EDP Sciences logo
Subscriber Authentication Point
Search
Menu
Home All issues Volume 686 (June 2024) A&A, 686 (2024) A37 Full HTML
Open Access
Issue
A&A
Volume 686, June 2024
Article Number A37
Number of page(s) 24
Section Stellar atmospheres
DOI https://doi.org/10.1051/0004-6361/202347327
Published online 28 May 2024

© The Authors 2024

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

The bottom of the initial mass function is populated by brown dwarfs (BDs), substellar objects with masses ≲80 MJup. The lightest among them are the so-called free-floating planetary-mass objects (PMOs) with masses below the deuterium-burning limit at ~12 MJup. Their existence have been confirmed by various deep surveys in nearby young star-forming regions (SFRs; Luhman et al. 2009; Scholz et al. 2012; Peña Ramírez et al. 2012; Lodieu et al. 2018; Miret-Roig et al. 2022; Bouy et al. 2022). However, details of the free-floating PMO formation process are still largely unknown and represent one of the important missing pieces in our understanding of star formation. Broadly speaking, populations of isolated PMOs in star clusters may form via cloud fragmentation and core collapse (star-like formation) or in protoplanetary disks (planet-like formation), followed by ejection due to encounters with other stars or planet-planet interactions (see Miret-Roig 2023 and references therein). The relative importance of these two scenarios is not known, mainly due to the lack of observational constraints; given their intrinsic faintness, studying PMOs is challenging with most of the current facilities. In total, only a few dozen young isolated PMOs have been spectroscopically confirmed so far and all of them have estimated masses above ~5 MJup. This situation is expected to change drastically in the near future with the advent of the James Webb Space Telescope (JWST), which will provide, for the first time, a robust census of young (1–3 Myr) free-floating PMOs in the 1–15 MJup range (Scholz et al. 2022; Pearson & McCaughrean 2023). These objects are expected to have spectral types (SpTs) in the range between ~L0 and early-to-mid T. Evolutionary models predict that a Jupiter-mass object at 1–3 Myr should have an effective temperature in the range 700–900 K; the modeled atmospheres of such objects show typical T-type features (Marley et al. 2021).

Spectral characteristics of BD atmospheres are strongly influenced by gravity, which has been extensively used to remove field contaminants from spectroscopic samples of BD candidates in SFRs (Bayo et al. 2011; Mužić et al. 2015; Esplin & Luhman 2020, to name only a few). In the near-infrared (NIR), gravity-sensitive features include the alkali lines (K I and Na I) and FeH absorption bands, which are less pronounced in young, low-gravity atmospheres, as well as the broadband form of the H band, which is shaped by water absorption and appears sharply triangular in young atmospheres, as opposed to the rounder shape seen in field dwarf atmospheres (Gorlova et al. 2003; McGovern et al. 2004; Allers & Liu 2013; Manjavacas et al. 2020; Almendros-Abad et al. 2022). The assessment of youth (and therefore the membership in a cluster or a SFR) and the SpT estimation are typically performed through a comparison of candidate spectra with a set of spectral templates or with the help of numerous spectral indices derived for both purposes (see Almendros-Abad et al. 2022 and references therein). But while the LT-type field sequence has been robustly defined (Kirkpatrick et al. 2010; Burgasser et al. 2006a), high-quality spectra of young objects are rare, making the existing young set of templates sparse for the L types (Luhman et al. 2017) and basically nonexistent for the T types. Furthermore, L-type field dwarfs exhibit significantly bluer colors than their young counterparts (Faherty et al. 2016; Schneider et al. 2017), making them unsuitable for deriving the SpTs of young objects.

The current sample of spectroscopically confirmed L-type members in nearby SFRs and young clusters with ages ⪝5 Myr, apart from being small, only extends down to SpTs L3–L4 (e.g. Alves de Oliveira et al. 2012; Mužić et al. 2015). Moreover, these members typically suffer from significant extinction, which may introduce additional uncertainties into the spectral fitting process. Due to both their vicinity and lack of extinction, it is useful to turn to nearby young moving groups (NYMGs) in order to search for a sample of relatively young objects (10–600 Myr) that span a large range of SpTs. Although due to differences in age not all the LT objects in NYMGs will be in the planetary-mass regime, they are still expected to share various spectral properties with younger objects at the same effective temperature. This kind of sample can be used in the comparison with JWST spectra of young free-floating PMO candidates, as well as with directly imaged giant planetary-mass companions (Faherty et al. 2016).

This paper builds upon the analysis presented in Almendros-Abad et al. (2022), who analyzed a large sample of NIR spectra of young and old cool dwarfs in the spectral range M0 to ~L3. They inspected a number of SpT- and gravity-sensitive spectral indices available in the literature and defined several new ones. Using machine learning models, they confirm that the broadband shape of the H band represents the most relevant feature when it comes to distinguishing low- from high-gravity atmospheres, followed by the FeH absorption bands and alkali lines in the J band portion of the spectrum. The main goal of this work is to extend the analysis of Almendros-Abad et al. (2022) to later SpTs and provide “prescriptions” to obtain an accurate SpT and gravity classification for mid-to-late L dwarfs, and eventually the T dwarfs. To that end, we present a collection of NIR spectra of LT members of various NYMGs, SFRs, and clusters with ages ⪞10 Myr. This is a starting point toward constructing a more complete set of templates of young LT dwarfs.

There are several other works in the literature that also deal with the classification and youth assessment of solar-neighbourhood L (Allers & Liu 2013; Gagné et al. 2015b; Martin et al. 2017; Cruz et al. 2018) and T dwarfs (Zhang et al. 2021). The SpT coverage of the proposed templates extends down to ~L4 but is quite patchy at later SpTs. Any new spectra would therefore help us fill in the gaps. A large fraction of the abovementioned spectral compendium comes from the northern facilities; it is therefore useful to turn to the south, where the sky is particularly rich in NYMG members. Furthermore, a large fraction of the available spectra have either very low spectral resolution (R of a few hundred) or a limited wavelength coverage (e.g. the J band in Martin et al. 2017). A wide spectral coverage in the NIR is important for properly assessing the extinction, which is in turn fundamental for the characterization of new planetary-mass members of SFRs. At medium spectral resolutions, atomic and molecular features are more easily separated, which both provides us with better line width measurements and increases the number of lines that can be used for youth assessment. Also, a medium spectral resolution enables an in-depth analysis of the objects’ atmospheric parameters (e.g., effective temperature, gravity, and metallicity) and an assessment of their cloud properties and disequilibrium chemistry (Petrus et al. 2023; Hood et al. 2023). Although a retrieval-type analysis beyond the scope of this paper, another goal of this work is to provide spectra as the basis for future studies and thereby help improve BD and planetary atmosphere models.

In Sect. 2, we present our main dataset and complementary ones, along with the details of the reduction process. We investigate the objects’ positions in a color-magnitude diagram (Sect. 3), derive their SpTs and extinctions (Sect. 4), and evaluate various features related to gravity classification (Sect. 5). In Sect. 6, we define a preliminary list of objects that can be used as standards for the spectral typing of young L and T dwarfs. A summary and our conclusions are presented in Sect. 7.

2 Dataset and data reduction

2.1 Sample selection

Young PMOs with ages of ~ 1–5 Myr and masses ~1–15 Mjup should have spectral classifications equal to ~L0 or later (Scholz et al. 2022). With this in mind, and in order to be complete, the selection criteria for the dataset used in this work are objects located in NYMGs, nearby SFRs and clusters, with NIR SpTs equal to M8 or later and with publicly available spectra from the X-shooter spectrograph at ESO’s Very Large Telescope (VLT). The sources were chosen from the list of UltraCoolDwarfs1, as well as from Allers & Liu (2013); Best et al. (2015); Burgasser et al. (2016); Kellogg et al. (2015); Manjavacas et al. (2020); Marocco et al. (2015); Naud et al. (2014); Schneider et al. (2016) and their NIR spectral classifications from the literature are between M8 and T5.5. We find that 56 objects match the described criteria. Of these, 34 objects belong to NYMGs — β Pictoris Moving Group (BPMG), AB Doradus Moving Group (ABDMG), Argus (ARG), Carina-Near (CARN), Castor (CAS), Columba (COL), Tucana-Horologium (THA), and TW Hydrae (TWA) — and 22 belong to nearby SFRs and clusters — Upper Scorpius (USCO), 32 Orionis (32OR), Praesepe (PRA), and Pleiades (PLE). Table A.1 summarizes the details of our sample and Table 1 the information regarding the mentioned regions.

In the analysis presented in the following sections, we divided the sample into four different classes based on age:

  • 10 Myr < Age ≤ 30 Myr: objects belonging to USCO, TWA, BPMG, and 32OR;

  • 30 Myr < Age ≤ 100 Myr: objects belonging to ARG, COL, and THA;

  • 100 Myr < Age ≲ 300 Myr: objects belonging to ABDMG, CARN, CAS, and PLE. All the objects in this age class have ages between 100 and 200 Myr except object #27, which is considered a member of the Castor Moving Group (~320 Myr)2;

  • Age ~ 600 Myr: objects belonging to PRA.

Table 1

NYMGs, young clusters, and SFRs included in this work.

2.2 Data reduction

X-shooter (Vernet et al. 2011) is an intermediate-resolution spectrograph that covers a wide wavelength range, ~300 to 2500 nm, with its three independent arms simultaneously: UVB arm (300–550 nm), VIS arm (550–1000 nm), and NIR arm (1000–2480 nm). The NIR arm is of the utmost relevance for this study, given the cool nature of the objects in our sample.

The publicly available data taken with X-shooter have been downloaded from the ESO archive, along with the appropriate calibrations. The EsoReflex environment (Freudling et al. 2013) was used to run version 3.5.3 of the X-shooter pipeline (Modigliani et al. 2010) for the data reduction. The data reduction steps encompass dark current subtraction, flat-field correction, order tracing, wavelength calibration, sky subtraction, and flux calibration. The final 1D spectrum was extracted from the rectified and order-merged 2D data product obtained from the pipeline. For that, we used the task apall, implementing the optimal spectral extraction algorithm, from the Image Reduction and Analysis Facility3 (IRAF; Tody 1986). The noise of the final 1D spectrum is extracted with the help of the 2D uncertainty map supplied by the pipeline. We compared the IRAF extraction with the final 1D extracted spectra obtained from the X-shooter pipeline and found that for the majority of our sample, the IRAF extraction returned spectra with a better signal-to-noise ratio (S/N). The telluric correction was performed using the Molec-fit software (Smette et al. 2015; Kausch et al. 2015). This tool computes a synthetic transmission spectrum of the Earth’s atmosphere considering the atmospheric conditions at the time of the observations.

For each object, the observing sequence typically provides several spectra, which were individually corrected. For the objects that have been observed on more than one night, we median-combined the telluric-corrected observations from each night. In each of these cases, we inspected whether the S/N of the combination improved that of an individual observation. If so, the final spectrum is the median-combined spectrum, if not, we kept the observation with a higher S/N.

In the NIR range, for spectra reduced and corrected for telluric absorption, the dominant noise contribution comes from OH line residuals from the sky subtraction process. We used the list of wavelengths corresponding to the sky emission lines from EsoReflex to remove the residuals of OH line subtraction above a defined threshold. The removed values were replaced by the mean flux in a narrow window around that wavelength. The wavelength ranges 1.138–1.143 μm, 1.165–1.185 μm, and 1.235–1.260 μm were not corrected in order not to interfere with the spectral lines present in this range: Na I doublet at ~1.14 μm and the KI doublets at 1.169, 1.177 μm and 1.243, 1.252 μm (Allers & Liu 2013), respectively.

