0705.2602/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: 
3: %% manuscript produces a one-column, double-spaced document:
4: 
5: %%\documentclass[manuscript]{aastex}
6: 
7: %% preprint2 produces a double-column, single-spaced document:
8: 
9: %%\documentclass[preprint2]{aastex}
10: 
11: %% Sometimes a paper's abstract is too long to fit on the
12: %% title page in preprint2 mode. When that is the case,
13: %% use the longabstract style option.
14: 
15: %%\documentclass[preprint2,longabstract]{aastex}
16: \def\wig#1{\mathrel{\hbox{\hbox to 0pt{%
17:           \lower.5ex\hbox{$\sim$}\hss}\raise.4ex\hbox{$#1$}}}}
18: 
19: \slugcomment{Accepted for publication in ApJ, May 15 2007}
20: 
21: 
22: \shorttitle{}
23: \shortauthors{Leggett et al.}
24: 
25: \begin{document}
26: 
27: \title{Physical and Spectral Characteristics of the T8 and Later-Type Dwarfs}
28: 
29: \author{S. K. Leggett}
30: \affil{Gemini Observatory,  670 N. A'ohoku Place Hilo HI 96720}
31: 
32: \author{M. S. Marley}
33: \affil{MS 245-3, NASA Ames Research Center, Moffett Field, CA 94035}
34: 
35: \author{R. Freedman}
36: \affil{SETI, 515 N. Whisman Road, Mountain View, CA 94043}
37: 
38: \author{D. Saumon}
39: \affil{Los Alamos National Laboratory, MS F663, Los Alamos, NM 87545}
40: \email{dsaumon@lanl.gov}
41: 
42: \author{Michael C. Liu}
43: \affil{Alfred P. Sloan Research Fellow, Institute for Astronomy, University of Hawaii, 2680 Woodlawn Drive, Honolulu HI 96822}
44: 
45: \author{T. R. Geballe}
46: \affil{Gemini Observatory,  670 N. A'ohoku Place, Hilo HI 96720}
47: 
48: \author{D. A. Golimowski}
49: \affil{Department of Physics \& Astronomy, Johns Hopkins University, 
50: 3400 North Charles Street, Baltimore, MD 21218}
51: 
52: \and
53: 
54: \author{D. C. Stephens}
55: \affil{BYU Department of Physics \& Astronomy, N-145 ESC, Provo, UT 84602}
56: 
57: 
58: \begin{abstract}
59: 
60: We use newly observed and published near--infrared spectra, 
61: together with synthetic spectra obtained from model atmospheres,
62: to derive physical properties of three of the 
63: latest-type T dwarfs.  A new $\rm R \approx 1700$ spectrum of the T7.5 dwarf HD~3651B, 
64: together with existing data, 
65: allows a detailed comparison to the well-studied and very similar dwarf, Gl~570D.  We find that HD~3651B 
66: has both higher gravity and metallicity than Gl~570D, with best-fit atmospheric parameters of $T_{\rm 
67: eff}=820 - 830\,\rm K$, $\log g= 5.4 - 5.5$, [m/H]$=+$0.2 and $K_{zz}=10^4$~cm$^2$s$^{-1}$.  Its  
68: age is 8 -- 12~Gyr and its implied mass is 60 -- 70~M$_{\rm Jupiter}$. We perform a similar analyis of the T8 and 
69: T7.5 dwarfs 2MASS J09393548-2448279 and 2MASS J11145133-2618235 using published data, comparing them to the 
70: well-studied T8, 2MASS J04151954-0935066.  We find that these two dwarfs have effectively the same 
71: $T_{\rm eff}$ as the reference dwarf, and similar or slightly higher gravities, but lower metallicities.
72: The derived parameters are $T_{\rm eff}=$725 - 775~K 
73: and [m/H]$=-$0.3; $\log g=5.3 - 5.45$ for 2MASS J09393548-2448279 and $\log g=5.0 - 5.3$ for 2MASS 
74: J11145133-261823. The age and mass are $\sim$10~Gyr and 60~M$_{\rm Jupiter}$ for 2MASS 
75: J09393548-2448279, and $\sim$5~Gyr and 40~M$_{\rm Jupiter}$ for 2MASS J11145133-261823. 
