Open Access
Issue
A&A
Volume 671, March 2023
Article Number A140
Number of page(s) 16
Section Stellar atmospheres
DOI https://doi.org/10.1051/0004-6361/202245427
Published online 17 March 2023

© The Authors 2023

Licence Creative CommonsOpen Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This article is published in open access under the Subscribe to Open model. Subscribe to A&A to support open access publication.

1 Introduction

Stellar metallicity, [M/H]1, may be a factor that plays a fundamental role in planet formation. Soon after the discovery of the first exoplanets around solar-type stars, a trend linking the presence of planets to stars with higher [M/H] values became apparent (Gonzalez 1997; Santos et al. 2000, 2001). As the number of detected exoplanets increased, further evidence confirmed and strengthen the today well-known “planet-metallicity” correlation for FGK stars (see, e.g., Adibekyan 2019; Osborn & Bayliss 2020, and references therein). However, the case of A-type stars is still far from clear. Less than 20 planets have been confirmed around stars with effective temperatures between 8000 and 12 000 K2, which makes a robust statistical study unfeasible at this stage. Indeed, the most common methods for detecting exoplanets are based on weak spectroscopic velocity signals and photometric transits, for which the relatively small number of photospheric signatures (typically large rotational velocities and high luminosities of hot stars) make them difficult targets to probe.

Concerning the precursors of main-sequence A and B stars, recent studies have been carried out to characterize large samples of young, intermediate-mass Herbig Ae/Be objects (HAeBes, Vioque et al. 2018; Arun et al. 2019; Wichittanakom et al. 2020; Guzmán-Díaz et al. 2021) and to significantly increase the number of potential members belonging to the class (Vioque et al. 2020, 2022; Zhang et al. 2022; Kuhn et al. 2023). However, most studies devoted to HAeBes do not usually analyze [M/H] and the solar value is adopted by default. Only a handful of works have considered the chemical peculiarities that HAeBes can show (such as the λ Boötis phenomenon; see e.g., Gray & Corbally 1998), determining abundances for relatively small sub-samples (e.g., Acke & Waelkens 2004; Guimarães et al. 2006; Montesinos et al. 2009; Folsom et al. 2012).

A major study involving [M/H] in HAeBes was carried out by Kama et al. (2015). These authors proposed that the HAeBes showing a deficit of refractory elements and, thus, low values of [M/H] could be linked to the presence of Jupiter-like size exoplanets in their protoplanetary disks. This scenario does not necessarily imply that A and B stars harboring giant planets are globally metal-poor, which would be in opposition to the planet-metallicity correlation found in late-type stars. According to Kama et al. (2015), planets undergoing formation mainly trap the metal-rich material, with the remaining accreted by the central star being metal-depleted. Given that A and B stars have radiative envelopes, the mixing timescale with the interior is on the order of ~Myr, much slower than that for lower mass stars with convective sub-photospheric regions. Therefore, [M/H] measurements in HAeBes refer to freshly accreted material and only reflect the stellar surfaces. Additional details on the theoretical model behind the previous scenario are described in Jermyn & Kama (2018), and similar views proposed to explain the properties of some post-AGB stars and solar twins can be consulted, for instance, in Oomen et al. (2019), Booth & Owen (2020), Kluska et al. (2022), and references therein. Other scenarios involving changes in the stellar metallicity have been explored, for instance, in the context of stars with planets and debris disks (Maldonado et al. 2012, 2015) or on galactic scales (e.g., Adibekyan et al. 2014; Hawkins 2022, and references therein).

On the other hand, the proposal from Kama et al. (2015) hinges on the fact that the deficit in [M/H] is observed in stars belonging to group I in the classification by Meeus et al. (2001). This is based on the shape of the spectral energy distribution (SED) in the infrared and submillimeter regions. Such a classification has been related (among others properties) with the morphology of the circumstellar disks, where group I and group II sources are associated with “flared” and self-shadowed, “flattened” disks, respectively (e.g., Meeus et al. 2001; Dullemond 2002; Dullemond & Dominik 2004). Demonstrating what is most relevant to the hypothesis by Kama et al. (2015), studies based on high-resolution imaging suggest that disks belonging to group I sources show cavities potentially carved out by giant planets (e.g., Maaskant et al. 2013; Honda et al. 2015; Garufi et al. 2017; Stapper et al. 2022).

Although the scenario proposed by Kama et al. (2015) could have major implications for our understanding of planet formation, it was based on a study of a relatively small sample of 22 HAeBes. In fact, these were the only ones having measured [M/H] values and disk structure classification from SEDs by that time, which were based on heterogeneous studies from the literature. Our main aim is to solve these issues by providing a robust test on whether the [M/H] of HAeBes can be related to a specific SED-based disk structure from the Meeus et al. (2001) scheme or not. Given that the answer we find is affirmative, the potential implications in terms of the presence of giant planets are further analyzed. The departure point is our previous work in Guzmán-Díaz et al. (2021), where the stellar parameters and the circumstellar morphology in terms of the Meeus et al. (2001) groups were homogeneously derived for essentially all classical, historically well-known HAeBes. Here, we derive [M/H] for all such objects for which this parameter can be measured from suitable spectra, extending the sample analyzed by Kama et al. (2015) by a factor ~3. On top of the new [M/H] values for all the stars, projected rotational velocities (υ sin i) and surface gravities (log ɡ) will also be provided as valuable outcomes of our work. Section 2 describes the sample and the observations. The process to estimate [M/H] and the rest of stellar parameters is described in Sect. 3. An analysis of the previous results is in Sect. 4, which includes a statistical study and discussion regarding the connection between [M/H] and the SED classification. Finally, a brief summary and our main conclusions are given in Sect. 5.

2 Sample and observations

We started by selected an initial sample of HAeBes with effective temperatures, Teff, below 12 000 K (spectral types later than B8) from Guzmán-Díaz et al. (2021) and with one-dimensional (1D) spectra available at the ESO Science Archive Facility3. Accordingly, the stellar mass ranges typically between 1 and 5 M and the ages are mostly <15 Myr. We discarded emissiondominated stars that do not show enough absorption lines (i.e., mainly HBe stars with Teff > 12 000 K). Sources with spectra that do not have a high enough signal-to-noise ratio (S/N typically ≥100 for all stars except for four listed in Appendix B, with S /N ≥ 50) or spectral resolution (see below) were also discarded. The final sample is comprised of 67 sources, whose main properties are summarized in Table A.1. All data listed in that table were compiled from the photometric study of Guzmán-Díaz et al. (2021), except for the spectroscopically determined Teff values taken from Wichittanakom et al. (2020), whenever available. The final sample covers a range of luminosities - in log(L*/L)-between 0.4 and 3.2, masses between 1.4 and 7.0 M, and ages between 0.2 and 20 Myr, representing almost 50% of all classical, late-type HAeBes studied in Guzmán-Díaz et al. (2021). Regarding the SEDs studied in that work, 28 sources were classified as group I and 34 as group II. Although five stars could not be classified with regard to the SED groups, they have also been included in the current sample for completeness, given that their [M/H] values could also be determined from the available spectra.