We note that some spectra seem to present a slight discontinuity around ~2.27 μm, which coincides with one of the CO band-heads. This is probably due to an illumination (vignetting) problem at the edge of the X-shooter 11th order (Gonneau et al. 2016; Lodieu et al. 2018).

2.3 Complementary spectroscopic dataset

In several parts of the analysis presented in this paper, we used the sample of approximately 2760 publicly available NIR spectra of cool dwarfs compiled by Almendros-Abad et al. (2022, hereafter AA22). The spectra span SpTs between M0 and L5, and were obtained with a variety of telescopes and instruments. This dataset is divided according to age into the following classes: young (≲10 Myr), mid-gravity (objects belonging to NYMGs older than 10 Myr, or those classified as INT-G by Allers & Liu 2013), and field objects (older objects with no youth spectro-scopic features). The AA22 dataset was complemented by the objects with the literature SpT later than L5 from the SpeX Prism Spectral Library4 and the Montreal Spectral Library5. In total, 98 spectra were included. To have a homogeneous spectral classification for the entire complementary sample, we re-derived the SpTs for the added objects using the same method as in AA22, considering the T-type field templates from Burgasser et al. (2006a). Of the 98 objects, 18 belong to the mid-gravity age class from AA22, whereas the remaining is inserted in the field class. In several figures presented in this paper, we maintain the same color coding for the three age classes as in AA22: orange for the young objects, blue for the mid-gravity, and gray for the field.

3 Color-absolute magnitude diagram

We started our analysis by examining the properties of our sample in the color-absolute magnitude diagram. The magnitudes in the J, H and K bands were retrieved from 2MASS (Cutri et al. 2003) and complemented with catalogs from several other authors (Luhman & Esplin 2020; Bouy et al. 2022, 2015; Miret-Roig et al. 2022; Medan et al. 2021; Faherty et al. 2016; Smart et al. 2017; Burgasser et al. 2016; Liu et al. 2016; Marocco et al. 2015; Lodieu et al. 2018). The parallaxes were taken from Gaia Early Data Release 3 (EDR3; Gaia Collaboration 2016, 2021) using a matching radius of 2″ and were complemented with measurements from several different authors (Liu et al. 2016; Best et al. 2021; Galli et al. 2017; Manjavacas et al. 2019). Only 28 of the initial 56 objects have trigonometric parallax measurements. For objects located in USCO and PRA that do not have a parallax measurement, we assumed the distance to be equal to the distance to the respective young association. We calculated the mean distance to USCO using Gaia DR3 parallaxes and the latest literature census of members of this region (Luhman 2022). We find a mean distance to USCO of 145±9 pc (see Table 1), which is in agreement with previous works (145±2 pc, Cook et al. 2017; 130±20 pc, Gagné et al. 2018). Object #63 is the only one with a parallax measurement but no available 2MASS photometry. We implemented the magnitude system conversions from Leggett et al. (2006) to convert the MKO magnitudes (Zhang et al. 2021) to 2MASS.

Figure 1 presents the 2MASS absolute magnitude in the J band as a function of (J–KS) for 44 of our initial 56 objects. We overplotted young (ages ≲ 10 Myr) and field objects from AA22 and L dwarfs, with gravity classifications of VL-G (very low-gravity) and FLD-G (field gravity), and T dwarfs from the Database of Ultracool Parallaxes6 (Dupuy & Liu 2012; Liu et al. 2016; Dupuy & Kraus 2013). The objects belonging to the young class of AA22 and those with gravity classification equal to VL-G are shown as orange dots whereas the FLD-G L dwarfs and T dwarfs as gray dots. The information regarding magnitudes and parallaxes for the objects in AA22 were also retrieved from 2MASS and Gaia EDR3, respectively, considering the same matching radius as for our sample. For the case of the remaining objects, their 2MASS absolute magnitudes in the J band and their (J–KS) colors are provided in the respective paper. Contrary to the objects in our sample, these are corrected for extinction.

The NIR colors of M- and L-type dwarfs are directly linked to their surface temperature: L dwarfs have a lower surface temperature thus they will exhibit redder colors. Moreover, young L dwarfs are typically redder than their field counterparts of the same SpT (Kirkpatrick et al. 2008; Cruz et al. 2009; Faherty et al. 2016), thought to be a consequence of their lower surface gravity leading to the formation of clouds in the upper layers of their atmospheres (Madhusudhan et al. 2011), or, alternatively, of thermochemical instabilities in cloudless atmospheres (Tremblin et al. 2016). As we move toward lower temperatures, into the T-dwarf sequence, we can observe a dramatic shift toward bluer J–KS colors, often interpreted as a transition toward “cloudless” spectra, that is to say, the clouds in these objects are expected to form so deep in the atmosphere that they have little or no influence on the final spectrum (Faherty et al. 2016).

The ellipses shown in Fig. 1 mark a rough location of different spectral classes. We can observe that the majority of the objects with expected SpT classifications of M and L appear to follow the young sequence. Nonetheless, some objects appear to be more in agreement with the field sequence (e.g., objects #2, #4, #27, #59, to name a few), which might suggest older ages.

thumbnail Fig. 1

(MJ, J–KS) color-absolute magnitude diagram of our objects, very low-gravity objects (orange circles), and field objects (gray circles) from the literature (AA22; Database of Ultracool Parallaxes). Our sample is divided according to the age of the association to which the objects probably belong. Green stars represent objects in regions with ages <30 Myr, dark gray pentagons with ages between 30 and 100 Myr (included), red diamonds with ages between 100 and 300 Myr (included), and white squares with ages ~600 Myr. For data points where the error bar is not visible, the bar is smaller than the marker size. The three ellipses approximately highlight where the M, L, and T dwarfs are located in this diagram. The arrow indicates what effect the addition of five magnitudes of extinction would have.

4 Spectral type and extinction

In this section, we derive the SpTs for the objects in our sample. To that end, compared the X-shooter spectra to a set of spectral templates (Sect. 4.1), as well as a variety of SpT-sensitive indices (Sect. 4.3). We also performed a methodical search for potential unresolved binaries in our sample (Sect. 4.2).

4.1 Comparison to spectral templates

The first method we used to derive the SpT is a direct comparison of the spectra with both young and field NIR spectral templates defined in the literature. The young spectral standards are from Luhman et al. (2017), whereas the field dwarf templates come from Kirkpatrick et al. (2010) and Burgasser et al. (2006a). The L-type spectral sequence (Luhman et al. 2017) has been constructed using the solar-neighborhood objects with an age of ≲ 100 Myr, a subset of which have been associated with NYMGs. Table 2 summarizes the different sets of templates, along with the SpT ranges for which they were defined. The regions affected by the telluric absorption were excluded from the fitting process (1.3–1.5 μm and 1.78–2.0 μm), and the spectra were resampled to have the same wavelength grid of the template that they were being compared to.

In the fitting process, extinction may also be used as an additional free parameter. Within our sample, this is only relevant for the objects belonging to USCO, which suffer low, but non-negligible, amount of interstellar reddening (Ardila et al. 2000; Rizzuto et al. 2015). In this case, the extinction AV was varied between 0 and 2 mag (Lodieu et al. 2008), with a step of 0.2 mag. We used the extinction law from Cardelli et al. (1989) with Rv = 3.1 to redden the spectral templates. For the remaining objects belonging to NYMGs (located within the so-called Local Bubble, d ≲ 100 pc) and the two open clusters (both PLE and PRA have low to negligible foreground reddening: E(B – V)PLE = 0.04 mag, Breger 1986; E(B – V)PRA = 0.027 ± 0.004 mag, Taylor 2006), we set AV = 0 (Bell et al. 2015; Pecaut & Mamajek 2013; Lodieu et al. 2019).

Similarly to the method implemented in AA22, the fitting process has three variables – SpT, extinction, and normalization wavelength. The normalization wavelength is searched for on a grid from 1.55 to 1.78 μm with a step of 0.02 μm. For objects in associations other than USCO, there are only two variables (SpT and normalization wavelength). The best-fit template is the one that minimizes the χ2 parameter, defined as χ2=1N−m∑i=1N(Oi−Ti)2σi2,${\chi ^2} = {1 \over {N - m}}\sum\limits_{i = 1}^N {{{{{\left( {{O_i} - {T_i}} \right)}^2}} \over {\sigma _i^2}}} ,$(1)

where N is the number of points included in the fit (number of wavelength values), m is the number of free parameters (m = 3 for USCO and m = 2 for the rest), O and σ are the object spectrum and its associated noise, respectively, and T is the template spectrum. All spectra were individually compared with each one of the field and young templates. Next, we compared the χ2 values of the two comparisons and used the template (young or field) associated with the lower χ2 value to derive the SpT and extinction, where applicable. As an estimate of the uncertainty on the derived SpT, we used the SpT step at which the correspondent set of templates is defined. In the case of young templates, if the SpT derived is L0 or L2 we assumed an uncertainty of ±2.0 sub-SpTs, −2.0+3.0$_{ - 2.0}^{ + 3.0}$ sub-SpTs for L4 and ±3.0 sub-SpTs for L7, due to the lack of young L1, L3, L5, L6, L8, and L9 young templates.

In several cases, the attributed best-fit SpT in the M- and L-type spectral range was that of a field template, which is somewhat unexpected given the youth of the sample. For these objects, we performed an additional check in order to visually inspect the quality of the best-fitted young and field templates.

Figure 2 shows the affected spectra, along with the best young (red) and field templates (purple). A particularly useful feature to observe here is the broadband shape of the H band, which has a well-defined triangular shape in the young templates. Objects #57 and #58 are located in PRA, which is by far the oldest region considered here, and are better represented by the field templates. Furthermore, field templates also appear to work well for the somewhat younger sources, such as #27 located in CAS, and #59 and #60 in PLE. For these objects, we retained the SpT attributed by the fitting process. The remaining spectra visually appear to be better represented by the young templates, and therefore we took the best-fit SpT derived using the young template as the final one.

The reason why these spectra originally appeared to be better fitted by the field templates is probably the sparsity of the young template grid. For the same reason, we decided to visually inspect the L-type objects to determine if some of them have a SpT that would appear to be in between the SpTs defined by the grid. In Fig. 3, we show a set of objects with SpTs in the range L0-L2, and in Fig. 4 those with SpTs potentially in the range L2-L4. From Fig. 3, we see that several spectra appear to be in between the L0 and L2 templates, and we therefore reassigned them a SpT L1±1 (objects #23, #24, #31, #34 and #35; see Table A.2). The range of L2-L4 (Fig. 4) is more difficult to judge visually. In most cases, at least two bands are very well fitted by one of the two templates, while the third one (most often the K band) appears a bit off. However, we cannot say either that something in between would necessarily provide a better fit. For these objects, we therefore retained the young SpT as returned by the fitting procedure.

