Mid-infrared followup of cold brown dwarfs : Diversity in age , mass and metallicity

We use Spitzer IRAC 3.6–8.0 m photometry of late-type T dwarfs to investigate various trends which can aid the planning and interpretation of infrared (IR) surveys for the coldest T or Y dwarfs. Brown dwarfs with effective temperature (Teff)<700 K emit>50% of their flux at >3 m, and the ratio of the midIR to the near-IR flux becomes very sensitive to Teff. The color H − [4.5] is a good indicator of Teff with a weak dependence on metallicity ([m/H]) and gravity (g) while H − K and [4.5] − [5.8] are sensitive to [m/H] and g. Thus Teff and g can be constrained and mass and age can then be determined from evolutionary models. There are 12 dwarfs known with H − [4.5] > 3.0 and 500 Teff K 800, which we examine in detail. The ages of these dwarfs range from very young (0.1–1.0 Gyr) to old (3–12 Gyr). The mass range is possibly as low as 5 MJup to 70 MJup, and [m/H] also spans a large range of ∼ −0.3 to ∼ +0.3. The T8–T9 dwarfs found so far in the UKIRT IR Deep Sky Survey are unexpectedly young and low-mass. Extensions to the warm Spitzer and WISE space missions are needed to obtain mid-IR data for cold brown dwarfs, and to discover more of these rare objects.


INTRODUCTION
The last decade has seen a remarkable increase in our knowledge of the bottom of the main-sequence and of the low-mass stellar and sub-stellar (brown dwarf) population of the solar neighbourhood.Two new classes have been added to the spectral type sequence following M: L and T. T dwarfs with effective temperatures (T eff ) as low as ∼500 K are now known (Warren et al. 2007; Burningham et al. 2008;Delorme et al. 2008; Leggett et al. 2009) and we are truly finding objects that provide the link between the low-mass stars and the giant planets.
As T eff decreases, brown dwarfs emit significant flux in the mid-infrared(IR) (e.g.Burrows et al. 2003; Leggett et al. 2009).Figure 1 demonstrates the rapidly increasing importance of this region for late-type T dwarfs; for dwarfs cooler than 700 K more than half the flux is emitted at wavelengths longer than 3 m.Here we investigates photometric trends seen in very late-type dwarfs at mid-IR wavelengths, with the expectation that these longer wavelengths will be crucial for both the discovery and the understanding of the coolest T-type dwarfs, and even more so for the proposed cooler Y-type dwarfs.Our group was awarded time on the Spitzer Space Telescope (Werner et al. 2004) to obtain IRAC (Fazio et al. 2004) photometry of late-type T dwarfs found in the Large Area Survey (LAS) component of the UKIRT IR Deep Sky Survey (UKIDSS; Lawrence et al. 2007).Identification and classification of the LAS T dwarfs is described by Pinfield et al. (2008) a e-mail: sleggett@gemini.eduThis is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial License 3.0, which permits unrestricted use, distribution, and reproduction in any noncommercial medium, provided the original work is properly cited.

Figure 1.
Percentage of total flux emitted at >3 m, as a function of T eff , calculated from solar metallicity log g = 5.0 models which include vertical gas transport.Cloudy models were used for T eff > 1000K and cloudfree for cooler temperatures.

THE SAMPLE
Figure 1 demonstrates the rapidly increasing ratio of the mid-IR flux to the near-IR flux for late-type T dwarfs.This means that a large wavelength baseline near-IR to mid-IR color should be a good indicator of T eff , and the H − [4.5] color has been shown to be optimum (Warren et al. 2007;Leggett et al. 2009;Stephens et al. 2009).Leggett et al. (2009) show that expected variations in gravity (g) and metallicity ([m/H]), and in the gas transport efficiency, each impact H − [4.5] by ∼0.2 magnitudes.This is small given the large degree of sensitivity to T eff : (H − [4.5]) ≈ 0.7 magnitudes for T eff = 100 K at T eff = 600 K for example.
We adopt H − [4.5] as a primary reference color and have selected T dwarfs with IRAC photometry that have H MKO − [4.5] > 3.0, which our models imply have T eff 800 K. Including the LAS T dwarfs with new IRAC photometry, 12 T dwarfs fall into this group.The dwarfs are listed in Table 1 together with some of their fundamental properties.All dwarfs with IRAC photometry that are T7.5 or later are included in this sample, except for the metal-rich T7.5 dwarf 2MASS J1217-03 (Saumon et al. 2007).Its exclusion is due to the (small) dependency of H − [4.5] on [m/H].Similarly, the relatively early-type very metal-poor dwarfs 2MASS J0937+29 (T6p) and 2MASS J1237+65 (T6.5e) have been included in our color-selected sample.