Most spectra in this study were taken with the X-shooter/VLT spectrograph. The spectral resolution, R = λλ, varies between ~3200 and ~18 400 depending on the wavelength range and the slit width. Spectra taken with UVES/VLT and HARPS/La Silla 3.6-m, were also used, with R ~ 41 000–110 000 and ~80 000–115 000, respectively. Finally, the spectrum of HBC 222 was taken with GIRAFFE/VLT (R ~ 5500–65 000). The spectrographs used to obtain the spectra of each source, and their resolution measurements at ~5000 Å are listed in Cols. 2 and 3 of Table A.2. The first initials indicate the instruments and resolutions corresponding to the spectra used to estimate υ sin i, and the second to those used for measuring the widths of the Balmer lines (i.e., log ɡ) and [M/H] (see Sect. 3). In the case where several spectra from the same instrument were available for a given star, priority was given to the one with the highest spectral resolution.

The spectra used in this work are the final products of the processing of the raw data – and the corresponding calibration files – through the pipelines developed by the instruments’ teams. Special care was taken to normalize the spectra; narrow regions free from lines were chosen in the original spectra and the mean intensities of each small interval were used as clips to build a full continuum using cubic splines. The RASSINE program (Cretignier et al. 2021) has been used in automatic mode to improve the normalization of a few UVES and HARPS spectra. Telluric contamination in our spectra is significant within the ranges of ~5850–6000 Å and ~6450–6600 Å, but these ranges are not used in the analysis (Sect. 3). Therefore, although telluric correction was not applied this does not affect our results.

3 Results

This section describes how the [M/H] values were obtained from the comparison of the observed spectra with photo-spheric models, with υ sin i and log ɡ additional outcomes of the process. Synthetic Castelli-Kurucz spectra computed with the suite of codes ATLAS9 (Castelli & Kurucz 2003) were used throughout. These models do not consider disk-to-star accretion, which can produce a continuum excess – and possible line veiling in the coldest sources – mainly at wavelengths shorter than studied in this work (e.g., Muzerolle et al. 2004; Mendigutía et al. 2011, 2014).

3.1 Projected rotational velocities

The first step was to estimate the υ sin i of each star. Although the derivation of [M/H] does not strictly depend on the υ sin i values – as rotation does not have an effect on equivalent width, which is extensively used throughout this work (see Sect. 3.2) – this parameter facilitates the visual comparison between observations and models.

Before describing the procedure for estimating υ sin i, some words of caution are pertinent. In some cases, when the resolution of the spectrograph is not particularly high (say, below 30 000) this parameter puts some limits to the estimated values of υ sin i. The widths of some telluric or arc lines give an idea of the instrumental response of the spectrograph; we may use σinst to refer to that value. Assuming Gaussian profiles, the actual stellar line profiles with widths, σreal, and the instrumental σinst can be combined according to the expression , where “obs” stands for “observed.” For the case where σreal = 2 σinst we would have . Identifying συ, namely, widths and velocities, the relative error (υobsυreal)/υreal would be ~11%. For the particular case of all the X-shooter spectra obtained with R = 9900, σinst ≃ 30 km s−1, that is, values of υ sin i below twice the aforementioned value would be affected by larger uncertainties and must be taken as upper limits, with the situation becoming more critical for the few stars whose spectra were obtained with even lowerresolutions.

The broad, isolated Mg II feature at 4481 Å was generally used to estimate the υ sin i of the objects. For each star, the full width at half depth (FWHD) of that line was varied by changing the υ sin i in the synthetic model with the corresponding Teff from Table A.1. The v sin i that provided a FWHD equal to that of the observed spectra within 5% was adopted as the final value. Column 4 of Table A.2 lists the υ sin i values obtained for the sample stars. Alternative photospheric lines were used for the few stars showing a Mg II line with a peculiar profile that makes the estimation of υ sin i challenging. In addition, previous υ sin i values from the literature were taken as a departure point and refined based on the procedure described above for the following stars: CO Ori (Herbig & Bell 1988; Mora et al. 2001; Glebocki & Gnacinski 2005), HBC 217 (McGinnis et al. 2018), HBC 222 (McGinnis et al. 2018), HD 101412 (Cowley et al. 2010), HD 104237 (da Silva et al. 2009), HD 143006 (Jönsson et al. 2020), HD 169142 (Alecian et al. 2013), PX Vul (Pereyra et al. 2009) and BP Psc (Torres et al. 2006). Finally, the value of υ sin i for BF Ori was directly adopted from Mora et al. (2001).

3.2 Surface gravities and metallicities

Figure 1 illustrates the steps followed in the estimation of log ɡ and [M/H]. First, for each star, we generated an initial grid of Kurucz models, with the corresponding Teff and υ sin i listed in Tables A.1 and A.2, log ɡ = 3.5,4.0,4.5 and [M/H] = −2.5, −2.0, −1.5, −1.0, −0.5, 0.0, +0.2, and +0.5. From this point on, two different paths were followed depending on the Teff of the source.

For stars with Teff > 8000 K, log ɡ was derived by comparing the observed widths of the Balmer lines Hδ, Hγ, and Hβ at intensity 0.80 with those from the models (Gray & Corbally 2009). The initial grid was interpolated in order to obtain the log ɡ values that provide the best fit for each line, the final value being the average of the previous. Uncertainties were calculated from the standard deviation of the individual log ɡ values obtained in each of the considered Balmer lines. Lines affected by strong emission components that do not allow for the wings of the absorption profiles to be correctly traced were excluded from the analysis (see the example in the top-right panel of Fig. 2).

Once the value of log ɡ was determined, the observed and modeled pseudo-equivalent widths (pEW) were compared to each other for the different [M/H] values, providing a final value of [M/H] based on a χ2 test. A pEW is defined exactly in the same way as the classical EW, but instead of characterizing a single spectral line, the measurement extends over several lines within a wavelength range [λ1, λ2]. The reason for using pEW instead of EW is that the former is less limited by spectral resolution. Indeed, most stars in the sample show large values of υ sin i, which makes an individual analysis of the spectral lines difficult due to blending. Moreover, the lower spectral resolution of the X-shooter spectra, which are intensively used, prevents us from measuring EWs of individual lines accurately, even in objects with moderately small values of υ sin i. A similar method to estimate metallicities was developed by Gray et al. (2002) but in a different context (i.e., G and K giants; see, e.g., Lillo-Box et al. 2014, for a practical application of that formalism).

The wavelength region used to calculate the pEWs is [5000, 7000] Å. The windows initially selected were 80-Å wide, excluding the broad interval containing Hα, and the number of such windows per object was about six. However, for each source, a careful selection of what intervals within these windows were usable was carried out. Several factors prevented us from selecting the same regions homogeneously for all the stars, namely: (i) the presence of emission lines; (ii) the normalization of particular sections of the stellar spectrum; (iii) a low S/N, which causes some lines not to be discernible and difficulties to trace the continuum; (iv) scarcity of spectral lines (especially in very hot stars); (v) circumstellar contribution in absorption, superimposed to the photospheric lines; and (vi) variability. Regarding the last point, when multi-epoch spectra are available for a given source, they were checked to avoid regions where variability in the profiles of the metallic lines was apparent. It should also be mentioned that some spectra had to be renormalized manually in order to obtain a better estimate of [M/H].