In Fig. 5, we show the objects best fitted by the young L4 or L7 templates. Four of the shown objects (#21, #28, #42, and #62) are closer to the L4 templates, although most of them do not consistently follow the template shape in all bands. In the case of object #21, its shape appears to be in between L4 and L7, but closer to L4. We therefore assigned it a SpT of L5. Four objects were classified as L7 young by direct comparison with spectral templates. The spectra of objects #25 (2MASS J03552337+1133437), #47 (2MASS J11193254-1137466), and #48 (2MASS J114724.10-204021.3) seem well fitted by a L7 young template, although in the case of #25, both the J and K bands appear somewhat fainter relative to the H band when compared to the template. As for object #20 (PSO J318.5-22), Fig. 6 compares its spectrum (in black) with that of object #56 (SIMP J013656.5+093347.3, in gray), which is classified as T2, and that of an L7 young (in red) and an L8 field templates (in purple). Compared to the young L7 template, object #20 appears to be of a slightly later SpT; however, it is difficult to assign an exact SpT to it, given the lack of appropriate comparison templates. The L-type field templates are too blue compared to young L types and thus completely inadequate for spectral typing in this age regime. We show object #56 from our sample, which is a ~200 Myr old T2 dwarf, to demonstrate the rapid change in the spectral form at the LT transition. We assign this object a SpT L8±1, given that it appears later than L7, but is also clearly not a T dwarf.

Young, low-gravity mid-L and late-L dwarfs typically show redder NIR colors than their older, higher-gravity counterparts with the same SpT (Faherty et al. 2016; Liu et al. 2016). For several T dwarfs in their sample, Zhang et al. (2021) report J – K colors that are 0.4–0.8 mag higher than what would be expected for field dwarfs. We observe a similar trend for four out of five T dwarfs in our sample (see Fig. 7). Several of these objects also appear as potential binary candidates. As we discuss in Sect. 4.2, it seems that the low-gravity behavior in T dwarfs may be mimicking binarity when assessed using the field T-dwarf templates.

Finally, we did not perform a similar visual inspection for the L-type objects located in USCO since they may have some degree of extinction, which makes it more difficult to judge visually if some intermediate SpT would be more appropriate (SpT+AV degeneracy). The final spectral classifications obtained by comparison with spectral templates and posterior visual inspection, are summarized in Table A.2, column “SpT Templates”.

In Fig. 8, we compare the SpTs derived in this section and those obtained from the literature. The solid lines indicate a perfect agreement, while the dotted lines mark ±2 sub-SpT uncertainty. We see that the majority of our results agree well with the literature classification, except for three objects. Object #16 belongs to USCO and its literature classification comes from Lodieu et al. (2018), which does not consider extinction when deriving SpTs. The effect of increasing the extinction in the spectra is similar to the effect of decreasing the effective temperature: by ignoring the extinction, one will obtain systematically later SpTs7. As for object #37 (J0357-4417), it is a known unresolved binary system identified by Bouy et al. (2003). Marocco et al. (2013) estimated the NIR SpTs of its components by fitting synthetic binary templates and the best-retrieved fit was that of an L4.5+L5(± 1) composite. On the other hand, in the optical, the SpTs derived for this object were M9+L1 (Martin et al. 2006). The difference between classifications was assigned to the fact that the NIR composite templates were created from field dwarfs and object #37 exhibits signs of youth. The assigned unresolved NIR SpT in the literature was L2pec (Marocco et al. 2013), based on the best fit in the J band. Our composed SpT derivation is M9.5±0.5. Nonetheless, we note that the template does not reproduce well the object’s spectrum (see Fig. B.1). Regarding the one object located below the lower dotted line (object #25), it was classified as an L3 dwarf in Allers & Liu (2013) by combining the information from direct comparison with spectral templates and index classifications. However, when analyzing individually the comparison between moderate-resolution spectra and templates for the J and K bands separately, they obtained the best match of L7 and L2, respectively. In this work, we estimate a spectral classification of L7.0± 1.0 by direct comparison with spectral templates. Nonetheless, we note that although the H band is well fitted, both the J and K bands are redder than the template fitted. An important point to note is that different normalization wavelengths can yield slightly different spectral classifications. In this work, we use the H band; however, some other works compare their template fittings by normalizing the spectra in the J band (e.g., Kirkpatrick et al. 2010).

Table 2

Sets of spectral templates used for the NIR SpT derivation.

thumbnail Fig. 2

Comparison between the best-fit young (red) and field (purple) templates for dwarfs where the spectral classification made via a direct comparison with templates yields a result of an M or L field. The ID and young association are shown above the correspondent spectrum (in black). All spectra are normalized at 1.67 μm. An offset between the spectra was added for clarity.
thumbnail Fig. 3

Spectra of objects in the SpT range L0–L2 (in black) compared with L0 (red) and L2 (purple) young templates. The spectra were normalized at 1.67 μm. The ID of each object is listed above the respective spectrum. An offset between the spectra was added for clarity.
thumbnail Fig. 4

Spectra of objects in the SpT range L2–L4 (in black) compared with L2 (red) and L4 (purple) young templates. The spectra were normalized at 1.67 μm. The ID of each object is listed above the respective spectrum. An offset between the spectra was added for clarity.
thumbnail Fig. 5

Spectra of objects in the SpT range L4–L7 (in black) compared with L4 (red) and L7 (purple) young templates. The spectra were normalized at 1.67 μm. The ID of each object is listed above the respective spectrum. An offset between the spectra was added for clarity.
thumbnail Fig. 6

Spectrum of object #20 (PSO J318.5-22, in black) along with the L7 young template (in red), L8 field template (in purple), and the spectrum of object #56 (SIMP J013656.5+093347.3, in gray), which was classified as T2 via a direct comparison with spectral templates. This comparison is shown to underline the significant change in the spectral morphology associated with the L-T transition. All spectra are normalized at 1.67 μm.
thumbnail Fig. 7

Spectra of objects classified as T-type dwarfs along with their best-fit template (in red). The ID, derived SpT, and region of each object are shown next to the corresponding spectrum. All the spectra are normalized at 1.57 μηι. An offset between the spectra was added for clarity.
thumbnail Fig. 8

Comparison between the derived NIR SpT via a direct comparison with spectral templates (see Sect. 4.1) and the literature NIR SpT for the X-shooter dataset. The solid black line indicates a perfect agreement between classifications. The dotted black lines represent the ±2 sub-SpT range. The ID of the objects for which their classifications do not agree within ±2 sub-SpTs is shown.
thumbnail Fig. 9

Spectral index criteria applied for unresolved binary candidate selection as defined in Burgasser et al. (2010). The top three panels and the first bottom panel compare pairs of index values, whereas the other panels compare index ratios to the NIR SpT derived in Sect. 4.1. The objects from our dataset are represented as black dots. Strong binary candidates are highlighted with red diamonds and weak binary candidates with red circles.

4.2 Unresolved binaries

Some objects in our sample show peculiar spectra that do not provide a good fit when compared to any of the spectral templates. One of the main reasons that might explain the observed peculiarity of a spectrum is multiplicity. The transition from L- to T-type objects supposes significant changes in spectral properties, caused by pronounced, deep molecular absorption bands (e.g., CH4 and H2O) developing at the lower temperatures. Hence, a blended-light spectrum from an L/T binary combination will appear peculiar when compared to spectra of other ultracool dwarfs (e.g., Bardalez Gagliuffi et al. 2014; Ashraf et al. 2022).

To test whether our sample harbors potential unresolved binaries, we used the method defined in Burgasser et al. (2010), based on six different combinations of index-index and index-SpT diagrams that segregate possible unresolved LT-dwarf binaries. If an object matches two of the six criteria it is selected as a weak binary candidate, whereas satisfying at least three criteria, places the object among the strong candidates. This method is especially sensitive to companions that differ by more than one SpT (i.e. a combination of an L-type and a T-type object).

The selection process is shown in Fig. 9, where six criteria for identifying potential binarity are delineated by dashed lines. We have identified five binary candidates: objects #52 and #55 were selected by six indices and object #63 by three indices therefore these are considered strong candidates; the weak candidates are objects #51 and #53 which were selected by two indices each. The strong and weak binary candidates are highlighted as red diamonds and circles, respectively.

To estimate the SpT of both companions of the binary candidates, we used synthetic binary templates, similar to the technique described in Day-Jones et al. (2013). For this exercise, we combined the T-type templates with the L-type field templates, since the young L-type templates do not have a complete sequence. The templates were normalized to 1 at 1.28 μm and scaled to a common flux level using the MJ-SpT relation from Faherty et al. (2016). Figure 10 exhibits the best-fitted individual templates (in magenta) along with the best-fitted composite templates (in green). The two different components of the synthetic binary templates are shown in yellow and blue. Table 3 summarizes the SpTs derived by these two different comparisons, complemented with the respective χ2 parameters.

The three strong binary candidates have previously been reported as such. Object #52 has been disentangled into an L8+T7 binary system by Day-Jones et al. (2013, BLRT15), in agreement with the SpTs estimated here. #55 has been reported as an L6+T5 composite in Marocco et al. (2015, BLRT203), with both SpTs slightly earlier than those found here. It was also selected as a strong binary candidate by Zhang et al. (2021) with a composite spectral classification of L7+T6. In their discovery paper, Naud et al. (2014) identified the T-type companion to the young M3 star GU Psc (#63 here) as a weak binary candidate, following the same set of criteria as we did. Its spectrum appears to be more similar to the confirmed T1+T5 tight binary SDSS J102109.69-030420.1 (Burgasser et al. 2006b) and a T2+T6 binary candidate J121440.95+631643.4 (Geißler et al. 2011) than to any of the single templates.

Similar to #63, the two weak candidates (#51 and #53) appear as composites of two T-type objects8. Particularly, in the case of object #53, appearing to be a combination of two objects with similar SpTs (T2 and T4), the improvement in the fit is only marginal, making its binary status highly questionable. We decided to retain these three objects for the remainder of the analysis since they represent one of the very few known T-type objects with low surface gravity. Our understanding of these cool atmospheres at young ages is rather limited, and, since all the claims of the composite nature of the objects come from the field templates, one cannot exclude that their slight peculiarity with respect to the standards may be related to the surface gravity. Three out of five T dwarfs in our sample appear redder than their field counterpart(#51, #63 and #50). Objects #51 and #63 were selected as binary candidates, along with object #53. The remaining object, #50, has a later SpT (T5), which is not included in the binarity criteria of Burgasser et al. (2010). Interestingly, high-resolution imaging reveals that binary factions among the T dwarfs are of the order 10% (Burgasser et al. 2003, 2006b; Fontanive et al. 2018), making it highly unlikely that at least 60% of our sample of T dwarfs, however small, should indeed be binaries. Considering this, the lower surface gravity of the T dwarfs in our sample might be mimicking binarity.

This selection method was defined to select binary candidates that straddle the L/T transition. Consequently, there are other objects in our sample that can be potential binary candidates but are not selected as such as a result of their similar SpTs. As discussed previously in Sect. 4.1, object #37 is one such example. Another case is object #47 (J1119-1137), a tight binary where the components have approximately equal magnitudes resulting in an unresolved SpT of L7 (Best et al. 2017). As the spectral composition of objects with L classifications does not result in such spectral morphological changes as an L+T dwarf composition, we did not perform a methodical search for them. Object #37 and #47 are therefore also retained in the following analysis such as the weak selected binary candidates.

thumbnail Fig. 10

Strong (objects #52, #55 and #63) and weak (objects #51 and #53) binary candidates (in black) overplotted with the best-fit individual templates (magenta) and the best-fit composite templates (green). The two components of the synthetic binary templates are also plotted in yellow and blue, respectively.
thumbnail Fig. 11

Selected SpT indices from the literature versus the NIR SpT derived via a direct comparison with spectral templates. The classes young, mid-gravity and field from the dataset of AA22 are represented as orange, blue and gray circles, respectively. The black solid lines represent the sensitivity range of each index (see Table 4). The red dashed line in the spectral index ωD represents the best fit if the sensitivity range is considered from M4 down to T8. The black arrows demonstrate the effect of 5 magnitudes of extinction on the index values.
Table 3

SpT derived for the strong and weak binary candidates via a direct comparison with single and composite templates.

