A NEW Hα EMISSION-LINE SURVEY IN THE ORION NEBULA CLUSTER

2.1.1. Slitless Grism Spectroscopy

We observed the ONC, centered on the Trapezium, with the Wide Field Grism Spectrograph–2 (WFGS2) installed on the University of Hawaii 2.2 m telescope, on the five nights of 2010 December 31, 2011 January 2, 2011 January 25, 2011 February 27, and 2011 October 16. We covered an area of almost 1 deg² centered on R.A. (2000) = 5^h35^m164 and Decl. (2000) = −5°23'25'' with a mosaic of 25 overlapping fields. We used a 300 line mm⁻¹ grism blazed at 6500 Å and providing a dispersion of 3.8 Å pixel⁻¹ and a resolving power of 820. The Hα filter had a 500 Å passband centered near 6515 Å. The detector for WFGS2 was a Tektronix 2048 × 2048 CCD, whose pixel size of 24 μm corresponded to 0.34 arcsec on the sky. The field of view was 115 × 115. For each field, we took a short, 60 s exposure in order to detect the Hα line in bright stars and avoid saturation of the central parts of the nebula. Then for 21 fields we obtained two frames with 480 s exposure time, and for the remaining four fields three frames of 300 s exposure.

Bias subtraction and flat-field correction of the images were done in IRAF.³ Then we used FITSH, a software package for astronomical image processing⁴ (Pál 2012), to remove cosmic rays, coadd the long-exposure images, identify the stars on the images, and transform the pixel coordinates into the equatorial coordinate system. A representative example of the reduced and coadded images can be seen in Figure 1.

Zoom In Zoom Out Reset image size

We examined all spectra visually for presence of the Hα emission line and found 587 stars with Hα emission in the whole observed area. We associated our emission sources with Two Micron All Sky Survey (2MASS; Cutri et al. 2003) point sources. We found matches unambiguously within 2'' for all but seven stars. Due to small separation, three pairs of Hα stars are associated with single 2MASS sources (J05342073-0530461, J05342207-0501342, and J05355535-0547017). Another Hα star, V1504 Ori, has no associated 2MASS point source. Since each of these stars can be found in the UKIDSS–DR8 Galactic Clusters Survey (Lawrence et al. 2007), we label them with their UGCS identifiers.

The equivalent width (EW) of the Hα emission line was computed as follows: first, small stamps with size of 30 × 30 pixels were cut around the core of the emission line. The size of this stamp was derived from the FWHM of the profile and the characteristic lengths of the structures in the grism frames and found to be stable throughout our observations. In each stamp, the background level (which varied significantly due to the intrinsic background of the Orion Nebula complex) was obtained from the upper 30 × 5 and lower 30 × 5 pixels by fitting either a constant level or a plane, depending on the variation level of the background. We found that in most cases, both fits yielded a smooth background; however, in some cases the two methods led to different results. In such cases we accepted the more "stable" solution. The uncertainty of the background level was determined by a simple statistic, i.e., from the fit residuals divided by the square root of the number of involved pixels. Second, these background-subtracted stamps were stacked via the y axis (i.e., almost exactly perpendicular to the grism dispersion), yielding 30 pixel long spectra. The uncertainty of each spectral point was estimated from the photon noise statistic (taking into account the effective camera gain 1.780 e⁻ ADU⁻¹) and by adding quadratically the scaled background scatter. Third, these spectra were fitted with a Gaussian, parameterized by the four free parameters of zero-point level, center, amplitude, and standard deviation. The EW of the emission line was then simply derived from these fitted parameters, i.e., by dividing the integral of the Gaussian (without the zero-point level) by the zero-point level (which is proportional to the instrumental continuum level). The uncertainty of the EWs was obtained from linear error propagation using the fitted results.

Applying the above described method, we could determine the EWs for 452 stars. For the remaining stars in which the Hα emission was clearly detected, this method could not be applied due to faint continuum or the bright and variable nebulous background. In such cases we estimated the EW using the IRAF "splot" package. However, the EWs could not be determined for 28 stars. Due to overlapping fields, 50 of the stars were observed more than once.