As a reference for the final [M/H] value, we used that given by the χ2 test, while the corresponding uncertainty was estimated from the individual [M/H] best-fitting value of each spectral region considered. An average of the absolute errors was computed, assuming a minimum uncertainty of 0.10 dex.

A different procedure was followed for stars with Teff ≤ 8000 K. The widths of the Balmer lines are insensitive to changes in surface gravity for such sources (Gray & Corbally 2009); spectral indicators for this kind of objects based on lines ratios are difficult to apply, given their typically large values of υ sin i. Therefore, in these cases moved directly to the step where the pEWs are calculated in the observed spectrum and in the whole collection of models (Fig. 1). Hence, the result of the χ2 test yields a value of [M/H], but with a degeneracy in log ɡ. In fact, the three best models with log ɡ = 3.5, 4.0, 4.5 fit the regions of the observed spectrum practically in the same way, hardly showing any variation in the pEWs of the regions explored. In order to find out which log ɡ is closer to the stellar value, we have used an indirect method that consists of determining which of these log ɡ is consistent with the distance from Gaia (E)DR3 (Lindegren et al. 2021; Gaia Collaboration 2023, see also Table B.1 of Guzmán-Díaz et al. 2021).

Let us denote d3.5, d4.0, and d4.5, the distances implied from the three values of log ɡ. These distances are derived from the expression , where L* is the stellar luminosity, and F* is the total observed flux from the star. The fluxes were calculated using the VOSA4 tool by integrating the -dereddened-Kurucz ODFNEW/NOVER models (Castelli et al. 1997) computed for Teff, [M/H] from the χ2 test, and the three values of log ɡ. Although the visual extinction (Av) was left virtually free in the fits, their final values obtained for each log g are similar to those shown in Table B.1 of Guzmán-Díaz et al. (2021). The stellar luminosities were derived by translating the points (Teff, log ɡ; with log ɡ = 3.5,4.0,4.5) from the Teff - log ɡ HR diagram to the Teff - log L* diagram. PARSEC V2.1s5 evolutionary tracks (Bressan et al. 2012) were used. Each one of the three points (Teff, log ɡ) has a one-to-one corresponding point (Teff, log L*); once d3.5, d4.0 and d4 5 are computed, the stellar gravity is the one whose derived distance matches that from Gaia (E)DR3. Interpolation between distances was used when required. A typical error of 0.05 dex on log ɡ was adopted.

In some stars, two different values of [M/H] provide similar values of χ2. For these cases, we generated new models by interpolating the values of [M/H] and repeated the χ2 test in order to check if a better fit was obtained. Should this occur, we assumed this new value; otherwise, the [M/H] with the lowest χ2 retrieved from the previous test was kept.

Table A.2 lists the final values inferred for log ɡ and [M/H] and their corresponding errors in Cols. 5 and 6. Figure 2 shows a representative example of the best fit obtained applying the procedure described above for HD 244314.

thumbnail Fig. 1

Flow chart summarizing the process followed in the estimation of the log ɡ and [M/H] values (see Sect. 3).

thumbnail Fig. 2

Fits of the Balmer lines Hδ>, Hγ, Hβ, and six regions of the HD 244314 spectrum (black solid line) with the best model obtained in the χ2 test (red solid line). The model has been generated with a Teff = 8500 K, log ɡ = 4.08, υ sin i = 55 km s−1 and [M/H] = 0.0. In this example, the line Hβ has not been considered in the estimation of log ɡ as it shows a strong emission.

thumbnail Fig. 3

Model with Teff = 8000 K, υ sin i = 50 km s−1, and [M/H] = 0.0 in the 5000–5400 Å region is represented in black, while only the C, O, and S lines generated with the same parameters are shown in cyan. The result of subtracting the volatile elements from the first model is shown in red. The models were shifted on the y-axis for clarity.

3.3 Interpretation of the [M/H] derived in this work

In order to interpret the results, it is important to emphasize explicitly that the method used does not take into account variations of individual abundances of specific elements, that is, for a given [M/H], all abundances are scaled in the same manner to the solar one by a factor 10[M/H], the whole procedure providing an average metal abundance for each star.

Since the hypothesis by Kama et al. (2015) linked a deficit of refractory elements to the presence of dust traps and cavities potentially caused by Jovian planets, we carried out the following exercise: all lines of C, O, N, and S in the interval 5000–7000 Å were identified and the pEWs were computed in the same regions both in the stellar spectra and in the models, avoiding the small windows where lines of those volatile elements appear. Given that the number of lines of volatiles is much less than that of refractories, it turned out that the new [M/H] values do not show significant changes when compared with the initial calculations.

The previous test was done for six stars of the sample with different values of Teff, υ sin i, and [M/H] to cover a representative set of parameters, finding no significant differences with respect to the values already listed in Table A.2. Figure 3 shows an example of how the subtraction of the C, O, N, and S lines from a synthetic model hardly produces any change in the 5000–5400 Å region, which serves to visualize why the [M/H] measurements are mainly determined by the abundances of refractory elements regardless of the abundances of volatiles.

4 Analysis and discussion

4.1 Consistency tests

4.1.1 Comparison with previous results

The top panel of Fig. 4 compares the values estimated here for υ sini with previous values from the literature, which are available for 63% of the stars in our sample. For those values in the literature where no associated error was found, an uncertainty of 5% was assumed. Results are in good agreement for the majority of the stars, being the mean relative error ~7%. The previous analysis excludes the ten stars (shown in Table A.2) with υ sin i values based or directly adopted from the literature (Sect. 3.1).

Concerning log ɡ, our values and previous ones based on the literature are plotted against each other in the middle panel of Fig. 4. In this case, the comparison has been possible for the whole sample; the log ɡ values not available in Wichittanakom et al. (2020) have been computed from the expression , taking the stellar mass, M*, and radius, R*, from Guzmán-Díaz et al. (2021). The corresponding uncertainties have been calculated by error propagation. The absolute error for most objects is below 0.20 dex, with 18% of them showing an error larger than that value.