4.3 Spectral type indices

An alternative way to derive SpT is by using spectral indices that correlate well within a given SpT range. This approach is of great interest when, for example, there is no access to the entire NIR wavelength range of the object’s spectrum, leading to a decrease in the accuracy of the SpT derivation by direct comparison with spectral templates. Moreover, quantifying specific spectral features does not require an age estimate (if defined as gravity-insensitive) and is less computationally costly. Therefore, it is of utmost relevance to test several indices in order to inspect which are their sensitivity ranges and whether they can be used for spectral classification. One thing to consider is that the main disadvantage regarding the use of spectral indices is the fact that the majority of them are extinction-dependent. Hence, extinction needs to be estimated using another procedure.

In total 15 indices matched the criteria mentioned above: H2OB, K1, H2O-C, H2O-H, H2O-1.5, ω0, ωD, sHJ, CH4A, CH4B, K2, CH4-B, CH4-C, CH4-H and CH4-K. Figures 11 and 12 show the indices selected along with their sensitivity range (solid black line) and the effect that an additional extinction of five magnitudes would have on the values of the indices (black arrow). Table 4 summarizes the ranges of applicability, the type of polynomial fit along with the respective coefficients, and the root mean square error of each selected SpT index.

We then proceeded to the computation of the selected indices for each object of our sample. For the objects in USCO, we considered the extinction value estimated in Sect. 4.1 to correct the spectra for reddening, using the Cardelli extinction law with RV = 3.1. The SpT and its uncertainty were calculated as the weighted mean and weighted standard deviation of the values obtained from various indices, which depend on each index’s SpT range of applicability. The final values have been rounded to the next 0.5 sub-SpT, and are presented in Table A.2, column “SpT Indices”.

The plots shown in Figs. 11 and 12 show a good correlation with SpTs earlier than ~L6, and later than ~T0. The LT transition range is not well populated and also does not show good behavior in the sense that most indices appear rather flat and/or with high dispersion. The only potentially valuable index for this range may be ωD (lower-right panel of Fig. 11), which has also the advantage of being extinction-independent. For this reason, we try to fit a single polynomial to the entire range of the plot (red dashed line). However, in the range L6-T1 we notice that basically all the points are below the fit line, which means that the index tends to provide systematically earlier SpTs in this spectral range. The SpT range L6-T0 is therefore not included in any sensitivity range of the SpT indices selected.

In Fig. 13, we compare the estimated SpTs obtained with both classification methods: comparison with spectral templates and spectral indices. The classifications seem to agree well within ±2 sub-SpTs, except for 6 objects in our sample. Three of these six objects (IDs #20, #25, and #47) are classified as late-L dwarfs (L8±1, L7±3, and L7±3, respectively) by direct comparison with spectral templates and posterior visual inspection but have an earlier SpT derived with the index classification method (L4±3, L2.5±1.0, and L4.5±3.0, respectively). This is not unexpected, given the behavior of the indices for the late-L types, as discussed above. Objects #11 and #13 are located in USCO and exhibit discrepant classifications by more than two sub-SpTs as well (L4−2+3${\rm{L}}4_{ - 2}^{ + 3}$, L2±2 using templates and M9±1, M9.5±1.0 using SpTs indices, respectively). On the other hand, the SpT retrieved from the comparison with spectral templates (L0±2) for object #38 is earlier than the one computed using the spectral indices (L2.5±0.5). The spectrum of object #11 is quite noisy in several regions important for index calculation, which is probably responsible for the discrepancy. In the case of the objects #13 and #38, we notice that the J-band part of the spectrum is not that well reproduced by the best-fit spectral template fits.

thumbnail Fig. 12

Selected SpT indices from the literature versus the NIR SpT derived via a direct comparison with spectral templates. The classes young, mid-gravity and field from the dataset of AA22 are represented as orange, blue and gray circles, respectively. The black solid lines represent the sensitivity range of each index (see Table 4). The black arrows demonstrate the effect of 5 magnitudes of extinction on the index values.
Table 4

Spectral type indices selected.

thumbnail Fig. 13

Comparison between the derived NIR SpT using spectral indices (see Sect. 4.3) and the SpT derived via a direct comparison with spectral templates (see Sect. 4.1). The solid black line indicates perfect agreement between classifications. The dotted black lines represent the ±2 sub-SpT range. The ID of the objects for which their classifications do not agree within ±2 sub-SpTs is shown.

5 Evaluation of youth-related features

Visual comparison of NIR spectra of young dwarfs to their field counterparts has revealed a wealth of spectral (atomic and molecular) features that differ with age, directly related to differences in surface gravity of these objects (e.g, Lucas et al. 2001; Gorlova et al. 2003; Luhman et al. 2017; Allers & Liu 2013). AA22 applied machine learning methods to the low-resolution spectra in the range between M0 and L3, and find that the features with the most importance when it comes to separating low-gravity young dwarfs from their field counterparts are the broadband shape of the H band, the FeH bands, and the lines of the alkali elements (KI and NaI). In this section, we evaluate and discuss various spectral indices sensitive to gravity and their applicability to the spectral range >L3. We also measure the pseudo-equivalent widths (pEWs) of various prominent alkali lines and study their behavior with SpT and age. In the following analysis, we assume the SpT of the sample to be that estimated by comparison with spectral templates (see Sect. 4.1).

5.1 Gravity-sensitive NIR spectral indices

The gravity-sensitive indices that we visually inspected from the literature are listed in Table D.2 along with their functional forms, spectral features targeted and references. Some of the indices inspected in Sect. 4.3 also demonstrated a gravity-sensitive behavior, such as sH2OJ, CH4 (1.6 μm), WH, QH, sH2OH1, K2, CH4-B, CH4-C, CO, CH4-H and CH4-K. Where applicable, the spectra were corrected for the extinction derived in Sect. 4.1 using Cardelli’s extinction law with RV = 3.1. Given the scarcity of low-gravity objects in the late L and T-type regime, this discussion is limited to the SpTs down to L7.

In Fig. 14, we show a set of gravity-sensitive indices, using the dataset from AA22, along with the objects analyzed in this work. The orange dots show objects that are typically younger (< 10 Myr) than our sample, but we see that they anyway separate well from the field objects. This means that some of these indices will be useful to classify the young members of SFRs in future deep spectroscopic surveys.

Both HPI and TLI-g are based on the H2O absorption in the H band. HPI correlates well with the SpT for the younger and field populations from SpT M6 while still being gravity-sensitive. The fact that a correlation with SpT exists between the different age classes provides an easier way to separate them. TLI-g shows a clear separation from, at least, M2 and the intermediate-gravity objects populate the region between the young and field populations. Our objects seem to overlap mostly with the intermediate-age class. The youngest objects from our sample have at least 10 Myr and the young class from AA22 has ages ≲ 10 Myr. Therefore, the separation might remain if late-L objects from SFRs with ages of only a few megayears follow the trend visible down to L2 for young dwarfs. The objects in the <30 Myr age class later than L4 seem to indicate that this may be so. Another water-based index, H2O-J, does not appear gravity-sensitive in the M spectral range but shows a progressively more gravity-sensitive behavior toward the late L types.

The gravity-sensitive indices that measure the methane features, both in the H and K bands, were defined as SpT indices to classify objects in the T spectral sequence. Interestingly, they can have a dual role, because in L dwarfs the spectral regions considered to compute these indices are actually quantifying the slope of the blue end of each band caused by the water absorption features. The CH4–1.6 μm index considers both absorption peaks in the H band of T dwarfs enhanced by the methane absorption bands, as does CH4–H but with the inverse formulation of CH4–1.6 μm. Both indices increase monotonically through the T sequence and thus can be used to classify T-type objects. However, for SpTs earlier than L6, they demonstrate a good performance as gravity-sensitive indices. The young and field populations can be easily separated since M0, having some overlap within M3 and M5, but for SpTs>M6 the separation is very pronounced. WH measures the water absorption feature on the blue side of the H band and has a similar behavior to CH4–1.6 μm index. CH4–K measures the 2.2 μm methane absorption feature and it does not exhibit a similar behavior as the other selected indices. It does not perform so well in the separation of the different age classes for the early- to mid-M sequence. However, for SpTs later than M8, the trend of the younger and older objects starts to increase and decrease, respectively. This behavior is interesting and could be extremely useful to disentangle young and field populations in the L sequence.

An equivalent analysis of gravity-sensitive indices would be extremely interesting for the T-dwarf range, in light of the upcoming JWST results. This has, so far, been basically impossible given the lack of known young T-type objects. We note that the T dwarfs included in this study appear in agreement with the field dwarf positions in plots like that shown in Fig. 14.

thumbnail Fig. 14

Selected gravity-sensitive spectral indices from the literature versus the NIR SpT derived in Sect. 4.1. Our sample is divided into the different age classes presented in Sect. 2.1 where the classes <30 Myr (USCO, TWA, BPMG, and 32OR), 30–100 Myr (ARG, COL, and THA), 100–300 Myr (ABDMG, CAS, CARN, and PLE), and 600 Myr (PRA) are shown as green stars, black pentagons, red diamonds, and white squares, respectively. The complementary dataset is divided into young, mid-gravity, and field objects represented as orange, blue, and gray dots, respectively.

5.2 Pseudo-equivalent widths of alkali lines

The alkali lines (Na I and K I) are the dominant atomic spectral features in the J-band spectra of ultracool dwarfs but are also found in the H and K bands. The strength of the absorption of the alkali lines is primarily affected by the surface temperature of the object, but pressure broadening plays a major role in creating their final shape. They have been found to be a very good probe for differences in surface gravity for late-type objects (Gorlova et al. 2003; Schlieder et al. 2012; Allers & Liu 2013; Martin et al. 2017; Manjavacas et al. 2020). In this section, we study the pEWs of the alkali lines for the objects in our sample and compare them with that of field objects. We focused on the K I doublets at 1.17, 1.24 and 1.51 μm and the Na I doublet at 1.14 μm. The NaI doublet in the K band (2.21 μm) basically disappears at L types (McLean et al. 2003; Cushing et al. 2005), which is our principal SpT range of interest in this work, and thus was not considered.

The pEWs were measured using the same method as in Manara et al. (2013), which integrates the Gaussian fit of the line, previously normalized to the local pseudo-continuum at the edges of the line. The errors in the pEW measurements are estimated from the propagation of 1σ of the estimated continuum. We visually checked that the extent of all the lines is correctly estimated, and corrected them if needed. We also inspected the appearance of the alkali lines. Some sources presented contamination from single-pixel spikes within the extent of the line: the pEW of those lines is not considered. We also measured the pEWs of the alkali lines for a sample of field dwarfs with SpTs M6-T8 with a somewhat lower spectral resolution than that of X-shooter (R ≈ 2000), coming from different public surveys (McLean et al. 2003; Cushing et al. 2005).

