In this section we review discoveries over the past ∼10 years that highlight the current
breadth and scope of the field from the observational perspective. This treatment is brief and not meant as all-inclusive.
Surface Gravity. Surface gravity is an important consideration in the treatment of VLM stars and brown dwarfs due to their small radii and compact nature. While surface gravity and its diagnostic features in higher mass stars are useful for distinguishing main sequence stars from evolved giant stars, giant L and T dwarfs are generally not thought to exist. Variations in surface gravity occur mainly as a result of early contraction, with low gravity being associated with youth. Gravity considerations are particularly relevant for L dwarfs, where depending on mass, age, and metallicity, a given spectral subtype may
2The ongoing development of evolutionary models incorporatingBT-Settlwas announced by F. Allard in
correspond to a stellar object, an intermediate age high mass brown dwarf, a young low mass brown dwarf, or even a very young planetary mass object.
The use of gravity sensitive spectral features for identifying youth in brown dwarfs was first demonstrated by Mart´ın et al. (1999) and emphasized by McGovern et al. (2004), who analyzed the infrared and optical spectra of the young brown dwarfs G196-3B, KPNO Tau
4, σ Ori 47, and σ Ori 51. They note that the widths of absorption lines due to neutral
alkali elements, in particular K, Na, Cs, Rb, as well as the strengths of absorption bands due to VO, TiO, CaH, and other metal hydrides can be used as relative measures of surface gravity, and hence youth. Because of the important role of surface gravity in determining the nature of L dwarfs, Cruz et al. (2009) have proposed a two-dimensional spectral typing
sequence from L0 to L5 where the suffixes α, β, and γ indicate normal gravity, low gravity,
and very low gravity, respectively. The L0 to L5 spectral range is extremely diverse as far as the nature of its members is concerned, and a two-dimensional classification scheme such as the ones proposed by Kirkpatrick (2005), Cruz et al. (2009), and Allers & Liu (2013) would be of significant value in linking spectral classification to underlying physics. Unfortunately, this scheme has not been universally adopted. Comparison to model atmospheres (e.g.
Allard et al. 2013) indicates that “normal” L dwarfs have surface gravities log g ∼ 5.0
(where g is the acceleration due to gravity in cgs units), whereas very low gravity (i.e.,
young) objects may have surface gravities as low as log g ∼3.5.
Metallicity. Several metal-poor L subdwarfs have been identified to date (e.g., Burgasser et al. 2003b, 2008b, 2009; Cushing et al. 2009; Sivarani et al. 2009; Lodieu et al. 2010). These ob-
jects appear to be analogous to subdwarfs of earlier spectral classes, where low metallicity is usually observed along with kinematics indicative of an older population. As an example, 2MASS J0616-6407, an sdL5 (Cushing et al. 2009), has spectroscopic signatures of low metal
content and an impressive radial velocity of 454±15 km s−1. There have also been sugges-
tions of particularly metal-rich L dwarfs (e.g., Looper et al. 2008), in which the spectroscopic effects of high metallicity may mimic the effects of low surface gravity.
The Blue L Dwarfs. Several L dwarfs have unusually blue near infrared spectra, and comprise a distinct group known as the blue L dwarfs (e.g., Cruz et al. 2007; Burgasser et al. 2008a; Bowler et al. 2010; Schmidt et al. 2010; Cushing et al. 2010). The blue excess in the infrared is most likely due to a thinner than usual cloud layer, which is likely the result of low metallicity, high surface gravity causing high sedimentation rate, high vertical mixing in the atmosphere, or a combination of these factors (Cushing et al. 2010). The blue L dwarfs highlight how grain formation and sedimentation produce complex atmospheric physics, and how the effects of these basic parameters are manifested in interconnected ways.
Multiplicity. Multiplicity across the stellar/substellar boundary is one of the main topics of this thesis. In keeping with this section’s goal of highlighting the diversity of VLM stars and brown dwarfs, an overview of the topic is presented here. A detailed discussion of topics such as the “brown dwarf desert” including presentation of new results, is found in Chapter 4. The results discussed there can be summarized by stating that binaries in which the primary component is of spectral type M or earlier and the secondary component is an L or a T dwarf are infrequent, with an occurrence rate of only a few percent. This
result is largely invariant across different stellar ages, mass ratios, and separations. In the following discussion we also include the M dwarfs because many multiplicity studies have focused simply on low mass objects, which includes the M, L, and T dwarfs.
Early infrared speckle studies established the multiplicity of stellar companions around
M dwarfs to be 32±11% (Henry & McCarthy 1990; Henry 1991). Subsequent work by
Fischer & Marcy (1992) indicated a slightly higher multiplicity rate of 42±9%. In a synthesis
of several multiplicity studies (Reid & Gizis 1997; Oppenheimer et al. 2001; Hinz et al. 2002;
Reid et al. 2003; Delfosse et al. 2004), Burgasser et al. (2007) derive an overall Multiplicity
Fraction (MuF) of 27+5−4% for M dwarfs. The same study finds that the observed MuF for
VLM systems with primaries of spectral type M6V or later is ∼20%; however, they note
that due to to selection effects the real MuF is likely somewhat lower. Taking the results of several studies into consideration, Burgasser et al. (2007) note that the MuF for VLM stars and brown dwarfs could range from as low as 10% to as high as 30%, and that the binary population is characterized by closely separated and nearly equal mass systems. In
particular, they note that 93% of the systems have separation<20 AU, and 77% have mass
ratios >0.8, where 1.0 denotes an equal mass binary.