76: 
77: A serious 
78: limitation to such analyses is the incompleteness of the line lists for transitions of CH$_4$ and NH$_3$ at 
79: $\lambda \leq 1.7 \mu$m, which are also needed for synthesizing the spectrum of the later, cooler, Y type.
80: Spectra of Saturn and Jupiter, and of laboratory CH$_4$ and NH$_3$ gas, suggest that NH$_3$ features
81: in the $Y$- and $J$-bands may be useful as indicators of the next spectral type, and not  features in the $H$- and
82: $K$-bands as previously thought.  However, until cooler objects are found, or the line lists improve,
83: large uncertainties remain, as the abundance of NH$_3$ is likely to be significantly below the
84: chemical equilibrium value. Moreover 
85: inclusion of laboratory NH$_3$ opacities in our models predicts band shapes that are discrepant with 
86: existing data.  It is possible that the T spectral class will have to be extended to temperatures 
87: around 400~K, when water clouds condense in the atmosphere and dramatically change the spectral energy distribution 
88: of the brown dwarf.
89: 
90: 
91:  
92: 
93: 
94: 
95: 
96: \end{abstract}
97: 
98: \keywords{
99: infrared: stars --- stars: low-mass, brown dwarfs --- 
100: stars: individual (2MASS J04151954-0935066, 2MASS J09393548-2448279, 2MASS J11145133-2618235,
101: HD~3651B, Gl~570D)}
102: 
103: \section{Introduction}
104: 
105: The first of the T- or methane-type ultracool brown dwarfs, Gl~229B, was discovered as a companion
106: to an M dwarf by Nakajima et al. in 1995.  It took four more years for other examples to be found,
107: and these were revealed primarily as a result of  far-red and near-infrared large area sky surveys:   
108: the Sloan Digital Sky Survey (SDSS; York et al.\ 2000) and the Two Micron
109: All Sky Survey (2MASS; Beichman et al.\ 1998, Skrutskie et al.\ 2006).  There are now over 100
110: spectroscopically confirmed and published T dwarfs (see the on-line compendium at DwarfArchives.org).
111: A near-infrared classification scheme for T dwarfs, using H$_2$O and CH$_4$ absorption bands, 
112: has been defined by the 2MASS and SDSS groups
113: (Burgasser et al. 2006a).  Gl~229B is now classified as a peculiar T7
114: dwarf, and the other  known T dwarfs range from T0 to T8.  
115: 
116: In this paper we study three of the six known dwarfs later than T7 published as of February 2007. A summary of the 
117: astrometric properties and previously derived physical parameters of all six is given in Table 1. The uncertainty 
118: in spectral type classification is typically 0.5 of a subclass; hence by including T7.5 dwarfs this sample is representative 
119: of a ``T8'' sample. New and fainter very-late T dwarfs are now being discovered in the UKIRT Infrared Deep Sky Survey 
120: (UKIDSS; Hewett et al. 2006, Lawrence et al. 2007, Warren et al. 2007a).  Lodieu et al. (2007) present eight new T dwarfs, 
121: of which the latest is classified as a T7.5.  Warren et al. (2007b) present the latest-type T dwarf to date, a T8.5 dwarf.  
122: These two dwarfs are not discussed further, and the reader is directed to the discovery papers.
123: 
124: Three of the six dwarfs listed in Table 1, 2MASS J04151954-0935066 (T8, Burgasser et al. 2002, hereafter referred to as 
125: 2MASS0415-09), 2MASS J12171110-0311131 (T7.5, Burgasser et al. 1999, hereafter 2MASS1217-03) and Gl~570D (T7.5, Burgasser et 
126: al. 2000), have been recently examined in detail by Saumon et al. (2006, 2007) using synthetic spectra calculated from model 
127: atmospheres, and ground- and space-based observational data covering the wavelength range 0.9---15~$\mu$m. Saumon et al. 