In Fig. 15, we present the pEW of the alkali lines as a function of the SpT. The sources from our X-shooter sample are separated into age classes following the same color scheme used in the rest of this work. Field objects are represented with purple circles. In order to get a clearer view of potential trends in the data, on the right side of each plot we show the distribution for each age class smoothed with a kernel density estimator (KDE). In the L-type range, our objects are clearly distinguished from the field ones, having lower pEW values, as expected given their low-gravity nature. A similar trend has previously been reported by Martin et al. (2017) down to L0 and Manjavacas et al. (2020) down to L4, and here we show that the trend maintains also for later L types for most of the lines. The exception is the K I 1516.7 nm and NaI 1139 nm lines (the two rightmost panels of Fig. 15), where we see that the distinction between the objects in our sample and the field dwarfs maintains only down to ~L4. Similar behavior for the Na I line can be observed in Fig. 21 of Allers & Liu (2013). Regarding the KI 1516.7 nm line, field sources with SpT earlier than L5 present larger pEWs, whereas young sources present very low pEWs (<1.5 Å).

For the K I lines located in the J band, we also observe that the youngest objects in our sample (<30 Myr; green stars) have on average lower pEW than the objects belonging to the next age group (30–100 Myr; gray pentagons), which is in turn not well distinguished from even older objects in the X-shooter sample. A similar observation has been previously made for the K I lines, but also for other lines in the optical portion of the spectra by Manjavacas et al. (2020). They found that the pEWs do not significantly change after the age of PLE, meaning that low-mass stars and more massive BDs have nearly approached their final radii after their first ~ 100 Myr, which is also in agreement with the evolutionary models (Baraffe et al. 2015). However, although the trend of lower pEWs for young with respect to the field dwarfs seems to maintain throughout the SpT L, we also observe a general drop in pEW values for all objects with SpT in the range L6–L8, also seen in plots of Allers & Liu (2013). At the transition from L to T-type field dwarfs, we see either a plateau in pEW values or an increase, followed by a decrease toward the later T types.

Although we have only a few young T dwarfs in our sample, it seems that they maintain the trend of lower pEWs at least in two lines (KI at 1169.2 nm and 1177.8 nm). In other cases, the distinction from the field objects is not that clear. Considering the field age class, the K I doublets in the J band seem to decrease drastically for later T types (SpT ⪞ T4/T5). This behavior was also observed in previous works (e.g., Martin et al. 2017). Clearly, it is an imperative to populate this SpT region with younger T dwarfs to be able to draw further conclusions.

However, the considerable decrease in pEW values in field late-T dwarfs might suggest that differences in surface gravity through the K I doublets are no longer distinguishable for late T types.

It is important to note that the K I 1243.7 nm line is partially blended with a FeH feature that may affect the measurement of the pEWs. Given that the field sources we have used present a lower spectral resolution, their pEWs may artificially have larger values than those measured from the X-shooter spectra. This is the reason why previous studies do not include this line in the analysis (Allers & Liu 2013; Martin et al. 2017).

Another way to observe the age trends is to plot the pEWs of the alkali lines as a function of the age group, as shown in Fig. 16. For this exercise, we remove all the objects in the T-type range. As expected from Fig. 15, the pEW on average increases as a function of age (surface gravity), which can be appreciated in the KDEs on the right-hand side of each panel, with the peaks of the distribution shifting upward with age.

thumbnail Fig. 15

Pseudo-equivalent widths of the alkali lines in the J-band as a function of the SpT derived in Sect. 4.1. The sample is divided according to the age classes defined in Sect. 2.1. The color and marker scheme for our sample is the same as Fig. 14. The purple circles are field objects from McLean et al. (2003) and Cushing et al. (2005). In the right panels of each sub-figure, there are the KDE for each age class (color-coded accordingly).
thumbnail Fig. 16

Pseudo-equivalent widths of the alkali lines as a function of the age. The sample is divided according to the age classes defined in Sect. 2.1. The color and marker scheme for our sample is the same as Fig. 14. The purple circles are field objects from McLean et al. (2003) and Cushing et al. (2005). In the right panels of each sub-figure, there are the KDEs for each age class (color-coded accordingly).

6 Suggested low-gravity spectral standards

There is plenty of spectra of old field dwarfs, with comparable temperatures and SpTs, which can be combined to create spectral templates. On the other hand, spectra of young late-type objects are rare and have a heterogeneous nature in the sense that they encompass a large spread of NIR colors for a fixed optically based SpT (Allers & Liu 2013; Cruz et al. 2018). As a result, a well-covered sequence of young LT templates is still lacking. The sequence used for template fitting in this work is that of Luhman et al. (2017), covering SpTs L0, L2, L4 and L7, at low spectral resolution. It is important to work toward constructing a more complete sequence that will allow the classification of the new ultracool spectra obtained by JWST and other upcoming facilities.

With this in mind, in Table 5, we present preliminary spectral standards for spectral classification of low-gravity LT dwarfs that should be used as a starting point toward achieving a more accurate young template sequence. In the selection process, we did not take into account objects in USCO, since these can be affected by extinction, or in PRA, since this association is the oldest in our sample. When more than one spectrum seemed suitable as a standard, preference was given to those located in the youngest regions considered in this work. We also prioritized objects with similar SpTs estimated by direct comparison with spectral templates and using SpT indices, within the respective error range. For some subtypes, there are more than one spectral standards that agree with the selection criteria and thus are proposed. In the end, the standard sequence suggested covers SpTs from L0 down to ~T5 but the subtypes L3, L6, L9, T0, T1, and T4 are lacking since none of the objects in our sample were classified as such.

We note that object #21 was classified visually as L5 since it appears to be later than L4 and earlier than L7, but closer in shape to the former. Object #20 is clearly later than L7 but at the same time lacks T-type characteristics; therefore, we proposed the L8 classification. This classification can be reiterated once a larger number of spectra in this spectral range are known. TObject #53 (T3) was selected as a weak binary candidate in Sect. 4.2, but we nevertheless decide to include it here because, even if it is a spectral binary, the components seem to have a very similar SpT that will not translate into a peculiar unresolved spectrum. The same argument applies to object #47, which, as already mentioned in Sect. 4.2, is a tight binary where the components have nearly equal luminosity (Best et al. 2017).

Regarding ages, for L types, all of the sources selected as possible spectral standards are younger than ~50 Myr. The proposed T3 standard, object #53, also has comparable age (located in ARG, 40–50 Myr). For the remaining two T-type dwarfs, older objects had to be selected: object #56 located in CARN (200±50 Myr), and object #50 in ABDMG (149−19+51 Myr$149_{ - 19}^{ + 51}{\rm{ Myr}}$). Ideally, one would want a set of templates with ages below 10 Myr to classify the PMO samples that JWST will detect but, at the moment, such data is lacking. What we present here should be seen as a step toward classification of the young late-L to T dwarfs in SFRs that should be revised once new data are acquired.

Table 5

Proposed low-gravity LT spectral standards.

7 Summary and conclusions

We present NIR VLT/X-shooter archival spectra of 56 ultracool dwarfs — probable members of NYMGs, young clusters, and SFRs — with SpTs in the range M8 to T6. The expected ages of the analyzed objects are between 10 and 600 Myr. We redetermined their SpTs through a comparison with a set of literature young and field templates, and a set of SpT-sensitive indices. With the help of a complementary spectroscopic sample comprising more than 3000 MLT dwarfs of all ages (from nearby SFRs to field), we identified 15 spectral indices that are useful for spectral-typing of LT dwarfs and provide their scaling relations with SpTs.

As one of the important goals of this paper is to provide means to distinguish young from old LT dwarfs based on their NIR spectra, we identified several spectral indices from the literature suitable for this purpose in the L-type spectral range. In the future, it will be crucial to establish a similar framework for the T-type objects, which is currently not possible due to the lack of known young objects in this spectral range. We find that the T-dwarf spectra in our sample consistently show redder colors when compared to their field counterparts, which has been interpreted as a signature of youth. Given their peculiarity when compared to field objects, these spectra also tend to be recognized as potential unresolved spectral binaries. However, we find this interpretation unlikely given the low prevalence of binaries in high-spatial-resolution surveys of T dwarfs and conclude that the red colors are probably caused by the effects of low surface gravity.

We also inspected the behavior of several alkali lines in the J and the H bands, observing the correlations of their widths with the SpT and age. The pEWs of the K I lines in the J band typically increase from M8 to ~L4, displaying a turnover toward later L types. The K I line in the H band (1516.7 nm) and the Na I J-band line (1139 nm) show a constant decrease in pEWs with increasing SpT. These observations are particularly valid for the field dwarfs, while the pEWs of the objects in our sample that are significantly younger are lower and do not show strong trends. The pEWs of the alkali lines show clear trends with age, as a consequence of changes in the surface gravity.

Finally, we identify 12 objects as preliminary spectral standards for young L and T dwarfs. They should be useful for spectral-typing of the objects identified with the new facilities, such as the JWST. More high-quality spectra of young ultracool dwarfs, in particular those with SpTs in the late L- and the T-type regime are needed to establish a complete set of templates for the characterization of young BDs and PMOs.

Acknowledgements

We thank the anonymous referee for the constructive comments, and we are grateful to Carlo Manara for providing the code to calculate the pseudo-equivalent widths of the alkali lines. L.P., K.M., and V.A. acknowledge funding by the Science and Technology Foundation of Portugal (FCT), grants PTDC/FISAST/7002/2020, UIDB/00099/2020, PTDC/FIS-AST/28731/2017, IF/00194/2015, SFRH/BD/143433/2019, 2022.03809.CEECIND, UIDB/04434/2020 and UIDP/04434/2020. L.P. acknowledges financial support from the CSIC project JAEICU-21-ICE-09, from Centro Superior de Investigaciones Cientificas (CSIC) under the PIE project 20215AT016, and the program Unidad de Excelencia Maria de Maeztu CEX2020-001058-M. This work has made use of data from the European Space Agency (ESA) mission Gaia (https://www.cosmos.esa.int/gaia), processed by the Gaia Data Processing and Analysis Consortium (DPAC, https://www.cosmos.esa.int/web/gaia/dpac/consortium). Funding for the DPAC has been provided by national institutions, in particular the institutions participating in the Gaia Multilateral Agreement. This research has benefitted from the SpeX Prism Spectral Libraries, maintained by Adam Burgasser at https://cass.ucsd.edu/ ajb/browndwarfs/spexprism/index.html. This research has benefitted from the Montreal Brown Dwarf and Exoplanet Spectral Library, maintained by Jonathan Gagné. This work has benefited from The UltracoolSheet at http://bit.ly/UltracoolSheet, maintained by Will Best, Trent Dupuy, Michael Liu, Aniket Sanghi, Rob Siverd, and Zhoujian Zhang, and developed from compilations by Dupuy & Liu (2012), Dupuy & Kraus (2013), Deacon et al. (2014), Liu et al. (2016), Best et al. (2018), Best et al. (2021), Sanghi et al. (2023), Schneider et al. (2023).

Appendix A Additional tables

In Table A.1 we summarize basic observational parameters of our sample presented in Sect. 2.

Table A.2 shows the spectral classification results discussed in Sect. 4.