Figure 2 shows the distribution of our Hα emission stars over the observed area centered on the Trapezium. Asterisks indicate the classical T Tauri stars (CTTSs), diamonds mark the weak-line T Tauri stars (WTTSs), while triangles represent sources of uncertain nature (see Section 3.1). Figure 3 shows the histogram of EWs.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

2.1.2. Photometry

Direct images of each field of the mosaic with the same instrument were obtained through r' and i' filters immediately before the spectroscopic exposures. One exposure was taken at each position and in each filter with integration time of 60 s. Aperture photometry was carried out with the PHOT task in the DAOPHOT package. Several stars from the SDSS Photometric Catalog, Release 8 can be found in every frame. In order to transform instrumental magnitudes into the Sloan Digital Sky Survey system, we applied those stars and the following transformation equations:

$\begin{equation*} r_{{\rm SDSS}}=C_{1} \times r^{\prime }_{{\rm instr}}+C_{2}\times {\rm airmass}+C_{3}, \end{equation*}$

$\begin{equation*} r_{{\rm SDSS}}-i_{{\rm SDSS}}=C_{4}\times (r^{\prime }-i^{\prime })_{{\rm instr}}+C_{5}\times {\rm airmass}+C_{6}. \end{equation*}$

The C₁...C₆ constants are derived from the fitting procedure. Photometric errors of typically 0.1 mag were derived from quadratic sums of the formal errors of coefficients of the transformation equations. Figure 4 shows the histogram of the r' magnitudes of our Hα emission stars, and the r' versus r' − i' color–magnitude diagram is displayed in Figure 5.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

To characterize the nature of the Hα emission stars and investigate possible correlations between the EW of the Hα line and other stellar properties, we used data available via the VizieR catalog access tool. Equatorial coordinates were determined using the 2MASS All-Sky Catalog of Point Sources (Cutri et al. 2003) as reference. JHK_s magnitudes for all but five stars are found in the same catalog. I magnitudes can be found in the DENIS (Epchtein et al. 1997) data release, and mid-infrared (W1 = 3.4, W2 = 4.5, W3 = 12, and W4 = 22 μm) data in the WISE All-Sky Data Release (Cutri et al. 2012). Spitzer IRAC data can be found for 345 of our Hα emission stars in Megeath et al. (2012). Spectral types for 346 stars can be found in Da Rio et al. (2009), Rebull et al. (2000), Rebull (2001), and Hillenbrand (1997).

A total of 496 of our stars are located within the area observed by the XMM-Newton X-ray observatory, and 257 of them were detected. Their X-ray data can be found in the XMM-Newton Serendipitous Source Catalogue 2XMM-DR3 (Watson et al. 2009). Rotation period measurements are available for 214 stars in Herbst et al. (2002), Cieza & Baliber (2007), and Flaccomio et al. (2005).

CTTSs are pre-main-sequence stars with an accretion disk. Material from the circumstellar disk accretes onto the star via magnetic funnel flows. This accretion process causes broad, often asymmetric Hα emission lines. A WTTS can be either a TTS that evolved and no longer has much circumstellar material to accrete from or a TTS that for some reason is not accreting much at the moment. Their Hα emission lines are much weaker, and the origin of the line is mostly chromospheric. There are several methods to distinguish accreting and non-accreting young stars from each other. The most widely used method (e.g., Martín 1998) is based on the EW of Hα and classifies each young star with EW(Hα) > 10 Å as a CTTS (accreting). Barrado y Navascués & Martín (2003) defined an empirical classification scheme that takes into account the spectral type as well, while White & Basri (2003) distinguished accreting and non-accreting TTSs based on the width of the Hα emission lines.