128: (2007) show that 2MASS1217-03 is more metal-rich and hotter than the other two T dwarfs, with effective temperature $T_{\rm 
129: eff}\approx$900~K.  Gl~570D has an effective temperature of $T_{\rm eff}\approx 800\,\rm K$ and 2MASS0415-09 has a lower 
130: $T_{\rm eff}$ of around 750~K. In this paper we use Gl~570D and 2MASS0415-09 as reference objects for the other 
131: three dwarfs listed in Table 1, 2MASS J09393548-2448279 and 2MASS J11145133-2618235 (Tinney et al. 2005, hereafter 
132: 2MASS0939-24 and 2MASS1114-26), and HD~3651B (Murgauer et al. 2006, Luhman et al. 2007).
133: 
134: Burgasser, Burrows \& Kirkpatrick (2006b, hereafter BBK) present a method for determining the physical parameters of the 
135: coldest T dwarfs using low-resolution near-infrared spectra. BBK use an index measuring the strength of the H$_2$O feature 
136: at 1.15~$\mu$m, H$_2$O$-J$, as a measure of $T_{\rm eff}$, while the ratio of the 2.1~$\mu$m flux to the 1.6~$\mu$m flux, 
137: $K/H$, is used to measure gravity. They use modelled values of these ratios, calibrated by the $T_{\rm eff}$ and gravity of 
138: the well-studied dwarf Gl~570D, to determine $T_{\rm eff}$ and gravity for a sample of T6 to T8 dwarfs.  The values for the 
139: dwarfs in our sample are given in Table 1.  BBK suggest that 2MASS0939-24 and 2MASS1114-26 are the coolest 
140: dwarfs known, with $T_{\rm eff}$s cooler than that of 2MASS0415-09 by $\gtrsim$50~K. However, the gravity index of BBK is also 
141: sensitive to metallicity, as shown for example by Liebert \& Burgasser (2007) and Liu, Leggett \& Chiu (2007, hereafter 
142: LLC).  This could affect the conclusions drawn by BBK; thus here we re-examine the properties of these two very-late T 
143: dwarfs.
144: 
145: We also examine in detail the atmospheric properties of HD~3651B.  Luhman et al. (2007), Burgasser (2007) and LLC use 
146: near-infrared spectra and (for Luhman et al.) IRAC 3--8~$\mu$m photometry, to determine the physical parameters of 
147: this dwarf. The values of LLC and BBK are given in Table 1 (the LLC values agree well with those of Luhman et al.).  
148: This  dwarf is of particular interest, as its metallicity and age can be constrained from its primary star, which is only 
149: possible for  one other of the very late T dwarfs, Gl~570D. HD~3651B is also of interest as it is a wide companion to a star 
150: which hosts a planet (Fischer et al. 2003).
151: 
152: Finally, as several groups continue to search for dwarfs later and cooler than these (e.g. UKIDSS), we also explore what 
153: such objects may look like, using the observed spectra of the solar system gas giant planets Jupiter and 
154: Saturn, and laboratory spectra of CH$_4$ and NH$_3$ gas.
155: 
156: 
157: \section{Known Properties of the Sample}
158: 
159: \subsection{The K~V $+$ T7.5 Systems: Gl~570 and HD~3651}
160: 
161: Figure 1 shows previously published low-resolution near-infrared spectra for the T7.5 dwarfs
162: Gl~570D and HD~3651B. It can be seen that the spectra are quite similar, with HD~3651B being
163: slightly brighter in the 1.6~$\mu$m $H$-band and more significantly brighter in the 2.1~$\mu$m
164: $K$-band.
165: 
166: The Gl~570 system consists of a K dwarf, two early-M dwarfs, and the T dwarf.  The last was discovered in the 2MASS database 
167: by Burgasser et al. (2000) as a very widely separated 258$\farcs$3 (1525~AU projected separation) companion to the system.  