Table A.1

Objects in our dataset, along with some basic observational parameters.

Table A.2

Spectral classification results.

Appendix B X-shooter spectra

In Fig. B.1 we show the remaining X-shooter spectra not present in the paper, along with their best-fit spectral template (see Sect. 4.1). All the retrieved best-fits were accomplished with young templates (in red).

thumbnail Fig. B.1

Remaining spectra of our sample along their best-fit template in red (see Sect. 4.1). For all of the objects presented here, the best-fit was accomplished with a young template. The ID, young association, final SpT and extinction derived (for objects in USCO) are shown above the correspondent spectrum (in black). All spectra are normalized at 1.67 μm. An offset between the spectra was added for clarity.

Appendix C Discussion on specific objects

In this section, we present a short discussion on the membership of some specific objects. To do that, we used the Bayesian Analysis for Nearby Young Associations (BANYAN) Σ, a tool that determines the membership probability of an object belonging to a nearby young association (Gagné et al. 2018). It requests as mandatory inputs the coordinates (RA and Dec, in decimal degrees), the proper motion (in mas/yr), and respective errors, both in right ascension (RA) and declination (Dec). Optionally, the radial velocity (in km/s) and parallax (in mas) of the objects, along with the associated errors, can also be given. If such information is lacking, the membership probability will still be computed and the algorithm will estimate the most likely distance to the object and the radial velocity that it should have if it was a member of a specific nearby young association.

Appendix C.1 Object #19 (J0032-4405)

As stated by Dupuy et al. (2018), object #19 has two parallax values in the literature: the one from Faherty et al. (2012), which was considered in Gagné et al. (2014c) and results in membership to BPMG, and the one from Faherty et al. (2016), which leads to ABDMG.

In the color—absolute magnitude diagram (see Fig. 1), object #19 is in-between both the young and the field sequences although it appears to be more in agreement with the field sequence. When comparing with spectral templates, the best fit was initially achieved with an L2 field template (see Sect. 4.1). However, after visual inspection (see Fig. 2), we maintained the classification obtained by the best young template (L0±2) since it appeared to better reproduce the K band. Nonetheless, we recognize that a subtle “shoulder” at ~1.57 μm is noticeable, which is typical of intermediate-gravity objects (Allers & Liu 2013) and thus suggests a slightly older age than some other objects of the sample. Moreover, in the TLI-g diagram in Fig. 14, object #19 is clearly in agreement with the mid-gravity age class of AA22 if we consider its SpT to be L0 (considering a young template), and even more so if we consider its SpT to be L2 (the first estimate obtained with field-template best fit in Sect. 4.1). Finally, its alkali lines are broader than those of the age class <30 Myr which is where the objects located in BPMG are included. Hence, all points toward a membership to ABDMG.

To reanalyze the nature of object #19, we implemented the BANYAN Σ tool considering the parallax and proper motion values from Gaia DR3 (π = 28.9542 ± 0.4217″, μα = 127.841 ± 0.284 mas/yr, μδ = −96.832 ± 0.314 mas/yr) and the radial velocity of Faherty et al. (2016) (rV = 12.95 ± 1.92 km/s). The result corroborates a membership to ABDMG (~150 Myr) with a probability of 99.4%, where the remaining 0.6% suggest the field population. This result is in agreement with previous works such as Ujjwal et al. (2020).

Appendix C.2 Object #62 (2MASS J05120636-2949540)

The derived SpT for object #62 was L4−2+3${\rm{L}}4_{ - 2}^{ + 3}$, considering a young template (see Sect. 4.1). Throughout this work, we considered that object #62 was a member of BPMG (age 24±2Myr, Bell et al. 2015). However, it shows high pEWs values for the alkali lines compared to the remaining objects within the <30 Myr age class. Its values seem to be more in agreement with those of the objects located in the age class 100–300 Myr. It is important to state that this object’s spectrum has an artifact in the range of calculation of the 1243.7 nm KI line; therefore, it was not included in the panels corresponding to this line in Figs. 15 and 16.

Regarding the gravity-sensitive indices, object #62 has not a trivial behavior. In some indices, such as HPI, it has a value that is in agreement with the mid-gravity class, toward the young sequence. On the other hand, its TLI-g and CH4-K index values still agree better with the intermediate-gravity age class but this time toward the overlapping region with the field objects of AA22. This behavior might suggest that object #62 does not belong to BPMG, which is among the youngest NYMGs in our sample.

We implemented the BANYAN Σ tool considering: μα = −5.0 ± 5.1 mas/yr, μδ = 84.5 ± 6.6 mas/yr, rv = 19.8 ± 1.5 km/s (Gagné et al. 2015c), and π = 49.6 ± 2.8″ (Best et al. 2021). The algorithm presented a membership probability that completely favors the field population. We conclude that object #62 might not be as young as members of BPMG. But since it exhibits signs of youth (e.g., a triangular H band), it does not seem to be as old as objects from the field population. Therefore, its membership is uncertain; it may belong to an yet undefined young association, or be an isolated object that was formed in a protoplanetary disk and ejected early in its evolution.

Appendix D NIR spectral indices

In Table D.1 we summarize the functional forms of all the SpT indices inspected in Sect. 4.3 along with the respective references and targeted spectral features. Some of these indices demonstrated a gravity-sensitive behavior in a specific SpT range, but they were all defined in the literature to be spectral indices that correlate well with SpT for a specific range initially designated.

There are spectral indices in the literature that were, in fact, defined to be gravity-sensitive, and the information regarding those is presented in Table D.2.

Table D.1

Spectral type indices inspected.

Table D.2

Gravity-sensitive spectral indices inspected.