To distinguish accreting and non-accreting stars, we applied the criteria by Barrado y Navascués & Martín (2003) for the 346 stars having spectral types in the literature (Da Rio et al. 2009; Rebull et al. 2000; Rebull 2001; Hillenbrand 1997) and adopted EW = 13 Å as a boundary between accreting and non-accreting stars for the unclassified population, which is fainter on average than the classified subsample (〈J〉 = 12.39 and 〈J〉 = 13.06 for the classified and unclassified group, respectively). This EW value corresponds to the saturation limit of chromospheric activity at M3–M4 type.

We list the 587 sources in Tables 1–3. Table 1 contains the 372 stars classified as CTTSs, Table 2 lists the 187 WTTSs, and Table 3 contains the 28 stars without measured EWs. Column 1 of each table gives their 2MASS designation of the star. The newly identified Hα emission stars are marked by an asterisk following their 2MASS identifier. EWs, measured on the dates given in the column head (yyyymmdd), are listed in Columns 2–6, Column 7 gives the r' magnitudes, Column 8 contains the r' − i' color index, while the last column contains cross-identification from the General Catalogue of Variable Stars (Samus et al. 2007–2012), Wiramihardja et al. (1991), Hillenbrand et al. (1998), and Fűrész et al. (2008).

To test whether our WTTS and CTTS samples indeed represent physically different objects, we applied a discriminant analysis. It is a common ingredient of professional statistical packages. We used SPSS⁵ in the computations. Discriminant analysis aims to find differences between groups in the multivariate parameter space, orders membership probabilities, and one may use this scheme for classifying additional cases not having assigned group memberships. During this process, we used those variables in which a significant number of cases remain after removing missing data. To fit with this requirement, we used the I − J, J − H, H − K_s, K_s −[W1], [W1]−[W2], [W2]−[W3], and [W3]−[W4] color indices, available for 241 of the 372 CTTSs and 134 of the 187 WTTSs. Using the analysis, we look for the linear combination of the selected color indices, the discriminant variable, which gives maximal separation between the groups of the cases.

The program calculated the means and standard deviations of quantities included in the analysis, along with the number of cases having valid measurements in all of the variables used in the computation. The results are shown in Table 4. A quick look at the table gives an impression that both the means and standard deviations are greater in the CTT group. The statistical tests show that, except for [W2]−[W3] and [W3]−[W4], the difference between the WTT and CTT groups is highly significant.

We found that the strongest contribution to the discriminant variable is given by [W1]−[W2], while those of [W2]−[W3] and [W3]−[W4] are negligible. From an astrophysical point of view this result can be interpreted in the following way. The spectral energy distributions (SEDs) of both the WTT and CTT classes are built up as a superposition of the photospheric and circumstellar thermal emission over the wavelengths covered by the near- and mid-infrared colors. At shorter wavelengths the contribution of the photospheric emission is the strongest and the opposite is true at the other end of the wavelength domain of the input variables. The wavelengths of W1 and W2 are between these extremes. The larger average I − J and J − H color indices of the CTT group possibly indicate greater extinction suffered by these stars, since infrared excesses are less significant in these colors.

Based on the canonical discriminant function (the discriminant variable) resulting from the analysis, one can control the reliability of the original (a priori) classification based on the EW of the Hα emission line. The histograms of the discriminant function for the WTT and CTT groups demonstrate that these groups are not disjoint, but rather overlapping with respect to zero, the a posteriori boundary between the two types of TTSs. We found that 25 of our 187 WTTSs have IR colors similar to CTTSs, and 90 of the 372 CTTSs exhibit infrared properties similar to WTTSs.

The discriminant function allowed us to estimate the probable types of the 28 unclassified stars. Although the 2MASS (J,H,K_s) colors of the unclassified sources are complete, a significant fraction of the WISE measurements are missing. The Expectation and Maximization (EM) algorithm of the Missing Value Analysis procedure, implemented in the SPSS statistical package, was used to generate a maximum likelihood estimation for the missing values. The discriminant analysis suggests that 13 stars from the 28 unknown sources are WTTSs, while the remaining 15 sources can be classified as CTTSs.