Finally, the comparison between the [M/H] values from this work and from the literature is included in the bottom panel of Fig. 4, which was feasible for 36% of the stars in the sample. Literature values come from Montesinos et al. (2009) and Kama et al. (2015). In turn, Kama et al. (2015) compiled individual abundances of certain elements from Acke & Waelkens (2004) and Folsom et al. (2012), except for HD 34282 (Merín et al. 2004), HD 100546 (Kama et al. 2016), and HD 142527 (Holmberg et al. 2009). In order to convert such abundances into metallicitiy values, we followed the process described in the appendix of Montesinos et al. (2009), considering only the abundances of Fe, Si, and Mg given in Kama et al. (2015). There is a reasonable agreement between our results and those from the literature, and the sources identified as having low- and high-[M/H] coincide. The mean absolute difference is approximately ±0.15 dex, excluding the sources with lower and upper limits. Notes on specific sources where the differences between our [M/H] values and those from the literature are relevant are presented in Appendix B.

thumbnail Fig. 4

υ sin i, log ɡ and [M/H] are compared with the values found in the literature (as indicated in the legends). Error bars represent the uncertainties in these parameters, while the lower and upper limits are denoted by right- and left-pointing triangles, respectively. The black dashed line indicates equal values.

thumbnail Fig. 5

Ca II K lines of the spectra of the stars (black solid lines) HD 39014 (left), HD 132947 (middle), and HD 31648 (right), and the corresponding best-fitting models (red solid lines). A good fit is only achieved in the first case, where no circumstellar contribution or peculiar profiles are apparent.

4.1.2 The Ca II K line

The depth of the Ca II K line is used in main sequence A-type stars as the prime estimator of their effective temperature (Gray & Corbally 2009). However, following other works devoted to the determination of stellar parameters of HAeBes (e.g., Acke & Waelkens 2004; Folsom et al. 2012), this line has also been excluded from the whole procedure devised in this work. The reason is that most stars in the sample show Ca II K lines with circumstellar contributions and/or peculiar profiles that pure photospheric models are not able to reproduce. In addition, the blue part of the spectrum would be affected by veiling in those cases where the accretion contribution to the total flux starts to be comparable to the photospheric flux.

With the previous caveat in mind, we checked how the best-fitting models from the χ2 tests match the observed Ca II K profile for stars with Teff between 8000 K and 10 500 K (A5 -B9). For the few stars with normal, symmetric profiles, the models reproduce the line profile with a reasonable level of accuracy, which constitutes an additional proof of self-consistency. Figure 5 shows three representative examples. For HD 39014 the model fits the Ca II K line reasonably well, HD 132947 shows a circumstellar component superimposed to the photospheric profile, and, finally, HD 31648. presents a peculiar profile that the model is unable to reproduce.

thumbnail Fig. 6

Histograms comparing the distributions of metallicities and SED classification, as indicated in the x-axis and the legend. Vertical error bars represent the variations of the number of stars contained in each bin considering the individual uncertainties in [M/H], whose typical value is represented by the horizontal error bar below the legend.

4.2 [M/H] and SED groups

Figure 6 shows the [M/H] distribution of the sample analyzed, where the blue and orange histograms refer to group I and group II sources, respectively. Although most HAeBes tend to show [M/H] values close to solar, the relative fraction of group I sources having sub-solar values is larger (~55%) than that for group II (~10%). In particular, the median [M/H] is −0.10 for group I and +0.14 for group II HAeBes. We carried out a Kolmogorov-Smirnov (KS) test to verify whether the group I and group II HAeBes have different [M/H] distributions or not. The result obtained is that there is a negligible probability (p-value = 0.003) that both samples are drawn from the same parent distribution at a 1% significant level. An Anderson-Darling (AD) test, more sensitive to the tails of the distributions than the KS test, also rejects the null hypothesis that both group I and group II samples are drawn from the same parent distribution at the same significance level. Whether the upper and lower limits in the histograms are considered or not does not significantly alter the results provided by the KS and AD tests.

In short, our enlarged sample of HAeBes, with the homogeneously derived [M/H] values and SED classification according to the Meeus et al. (2001) scheme, statistically confirms the finding by Kama et al. (2015). Specifically, group I and group II sources tend to show different distributions of [M/H] values, with group I sources being less metallic than group II.

thumbnail Fig. 7

2D representation of the distribution in the Galaxy of the HAeBes studied in this work. The Galactic azimuthal angle in the y-axis is plotted against the distance to the Galactic center in the x-axis. The colour bar indicates their [M/H] values. The most crowded region of the plot has been zoomed in. The values for the sun – with [M/H] = 0 – are also indicated with the vertical and horizontal dashed lines, for reference.

4.3 A possible link with planet formation

In the following sections, we aim to interpret the previous observational trend linking low stellar metallicities with group I sources, focusing on the plausibility of the scenario in which giant planets trapping refractory elements are more frequent around HAeBes with such a type of SED (Kama et al. 2015).

4.3.1 Galactic distribution of group I and group II HAeBes

The first question to be addressed is whether different stellar metallicities are actually associated with different SED groups or whether these could instead be driven by environmental effects. Given that the mixing timescale for radiative sub-photospheric regions such as those of HAeBes is similar to their typical ages (~Myr), the second possibility cannot be neglected. For instance, we may wonder if the low- and high-[M/H] stars are located in different clusters or regions within the Galaxy, in which case the stellar metallicities may reflect the local, initial conditions. However, no significant difference is found concerning the clustering properties of the low- and high-[M/H] stars in our sample, which appear mixed together and sharing common regions (Fig. 7). A similar result is obtained for the whole sample classified by Guzmán-Díaz et al. (2021) in 112 group I and 70 group II HAeBes.

We also tested whether the trend linking lower metallicities to group I sources could be related to the well-known empirical relations showing that stellar abundances depend on both the distance to the Galactic center and to the Galactic plane (e.g., Adibekyan et al. 2014; Hawkins 2022, and references therein). The top and bottom panels of Fig. 8 show the [M/H] estimated in this work versus the galactocentric radius and the vertical distance to the galactic plane, respectively. The scatter observed in both panels indicates that the observed differences between the typical metallicities of group I and group II sources are not driven by [M/H] gradients on a Galactic scale, at least for the HAeBes analyzed here. We can therefore conclude that the [M/H] values of the stars in our sample do not depend on where they were born, retaining the hypothesis from Kama et al. (2015) as a possible cause for the different metallicities between group I and group II HAeBes.

thumbnail Fig. 8

Metallicity of the stars in our sample versus the distance to the Galactic center (top) and to the galactic plane (bottom). Group I and II sources are color coded as indicated in the legends. The lower and upper limits are indicated by the up and down triangles. The values of the sun are also indicated with the vertical and horizontal dashed lines, for reference.

4.3.2 [M/H] and transitional disks

The last column of Table A.1 lists the type of SED shown by each star in the sample according to the “JHK classification” carried out in Guzmán-Díaz et al. (2021). According to this classification, stars are divided in four groups (J, H, Ks, and >Ks), indicating the shortest near-infrared band where excess emission over the photosphere is already present. As detailed in Guzmán-Díaz et al. (2021), and references therein, such a classification can be related to the size of the holes in the inner dust disks, being the Ks and >Ks sources associated with so-called “transitional disks” with relatively large inner holes.

Figure 9 shows the [M/H] distribution of the sample analyzed, where this time the red and cyan histograms refer to non-transitional and transitional disks as inferred from the JHK classification. Based on the KS and AD tests, the null hypothesis that both sub-samples are drawn from the same parent distribution cannot be rejected, contrasting with the analogous analysis in Sect. 4.2 concerning the SED groups.