References

  1. Allers, K. N., & Liu, M. C. 2013, ApJ, 772, 79 [NASA ADS] [CrossRef] [Google Scholar]
  2. Allers, K. N., Jaffe, D. T., Luhman, K. L., et al. 2007, ApJ, 657, 511 [Google Scholar]
  3. Almendros-Abad, V., Mužić, K., Moitinho, A., Krone-Martins, A., & Kubiak, K. 2022, A&A, 657, A129 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  4. Alves de Oliveira, C., Moraux, E., Bouvier, J., & Bouy, H. 2012, A&A, 539, A151 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  5. Ardila, D., Martín, E., & Basri, G. 2000, AJ, 120, 479 [NASA ADS] [CrossRef] [Google Scholar]
  6. Artigau, É., Doyon, R., Lafrenière, D., et al. 2006, ApJ, 651, L57 [NASA ADS] [CrossRef] [Google Scholar]
  7. Ashraf, A., Bardalez Gagliuffi, D. C., Manjavacas, E., et al. 2022, ApJ, 934, 178 [CrossRef] [Google Scholar]
  8. Baraffe, I., Homeier, D., Allard, F., & Chabrier, G. 2015, A&A, 577, A42 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  9. Bardalez Gagliuffi, D. C., Burgasser, A. J., Gelino, C. R., et al. 2014, ApJ, 794, 143 [NASA ADS] [CrossRef] [Google Scholar]
  10. Bayo, A., Barrado, D., Stauffer, J., et al. 2011, A&A, 536, A63 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  11. Bell, C. P. M., Mamajek, E. E., & Naylor, T. 2015, MNRAS, 454, 593 [Google Scholar]
  12. Best, W. M. J., Liu, M. C., Magnier, E. A., et al. 2015, ApJ, 814, 118 [CrossRef] [Google Scholar]
  13. Best, W. M. J., Liu, M. C., Dupuy, T. J., & Magnier, E. A. 2017, ApJ, 843, L4 [NASA ADS] [CrossRef] [Google Scholar]
  14. Best, W. M. J., Magnier, E. A., Liu, M. C., et al. 2018, ApJS, 234, 1 [Google Scholar]
  15. Best, W. M. J., Liu, M. C., Magnier, E. A., & Dupuy, T. J. 2021, AJ, 161, 42 [Google Scholar]
  16. Bouy, H., Brandner, W., Martín, E. L., et al. 2003, AJ, 126, 1526 [CrossRef] [Google Scholar]
  17. Bouy, H., Bertin, E., Sarro, L. M., et al. 2015, A&A, 577, A148 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  18. Bouy, H., Tamura, M., Barrado, D., et al. 2022, A&A, 664, A111 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  19. Brandt, T. D., Kuzuhara, M., McElwain, M. W., et al. 2014, ApJ, 786, 1 [Google Scholar]
  20. Breger, M. 1986, ApJ, 309, 311 [NASA ADS] [CrossRef] [Google Scholar]
  21. Burgasser, A. J., Kirkpatrick, J. D., Brown, M. E., et al. 2002, ApJ, 564, 421 [NASA ADS] [CrossRef] [Google Scholar]
  22. Burgasser, A. J., Kirkpatrick, J. D., Reid, I. N., et al. 2003, ApJ, 586, 512 [Google Scholar]
  23. Burgasser, A. J., Geballe, T. R., Leggett, S. K., Kirkpatrick, J. D., & Golimowski, D. A. 2006a, ApJ, 637, 1067 [Google Scholar]
  24. Burgasser, A. J., Kirkpatrick, J. D., Cruz, K. L., et al. 2006b, ApJS, 166, 585 [NASA ADS] [CrossRef] [Google Scholar]
  25. Burgasser, A. J., Cruz, K. L., Cushing, M., et al. 2010, ApJ, 710, 1142 [Google Scholar]
  26. Burgasser, A. J., Lopez, M. A., Mamajek, E. E., et al. 2016, ApJ, 820, 32 [NASA ADS] [CrossRef] [Google Scholar]
  27. Canty, J. I., Lucas, P. W., Roche, P. F., & Pinfield, D. J. 2013, MNRAS, 435, 2650 [NASA ADS] [CrossRef] [Google Scholar]
  28. Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245 [Google Scholar]
  29. Cook, N. J., Scholz, A., & Jayawardhana, R. 2017, AJ, 154, 256 [NASA ADS] [CrossRef] [Google Scholar]
  30. Covey, K. R., Lada, C. J., Román-Zúñiga, C., et al. 2010, ApJ, 722, 971 [NASA ADS] [CrossRef] [Google Scholar]
  31. Cruz, K. L., Kirkpatrick, J. D., & Burgasser, A. J. 2009, AJ, 137, 3345 [NASA ADS] [CrossRef] [Google Scholar]
  32. Cruz, K. L., Núñez, A., Burgasser, A. J., et al. 2018, AJ, 155, 34 [NASA ADS] [CrossRef] [Google Scholar]
  33. Cushing, M. C., Tokunaga, A. T., & Kobayashi, N. 2000, AJ, 119, 3019 [NASA ADS] [CrossRef] [Google Scholar]
  34. Cushing, M. C., Rayner, J. T., & Vacca, W. D. 2005, ApJ, 623, 1115 [Google Scholar]
  35. Cutri, R. M., Skrutskie, M. F., van Dyk, S., et al. 2003, VizieR Online Data Catalog: II/246 [Google Scholar]
  36. Dahm, S. E. 2015, ApJ, 813, 108 [Google Scholar]
  37. Day-Jones, A. C., Marocco, F., Pinfield, D. J., et al. 2013, MNRAS, 430, 1171 [NASA ADS] [CrossRef] [Google Scholar]
  38. Deacon, N. R., Liu, M. C., Magnier, E. A., et al. 2014, ApJ, 792, 119 [NASA ADS] [CrossRef] [Google Scholar]
  39. Dupuy, T. J., & Kraus, A. L. 2013, Science, 341, 1492 [NASA ADS] [CrossRef] [Google Scholar]
  40. Dupuy, T. J., & Liu, M. C. 2012, ApJS, 201, 19 [NASA ADS] [CrossRef] [Google Scholar]
  41. Dupuy, T. J., Liu, M. C., Allers, K. N., et al. 2018, AJ, 156, 57 [NASA ADS] [CrossRef] [Google Scholar]
  42. Esplin, T. L., & Luhman, K. L. 2020, AJ, 159, 282 [Google Scholar]
  43. Faherty, J. K., Burgasser, A. J., Walter, F. M., et al. 2012, ApJ, 752, 56 [Google Scholar]
  44. Faherty, J. K., Riedel, A. R., Cruz, K. L., et al. 2016, ApJS, 225, 10 [Google Scholar]
  45. Fontanive, C., Biller, B., Bonavita, M., & Allers, K. 2018, MNRAS, 479, 2702 [Google Scholar]
  46. Freudling, W., Romaniello, M., Bramich, D. M., et al. 2013, A&A, 559, A96 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  47. Gagné, J., Faherty, J. K., Cruz, K., et al. 2014a, ApJ, 785, L14 [CrossRef] [Google Scholar]
  48. Gagné, J., Lafrenière, D., Doyon, R., et al. 2014b, ApJ, 792, L17 [CrossRef] [Google Scholar]
  49. Gagné, J., Lafrenière, D., Doyon, R., Malo, L., & Artigau, É. 2014c, ApJ, 783, 121 [Google Scholar]
  50. Gagné, J., Burgasser, A. J., Faherty, J. K., et al. 2015a, ApJ, 808, L20 [CrossRef] [Google Scholar]
  51. Gagné, J., Faherty, J. K., Cruz, K. L., et al. 2015b, ApJS, 219, 33 [CrossRef] [Google Scholar]
  52. Gagné, J., Lafrenière, D., Doyon, R., Malo, L., & Artigau, É. 2015c, ApJ, 798, 73 [Google Scholar]
  53. Gagné, J., Faherty, J. K., Burgasser, A. J., et al. 2017, ApJ, 841, L1 [CrossRef] [Google Scholar]
  54. Gagné, J., Mamajek, E. E., Malo, L., et al. 2018, ApJ, 856, 23 [Google Scholar]
  55. Gaia Collaboration (Prusti, T., et al.) 2016, A&A, 595, A1 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  56. Gaia Collaboration (Brown, A. G. A., et al.) 2021, A&A, 649, A1 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  57. Galli, P. A. B., Moraux, E., Bouy, H., et al. 2017, A&A, 598, A48 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  58. Geballe, T. R., Knapp, G. R., Leggett, S. K., et al. 2002, ApJ, 564, 466 [Google Scholar]
  59. Geißler, K., Metchev, S., Kirkpatrick, J. D., Berriman, G. B., & Looper, D. 2011, ApJ, 732, 56 [CrossRef] [Google Scholar]
  60. Gonneau, A., Lançon, A., Trager, S. C., et al. 2016, A&A, 589, A36 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  61. Gorlova, N. I., Meyer, M. R., Rieke, G. H., & Liebert, J. 2003, ApJ, 593, 1074 [NASA ADS] [CrossRef] [Google Scholar]
  62. Hood, C. E., Fortney, J. J., Line, M. R., & Faherty, J. K. 2023, ApJ, 953, 170 [NASA ADS] [CrossRef] [Google Scholar]
  63. Kausch, W., Noll, S., Smette, A., et al. 2015, A&A, 576, A78 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  64. Kellogg, K., Metchev, S., Geißler, K., et al. 2015, AJ, 150, 182 [NASA ADS] [CrossRef] [Google Scholar]
  65. Kirkpatrick, J. D., Cruz, K. L., Barman, T. S., et al. 2008, ApJ, 689, 1295 [NASA ADS] [CrossRef] [Google Scholar]
  66. Kirkpatrick, J. D., Looper, D. L., Burgasser, A. J., et al. 2010, ApJ, 190, 100 [NASA ADS] [Google Scholar]
  67. Knapp, G. R., Leggett, S. K., Fan, X., et al. 2004, AJ, 127, 3553 [Google Scholar]
  68. Kraus, A. L., & Hillenbrand, L. A. 2007, AJ, 134, 2340 [NASA ADS] [CrossRef] [Google Scholar]
  69. Leggett, S. K., Currie, M. J., Varricatt, W. P., et al. 2006, MNRAS, 373, 781 [NASA ADS] [CrossRef] [Google Scholar]
  70. Liu, M. C., Dupuy, T. J., & Allers, K. N. 2013a, Astron. Nachr., 334, 85 [NASA ADS] [CrossRef] [Google Scholar]
  71. Liu, M. C., Magnier, E. A., Deacon, N. R., et al. 2013b, ApJ, 777, L20 [NASA ADS] [CrossRef] [Google Scholar]
  72. Liu, M. C., Dupuy, T. J., & Allers, K. N. 2016, ApJ, 833, 96 [NASA ADS] [CrossRef] [Google Scholar]
  73. Lodieu, N., Hambly, N. C., Jameson, R. F., & Hodgkin, S. T. 2008, MNRAS, 383, 1385 [Google Scholar]
  74. Lodieu, N., Zapatero Osorio, M. R., Béjar, V. J. S., & Perla Ramirez, K. 2018, MNRAS, 473, 2020 [CrossRef] [Google Scholar]
  75. Lodieu, N., Pérez-Garrido, A., Smart, R. L., & Silvotti, R. 2019, A&A, 628, A66 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  76. Lucas, P. W., Roche, P. F., Allard, F., & Hauschildt, P. H. 2001, MNRAS, 326, 695 [NASA ADS] [CrossRef] [Google Scholar]
  77. Luhman, K. L. 2022, AJ, 163, 24 [NASA ADS] [CrossRef] [Google Scholar]
  78. Luhman, K. L., & Esplin, T. L. 2020, AJ, 160, 44 [Google Scholar]
  79. Luhman, K. L., Mamajek, E. E., Allen, P. R., & Cruz, K. L. 2009, ApJ, 703, 399 [NASA ADS] [CrossRef] [Google Scholar]
  80. Luhman, K. L., Mamajek, E. E., Shukla, S. J., & Loutrel, N. P. 2017, AJ, 153, 46 [Google Scholar]
  81. Luhman, K. L., Herrmann, K. A., Mamajek, E. E., Esplin, T. L., & Pecaut, M. J. 2018, AJ, 156, 76 [Google Scholar]
  82. Madhusudhan, N., Burrows, A., & Currie, T. 2011, ApJ, 737, 34 [Google Scholar]
  83. Mamajek, E. E., Bartlett, J. L., Seifahrt, A., et al. 2013, AJ, 146, 154 [Google Scholar]
  84. Manara, C. F., Beccari, G., Da Rio, N., et al. 2013, A&A, 558, A114 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  85. Manjavacas, E., Apai, D., Zhou, Y., et al. 2019, AJ, 157, 101 [Google Scholar]
  86. Manjavacas, E., Lodieu, N., Béjar, V. J. S., et al. 2020, MNRAS, 491, 5925 [NASA ADS] [CrossRef] [Google Scholar]
  87. Marley, M. S., Saumon, D., Visscher, C., et al. 2021, ApJ, 920, 85 [NASA ADS] [CrossRef] [Google Scholar]
  88. Marocco, F., Andrei, A. H., Smart, R. L., et al. 2013, AJ, 146, 161 [NASA ADS] [CrossRef] [Google Scholar]
  89. Marocco, F., Jones, H. R. A., Day-Jones, A. C., et al. 2015, MNRAS, 449, 3651 [NASA ADS] [CrossRef] [Google Scholar]
  90. Martin, E. C., Mace, G. N., McLean, I. S., et al. 2017, ApJ, 838, 73 [NASA ADS] [CrossRef] [Google Scholar]
  91. Martin, E. L., Brandner, W., Bouy, H., et al. 2006, A&A, 456, 253 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  92. McGovern, M. R., Kirkpatrick, J. D., McLean, I. S., et al. 2004, ApJ, 600, 1020 [NASA ADS] [CrossRef] [Google Scholar]
  93. McLean, I. S., McGovern, M. R., Burgasser, A. J., et al. 2003, ApJ, 596, 561 [Google Scholar]
  94. Medan, I., Lépine, S., & Hartman, Z. 2021, AJ, 161, 234 [NASA ADS] [CrossRef] [Google Scholar]
  95. Miret-Roig, N. 2023, Ap&SS, 368, 17 [NASA ADS] [CrossRef] [Google Scholar]
  96. Miret-Roig, N., Bouy, H., Raymond, S. N., et al. 2022, Nat. Astron., 6, 89 [NASA ADS] [CrossRef] [Google Scholar]
  97. Modigliani, A., Goldoni, P., Royer, F., et al. 2010, SPIE Conf. Ser., 7737, 773728 [Google Scholar]
  98. Mužić, K., Scholz, A., Geers, V. C., & Jayawardhana, R. 2015, ApJ, 810, 159 [Google Scholar]
  99. Naud, M.-E., Artigau, É., Malo, L., et al. 2014, ApJ, 787, 5 [Google Scholar]
  100. Peña Ramírez, K., Béjar, V. J. S., Zapatero Osorio, M. R., Petr-Gotzens, M. G., & Martin, E. L. 2012, ApJ, 754, 30 [CrossRef] [Google Scholar]
  101. Pearson, S. G., & McCaughrean, M. J. 2023, arXiv e-prints [arXiv:2318.81231] [Google Scholar]
  102. Pecaut, M. J., & Mamajek, E. E. 2013, ApJS, 208, 9 [Google Scholar]
  103. Pecaut, M. J., & Mamajek, E. E. 2016, MNRAS, 461, 794 [Google Scholar]
  104. Petrus, S., Chauvin, G., Bonnefoy, M., et al. 2023, A&A, 670, A9 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  105. Ribas, I. 2003, A&A, 400, 297 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  106. Rizzuto, A. C., Ireland, M. J., & Kraus, A. L. 2015, MNRAS, 448, 2737 [NASA ADS] [CrossRef] [Google Scholar]
  107. Rojas-Ayala, B., Covey, K. R., Muirhead, P. S., & Lloyd, J. P. 2012, ApJ, 748, 93 [Google Scholar]
  108. Sanghi, A., Liu, M. C., Best, W. M. J., et al. 2023, ApJ, 959, 63 [NASA ADS] [CrossRef] [Google Scholar]
  109. Schlieder, J. E., Lépine, S., Rice, E., et al. 2012, AJ, 143, 114 [NASA ADS] [CrossRef] [Google Scholar]
  110. Schneider, A. C., Windsor, J., Cushing, M. C., Kirkpatrick, J. D., & Wright, E. L. 2016, ApJ, 822, L1 [NASA ADS] [CrossRef] [Google Scholar]
  111. Schneider, A. C., Windsor, J., Cushing, M. C., Kirkpatrick, J. D., & Shkolnik, E. L. 2017, AJ, 153, 196 [NASA ADS] [CrossRef] [Google Scholar]
  112. Schneider, A. C., Munn, J. A., Vrba, F. J., et al. 2023, AJ 166, 103 [CrossRef] [Google Scholar]
  113. Scholz, A., Jayawardhana, R., Mužić, K., et al. 2012, ApJ, 756, 24 [NASA ADS] [CrossRef] [Google Scholar]
  114. Scholz, A., Mužić, K., Jayawardhana, R., Quinlan, L., & Wurster, J. 2022, PASP, 134, 104401 [NASA ADS] [CrossRef] [Google Scholar]
  115. Slesnick, C. L., Hillenbrand, L. A., & Carpenter, J. M. 2004, ApJ, 610, 1045 [NASA ADS] [CrossRef] [Google Scholar]
  116. Smart, R. L., Marocco, F., Caballero, J. A., et al. 2017, MNRAS, 469, 401 [NASA ADS] [CrossRef] [Google Scholar]
  117. Smette, A., Sana, H., Noll, S., et al. 2015, A&A, 576, A77 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  118. Taylor, B. J. 2006, AJ, 132, 2453 [NASA ADS] [CrossRef] [Google Scholar]
  119. Testi, L., D’Antona, F., Ghinassi, F., et al. 2001, ApJ, 552, L147 [NASA ADS] [CrossRef] [Google Scholar]
  120. Tody, D. 1986, SPIE Conf. Ser., 627, 733 [Google Scholar]
  121. Tokunaga, A. T., & Kobayashi, N. 1999, AJ, 117, 1010 [CrossRef] [Google Scholar]
  122. Tremblin, P., Amundsen, D. S., Chabrier, G., et al. 2016, ApJ, 817, L19 [NASA ADS] [CrossRef] [Google Scholar]
  123. Ujjwal, K., Kartha, S. S., Mathew, B., Manoj, P., & Narang, M. 2020, AJ, 159, 166 [NASA ADS] [CrossRef] [Google Scholar]
  124. Vernet, J., Dekker, H., D’Odorico, S., et al. 2011, A&A, 536, A105 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  125. Weights, D. J., Lucas, P. W., Roche, P. F., Pinfield, D. J., & Riddick, F. 2009, MNRAS, 392, 817 [NASA ADS] [CrossRef] [Google Scholar]
  126. Zhang, Z., Liu, M. C., Best, W. M. J., et al. 2018, ApJ, 858, 41 [NASA ADS] [CrossRef] [Google Scholar]
  127. Zhang, Z., Liu, M. C., Best, W. M. J., Dupuy, T. J., & Siverd, R. J. 2021, ApJ, 911, 7 [NASA ADS] [CrossRef] [Google Scholar]
  128. Zuckerman, B. 2019, ApJ, 870, 27 [Google Scholar]
  129. Zuckerman, B., Bessell, M. S., Song, I., & Kim, S. 2006, ApJ, 649, L115 [NASA ADS] [CrossRef] [Google Scholar]
  130. Zuckerman, B., Vican, L., Song, I., & Schneider, A. 2013, ApJ, 778, 5 [Google Scholar]
1

