A New Catalog of Am-type Chemically Peculiar Stars Based on LAMOST

The Transiting Exoplanet Survey Satellite Input Catalog (v8.2, hereafter TIC-v8.2; Paegert et al. 2021) uses the Gaia Data Release 2 (DR2) catalog as a base and merges a large number of other photometric catalogs containing spectroscopic quantities, photometry in the various passbands (including Two Micron All Sky Survey, UCAC4, APASS, Sloan Digital Sky Survey, Wide-field Infrared Survey Explorer, etc.) to estimate various stellar properties, proper motions, and a number of other ancillary data.

The effective temperatures (T_eff) and metallicity ([M/H]) were selected from the catalogs shown in the paper when available; Paegert et al. (2021) prefer spectroscopic T_eff if it is available with a reported error of less than 300 K; otherwise, they calculate T_eff from photometric colors via empirical relations and through dereddening correction, as described in the paper. The metallicity parameters are reported when available from spectroscopic catalogs, while the metallicities were not used in any relations or derived quantities.

The stellar radii were computed using the Gaia parallaxes according to the standard expression from the Stefan–Boltzmann relation as a function of T_eff. The above relation assumes that the object is a single star, and it will in general return biased values for the radius if it is a binary because the Gaia photometry will be affected by the companion. As such, the radii of the eclipsing Am binary candidates obtained from TIC-v8.2 were not adopted. The stellar mass can be inferred from T_eff for stars that are on the main sequence or not too far evolved from it. The T_eff–mass relation was only applied if the stellar radius placed the star below the red giant branch and above the white-dwarf sequence. Surface gravities (log g) are always calculated in TIC-v8 using the reported mass and radius; log g is not adopted from spectroscopic catalogs even when available so as to ensure internal consistency of log g with mass and radius. Accordingly, the log g values of eclipsing Am binary candidates were not adopted because of the unreliability of the radii. Bayesian distances, estimated from Bailer-Jones et al. (2018), were utilized wherever their relations involved the Gaia DR2 parallax. The apparent V magnitudes were computed for TIC stars that do not possess a measured V but which possess Gaia photometry using the relations provided by the Gaia team (Evans et al. 2018).

The MKCLASS code was used to search for Am stars through all the low-resolution early F-type (including F0, F1 and F2) and A-type spectra (with a S/N of the g filter larger than 50) of LAMOST DR8 v1.0, DR9 v0, and DR10 v0. All the spectra were normalized to the flux at 4503 Å in order to keep same style with the standard spectra of the library. Preprocessing was performed on the spectra before starting the MKCLASS program. We here present the MK spectral classifications for all the selected spectra. In total, 21,635 Am candidates were detected from the selected early F- and A-type spectra of LAMOST. Figure 1 shows three sample spectra with MK classifications. The new catalog of Am candidates is the largest catalog to date and will supply a wealth of important targets for continuing research into Am stars. The main catalog of Am candidates is presented along with their MK spectral type in Table 1. The parameters applied by LAMOST are also listed in the table. We further crossmatch the catalog of Am candidates with TIC-v8.2 to obtain more property parameters. The main crossmatched parameters are all listed in the new catalog of Am candidates. Considering there are too many columns to show in this paper, we only give a partial list of the columns of the example candidates in Table 1. The full table is available online.

Zoom In Zoom Out Reset image size

The following identifiers are given in the example Table 1: LAMOST data release version of each spectrum; the name of the original LAMOST fits file; the designation of the LAMOST target; the R.A. (J2000) and decl. (J2000) of the targets during observing; the S/N of the spectrum in the Sloan g band (λ_eff = 4686 Å) taken from the fits header; the peculiar classification MK spectral type shown with the K-line spectral type, the hydrogen-line type, and the metallic-line type; and an additional note on the quality of the spectrum. In order to ensure the reliability of the results, all candidates indicated with a "?" in the notes column have been abandoned. The "..." symbols denote omitted columns, which will shown in the full table online, such as the parameters obtained from LAMOST and TIC-v8.2.