The fact that [M/H] is not related to the presence of transitional disks (as inferred from SEDs) indicates that the stellar abundances are not generally affected by the different processes that may cause the appearance of dust inner holes. Such holes can be related to other mechanisms apart from the presence of planets, such as grain growth, dust settling, or photoevaporation (e.g., Espaillat et al. 2014, and references therein). Indeed, most transitional disks in HAeBes seem to be related with the latter process and the high stellar luminosities, being much more frequent for stars with the earliest spectral types (HBes) than for the rest (see the discussion in Guzmán-Díaz et al. 2021, and references therein). In contrast, the cavities associated with group I sources may be related to the potential presence of giant planets, which could cause the metal depletion of the central stars (Kama et al. 2015).

thumbnail Fig. 9

Histograms comparing the distributions of metallicities and JHK classification, as indicated in the x-axis and the legend. Vertical error bars represent the variations of the number of stars contained in each bin considering the individual uncertainties in [M/H], whose typical value is represented by the horizontal error bar below the legend.

4.3.3 Dust disk masses of group I and group II HAeBes

The formation of giant planets requires a large enough mass reservoir in the circumstellar environment. In fact, models predict better chances to host giant planets for the sources having larger disk masses, Mdisk, within a similar stellar mass range (e.g., Alexander & Armitage 2007; Andrews et al. 2013). Guzmán-Díaz et al. (2021) found that the stellar mass distributions of hundreds of group I and II HAeBes are similar, providing Mdisk based on (sub)mm continuum emission for many of them. Figure 10 shows the stellar versus dust disk mass (Mdust = Mgas/100 ≃ Mdisk/100) distribution for most of such sources. The stars from the sample of Guzmán-Díaz et al. (2021) with upper and lower limits in the continuum fluxes were discarded. The ones having excessive continuum emission leading to Mdust > 0.01 M, probably due to contamination from the surroundings, were not included in the plot either. It is apparent from the 27 group I and 22 group II HAeBes plotted in Fig. 10 that although both groups span over the same stellar mass range, Mdust (and hence Mdisk) for group I stars tend to be larger than those for group II. Indeed, the median Mdust for group I and II stars is (4±3) × 10−4 M and (1±0.9) × 10−4 M, respectively, with the uncertainties indicating the median absolute deviations. Moreover, anon-negligible fraction close to 20% of group II sources fall below the rough disk mass-stellar mass correlation found for most lower-mass stars (black and red lines in Fig. 10; see e.g., Andrews et al. 2013; Testi et al. 2022, and references therein).

A similar result indicating smaller Mdust in group II HAeBes was recently obtained by Stapper et al. (2022) from ALMA data. In addition to this result, Stapper et al. (2022) put forward an alternative explanation. This is that group I and group II sources may have similar Mdust, but with different spatial dust grain distributions. In particular, they hypothesized that in group II sources most of the dust is concentrated in the inner, compact regions of the disks. As a consequence, such regions would be optically thick, trapping the radiation and leading to the inference of a lower Mdust than they would actually be subject to. In contrast, the inner parts of the disks around group I sources would be optically thin due to presence of a cavity, allowing the radiation to escape and yielding larger Mdust than those in group II. In turn, Stapper et al. (2022) argued that the different dust grain distributions may be associated with the presence of giant planets in the disks of group I sources. Therefore, regardless of the interpretation the typically stronger (sub-)mm continuum emission of group I stars can currently be understood in terms of a higher probability to host giant planets, compared to group II HAeBes.

thumbnail Fig. 10

Dust disk mass versus stellar mass for most HAeBes having (sub)mm emission, as compiled in Guzmán-Díaz et al. (2021). Group I and II sources are color coded as indicated in the legend. Typical power-law trends with exponents 1 and 2 are also shown in black and red, respectively. The corresponding ±1 dex scatter are indicated with dashed lines.

4.3.4 Planet candidates, metallicities, and SED groups

The hypothesis that giant planets are more likely present in disks around group I sources with low stellar metallicities than in the rest of the HAeBes can only be unambiguously tested once a statistically significant sample of such planets are detected. However, confirming potential detections of forming planets in protoplanetary disks is still a challenging task (e.g., Mendigutía et al. 2018) and has only been possible around the low-mass T Tauri star PDS 70 (Keppler et al. 2018; Haffert et al. 2019). In addition, a possible forming planet has been recently discovered around the HAeBe star AB Aur (Currie et al. 2022, see also Zhou et al. 2022). The properties of this system are fully consistent with the above hypothesis: AB Aur belongs to group I (Guzmán-Díaz et al. 2021) and we derive a relatively small [M/H] = −0.35 ± −0.25 (based on the individual abundances in Kama et al. 2015, and the procedure indicated in Sect. 4.1.1).

In the rest of this section, we aim to provide a rough quantification about the possible presence of giant planets in relation with the SED groups and the stellar metallicities, based on observed disk structures from high-resolution techniques. It must be noted that such a quantification is difficult to assess, given that most (if not all) structures observed in protoplanetary disks can be potentially explained either invoking the presence of planets or different phenomena (for a recent example, see, e.g., Demidova & Grinin 2022).

An important exception may be the “velocity kinks”, that is, perturbations of the Keplerian orbits of the gas, most probably caused by the presence of giant planets (Pinte et al. 2018; Teague et al. 2018, but see also Norfolk et al. 2022). Using such a technique Pinte et al. (2020) reported two HAeBes with plausible planets in their disks: HD 143006 and HD 163296. The previous sample can be complemented by the one in Asensio-Torres et al. (2021), where the most promising candidate young stars hosting planets based on SPHERE data were analyzed. In that work, the masses of possible giant planets capable of generating the structures observed in scattered light (rings, cavities, spirals) were estimated for the HAeBes HD 100546, HD 135344B, HD 139614, HD 97048, HD 169142, CQ Tau, HD 34282, and HD 36112.

One of the most striking features when grouping together the HAeBes of both previous works (plus AB Aur, making a total of 11 stars) is that all sources, except HD 163296, are classified as group I (Guzmán-Díaz et al. 2021). Considering the previous numbers alone, the probability of detecting disks with substructures presumably related to the presence of giant planets is about ten times larger in group I sources than in group II stars. However, the previous estimate must be considered as an upper limit, given that high-resolution observations are still biased towards brighter and more extended group I sources (Garufi et al. 2018, 2022). Concerning [M/H], none of the mentioned stars exceed the solar one except HD 163296 with [M/H] ~ 0.2 (Table A.2) and HD 36112 with [M/H] ~ 0.09 (as inferred from the values in Kama et al. 2015, and the procedure of Sect. 4.1.1). The crude estimate presented above suggests that current high-resolution data indeed support that group I sources with relatively small values of [M/H] are better candidates to host giant planets than the rest of the HAeBes.