Jonathan Gagné’s List of Ultracool Dwarfs (https://jgagneastro.com/list-of-ultracool-dwarfs/)

2

The age of the Castor Moving Group (CAS) is not that well constrained and shows a large spread among its original and probable members (Brandt et al. 2014; Mamajek et al. 2013). But most of the stars in CAS seem to be older than 200 Myr (Zuckerman et al. 2013). In fact, it will be noticeable in the posterior analysis that object #27 has an age between 200 Myr and the oldest age class considered (PRA, age ~600 Myr).

3

IRAF is distributed by the National Optical Astronomy Observatories, which are operated by the Association of Universities for Research in Astronomy, Inc., under cooperative agreement with the National Science Foundation.

7

This in fact is the case for all other USCO objects as well, whose X-shooter spectra have first been analyzed in Lodieu et al. (2018). For this reason, as a literature value in Table A.2, we prefer the classification from Luhman et al. (2018) which considers extinction in their fitting process.

8

The disentanglement of these objects’ spectra into two T-type objects is in agreement with the combined classifications estimated in Zhang et al. (2021): object #51 was flagged as a weak binary candidate where its combined SpT was estimated as T2+T3, and object #53 was flagged as a strong candidate with T1.5+T3.5.

All Tables

Table 1

NYMGs, young clusters, and SFRs included in this work.

In the text
Table 2

Sets of spectral templates used for the NIR SpT derivation.

In the text
Table 3

SpT derived for the strong and weak binary candidates via a direct comparison with single and composite templates.

In the text
Table 4

Spectral type indices selected.

In the text
Table 5

Proposed low-gravity LT spectral standards.

In the text
Table A.1

Objects in our dataset, along with some basic observational parameters.

In the text
Table A.2

Spectral classification results.

In the text
Table D.1

Spectral type indices inspected.

In the text
Table D.2

Gravity-sensitive spectral indices inspected.

In the text

All Figures

thumbnail Fig. 1

(MJ, J–KS) color-absolute magnitude diagram of our objects, very low-gravity objects (orange circles), and field objects (gray circles) from the literature (AA22; Database of Ultracool Parallaxes). Our sample is divided according to the age of the association to which the objects probably belong. Green stars represent objects in regions with ages <30 Myr, dark gray pentagons with ages between 30 and 100 Myr (included), red diamonds with ages between 100 and 300 Myr (included), and white squares with ages ~600 Myr. For data points where the error bar is not visible, the bar is smaller than the marker size. The three ellipses approximately highlight where the M, L, and T dwarfs are located in this diagram. The arrow indicates what effect the addition of five magnitudes of extinction would have.
In the text
thumbnail Fig. 2

Comparison between the best-fit young (red) and field (purple) templates for dwarfs where the spectral classification made via a direct comparison with templates yields a result of an M or L field. The ID and young association are shown above the correspondent spectrum (in black). All spectra are normalized at 1.67 μm. An offset between the spectra was added for clarity.
In the text
thumbnail Fig. 3

Spectra of objects in the SpT range L0–L2 (in black) compared with L0 (red) and L2 (purple) young templates. The spectra were normalized at 1.67 μm. The ID of each object is listed above the respective spectrum. An offset between the spectra was added for clarity.
In the text
thumbnail Fig. 4

Spectra of objects in the SpT range L2–L4 (in black) compared with L2 (red) and L4 (purple) young templates. The spectra were normalized at 1.67 μm. The ID of each object is listed above the respective spectrum. An offset between the spectra was added for clarity.
In the text
thumbnail Fig. 5

Spectra of objects in the SpT range L4–L7 (in black) compared with L4 (red) and L7 (purple) young templates. The spectra were normalized at 1.67 μm. The ID of each object is listed above the respective spectrum. An offset between the spectra was added for clarity.
In the text
thumbnail Fig. 6

Spectrum of object #20 (PSO J318.5-22, in black) along with the L7 young template (in red), L8 field template (in purple), and the spectrum of object #56 (SIMP J013656.5+093347.3, in gray), which was classified as T2 via a direct comparison with spectral templates. This comparison is shown to underline the significant change in the spectral morphology associated with the L-T transition. All spectra are normalized at 1.67 μm.
In the text
thumbnail Fig. 7

Spectra of objects classified as T-type dwarfs along with their best-fit template (in red). The ID, derived SpT, and region of each object are shown next to the corresponding spectrum. All the spectra are normalized at 1.57 μηι. An offset between the spectra was added for clarity.
In the text
thumbnail Fig. 8

Comparison between the derived NIR SpT via a direct comparison with spectral templates (see Sect. 4.1) and the literature NIR SpT for the X-shooter dataset. The solid black line indicates a perfect agreement between classifications. The dotted black lines represent the ±2 sub-SpT range. The ID of the objects for which their classifications do not agree within ±2 sub-SpTs is shown.
In the text
thumbnail Fig. 9

Spectral index criteria applied for unresolved binary candidate selection as defined in Burgasser et al. (2010). The top three panels and the first bottom panel compare pairs of index values, whereas the other panels compare index ratios to the NIR SpT derived in Sect. 4.1. The objects from our dataset are represented as black dots. Strong binary candidates are highlighted with red diamonds and weak binary candidates with red circles.
In the text
thumbnail Fig. 10

Strong (objects #52, #55 and #63) and weak (objects #51 and #53) binary candidates (in black) overplotted with the best-fit individual templates (magenta) and the best-fit composite templates (green). The two components of the synthetic binary templates are also plotted in yellow and blue, respectively.
In the text
thumbnail Fig. 11

Selected SpT indices from the literature versus the NIR SpT derived via a direct comparison with spectral templates. The classes young, mid-gravity and field from the dataset of AA22 are represented as orange, blue and gray circles, respectively. The black solid lines represent the sensitivity range of each index (see Table 4). The red dashed line in the spectral index ωD represents the best fit if the sensitivity range is considered from M4 down to T8. The black arrows demonstrate the effect of 5 magnitudes of extinction on the index values.
In the text
thumbnail Fig. 12

Selected SpT indices from the literature versus the NIR SpT derived via a direct comparison with spectral templates. The classes young, mid-gravity and field from the dataset of AA22 are represented as orange, blue and gray circles, respectively. The black solid lines represent the sensitivity range of each index (see Table 4). The black arrows demonstrate the effect of 5 magnitudes of extinction on the index values.
In the text
thumbnail Fig. 13

Comparison between the derived NIR SpT using spectral indices (see Sect. 4.3) and the SpT derived via a direct comparison with spectral templates (see Sect. 4.1). The solid black line indicates perfect agreement between classifications. The dotted black lines represent the ±2 sub-SpT range. The ID of the objects for which their classifications do not agree within ±2 sub-SpTs is shown.
In the text
thumbnail Fig. 14

Selected gravity-sensitive spectral indices from the literature versus the NIR SpT derived in Sect. 4.1. Our sample is divided into the different age classes presented in Sect. 2.1 where the classes <30 Myr (USCO, TWA, BPMG, and 32OR), 30–100 Myr (ARG, COL, and THA), 100–300 Myr (ABDMG, CAS, CARN, and PLE), and 600 Myr (PRA) are shown as green stars, black pentagons, red diamonds, and white squares, respectively. The complementary dataset is divided into young, mid-gravity, and field objects represented as orange, blue, and gray dots, respectively.
In the text
thumbnail Fig. 15

Pseudo-equivalent widths of the alkali lines in the J-band as a function of the SpT derived in Sect. 4.1. The sample is divided according to the age classes defined in Sect. 2.1. The color and marker scheme for our sample is the same as Fig. 14. The purple circles are field objects from McLean et al. (2003) and Cushing et al. (2005). In the right panels of each sub-figure, there are the KDE for each age class (color-coded accordingly).
In the text
thumbnail Fig. 16

Pseudo-equivalent widths of the alkali lines as a function of the age. The sample is divided according to the age classes defined in Sect. 2.1. The color and marker scheme for our sample is the same as Fig. 14. The purple circles are field objects from McLean et al. (2003) and Cushing et al. (2005). In the right panels of each sub-figure, there are the KDEs for each age class (color-coded accordingly).
In the text
thumbnail Fig. B.1

Remaining spectra of our sample along their best-fit template in red (see Sect. 4.1). For all of the objects presented here, the best-fit was accomplished with a young template. The ID, young association, final SpT and extinction derived (for objects in USCO) are shown above the correspondent spectrum (in black). All spectra are normalized at 1.67 μm. An offset between the spectra was added for clarity.
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.