We compare our sample of Hα emission stars with those detected by the Kiso survey (Wiramihardja et al. 1991), Fűrész et al. (2008), and Da Rio et al. (2009) in Figure 6. The Kiso survey and the survey by Fűrész et al. (2008) covered the whole area of our survey, while the photometric Hα survey of Da Rio et al. (2009) covered just the central 34' × 34'. Due to the bright nebulosity, we could barely detect sources in the densest central region of the ONC, and for the few detected sources EW(Hα) could not be determined, therefore comparison with previous Hα surveys is justified only outside the central area of ∼5' in right ascension and ∼10' in declination.

Zoom In Zoom Out Reset image size

Based on the VizieR database, 92 Hα stars identified by the Kiso survey are located in the area covered by our observations. Out of these 92 stars 79 show emission in our images. Comparison of these two surveys suggests that most of the stars detected by Wiramihardja et al. (1991) were strong emitters. A total of 763 stars observed by Fűrész et al. (2008) are located in the area of our survey. Among those, we detected Hα emission in 337 stars. A detailed comparison shows that, outside the central region, we detected most stars for which Fűrész et al. (2008) measured EW(Hα) >10 Å, and their low-EW stars, when detected, appeared as WTTSs in our survey as well.

A total of 274 of the 638 Hα emission stars identified by Da Rio et al. (2009) coincide with our emission stars. Figure 6 shows that most of our Hα emission stars, located within the area observed by Da Rio et al. (2009), are found in their database. Those of their Hα stars that were missed by our survey are either located within the central 5' × 10' area or are fainter than V ∼ 19 mag.

A total of 99 of our Hα emission stars were detected by neither the above surveys nor earlier surveys. The distribution of these stars can be seen in Figure 7. Most of them are located outside the central, thoroughly studied region of the ONC. Forty-eight stars belong to the CTTS group, 44 to the WTTS group, and the remaining 7 stars to the third group of uncertain nature. Several of them are known variable stars or cluster members. In particular, 15 of the near-infrared variables studied by Carpenter et al. (2001) appear in this sample. Figure 8 shows the H versus H − K_s color–magnitude diagram of the newly identified Hα emission line stars, based on the 2MASS data. Diamonds indicate M-type cluster members identified by Hillenbrand et al. (1998). The diagram suggests that the new emission line stars may be pre-main-sequence stars located at the distance of the ONC.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The most complete census of the disk-bearing population of the ONC can be found in the Spitzer database (Megeath et al. 2012), whereas the diskless young stars down to 0.1–0.2 M_☉ were identified by the Chandra Orion Ultradeep Project X-ray survey (Preibisch et al. 2005a, 2005b). To test whether our Hα emission stars represent a typical subsample of the cluster population or not, we examined the number and ratio of the infrared and X-ray sources in the 17' × 17' field of view of Chandra, excluding the central 5' × 10' region where the nebulosity prevented us from detecting emission-line stars. There are 758 X-ray-selected young stars in this area, 369 (48.7%) of which are IR-selected disk-bearing stars. We detected 101 Hα emission stars, some 13% of the whole known young stellar population in the same area. This sample contains 86 CTTSs and 15 WTTSs. Our data show an apparent CTTS fraction of 85%, indicating that our survey, as expected, is biased toward disk-bearing stars.

Since both infrared excess and Hα emission in young stellar objects (YSOs) are disk indicators, overlapping of the samples of Hα emission stars and the infrared-excess stars of the same region is expected. In Figure 9 the surface distribution of our stars is displayed together with the mid-infrared excess stars identified by Megeath et al. (2012), based on Spitzer observations. It can be seen that numerous sources are in common, but of course there are sources that were detected in only one of the surveys. Both Figures 6 and 9 show that outside the cluster core our results overlap well with those of other surveys and reveal new cluster/association members in the outer, lower density regions of the ONC. Although the r' versus r' − i' color–magnitude diagram (Figure 5) indicates a limiting magnitude of r' ∼ 20 mag, Figure 4 shows that our sampling starts to be fairly incomplete beyond r' ∼ 16–17 mag.