5 Conclusions and final remarks

In this work, we derive estimates of [M/H] for 67 HAeBe stars. This is, to our knowledge, the largest sample of such objects for which these values have been derived homogeneously, complementing our previous work on the stellar and circumstellar properties of HAeBes. Values of log ɡ and υ sin i are additional byproducts of the analysis. Our new data have served to robustly confirm the proposal presented by Kama et al. (2015), namely, that HAeBes with group I SEDs typically show lower [M/H] values than those of group II. The hypothesis put forward by Kama et al. (2015) to explain this observational fact is that the presence of dust traps and cavities, potentially caused by Jovian planets trapping the refractory elements, are more frequent in disks of group I HAeBes. We provide further evidence that reinforces this hypothesis:

  • The [M/H] values of the HAeBes in our sample and the distinctions between group I and II sources do not depend on their location within the galaxy, suggesting that the observed differences in [M/H] are not caused by the local environment but are most likely connected with the cavities associated with group I stars. In contrast, we do not find any relation between [M/H] and SED-based transitional disks without infrared excess at least up to ~2.2 µm, indicating that not all mechanisms causing the associated holes (such as photoevaporation) affect the abundances of the central stars.

  • We confirm that the (sub-) mm continuum emission of group I stars is typically larger than that of group II. Regardless of the interpretation in terms of larger disk masses or different dust grain distributions, the previous finding is in line with the hypothesis that group I/low [M/H] sources are better candidates for hosting giant planets than the rest of the HAeBes.

  • Current high-resolution imaging data available from the literature support the hypothesis as well, with the most promising giant planet-hosting candidates being mainly group I sources with relatively low metallicities.

We note that our results do not rule out the presence of giant planets in disks around group II, high-metallicity sources. Strictly speaking, our results suggest that giant planets may trap the refractory material more frequently in group I sources than in group II, at least during the evolutionary stage represented by each group. Nevertheless, exoplanet synthesis models indicate that while the effect of metallicity in forming planets is relevant for low-mass stars, giant planet formation can occur in a low-metallicity (low dust-to-gas ratio) and high-mass proto-planetary disks surrounding higher mass sources (Alibert et al. 2011). Indeed, Maldonado & Villaver (2016) reported that giant stars with planets do not show the metal-rich signature. This fact could be explained by the more massive protoplanetary disks of their progenitors, since these sources are more massive than the typical FGK stars that are the main focus of the planet searches around main sequence stars.

An unambiguous confirmation of the relation between group I/low [M/H] HAeBes and the presence of giant planets in the process of formation would represent a major step towards our understanding about planet frequencies around A an B stars, bridging the gap with the FGK stars mainly explored by exo-planet surveys. The current situation is ideal for that purpose, with hundreds of HAeBes recently identified and thousands of new potential members thanks to Gaia data. However, such a confirmation requires the actual detection of forming planets around HAeBes, which is currently limited to the candidate around the group I/ low [M/H] star AB Aur.

Acknowledgements

J.G.-D. and I.M. are funded by a RyC2019-026992-I grant. J.G.-D., B.M., I.M. and E.V. acknowledge support from the “On the rocks II project” under grant PGC2018-101950-B-I00, and MDM-2017-0737 Unidad de Excelencia “Maria de Maeztu”- Centro de Astrobiología (INTA-CSIC), both funded by the Spanish Ministry of Science and Innovation/State Agency of Research MCIN/AEI. This research is based on data obtained from the ESO Science Archive Facility with DOIs: https://doi.org/10.18727/archive/27, https://doi.org/10.18727/archive/33, https://doi.org/10.18727/archive/50, https://doi.org/10.18727/archive/71. The authors acknowledge Carlos Eiroa for providing insightful suggestions based on a preliminary version of the manuscript. The authors also acknowledge the referee for her/his useful comments, which have served to improve the original manuscript.

Appendix A Tables

Table A.1

Sample and properties

Table A.2

Spectra used and stellar parameters derived in this work

Appendix B Notes on individual objects

  • HD 31648: Discrepancies in stellar parameters lead to different values of [M/H] in the literature. For instance, in Folsom et al. (2012) [M/H] = 0.23 is found for a Teff = 8800 K and log ɡ = 4.1. On the other hand, in Montesinos et al. (2009) the estimate is [M/H] = 0.0, being in this case Teff = 8250 K and log ɡ = 4.0. In this work, we derive a surface gravity similar to that of these works (log ɡ = 4.13) but a lower effective temperature is used (Teff = 8000 K), finally deriving [M/H] = −0.25.

  • HD 290380: The observed spectrum analyzed has S/N below 100 (but high enough; ≥ 50).

  • HD 287823: A value of log ɡ = 4.25, which is similar to that provided by Wichittanakom et al. 2020, was fixed. The assumed uncertainty in this parameter was the mean error of log ɡ from HAeBes with Teff > 8000 K, which is 0.11.

  • BF Ori: The v sin i of Mora et al. (2001) was considerered since the value of such parameter estimated from the process described in Sect. 3.1 does not fit properly the profiles of the photospheric lines of the selected regions in the interval 5000-7000 Å.

  • V599 Ori: The observed spectrum analyzed has a S/N below 100 (but high enough; ≥ 70).

  • HD 38120: The widths of the Balmer lines in the observed spectrum are larger than those estimated in the model generated with log ɡ = 4.50. We assumed this value of log ɡ, and the same error as HD 287823 for this object.

  • HD 87048: The value of log ɡ for each [M/H] was estimated trough PARSEC V2.1s evolutionary tracks giving the Teff from Wichittanakom et al. (2020), and the stellar mass and radius from Guzmán-Díaz et al. (2021). The considered error in log ɡ is the same as HD 287823.

  • HD 95881: The value of log ɡ for each [M/H] was estimated trough PARSEC V2.1s evolutionary tracks giving the Teff from Wichittanakom et al. (2020) and stellar mass and radius from Guzmán-Díaz et al. (2021). The considered error in log ɡ is the same as HD 287823.

  • HD 141569: The observed spectrum was challenging to analyze due to the scarcity of spectral lines, the high υ sin i, and S/N. Similar χ2 were obtained in the fits with models generated at [M/H] equal to -0.5 and 0.0. Nevertheless, for this source we decided to fix [M/H] = −0.5, as this value is similar to the one derived from Kama et al. (2015), namely, [M/H] = −0.69.

  • HR 5999: The widths of the Balmer series lines in the observed spectra are lower than those obtained in the models generated at log ɡ = 3.5. The log ɡ for each [M/H] was also estimated from PARSEC V2.1s evolutionary tracks, but no values below 3.5 were obtained. Therefore, a log ɡ = 3.5 with the same error as HD 287823 were adopted.

  • HD 149914: A log ɡ = 3.5 was assumed based on the same reason described in HR 5999.

  • HD 158643: A log ɡ = 3.5 was assumed based on the same reason described in the case of HR 5999.

  • PX Vul: The observed spectrum analyzed has a S/N below 100 (but high enough; ~ 50).

  • BP Psc: The observed spectrum analyzed has a S/N below 100 (but high enough; ≥ 70).