In this section, we will introduce the main catalogs related to Am stars. Renson et al. (1991) prepared a catalog of all the known Ap and Am stars, containing a total of 6684 objects, of which 3500 Am were stars (with only 80 well known Am stars). A few years later, Renson & Manfroid (2009) made a major update of their first catalog of CP stars, in which 4299 Am stars (or probable) were collected from a large sample of the literature. Of the Am stars, there are 116 Am stars which have been well studied and only 73 eclipsing Am binaries. Hou et al. (2015) attempted the first search of Am stars based on the low-resolution spectra of LAMOST Data Release 1. The empirical separation curve derived from the line index of the Ca ii K line and nine groups of iron lines were used to find 3537 Am candidates. They provided a much fainter list of the Am candidates with 10 relevant parameters. Gray et al. (2016) employed the MKCLASS automatic classification code to classify spectra in the LAMOST-Kepler field and in total obtained 1067 Am stars with hydrogen line types ranging between A4 and F1. A specific description of each spectral type of Am star is given, which is written as kA2hF0mF2, for example, where A2 represents the K-line spectral type, F0 the hydrogen-line type, and F2 the metallic-line type. Qin et al. (2019) carried out another search for Am stars in the low-resolution spectra of LAMOST Data Release 5. Using known Am spectra, six machine-learning algorithms were experimented with. A random forest algorithm was used to pick out 15,269 Am candidates, then manual identification was conducted to identify 9372 Am stars with spectral types between F0 and A4, finally. Of the Am stars, 357 classical Am stars and 76 marginal Am stars were obtained.

Except for the first-edition catalog of Renson et al. (1991), the other four catalogs related to Am stars were all used to crossmatch with the VSX index (Watson et al. 2006), which can generate a preliminary catalog of the eclipsing Am binaries.

The AAVSO International Variable Star Index (VSX; http://www.aavso.org/vsx) is a comprehensive relational database of known and suspected variable stars gathered from a variety of respected published sources and made available through a powerful web interface. The information access includes all known cross-references, basic parameters (such as period and variability type), and finding charts. The upload feature permits information updates on known variables (such as a new period) as well as entering new variable stars into the system.

According to the VSX, classification of eclipsing binaries based on light curves was adopted to chose the eclipsing Am binaries due to its merits of being simple, traditional, and suiting observers. The effective variability type can be shown as follows: E, eclipsing binary systems with high inclination i close to 90° so that the components periodically eclipse each other; EA, Algol (Beta Persei)–type eclipsing systems; EB, Beta Lyrae–type eclipsing systems; EW, W Ursae Majoris–type eclipsing variables. We adopt the four variability types to select the eclipsing Am binaries.

We crossmatch the known catalogs of Am stars and the new catalog of Am candidates based on LAMOST to the VSX, respectively, by using the CDS xMatch service ⁶ . The critical condition is that the radius of the positions of two targets be less than 2''. The CDS crossmatch service is a tool allowing astronomers to efficiently cross-identify sources between very large catalogs (up to 1 billion rows) or between a user-uploaded list of positions and a large catalog. A duplicate check for the results of the crossmatch was done seriously and carefully. If there were two or more different Am stars matched to the same variable star, the target with the lowest angle distance was selected. The catalog of the eclipsing Am binary candidates was also crossmatched with TIC-v8.2 (Paegert et al. 2021) to gain more parameters. As discussed by Zhang et al. (2019), the atmospheric parameters measured from the spectra approximately represent the brighter component star that dominates the light of the whole system. Thus, all the atmospheric parameters and the physical properties (such as mass and radius) in our tables of eclipsing Am binaries pertain chiefly to the primary components.

The two catalogs of eclipsing Am binaries obtained from the crossmatch results based on the two kind of catalogs of Am stars are listed as two independent tables. The catalog of the eclipsing Am binary candidates based on our new catalog of Am candidates is shown in Table 2. The full table is available online. The information and parameters in Table 2 can be divided into four parts: the observation information, the parameters based on LAMOST, the parameters collected from VSX, and the parameters obtained from TIC-v8.2. The observation information contains the fundamental state (such as data release version, DESIG, CLASS, SUBCLASS) and observation information, such as SNRG (S/N of the g filter), RA_OBS, and DEC_OBS (the R.A. and decl. in J2000). The atmospheric parameters based on LAMOST include effective temperature (T_eff), metallicity ([Fe/H]), and surface gravity (log g). They also contain the MK spectral type obtained by us through MKCLASS based on the spectra from LAMOST. The "angDist1" column shows the angle distance between the Am star and the target of VSX. The information and parameters regarding the binary characteristics (such as name, variable star identifier, variability type, period) were supplied by VSX. We also collected additional atmospheric parameters and basic physical properties through TIC-v8.2, such as V_mag, T_eff, log g, [M/H], radius, mass, distance, and so on. The "angDist2" column shows the angle distance between the Am star and the target of TIC-v8.2.