Zoom In Zoom Out Reset image size

A look at the surface distribution of the Hα emission stars in Figure 2 gives the impression of clustering of the stars. The surface distribution of the merged YSO samples (Figures 6 and 9) also suggests the presence of small apparent stellar groups in the outer regions of the field, sampled by our survey. We examined the reality of the clusterings in our sample by an appropriate statistical procedure. To test our impression, we generated a fully random case with the same numbers of CTTSs and WTTSs as in the real case and used the Fast Nearest Neighbor Search Algorithms and Applications (FNN) package of the R project⁶ to prove the reality of the clustering in the surface distribution. We compared the cumulative distributions of the nearest neighbor distances in the real and random samples applying a Kolmogorov–Smirnov (K-S) test. The results of K-S statistics indicate that the distribution of the Hα emission stars in the observed area indeed shows some clustering with respect to the fully random surface distribution.

To study the nature of the clustered and distributed population and reveal the possible difference between their measured properties, we first defined clusters in the surface distribution. Based on the maximum departure of the real and random distribution, we assigned an object to a cluster if its fourth nearest neighbor had a distance of d < 130''. Comparison of the cluster members and outliers has shown that the mean brightness of the cluster stars is greater at all wavelengths. The difference is increasing toward the longer wavelengths. It is about 0.5 mag in the near-infrared and reaches about 1 mag at W4. For testing the significance of the difference between the corresponding groups, we applied a statistical T-test. According to the applied T-test, these differences are highly significant. In contrast, however, there is no significant difference in EW(Hα).

Regarding the cluster members, a T-test demonstrates that the difference between the CTTSs and WTTSs in the cluster environment is smaller than the standard error of the mean in the K_s brightness. In the J band, however, the CTT class is fainter. This may be accounted for by the greater extinction of the circumstellar matter or, alternatively, may indicate that the weak Hα emission of the fainter WTTS population remained undetected.

The strength and shape of the Hα emission line can show temporal variability due to variable accretion rate and wind. EW(Hα) can change even in the absence of flux variations, due to the photometric variations of the stellar continuum. The subsample of Hα sources observed several times due to overlapping of the image fields allowed us to examine the short timescale (hours to months) variability. Seven stars were observed twice within a few hours. They showed significant short timescale variations in the Hα EW. Of these seven stars, six possess a flat or rising (Class I) SED (Figure 10). This suggests that these stars are very young and may still be surrounded by an envelope, and this can lead to significant variability of the accretion rate and accretion-related wind. Nine of the 11 stars that were observed twice or more within a few days showed no significant variation, and the remaining 2 exhibited small (some 30%) variations in EW(Hα). Stars observed on longer timescales (some weeks or months) show a wide range of variations in EW(Hα) (see Table 1).

Zoom In Zoom Out Reset image size

We compared the derived EWs with those published by Fűrész et al. (2008) and Da Rio et al. (2009). This comparison allowed us to study the changes of the EWs on timescales of a few years. Part of the observed differences may originate from the fact that the EWs in the compared works were determined with different methods. To check this point, we selected all of the sources common among these three surveys (Figure 11) and calculated the significance of correlations using Kendall's tau statistics. Based on the statistics, we find that the surveys of Fűrész et al. (2008) and Da Rio et al. (2009) are less correlated, while our survey and the survey of Da Rio et al. (2009) are well correlated. The comparison between the measurements of Fűrész et al. (2008) (high resolution) and our survey (low resolution) suggests a quite good correlation at lower EW values.