Photometric Indicators
Figure 2 shows a selection of color-magnitude diagrams for our sample, and Figure 3 color-color diagrams.Sequences are shown calculated by our cloud-free models which include chemical mixing by gas transport, for a range of g and [m/H].Although not perfect, the relative location of the objects in the plots is in agreement with our models.For example, metal-poor dwarfs are bright in Figure 2, and blue in H − K and [4.5] − [5.8] in Figure 3.The values of T eff indicated for the sample in Figures 2  and 3 are consistent with other estimates (Table 1).2MASS 1237+65 has very strong H emission.Liebert & Burgasser (2007) discount the possibility that it is young and low-g, which would also be inconsistent with our analysis.They suggest that the object may be a close double system, and if the companion fills the Roche lobe then it must be cooler than 650 K.As no significant [4.5] excess is seen in Figure 2, our models imply that any companion would have T eff < 500 K.The red H − [4.5] for the dwarf is due to its low metallicity and not to the detected presence of a cool companion.

Research, Science and Technology of Brown Dwarfs and Exoplanets
Figure 3 suggest that ULAS 1017+01 and ULAS 1238+09 are low gravity possibly metal-rich objects, similar to the other LAS very late-type T dwarfs ULAS 0034−07 and ULAS 1335+11, but not as cool.This low gravity for ULAS 1017+01 disagrees with the value determined by Burningham et al. (2008) based on near-IR data only and implies a very young age of 0.1-0.4Gyr.This dwarf is spectrally peculiar; if the object is an unresolved 900+700 K or 800+600 K binary, then our model colors indicate that g could be higher and the age would be ∼1 Gyr, more typical for a local disk dwarf.
Figures 2 and 3 allow us to constrain the metallicity of the Wolf 940 system: Burningham et al. ( 2009) give [Fe/H] = −0.06± 0.20 based on the V − K color of Wolf 940A but the photometric analysis presented here implies that [m/H] for the system is between 0 and +0.3.We can exclude the possibility that Wolf 940B is metal-poor.

The Low-Mass LAS Dwarfs
The 2MASS-selected dwarfs tend to be high-g and/or low-[m/H] because they are selected for blue H − K color (Figures 2 and 3).However the tendency for the LAS objects to be low-g (Table 1) is difficult to understand. Figure 3 shows that the Y J H colors used to select the LAS dwarfs are not sensitive to g or [m/H].There should also be no g/age bias introduced by a brightness selection, as field brown dwarfs have an almost constant radius so that luminosity is effectively a function of T eff only (Saumon & Marley 2008).
Simulations of the mass function (Burgasser 2004) show that the median field brown dwarf mass (and hence g) decreases to lower T eff because lower mass brown dwarfs start off cooler than higher mass brown dwarfs.However the simulations indicate that, for a flat mass function (e.g.Metchev et al. 2008, Pinfield et al. 2008) and constant birth-rate, at T eff ≈ 600 K the median mass would be 30-40 M Jup , age >6 Gyr and log g ≈ 5.The LAS T8-T9 dwarfs however appear to be generally younger than 2 Gyr and less massive than 20 M Jup .This puzzle will be re-examined when the sample of cold LAS brown dwarfs is larger.

CONCLUSIONS
The wavelength region beyond 3 m makes up most of the emitted flux for dwarfs cooler than 700 K and is crucial for analysis of their photospheres.This region is extremely challenging from the ground and extensions to the warm Spitzer and WISE missions are desirable.

06007-p.4
Research, Science and Technology of Brown Dwarfs and Exoplanets  ] are sensitive to metallicity and gravity.As more data are obtained and the models improve it may be possible to separate these effects for independent determinations of these parameters.

Figure 2 .
Figure 2. Color-magnitude plots for late T dwarfs.Small dots are a generic sample, known binaries are shown with ringed symbols.Larger symbols are the Table 1 dwarfs; largest to smallest circles represent log g ≈ 5.4, log g ≈ 5.2, log g ≈ 5.0 and log g ≈ 4.3.Dark to light grey circles are metal-rich, solar and metal-poor dwarfs; black are unknown [m/H].Model sequences are shown with parameters indicated in the top panel.Crosses along the sequences indicate T eff = 800 K and 600 K.

Figure 3 .
Figure 3. H − [4.5] against Y − J , J − H , H − K and [4.5] − [5.8].Symbols are as in Figure 2 with open circles as unknown g.Model sequences are shown with parameters as indicated on the right axes.The T eff values for the log g = 4.5 [m/H] = 0 model are indicated on the top axis, and crosses along the sequences indicate the 800 K and 600 K points for each model set.