The common characteristic of Am stars, i.e., Sp(K) < Sp(m), can be used to select Am stars from known catalogs. The spectral type of the Am stars in the catalog of Am, HgMn, and CP stars (Renson & Manfroid 2009) is marked with the K-line type and the metallic-line type, separated by a hyphen. The spectral type of the Am candidates in the catalog of Hou et al. (2015) is given with the K-line type and the metallic line. In the other two catalogs, of Gray et al. (2016) and Qin et al. (2019), the three spectral types of the Am candidates are given. The catalog of CP stars (HgMn, Ap/Bp, Am/Fm; Ghazaryan et al. 2018) marks the Am stars with "AmFm" in the CP type column. We collected all the Am candidates through the above five catalogs. Then, the Am candidates were also crossmatched to the VSX to obtain more eclipsing Am binary candidates. The catalog of the eclipsing Am binary candidates based on the crossmatch between the known Am catalogs and the VSX are shown in Table 3. Meanwhile, we also collected the spectra files, the atmospheric parameters, and the physical properties from LAMOST and TIC-v8.2 through a crossmatch. All the matched results are shown in Table 3. For the same reason, we also show a partial list of the main columns of Table 3 in this paper. The full table is available online. The information and parameters in Table 3 can also be divided into four parts: the information about the Am peculiarity, the parameters obtained from VSX, the parameters based on LAMOST, and the parameters obtained from TIC-v8.2. The information about the Am peculiarity shows the original catalog of the Am star and the spectral type or CP type, including the peculiarity probability note or confidence applied by the original catalog. The information and parameters of the remaining three main parts are similar to that of Table 2.

There are a total of 537 and 239 eclipsing Am binary candidates in Tables 2 and 3, respectively. In addition, there are 22 identical candidates in the two tables, which are listed in Table 4. Finally, a catalog of eclipsing Am binaries with about 754 candidates is obtained—the largest catalog of such targets so far. This new catalog of eclipsing Am binary candidates with additional parameters can supply more information about the observed targets and should help researchers to conduct further research about the eclipsing Am binaries and Am stars population.

The majority of the Am and eclipsing Am binary candidates have a plentiful list of physical properties obtained from LAMOST, VSX, and TIC-v8.2. The spatial distribution of Am stars on the Galactic coordinate plane is shown in Figure 2. The difference of spatial distribution of Am stars on the Galactic disk and on other regions is caused by the observation range of the LAMOST Spectroscopic Survey of the Galactic Anti-center (Luo et al. 2015), which covers a continuous sky area of ∼3400 deg², covering Galactic longitudes 150° ≤ ℓ ≤ 210° and latitudes ∣b∣ ≤30° (Yuan et al. 2015). The distributions of standard errors along with the main stellar atmospheric parameters were discussed, which are shown in the eight panels of Figure 3. The standard errors for about 90% of those stars with corresponding parameters are marked in each panel with dashed lines. The 90% standard errors are lower than 0.32 M_⊙ for Mass_tic, 9.06 L_⊙ for luminosity (Lum_tic), 0.084 dex for [M/H]_tic, 0.117 cm s⁻² for $\mathrm{log}g\_\mathrm{tic}$, 207 K for T_eff_tic, and 0.23 mag for V_mag_tic (V-band magnitude), respectively. We also plot the same distributions of 90% standard errors for the parameters obtained from LAMOST in Figure 4, i.e., effective temperature (marked as T_eff_lamost), metallicity (marked as [Fe/H]_lamost), and surface gravity (marked as $\mathrm{log}g\_\mathrm{lamost}$). The 90% standard errors are lower than 65 K for T_eff_lamost, 0.06 dex for [Fe/H]_lamost, and 0.10 cm s⁻² for $\mathrm{log}g\_\mathrm{lamost}$, respectively. The 90% standard errors of the temperature obtained from LAMOST is little more than that of the temperature obtained from TIC-v8.2. As such, we will mainly focus our analysis on the LAMOST temperature in the latter parts of this paper.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The distributions of the atmospheric parameters and stellar magnitude for both Am stars and eclipsing Am binaries (hereafter Am EB in the following figures) are plotted in Figure 5. The peak of $\mathrm{log}g\_\mathrm{tic}$ is 4.1 for the Am targets. The peaks of V_mag_tic of both candidates are about 13–14 mag. The peaks of [M/H]_tic are around −0.1 and −0.5 dex for Am stars and eclipsing Am binary candidates, respectively. The majority of both targets are distributed from −0.5 to 0.3 dex. The distributions of the [Fe/H]_lamost, $\mathrm{log}g\_\mathrm{lamost}$, and T_eff_lamost of both Am and eclipsing Am binary candidates are plotted in Figure 6. The peaks of $\mathrm{log}g\_\mathrm{lamost}$ are 4.0 and 4.1 for Am targets and eclipsing Am binary candidates, respectively. The peaks of T_eff_lamost of both candidates are around 7000 K. The peaks of [Fe/H]_lamost are around −0.1 dex and −0.3 to −0.2 dex for Am stars and eclipsing Am binary candidates, respectively. Through comparing Figure 6 with Figure 5, we can see that the same parameters in both figures have almost the same peak distribution. There are many targets with a temperature lower than 7000 K in our new catalogs of Am stars and eclipsing Am binary candidates, which is an interesting phenomenon and a great opportunity to explore the real relationship between the Am peculiarity and the stellar rotation.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