168: HD~3651B is also a companion to a K dwarf, at a distance of 42$\farcs$9 (476~AU projected separation), and the 
169: primary is known to be an exoplanet host. HD~3651B was discovered in near- and mid-infrared imaging searches for companions 
170: to exoplanet stars by Mugrauer et al. (2006) and Luhman et al. (2007).  The metallicities of both primaries are known: 
171: Gl~570A has [m/H]$=$ +0.00 to +0.07 (Feltzing \& Gustafsson 1998, Santos et al. 2005), and HD~3651A has [m/H]$=$ +0.09 to 
172: +0.16 (Gray et al. 2003, Santos et al. 2004, Valenti \& Fischer 2005).
173: 
174: 
175: Evolutionary models, constrained by the observed luminosity and further refined by comparison to synthetic 
176: spectra, show that the age of Gl~570D is 3 -- 5~Gyr, with $T_{\rm eff}= 800 - 820\,\rm K$, $\log g= 5.09 - 
177: 5.23$ and mass $38 - 47\,\rm M_{Jupiter}$, where the lower limit to $T_{\rm eff}$, log~$g$ and mass 
178: correspond to the youngest age (Saumon et al. 2006).  Similar analysis shows that the age of HD~3651B is 3 
179: -- 12~Gyr, with $T_{\rm eff}= 780-840\,\rm K$, $\log g= 5.1-5.5$ and mass $40-72\,\rm M_{Jupiter}$, where 
180: the younger age gives lower $T_{\rm eff}$, log~$g$ and mass (LLC).  However, LLC show that the age of the 
181: HD~3651 system {\it relative} to the Gl~570 system can also be determined.  The log($R'_{HK}$) 
182: chromospheric activity parameters suggest that the HD~3651 system is about five times older than the 
183: Gl~570 system; the X-ray luminosities suggest that the HD~3651 system is about twice as old as the Gl~570 
184: system; and isochrone analyses suggest that the HD~3651 system is at least 2.5 times older than the 
185: Gl~570 system.  Hence if the age of Gl~570D is 3 -- 5~Gyr, then the age of HD~3651B is 8 -- 12~Gyr and the dwarf 
186: has $T_{\rm eff}=$ 825 -- 840, log~$g=$ 5.4 -- 5.5 and mass 60 -- 72~M$_{Jupiter}$ (Figure 3 of LLC).
187: 
188: A higher value of $\log g$ for HD~3651B than Gl~570D would seem discrepant 
189: with the observed brighter $K$-band flux peak of HD~3651B (Figure 1),
190: as a higher gravity leads to stronger $K$-band H$_2$ absorption at this wavelength (\S 3).
191: Because HD~3651B is brighter than Gl~570D at $K$, 
192: Burgasser (2007), applying the BBK technique  to derive gravity, determines
193: a lower gravity for HD~3651B of $\log g= 4.7$, if the HD~3651 system has solar metallicity. 
194: However, using atmospheric models, Burgasser shows that increasing metallicity results in a
195: higher indices-implied gravity, and he derives values that agree with the luminosity- and age-driven
196: LLC values if the system has [m/H]$=+$0.24.
197: 
198: 
199: \subsection{2MASS0415-09, 2MASS0939-24 and 2MASS1114-26}
200: 
201: Figure 2 shows previously published low-resolution near-infrared spectra for the 
202: T8 dwarfs 2MASS0415-09 and 2MASS0939-24, and the T7.5 dwarf 2MASS1114-26.
203: It can be seen that the spectra are similar, but 2MASS0939-24 and 2MASS1114-26 
204: have broader 1.0~$\mu$m $Y$-peaks and are fainter at $K$ than 2MASS0415-09.
205: 
206: These three dwarfs  were all discovered as
207: isolated field T dwarfs in the 2MASS survey (Burgasser et al. 2002, Tinney et al. 2005).
208: Their ages  cannot be well constrained; however, the measured or estimated tangential 
209: velocities (see Table 1) suggest that they are not likely to be either very young, or members of the halo
210: (see for example the discussion of kinematics in the solar region by Eggen (1998), and references therein).
211: Their kinematics suggest ages of 1 -- 10~Gyr.  