In addition to the general population trends described above, there are also noteworthy subsets of very widely separated binaries and very widely separated hierarchical higher order multiples. The contrast between these subsets and the main binary population described
above may shed light into brown dwarf formation scenarios (§4.6.2). As an example of one
∼25000 AU may survive dynamic interactions in the young cluster environment and form stable field-aged systems. These systems have binding energies that are close to the limit for bound systems (Close et al. 2007). Several studies also note a disproportionate number of systems where two closely separated, nearly equal mass brown dwarfs are in a wide hi- erarchical orbit around a more massive main sequence star (e.g., Golimowski et al. 2004a; Burgasser et al. 2005; Caballero 2007; Law et al. 2010; Faherty et al. 2010). In particular, Faherty et al. (2010) note that the ratio of triple systems to binary systems is 2.4 times higher for VLM objects than for the general field population. It has also been noted that the reverse configuration, where the central star is a close binary and those two components are orbited by a single wide separation low mass companion, is also common (Allen et al. 2012). That study has suggested that formation scenarios based on conservation of angu- lar momentum where angular momentum is transferred from the binary component to the isolated wide component, therefore decreasing the separation of the binary component, are likely responsible for the high occurrence of triple systems.
Finally, very close binaries have also been detected via spectroscopic observations. Shkolnik et al. (2010) report an overall spectroscopic binary fraction of 16% for M dwarfs. More specifically, Blake et al. (2010) report the late M and L dwarf binary fraction to be
2.5+8−1..66% for separations smaller than 1.0 AU and Clark et al. (2012) report the binary frac-
tion for M dwarfs with separations smaller than 0.4 AU to be 3% to 4%.
Galactic Kinematics. Studies of the Galactic kinematics of L and T dwarfs have been primarily concerned with determining the sample’s age based on its Galactic velocity
distribution. The method relies on the idea that stars are born in a cluster environment with a small velocity dispersion, which then gradually increases as stars travel through the Galaxy and undergo dynamical interactions with other objects. The age dependence of the
velocity dispersion can be modeled as a power law of the formtα withα∼0.33 for the solar
neighborhood (Binney et al. 2000). Because the calculation of three-dimensional Galactic velocities must rely on known tangential velocities from trigonometric parallax observations as well as spectroscopic radial velocities, only a small subset of the known L and T dwarfs can be completely characterized at this time. Many more have been studied in a more restricted sense through the determination of proper motion only (Schmidt et al. 2007; Faherty et al. 2009).
There have been conflicting results in the literature regarding age determination through kinematic means. Zapatero Osorio et al. (2007) examined the space motions of 21 L and T dwarfs and concluded that the kinematics of these objects most closely resemble those of
hot F type stars. From the comparison, they derive ages ranging from 0.5−4 Gyr for the
substellar population, which is lower than the accepted age for low mass stars in the solar neighborhood. In contrast, subsequent larger studies have found no clear evidence that the low mass stellar population and the VLM and substellar populations are characterized by different ages. Schmidt et al. (2010) determined the three-dimensional Galactic velocities of 306 L dwarfs and concluded that while the population is best characterized by a mix of a younger component and an older component, the overall characteristics are no different than those noted for M dwarfs. Schmidt et al. (2010) find the mean tangential velocity to
be Vtan = 28 km s−1 with a dispersion of σtan = 25 km s−1. Seifahrt et al. (2010) examined
a sample of 43 L dwarfs with three-dimensional velocities and derived an age of ∼3 Gyr
for the L dwarfs, which is comparable to the accepted age for late M dwarfs in the solar neighborhood (Reiners & Basri 2009).
As pointed out by Seifahrt et al. (2010), the derivation of an age for the L dwarf popu- lation that is comparable to the age of the M dwarf population, as in Seifahrt et al. (2010) and Schmidt et al. (2010), poses problems for the interpretation of substellar evolution. Even without detailed knowledge of the spectral sub-type corresponding to the HBMM, it is accepted that the L dwarf population constitutes a mix of stellar and substellar objects. Because L type brown dwarfs comprise the hotter component of a permanently cooling pop- ulation, they should be on average younger than the stellar population, which does not cool with time. Seifahrt et al. (2010) speculate that perhaps their observed sample is not ade- quate for statistical treatment due to the non-Gaussian nature of their velocity dispersion, or that perhaps the initial velocity dispersion of substellar objects is higher than that of stellar objects. The latter consideration would mean that a different zero point velocity dispersion is necessary when estimating the age of brown dwarfs, and is consistent with the ejection
scenario of brown dwarf formation (§4.6.2).