For all eclipsing Am binaries the periods and the variability types were also analyzed. There are 741 systems in our eclipsing Am binary catalog, including 459 EA-type (about 383 confirmed), 93 EB-type (78 confirmed), 149 EW-type (142 confirmed), and 38 E-type targets. The distributions of the periods and the variability types are shown in the panels of Figure 7; the part distribution of all periods less than 10 days is plotted in the upper-right area of the top panel. One can see in the figure that there is an obvious peak in the interval from 0.5 to 1 day, which is different to the results of Renson & Manfroid (2009) and Smalley et al. (2014), whose histograms show a peaked distribution with a maximum around 2 days. The shorter-period systems (shorter than 1 day) in our study pose a challenge to the well-known absence of Am binaries with very short orbital periods (P_orb < 1.2 days; Budaj 1996; Iliev & Budaj 2008), because the synchronism in such systems should force the primary to rotate faster than 120 km s⁻¹. The relationship between the slow rotation and the formation of an Am star is not clear yet. These short-period systems will become significant targets for more observation and investigation. We will pay more attention to these important systems in future.

Zoom In Zoom Out Reset image size

The correlations between the parameters were also investigated. Only several fundamental relationships relating to the temperature and period (for eclipsing Am binary systems only) are exhibited in following figures. There are no obvious relationships between temperature–metallicity and temperature–gravity based on LAMOST for Am stars and the primary stars of eclipsing Am binary candidates. The relationships between [Fe/H]_lamost−T_eff_lamost and $\mathrm{log}g\_\mathrm{lamost}$−T_eff_lamost are shown in Figure 8.

Zoom In Zoom Out Reset image size

We also explored the correlations between properties (including [Fe/H]_lamost, T_eff_lamost, $\mathrm{log}g\_\mathrm{lamost}$, [M/H]_tic, Lum_tic, and Mass_tic) and the period of the eclipsing Am binary candidates, which are shown in Figure 9. Here, no obvious relationship is seen. Relationship diagrams are given in each panel of Figure 9 of eclipsing Am binaries with three different variability types (i.e., EA, EB, and EW) when the period is less than 10 days, which is also the main distribution area of the eclipsing Am binary systems. Since the EB samples are small, we only compare and discuss the short-period EA and EW ones. The log g values of the EW-type systems are a little larger than those of the EAs. This suggests that these EW-type systems have slightly undergone expansion compared with the EA samples. The metallicity ([Fe/H]_lamost or [M/H]_tic) of short-period EA-type systems is a little larger than that of the EWs. This means the former are younger systems while the latter are older ones. These phenomena may relate to the extensively discussed evolutionary channel of binaries evolving from EA to EW through angular momentum loss (Qian et al. 2018). However, there is no obvious correlation between these atmospheric parameters and orbital period.

Zoom In Zoom Out Reset image size

Based on the effective temperature and luminosity adopted from TIC-v8.2, a H-R diagram of the Am stars and the primary stars of eclipsing Am binary candidates is plotted in Figure 10. In the figure the evolutionary tracks and isochrones were calculated for Y = 0.279, Z = 0.017 by Bressan et al. (2012). A majority of the stars are in or around the main-sequence evolution stage, and we notice that some stars are evolving to main sequence while other stars have obviously evolved through the main sequence, while almost all the primary stars of eclipsing Am binary candidates are in the main-sequence stage. The red and green dotted lines denote the isochrone lines, which show that all the Am stars have an age between 2 and 180 Gyr, while the eclipsing Am binaries have an age between 4 and 30 Gyr.

Zoom In Zoom Out Reset image size