212: Evolutionary models, constrained by observed luminosity and comparisons with model spectra, 
213: show that the age of 2MASS0415-09 is 3 -- 10~Gyr, with
214: $T_{\rm eff}=$ 725 -- 775~K and log~$g=$ 5.00 -- 5.37, where the younger age corresponds to lower $T_{\rm eff}$
215: and  log~$g$ (Saumon et al. 2007).  Metallicity could not be well constrained by Saumon et al.
216: as both [m/H]=0 and +0.3  models gave good fits to the data. 
217: 
218: The spectral classification indices for all three of these dwarfs are very similar (Burgasser et al. 2006a), despite the 
219: fact that 2MASS1114-26 is classified 0.5 of a sub-type earlier. The H$_2$O$-J$- and 
220: H$_2$O$-H$ indices are essentially identical, while the CH$_4-J$, CH$_4-H$ and CH$_4-K$ indices place 
221: 2MASS0939-24 and 2MASS1114-26 at a very slightly earlier type than 2MASS0415-09 (see Figure 2).
222: 
223: 
224: \section{Model Atmospheres}
225: 
226: \subsection{Description of the Models}
227: 
228: Models of ultracool-dwarf atmospheres have been developed by members
229: of our team (Ackerman \& Marley 2001; Marley et al.\ 2002; Saumon et al.\ 2003). For this work,
230: a grid of models and synthetic spectra have been calculated  with
231: effective temperature $T_{\rm eff}=600$ -- 850~K in steps of 50~K; gravity $g=$300, 1000 and
232: 3000 m/s$^2$ (or in the commonly-used c.g.s units, log~$g=$4.5, 5.0 and 5.5); and metallicity
233: [m/H]$=-0.3$, 0 and $+0.3$.  To begin to explore lower temperatures, 500~K and 550~K solar metallicity models
234: with $g=$300 and 1000 m/s$^2$ have also been calculated.
235: 
236: These models yield temperature-pressure-composition structures under conditions of 
237: radiative-convective equilibrium using 
238: the thermal radiative transfer source function technique of Toon et al.\ (1989).  For the very low-temperature dwarfs 
239: considered here, cloud decks of iron and silicate condensates are expected to lie below the photosphere and do not impact 
240: the spectral energy distribution except by reduction of the abundances of O, 
241: Fe, Si, Mg, Ti, and other refractory elements.  Thus we neglect grain opacities in the atmospheric structure and spectral 
242: calculations, although grain formation is included in the calculation of the chemical equilibrium.
243: 
244: The chemical equilibrium calculations for the atmosphere models are described in Lodders \& Fegley (2002) 
245: and use the solar system  
246: abundances of Lodders (2003).  There is debate about the solar CNO abundance values (e.g. Ayres, Plymate \& Keller 2006, 
247: Allende Prieto 2007), which strongly impact brown dwarfs' spectral energy distributions (SEDs) as they are shaped by 
248: opacities involving these species. Studies of line formation in the solar atmosphere by Asplund et al. (2004) have 
249: determined carbon, oxygen and nitrogen abundances 0.13--0.17~dex (35--50\%) 
250: lower than those previously given by Grevesse \& 
251: Sauval (1998).  The Lodders (2003) CNO abundances are 0.07--0.12~dex higher than the Asplund et al. values, 
252: and 0.02--0.07 dex  lower than the Grevesse \& Sauval (1998) values. The total mass fraction of all heavy elements is 
253: 0.015 compared to Grevesse \& Sauval's 0.017 and Asplund et al.'s 0.012.
254: 
255: The gas opacity database includes the molecular lines of H$_2$O, CH$_4$, CO, NH$_3$, H$_2$S, PH$_3$, TiO, VO, CrH, FeH and 
256: CO$_2$, the atomic lines of the alkali metals (Li, Na, K, Rb and Cs), and continuum opacity from H$_2$ collisionally induced 
257: absorption (CIA), H$_2$, H and He Rayleigh scattering, bound-free opacity from H$^-$ and H$_2^+$, and free-free opacity from 
258: H$^-$, He, H$_2^-$, and H$_2^+$ (Freedman, 
259: Marley \& Lodders 2007).  For  further discussion of the opacities important in brown dwarf photospheres see Sharp \& 
260: Burrows (2007). For this work, the line list for CH$_4$ (which does not include transitions at $\lambda < 1.5 \mu$m) has been 
261: supplemented with the Strong et al. (1993) CH$_4$ laboratory 
262: data for 1.05--1.5~$\mu$m and the CH$_4$ opacity derived from the spectrum of Saturn by Karkoschka (1994) below 1$\,\mu$m.