References

  1. Acke, B., & Waelkens, C. 2004, A&A, 427, 1009 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  2. Adibekyan, V. 2019, Geosciences, 9, 105 [Google Scholar]
  3. Adibekyan, V.Z., González Hernández, J.I., Delgado Mena, E., et al. 2014, A&A, 564, L15 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  4. Alecian, E., Wade, G.A., Catala, C., et al. 2013, MNRAS, 429, 1001 [NASA ADS] [CrossRef] [Google Scholar]
  5. Alexander, R.D., & Armitage, P.J. 2007, MNRAS, 375, 500 [NASA ADS] [CrossRef] [Google Scholar]
  6. Alibert, Y., Mordasini, C., & Benz, W. 2011, A&A, 526, A63 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  7. Andrews, S.M., Rosenfeld, K.A., Kraus, A.L., & Wilner, D.J. 2013, ApJ, 771, 129 [NASA ADS] [CrossRef] [Google Scholar]
  8. Arun, R., Mathew, B., Manoj, P., et al. 2019, AJ, 157, 159 [Google Scholar]
  9. Asensio-Torres, R., Henning, T., Cantalloube, F., et al. 2021, A&A, 652, A101 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  10. Booth, R.A., & Owen, J.E. 2020, MNRAS, 493, 5079 [NASA ADS] [CrossRef] [Google Scholar]
  11. Bressan, A., Marigo, P., Girardi, L., et al. 2012, MNRAS, 427, 127 [Google Scholar]
  12. Castelli, F., & Kurucz, R.L. 2003, IAU Symp., 210, A20 [NASA ADS] [Google Scholar]
  13. Castelli, F., Gratton, R.G., & Kurucz, R.L. 1997, A&A, 318, 841 [Google Scholar]
  14. Cowley, C.R., Hubrig, S., González, J.F., & Savanov, I. 2010, A&A, 523, A65 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  15. Cretignier, M., Francfort, J., Dumusque, X., Allart, R., & Pepe, F. 2021, Astrophysics Source Code Library [record ascl:2102.022] [Google Scholar]
  16. Currie, T., Lawson, K., Schneider, G., et al. 2022, Nat. Astron., 6, 751 [NASA ADS] [CrossRef] [Google Scholar]
  17. da Silva, L., Torres, C.A.O., de La Reza, R., et al. 2009, A&A, 508, 833 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  18. Demidova, T.V., & Grinin, V.P. 2022, ApJ, 930, 111 [NASA ADS] [CrossRef] [Google Scholar]
  19. Dullemond, C.P. 2002, A&A, 395, 853 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  20. Dullemond, C.P., & Dominik, C. 2004, A&A, 417, 159 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  21. Espaillat, C., Muzerolle, J., Najita, J., et al. 2014, in Protostars and Planets VI, eds. H. Beuther, R.S. Klessen, C.P. Dullemond, & T. Henning (Tucson: University of Arizona Press), 497 [Google Scholar]
  22. Folsom, C.P., Bagnulo, S., Wade, G.A., et al. 2012, MNRAS, 422, 2072 [NASA ADS] [CrossRef] [Google Scholar]
  23. Gaia Collaboration (Vallenari, A., et al.) 2023, A&A, in press https://doi.org/10.1051/0004-6361/202243940 [Google Scholar]
  24. Garufi, A., Meeus, G., Benisty, M., et al. 2017, A&A, 603, A21 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  25. Garufi, A., Benisty, M., Pinilla, P., et al. 2018, A&A, 620, A94 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  26. Garufi, A., Dominik, C., Ginski, C., et al. 2022, A&A, 658, A137 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  27. Glebocki, R., & Gnacinski, P. 2005, VizieR Online Data Catalog: III/244 [Google Scholar]
  28. Gonzalez, G. 1997, MNRAS, 285, 403 [Google Scholar]
  29. Gray, D.F., Scott, H.R., & Postma, J.E. 2002, PASP, 114, 536 [CrossRef] [Google Scholar]
  30. Gray, R.O., & Corbally, C.J. 1998, AJ, 116, 2530 [NASA ADS] [CrossRef] [Google Scholar]
  31. Gray, R.O., & Corbally, Christopher, J. 2009, Stellar Spectral Classification (Princeton: Princeton University Press) [CrossRef] [Google Scholar]
  32. Guimarães, M.M., Alencar, S.H.P., Corradi, W.J.B., & Vieira, S.L.A. 2006, A&A, 457, 581 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  33. Guzmán-Díaz, J., Mendigutía, I., Montesinos, B., et al. 2021, A&A, 650, A182 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  34. Haffert, S.Y., Bohn, A.J., de Boer, J., et al. 2019, Nat. Astron., 3, 749 [CrossRef] [Google Scholar]
  35. Hawkins, K. 2022, MNRAS, submitted [arXiv:2207.04542] [Google Scholar]
  36. Herbig, G.H., & Bell, K.R. 1988, Third Catalog of Emission-Line Stars of the Orion Population : 3 : 1988 (Santa Cruz: Lick Observatory) [Google Scholar]
  37. Holmberg, J., Nordström, B., & Andersen, J. 2009, A&A, 501, 941 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  38. Honda, M., Maaskant, K., Okamoto, Y.K., et al. 2015, ApJ, 804, 143 [NASA ADS] [CrossRef] [Google Scholar]
  39. Jermyn, A.S., & Kama, M. 2018, MNRAS, 476, 4418 [NASA ADS] [CrossRef] [Google Scholar]
  40. Jönsson, H., Holtzman, J.A., Allende Prieto, C., et al. 2020, AJ, 160, 120 [CrossRef] [Google Scholar]
  41. Kama, M., Folsom, C.P., & Pinilla, P. 2015, A&A, 582, L10 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  42. Kama, M., Bruderer, S., van Dishoeck, E.F., et al. 2016, A&A, 592, A83 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  43. Keppler, M., Benisty, M., Müller, A., et al. 2018, A&A, 617, A44 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  44. Kluska, J., Van Winckel, H., Coppée, Q., et al. 2022, A&A, 658, A36 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  45. Kuhn, M.A., Saber, R., Povich, M.S., et al. 2023, AJ, 165, 3 [NASA ADS] [CrossRef] [Google Scholar]
  46. Lillo-Box, J., Barrado, D., Moya, A., et al. 2014, A&A, 562, A109 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  47. Lindegren, L., Klioner, S.A., Hernández, J., et al. 2021, A&A, 649, A2 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  48. Maaskant, K.M., Honda, M., Waters, L.B.F.M., et al. 2013, A&A, 555, A64 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  49. Maldonado, J., & Villaver, E. 2016, A&A, 588, A98 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  50. Maldonado, J., Eiroa, C., Villaver, E., Montesinos, B., & Mora, A. 2012, A&A, 541, A40 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  51. Maldonado, J., Eiroa, C., Villaver, E., Montesinos, B., & Mora, A. 2015, A&A, 579, A20 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  52. McGinnis, P., Dougados, C., Alencar, S.H.P., Bouvier, J., & Cabrit, S. 2018, A&A, 620, A87 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  53. Meeus, G., Waters, L.B.F.M., Bouwman, J., et al. 2001, A&A, 365, 476 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  54. Mendigutía, I., Calvet, N., Montesinos, B., et al. 2011, A&A, 535, A99 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  55. Mendigutía, I., Fairlamb, J., Montesinos, B., et al. 2014, ApJ, 790, 21 [CrossRef] [Google Scholar]
  56. Mendigutía, I., Oudmaijer, R.D., Schneider, P.C., et al. 2018, A&A, 618, L9 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  57. Merín, B., Montesinos, B., Eiroa, C., et al. 2004, A&A, 419, 301 [Google Scholar]
  58. Montesinos, B., Eiroa, C., Mora, A., & Merín, B. 2009, A&A, 495, 901 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  59. Mora, A., Merín, B., Solano, E., et al. 2001, A&A, 378, 116 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  60. Muzerolle, J., D’Alessio, P., Calvet, N., & Hartmann, L. 2004, ApJ, 617, 406 [CrossRef] [Google Scholar]
  61. Norfolk, B.J., Pinte, C., Calcino, J., et al. 2022, ApJ, 936, L4 [NASA ADS] [CrossRef] [Google Scholar]
  62. Oomen, G.-M., Van Winckel, H., Pols, O., & Nelemans, G. 2019, A&A, 629, A49 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  63. Osborn, A., & Bayliss, D. 2020, MNRAS, 491, 4481 [Google Scholar]
  64. Pereyra, A., Magalhäes, A.M., & de Araújo, F.X. 2009, A&A, 495, 195 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  65. Pinte, C., Price, D.J., Ménard, F., et al. 2018, ApJ, 860, L13 [NASA ADS] [CrossRef] [Google Scholar]
  66. Pinte, C., Price, D.J., Ménard, F., et al. 2020, ApJ, 890, L9 [CrossRef] [Google Scholar]
  67. Santos, N.C., Israelian, G., & Mayor, M. 2000, A&A, 363, 228 [Google Scholar]
  68. Santos, N.C., Israelian, G., & Mayor, M. 2001, ArXiv e-prints [arXiv: astro-ph/0109018] [Google Scholar]
  69. Stapper, L.M., Hogerheijde, M.R., van Dishoeck, E.F., & Mentel, R. 2022, A&A, 658, A112 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  70. Teague, R., Bae, J., Bergin, E.A., Birnstiel, T., & Foreman-Mackey, D. 2018, ApJ, 860, L12 [NASA ADS] [CrossRef] [Google Scholar]
  71. Testi, L., Natta, A., Manara, C.F., et al. 2022, A&A, 663, A98 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  72. Torres, C.A.O., Quast, G.R., da Silva, L., et al. 2006, A&A, 460, 695 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  73. Vioque, M., Oudmaijer, R.D., Baines, D., Mendigutía, I., & Pérez-Martínez, R. 2018, A&A, 620, A128 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  74. Vioque, M., Oudmaijer, R.D., Schreiner, M., et al. 2020, A&A, 638, A21 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  75. Vioque, M., Oudmaijer, R.D., Wichittanakom, C., et al. 2022, ApJ, 930, 39 [NASA ADS] [CrossRef] [Google Scholar]
  76. Wichittanakom, C., Oudmaijer, R.D., Fairlamb, J.R., et al. 2020, MNRAS, 493, 234 [NASA ADS] [CrossRef] [Google Scholar]
  77. Zhang, Y.-J., Hou, W., Luo, A.L., et al. 2022, ApJS, 259, 38 [NASA ADS] [CrossRef] [Google Scholar]
  78. Zhou, Y., Sanghi, A., Bowler, B.P., et al. 2022, ApJ, 934, L13 [NASA ADS] [CrossRef] [Google Scholar]