Zoom In Zoom Out Reset image size

The results are presented in Figures 12 and 13. The histogram with upward diagonal hatching (red in the online version) represents the CTTSs, while the histogram with horizontal hatching symbolizes the WTTSs. It can be seen that the amplitudes of variations are moderate in the greater part of the sample: most of the observed changes in EW are within a factor of two to three, whereas changes as high as ten- to twenty-fold appear in a few cases. The greater EW ratios in Figure 13 reflect the lower level of correlation between these data sets.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The classical and weak-line subclasses of TTSs are generally correlated with the infrared signature of the accretion disks. Figure 14 is the J − H versus H − K_s two-color diagram for the stars listed in Tables 1–3 with the colors from 2MASS Point Source Catalog. This figure shows that only a few of the low-EW stars (diamonds) exhibit near-infrared excess, while a significant part of the high-EW population (asterisks) and a few stars without EW data (triangles) are distributed over the region occupied by reddened stars associated with accretion disks. Stars that are lying below the CTTS locus and right from the zero-age main sequence can be explained by the circumstellar disk models of Lada & Adams (1992). The models take into account the surface temperature, disk inclinations, and radial temperature profiles. The modeled systems with stellar surface temperature from 3000 to 12,000 K can explain this part of the two-color diagram. Our sample mainly consists of stars of K and M spectral type, but we detected Hα emission in four known Herbig Ae/Be stars too.

Zoom In Zoom Out Reset image size

Figure 15 shows the WISE [3.4]–[4.6] versus [4.6]–[12] color–color diagram of our CTTS and WTTS sample. The boundaries dividing Class I, Class II, and Class III sources (Koenig et al. 2012) are indicated. Comparing this plot with the similar diagram of the Taurus star-forming region (Koenig et al. 2012) shows that the [4.6]–[12] color indices of our stars are unusually high. The reason for this may be that due to the wide point-spread function of WISE and the strong background of the Orion Nebula at 12 μm, fluxes measured in the W3 band are seriously contaminated by emission from small nebular clumps or faint neighboring stars (Koenig et al. 2012). To test this hypothesis, we examined the positions in the WISE [3.4]–[4.6] versus [4.6]–[12] color–color diagram of all Chandra sources (Getman et al. 2005) of the ONC without Spitzer detection. The strong X-ray emission and lack of mid-infrared excess ensure their diskless young star nature. We found that these X-ray-selected cluster members also exhibited high [4.6]–[12] color indices, supporting that the 12 μm flux probably originates from the environment. The [3.4]–[4.6] colors indicate that most of our CTTSs are Class II infrared sources. Most of the WTTSs exhibit no excess in these bands, but, in contrast with the case of the J − H versus H − K_s diagram, a significant subgroup of them (37 stars) exhibit quite large excesses in the [3.4]–[4.6] color.

Zoom In Zoom Out Reset image size

Based on the WISE [3.4]–[4.6] color (Koenig et al. 2012) or, where they are available, Spitzer IRAC [3.6]–[4.5] colors (Gutermuth et al. 2008), we identified a subsample of low-EW Hα emission stars exhibiting infrared excess indicative of an optically thick accretion disk. We checked the VizieR catalogs and omitted the stars having known neighbors within the angular resolution of the WFGS2 instrument. The stars, selected in this manner and listed in Table 5, are apparently not accreting, in spite of their primordial accretion disks (Figure 16). Most of them are known variables (see Table 2). Comparison of their EWs, measured by various surveys and shown in Columns 4–6 of Table 5, shows that 7 of the 37 stars exhibited much stronger Hα emission in 2005, at the epochs of the observations of Fűrész et al. (2008) and Da Rio et al. (2009), that is, their accretion and wind activities are highly variable on a few-year timescale.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

One star, exhibiting an extreme EW value around 500 Å in Figure 3, deserves attention. It is V421 Ori, associated with the proplyd 280-931 (Ricci et al. 2008). Da Rio et al. (2009) also measured very strong emission, EW(Hα) = 420 Å from this object, in which the emission from the externally ionized disk contributes to the observed Hα.