263: The line list for NH$_3$ does not include wavelengths shorter than 1.4~$\mu$m, and is incomplete at 1.4--1.9~$\mu$m. We 
264: discuss this further in \S 7.
265: 
266: The models have been further advanced to allow for the inclusion of turbulent vertical
267: mixing, which draws up long-lived chemical species from deep, hot layers into
268: the cool radiative photosphere, causing their observed abundances to be greater than expected for the case of
269: chemical equilibrium.  The mixing is parameterized using a diffusion coefficient 
270: $K_{zz}$. In the turbulent convection zone we use mixing length theory to
271: derive $K_{zz}$, but we treat it as a tunable parameter in the nominally stable radiative zone. 
272: The abundances of very stable species like
273: CO and N$_2$ are enhanced in the upper atmosphere relative to their
274: chemical-equilibrium values, while those of H$_2$O, CH$_4$ and NH$_3$
275: are reduced (see e.g. discussion in Saumon et al. 2007 and references therein).
276: For T dwarfs, non-equilibrium chemistry is very important at mid-infrared wavelengths as
277: the abundances of CO and NH$_3$ can be affected by an order of magnitude or more
278: (e.g. Figure 3 of Saumon et al. 2006),
279: and these species have strong features around 5~$\mu$m and 10~$\mu$m. 
280: The effect is not as important in the near-infrared  as the abundances of species producing
281: the majority of the opacity there, H$_2$O and CH$_4$, are not as strongly impacted.
282: 
283: \subsection{Comparison of Synthetic and Observed Spectra}
284: 
285: Figure 3 shows synthetic spectra calculated from model atmospheres with a range of $T_{\rm eff}$, log~$g$ and metallicity 
286: appropriate for the dwarfs studied here. Vertical mixing is not included in these spectra as the effects in the 
287: near-infrared are insignificant (see Figure 2 of Saumon et al. 2006). Temperature, gravity and metallicity do however 
288: impact the near-infrared spectral energy distribution. The CH$_4$ and H$_2$O absorption bands are so strong at these 
289: temperatures that the effect of changing $T_{\rm eff}$ is quite subtle; the largest effects are calculated to occur in the 
290: blue and red wings of the $H$-band flux peak, and at the peak of the $K$-band. Increasing metallicity 
291: has a similar effect on the $K$-band
292: to decreasing gravity, although the effect on the $Y$-band differs.  The 2~$\mu$m flux is very sensitive to 
293: gravity and metallicity because of the pressure-induced H$_2$ opacity; the 1~$\mu$m flux is controlled by the H$_2$O opacity 
294: which is more sensitive to metallicity than gravity.
295: 
296: Modelling these atmospheres is extremely complex. While great advances have been made (and continue to be made), there are 
297: known deficiencies and uncertainties in the models.  In particular, the molecular line lists for CH$_4$ and NH$_3$ are 
298: incomplete or even nonexistent at $\lambda\leq$1.7~$\mu$m (e.g. Saumon et al. 2007). Figure 4 tests the models by comparing 
299: the synthetic and observed spectra for the two well-studied dwarfs 2MASS0415-09 and Gl~570D.  Note that the synthetic 
300: spectra have not been scaled to fit, but instead the known distance and structural calculations of the radius (determined by 
301: $T_{\rm eff}$ and gravity) are combined to determine the flux at the Earth for each dwarf. The observed luminosity and 
302: robust evolutionary calculations can constrain $T_{\rm eff}$ and gravity for the dwarfs to a small allowed range as 
303: described in \S 2. The synthetic spectra in Figure 4 fall within these ranges although they may not be optimal at the 0.2 
304: dex level for gravity and metallicity.