1

[M/H] is defined as log [NM/NH]star− log [NM/NH], NM and NH being the abundances of all elements heavier than hydrogen and helium, and the abundance of hydrogen, respectively.

All Tables

Table A.1

Sample and properties

Table A.2

Spectra used and stellar parameters derived in this work

All Figures

thumbnail Fig. 1

Flow chart summarizing the process followed in the estimation of the log ɡ and [M/H] values (see Sect. 3).

In the text
thumbnail Fig. 2

Fits of the Balmer lines Hδ>, Hγ, Hβ, and six regions of the HD 244314 spectrum (black solid line) with the best model obtained in the χ2 test (red solid line). The model has been generated with a Teff = 8500 K, log ɡ = 4.08, υ sin i = 55 km s−1 and [M/H] = 0.0. In this example, the line Hβ has not been considered in the estimation of log ɡ as it shows a strong emission.

In the text
thumbnail Fig. 3

Model with Teff = 8000 K, υ sin i = 50 km s−1, and [M/H] = 0.0 in the 5000–5400 Å region is represented in black, while only the C, O, and S lines generated with the same parameters are shown in cyan. The result of subtracting the volatile elements from the first model is shown in red. The models were shifted on the y-axis for clarity.

In the text
thumbnail Fig. 4

υ sin i, log ɡ and [M/H] are compared with the values found in the literature (as indicated in the legends). Error bars represent the uncertainties in these parameters, while the lower and upper limits are denoted by right- and left-pointing triangles, respectively. The black dashed line indicates equal values.

In the text
thumbnail Fig. 5

Ca II K lines of the spectra of the stars (black solid lines) HD 39014 (left), HD 132947 (middle), and HD 31648 (right), and the corresponding best-fitting models (red solid lines). A good fit is only achieved in the first case, where no circumstellar contribution or peculiar profiles are apparent.

In the text
thumbnail Fig. 6

Histograms comparing the distributions of metallicities and SED classification, as indicated in the x-axis and the legend. Vertical error bars represent the variations of the number of stars contained in each bin considering the individual uncertainties in [M/H], whose typical value is represented by the horizontal error bar below the legend.

In the text
thumbnail Fig. 7

2D representation of the distribution in the Galaxy of the HAeBes studied in this work. The Galactic azimuthal angle in the y-axis is plotted against the distance to the Galactic center in the x-axis. The colour bar indicates their [M/H] values. The most crowded region of the plot has been zoomed in. The values for the sun – with [M/H] = 0 – are also indicated with the vertical and horizontal dashed lines, for reference.

In the text
thumbnail Fig. 8

Metallicity of the stars in our sample versus the distance to the Galactic center (top) and to the galactic plane (bottom). Group I and II sources are color coded as indicated in the legends. The lower and upper limits are indicated by the up and down triangles. The values of the sun are also indicated with the vertical and horizontal dashed lines, for reference.

In the text
thumbnail Fig. 9

Histograms comparing the distributions of metallicities and JHK classification, as indicated in the x-axis and the legend. Vertical error bars represent the variations of the number of stars contained in each bin considering the individual uncertainties in [M/H], whose typical value is represented by the horizontal error bar below the legend.

In the text
thumbnail Fig. 10

Dust disk mass versus stellar mass for most HAeBes having (sub)mm emission, as compiled in Guzmán-Díaz et al. (2021). Group I and II sources are color coded as indicated in the legend. Typical power-law trends with exponents 1 and 2 are also shown in black and red, respectively. The corresponding ±1 dex scatter are indicated with dashed lines.

In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.