YSOs we identify to 421, but we lack sufficient data to determine if they host disks or not.

Twenty-three XMM sources have UKIRT counterparts and matches to Spitzer or AllWISE, but have Gaia parallaxes that put them outside the cluster. These are a mix of 12 foreground sources, 2 background sources, and 9 sources where the range between the 5th and 95th percentile values of the distance distribution exceeds 1000 pc. The nine sources with highly uncertain distance estimates are almost certainly background sources; however, in some cases, their highly uncertain distance estimates overlap with the distance estimate for the cluster, so they are treated separately.

This leaves 20 sources with UKIRT matches, Spitzer or AllWISE data matched to the UKIRT source, and a parallax that does not exclude them from the cluster. In two of these cases, the AllWISE match is matched to more than one UKIRT source, and these are thus excluded (as in Section 3.2), to produce our final sample of 18.

It is worth noting that the distributions of the near-IR (K-band) brightness of the X-ray-detected cluster members with and without mid-IR data are distinct. The sources with mid-IR data are in general much brighter—the median UKIRT/WFCAM K magnitude of the sources with mid-IR data is 12.539, while the median K magnitude for the sources without mid-IR data is 18.304. This is not overly surprising, as one would expect brighter sources to be those detected with WISE in the regions where Spitzer data is unavailable. It does, however, indicate that the sources without mid-IR data are likely of lower mass than the sources with disks, likely are at the far side of the cluster rather than the near side, and thus likely are more embedded in the nebula as well.

To evaluate these X-ray sources independently of their match to UKIRT, we computed hardness ratios for each source, using the broad bands defined in the XMM-Newton pipeline (band 6 = 0.2–2 keV; band 7 = 2–12 keV):

$$\begin{eqnarray}&&{HR}=\displaystyle \frac{{B}_{H}-{B}_{S}}{{B}_{H}+{B}_{S}},\end{eqnarray} \tag{ 1 }$$

where B_H,S are the count rates in the hard and soft bands, respectively. These are shown in comparison to the WISE image in Figure 5.

Zoom In Zoom Out Reset image size

As expected, foreground sources are the softest sources. The background sources are not exclusively hard, though this varies positionally, with softer background sources appearing at areas of minimal dust emission, as expected for Galactic background sources. The cluster members show a wide variety of hardnesses; some are among the hardest sources in the field. These appear to be spatially coincident with the star-forming nebular cloud; harder cluster members typically fall along the dust band, indicating a higher degree of embeddedness. The unmatched sources vary in hardness as well, with some soft sources appearing in regions of low nebular emission, as well as the harder emission expected of extragalactic sources. The appearance of a set of soft unmatched sources along the nebula as it branches to the northwest (starting at approximately R.A. 21_h33_m42_s and decl. $57^\circ 32^{\prime} 06^{\prime\prime}$) suggests a potential additional population of cluster members that are not detected in the infrared data, either due to extinction from the nebula or intrinsic faintness.

We find that only ${17}_{-7}^{+10} \%$ of likely cluster members in the IC 1396-West field with X-ray detections and available mid-IR data have disks. While this is not statistically different from the ${29}_{-3}^{+4}$% found by Getman et al. (2012) or the ${20}_{-7}^{+10} \%$ found by Mercer et al. (2009), it qualitatively suggests that the disk frequency diminishes as a function of distance from the cluster core, as found for the ONC by Getman et al. (2014), though some of this difference is likely due to the difference in X-ray sensitivity (Sections 4.1 and 4.4). While our statistics are not robust enough to provide a quantitative analysis of how the disks are distributed, we can qualitatively evaluate the distribution of disks through IC 1396-West, as depicted in Figure 4, keeping in mind the selection effects of the observed distribution due to the targeted nature of the Spitzer observations. We can also qualitatively assess the distribution of X-ray-identified likely cluster members through the nebula in conjunction.

We clearly see that disks tend to fall along the dust bands in the AllWISE data, which is borne out in the central Spitzer field (which had coverage in both I1 and I2)—more sources are identified in the central pointing of the Spitzer data, which falls on the nebula, and they generally tend to be disk sources. We note that the higher density of sources detected by Spitzer is due to its deeper observations—the faintest Spitzer sources in this field are an order of magnitude fainter in observed brightness than the faintest AllWISE sources.

There is a void of disk hosts and non-disk sources in the northern portion of the XMM-Newton field, save for a small group of disk sources in the upper corner that are not detected in X-rays, despite seemingly low nebular emission compared to the main "trunk." As fewer disks are expected in regions of lower nebular emission, the presence of these sources is intriguing. The group of disks seems potentially connected to a cluster of three X-ray sources without IR data (one in the cluster, two others unmatched to UKIRT), suggesting a filament of ongoing formation.