305: 
306: Figure 4 illustrates the successes and remaining failures of current models. Given the complexity of the
307: model physics, the match to the observations is  good, but there are known problems.
308: Incompleteness in the CH$_4$ and NH$_3$ 
309: opacities most likely lead to the excess model flux in the $Y$, $J$ and $H$ bands (this is discussed further in \S 7).  The 
310: K~I doublet at 1.25~$\mu$m is too strong; this is probably due to remaining errors in the temperature-pressure profile for 
311: the atmosphere, to which the condensation of K~I into KCl is very sensitive (e.g. Lodders 1999 Figure 2).  The shortfall in 
312: flux in the $K$-band may be due to low CNO abundances (mimicking metal deficiency), or to errors in the 
313: pressure-induced H$_2$ opacity.
314: 
315: Using Gl~570D and 2MASS0415-09 as templates, and understanding the remaining model deficiencies, we now study the three 
316: very-late T dwarfs, HD~3651B, 2MASS0939-24 and 2MASS1114-26 in detail.  We start with HD~3651B, for which new data have been 
317: obtained.
318: 
319: 
320: 
321: \section{New Observations of HD~3651B}
322: 
323: The Gemini Near-Infrared Spectrograph (GNIRS, Elias et al. 2006) was used in cross-dispersed mode with a 
324: 0$\farcs$3 slit 
325: to obtain a 0.9 -- 2.5~$\mu$m R=1700 spectrum of HD3651B on 2006 September 23, through program GS-2006B-DD-5.  
326: The exposure time was 600~s, and HD3651B was nodded $6^{\prime\prime}$ along the slit in one ``ABBA'' pattern, for a total 
327: of 40 minutes on-source integration time. The F2V star HIP 118301 was used to remove telluric features, and 
328: flat fields and wavelength calibration 
329: were obtained using lamps in the on-telescope calibration unit.  The data were reduced using IRAF
330: and FIGARO software packages.  A slit position angle of 200 degrees was used, and no contamination 
331: was seen from the bright primary $43^{\prime\prime}$ to the north-west.
332: 
333: Figure 5 shows the final spectrum, compared to a similar resolution spectrum from McLean et al. (2003) for Gl~570D.  As  
334: in the low-resolution spectra for these dwarfs (Figure 1), the two spectra are very similar, and what looks like noise is in 
335: fact real structure due to, primarily, H$_2$O and CH$_4$ absorption.  The apparent feature seen in Figure 5 for HD~3651B 
336: around 1.27~$\mu$m is not seen in Figure 1 and is
337: most likely due to incomplete removal of the strong sky lines  at this wavelength.
338: The only significant difference between 
339: HD~3651B and Gl~570D appears in the $K$-band, where the absorption features of HD~3651B are stronger.
340: 
341: 
342: \section{Re-Examination of HD~3651B}
343: 
344: \subsection{Near-Infrared Spectra}
345:  
346: Figure 6 shows the observed low-resolution near-infrared spectrum of HD~3651B and compares it to synthetic 
347: spectra with $T_{\rm eff}=$ 800 and 850~K, log~$g=$ 5.0 and 5.5, and metallicity [m/H]$=$ 0 and +0.3,
348: covering the range of values determined for this dwarf by LLC (\S 2.1). 
349: These spectra have been scaled to the flux at the Earth, given the known distance to the dwarf and the 
350: calculated radii appropriate for the values of $T_{\rm eff}$, $g$ and [m/H]. Comparison of Figure 6 to the 
351: template model fits of Figure 4 shows that for either the 800~K or 850~K models, the log~$g=$ 5.5 and 
352: [m/H]$=$ 0 spectra produce too low a $K$-band flux, and hence this combination of gravity and metallicity 
353: can be excluded. An average of the log~$g=$5.0 and 5.5, [m/H]$=$ 0 and $+0.3$, 800~K spectra shown in the 
354: top panel, weighted towards the higher gravity spectrum for less flux at 1.2~$\mu$m, would produce a 
355: similar fit to the one seen for Gl~570D in Figure 4.  However the flux at the $H$ and $K$ peaks would be 
356: low, implying a slightly hotter temperature as suggested by the 850~K log~$g=$5.5 [m/H]$=+0.3$ comparison. 