To test whether the apparent clusters and voids in the surface distribution reflect patchy interstellar extinction, we plotted the cluster members and outliers, defined by their nearest neighbor distances, with different symbols onto the 500 μm Herschel SPIRE image of the Orion region, downloaded from the Herschel Science Archive. The data are part of the Guaranteed Time Key Program "Probing the origin of the stellar initial mass function: A wide-field Herschel photometric survey of nearby star-forming cloud complexes" (Herschel Gould Belt Survey; PI: P. André; André et al. 2010). The structures in the 500 μm radiation of the cold dust correlate well with those seen in the extinction maps of the region (Rowles & Froebrich 2009; Scandariato et al. 2011). We chose the far-infrared image to display the surface distribution of the absorbing dust since its angular resolution is higher than those of the available extinction maps. Figure 17 suggests that variable extinction alone cannot account for the surface inhomogeneities of the stars. An apparent void can be seen next to the Trapezium where the bright background prevented us from detecting any emission line. The clustered stars (those having at least four neighbors within 216) are found near the cold dust structures, suggesting that they are probably younger than the distributed population, and are closely associated with their natal clouds. Figure 17 suggests that the lower average brightness of the distributed stars cannot result from a greater extinction along their line of sight.

Zoom In Zoom Out Reset image size

The difference in brightness between the clustered and distributed stars suggests that we observe two populations of pre-main-sequence stars that differ from each other in both age and location in space. The lower average brightness of this scattered population suggests greater distance, higher age, or lower average mass of these stars. Since the presence of the massive molecular cloud, associated with the ONC, efficiently blocks the background stars, the greater distance is less probable than the older age and lower average mass. The increase in the brightness difference toward longer wavelengths points to a greater amount of circumstellar matter in cluster members with respect to the field stars, indicating a younger age of the clustered population. The presence of two overlapping populations toward the line of sight of ONC was suggested by the radial velocity survey of Fűrész et al. (2008) and is demonstrated by the recent work of Alves & Bouy (2012). The distribution of our Hα emission stars also reflects the presence of two, probably unrelated populations of pre-main-sequence stars.

4.3.1. Hα Equivalent Width and Infrared Excess

The study of the slope of a young star's SED offers a great possibility to characterize the disk population of young stars. This method was applied, e.g., by Lada et al. (2006) and Teixeira et al. (2012). They measured the slope of the SED between 3.6 μm and 8 μm and, using this so-called α_IRAC index, sorted the disks into an evolutionary sequence. To get an idea of the evolutionary stages of the disk population in our sample, we applied this method for the stars having Spitzer data in the literature. Figure 18 shows that most of our CTTSs have −1.8 < α_IRAC < −0.5, characteristic of optically thick accretion disks. The stars at α_IRAC ≲ −2.5 exhibit low EWs, suggesting an overlap between our CTTS and WTTS classes. The WTTS sample clearly splits into a group of diskless stars at α_IRAC < −2.0 and temporarily non-accreting disked stars at −1.8 < α_IRAC < −0.5.

Zoom In Zoom Out Reset image size

4.3.2. Hα Equivalent Width and Rotation Period

Theoretical predictions suggest that the rotation rate of young stars may be influenced by magnetic interactions with the accretion disk: the star is forced to corotate with the inner edge of the disk (see, e.g., Königl 1991). If this disk locking is real, CTTSs should have longer rotational periods than WTTSs.

Of the 587 TTSs listed in Tables 1–3, 214 were found to be periodic variables by Herbst et al. (2002), Cieza & Baliber (2007), and Flaccomio et al. (2005). Of the periodic variables, 144 are CTTSs and 70 WTTSs. The period histograms for the whole Hα emission star sample, CTTSs, and WTTSs are shown in the upper, middle, and lower panels of Figure 19, respectively, using the rotation data from Herbst et al. (2002), Cieza & Baliber (2007), and Flaccomio et al. (2005), binned into 1 day intervals. It can be seen clearly that there are two peaks at 2 and 8 days in the top panel. Herbst et al. (2002) showed a similar bimodal distribution for stars having masses greater than 0.25 M_☉ in their rotational study. There is only one weakly defined peak at 8 days in the middle panel, while in the bottom panel there is another peak at 2 days. Applying the statistical Welch's test, we found a significant difference between the rotation periods of CTTSs and WTTSs. The mean rotation period of the CTTS sample is statistically longer than that of the WTTS group, as expected.

Zoom In Zoom Out Reset image size

4.3.3. Hα Equivalent Width and X-Ray Emission