Surprisingly, many of the AllWISE-detected disk sources along the trunk are classified as Class I, the youngest type considered here. This suggests potential ongoing new star formation, extending the "triggered star formation" identified by Getman et al. (2012) to a farther distance from the cluster core. This supports the hypothesis of Pfalzner et al. (2014)—it will take longer for Class I YSO disks to evolve into Class II YSOs and then eventually dissipate their disks than it will take Class II YSO disks to dissipate, so the Class I disks identified here should remain present (albeit evolving) for an extended period of time.

There is also an intriguing band of X-ray-detected cluster members tracing the gap between the southern edge of the main trunk of the nebula and a smaller dust cloud. This range includes an X-ray-detected Class III source in a Spitzer field. The relative overabundance of X-ray-detected sources in this belt compared to the dust band above, combined with the higher concentration of disk sources along the belt, suggests that these sources are likely older, having already formed, evolved, and dissipated their disks. Similarly, the disks identified in this region are primarily Spitzer-detected Class II YSOs, suggesting a more evolved state of evolution than the Class I sources found along the dust band.

Pfalzner et al. (2014) hypothesized that disk fractions would be higher outside of the core of a massive cluster than in the cluster core, due in part to the enhanced contributions of external photoevaporation on the dissipation of the disk. However, Tr 37 and IC 1396 have an O-star binary at the core (HD 206267), which should contribute both to disk photoevaporation and dissipation of the nebula. We can see the nebular dissipation effects in the "bubble" appearance of the background gas and dust (Figure 1). However, the disk fraction we find here at a minimum indicates no increase in the disk fraction farther away from the cluster core. While our data are not robust enough to completely disprove this hypothesis, they do suggest that if there is a disk dissipation gradient, it is in the opposite direction of what Pfalzner et al. (2014) hypothesize.

To determine the X-ray sensitivity of the XMM-Newton observation presented here, we converted the observed X-ray fluxes to luminosities using distances derived from Gaia parallaxes. Only 15 of our identified X-ray-detected cluster members have measured parallaxes, including 13 with mid-IR data; for the other 43 sources, we make the assumption that they are at the estimated cluster distance of 931 pc. The X-ray source with the fewest counts detected of the X-ray-detected cluster members (i.e., sources with both near-IR and X-ray detections and are not outside the cluster per Gaia) has X-ray flux ∼1.3 × 10⁻¹⁵ erg s⁻¹ cm⁻² and distance ${914}_{-159}^{+329}$ pc, corresponding to an absorbed $\mathrm{log}{L}_{{\rm{X}}}\approx 29.11$. For comparison, the faintest source with a determined L_X in Getman et al. (2012) has an unabsorbed $\mathrm{log}{L}_{{\rm{X}}}=28.99\pm 0.17$ and an $\mathrm{log}({N}_{H})=20.26\pm 0.32$, which yields an absorbed $\mathrm{log}{L}_{{\rm{X}}}\approx 28.95$. This indicates that the XMM observations are less sensitive than the adjacent Chandra field by 0.16 dex. We present our absorbed X-ray luminosity function (XLF) for both sources with mid-IR detections and the full set in comparison to the Getman et al. (2012) XLF in Figure 6. We also compare these XLFs against those of the Orion Nebula Cluster (the "cool unobscured" sample of cool stars with minimal optical extinction and X-ray absorption from Feigelson et al. 2005) and IC 348 (Stelzer et al. 2012).

Zoom In Zoom Out Reset image size

Based on the comparison of our XLF to that of Getman et al. (2012), we estimate that the IC 1396-West sample is complete only to an unabsorbed $\mathrm{log}{L}_{{\rm{X}}}\approx 30.8$, compared to the $\mathrm{log}{L}_{{\rm{X}}}\approx 30.5$ completeness level in the IC 1396A field. Using the L_X −M relation of Telleschi et al. (2007), this yields a minimum stellar mass of 1.8 M_⊙, well above the 1.0 M_⊙ of Getman et al. (2012) and thus explaining the qualitative discrepancy between X-ray-detected disk fractions for the two data sets; indeed, Ribas et al. (2015) note a clear difference in the rate of protoplanetary disk evolution (and thus disk lifetime) for >2 M_⊙ stars compared to <2 M_⊙ stars. We also note the similarity of the position of the XLF for both this observation of IC 1396 and that of Getman et al. (2012) to the XLF for IC 348, albeit not as complete at low L_X . We note that based on the exposure time for this XMM-Newton observation, to achieve completeness down to $\mathrm{log}{L}_{{\rm{X}}}\approx 30.5$ would require a 174.5 ks observation, twice the length of the observation presented here.