357: A temperature as high as 850~K can be excluded by the excessive 1.60 -- 1.75~$\mu$m flux.  
358: 
359: The bottom panel compares the observed spectrum to a synthetic spectrum for $T_{\rm eff}=$ 825~K, 
360: log~$g=$5.5 and [m/H]$=+0.3$ (calculated by averaging the 800 and 850~K spectra). The
361: quality of the fit is indeed similar to that for Gl~570D in Figure 4, although a slightly lower 
362: metallicity would improve the match at $Y$, and this would then require a decrease in 
363: gravity to maintain the match at $K$.   Hence our 
364: best  fits have $T_{\rm eff}\approx$ 825 with log~$g=$5.4 -- 5.5 and  [m/H]$\approx +0.2$.
365: This metallicity is somewhat higher than the measured values of $+0.09$ to 
366: $+$0.16 (Gray et al. 2003, Santos et al. 2004, Valenti \& Fischer 2005).  However Valenti \& Fischer find 
367: a range for individual elements from $+$0.11 for [Ti/H] to $+$0.24 for [Na/H] and so [m/H]$\approx +0.2$ 
368: is not unreasonable.  Note that although there are uncertainties in the absolute values of the solar 
369: chemical abundances (\S 3.2), they do not affect the conclusion reached here that HD~3651B is 0.2 dex 
370: more metal-rich than Gl~570D, which has effectively solar chemistry (\S 2.1).
371: 
372: Figure 7 plots the observed medium-resolution $K$-band spectra of Gl~570D and HD~3651B with the synthetic 
373: spectra for the [$T_{\rm eff}$, log~$g$, [m/H]] models [800, 5.0, 0.0] and [800, 5.5, +0.3]. Although 
374: there are discrepancies between the data and the models, 
375: the relative depths of the observed and modelled features around 
376: 2.05~$\mu$m and 2.14 -- 2.18~$\mu$m (see also Figure 5) supports the conclusions reached above that HD~3651B has 
377: a higher gravity and higher metallicity than Gl~570D.  
378: 
379: 
380: \subsection{IRAC Photometry}
381: 
382: $Spitzer$ IRAC mid-infrared four-channel photometry (Fazio et al. 2004) is available for both Gl~570D 
383: (Patten et al. 2006) and HD~3651B (Luhman et al. 2007).  Saumon et al. (2006) show that the 6 -- 14~$\mu$m 
384: spectrum of Gl~570D obtained with the $Spitzer$ Infrared Spectrograph (IRS, Houck et al. 2004) matches 
385: very well the synthetic spectrum generated from a model with $T_{\rm eff}=$821~K, log~$g=$5.23, 
386: [m/H]$=$0.0 and $K_{zz}=10^2$~cm$^2$s$^{-1}$.  As the distances to the two dwarfs are known, we can 
387: compare the IRAC absolute magnitudes calculated by the models to the observed values (note that the IRAC 
388: magnitudes cannot be literally interpreted as a flux at the nominal filter wavelength, Cushing et al. 
389: 2006).
390: 
391: Figure 8 plots the observed IRAC absolute magnitudes for the two dwarfs, together with modelled values for 
392: $T_{\rm eff}=$825~K and vertical mixing coefficients $K_{zz}=0$ (i.e. equilibrium chemistry) and 
393: $K_{zz}=10^4$. The modelled values have been determined by interpolating between the 800 and 850~K 
394: synthesized photometry. Log~$g=5.25$ and [m/H]$=0$ values appropriate for Gl~570D are shown, as well as 
395: log~$g=5.5$ and [m/H]$=+0.3$ values, appropriate for HD~3651B. The agreement between observed and modelled 
396: values for both dwarfs is excellent for the non-equilibrium chemistry case, with an average deviation of 
397: 0.10 magnitudes. The deviation between the observed and modelled $K_{zz}=0$ magnitudes is on average 0.2 
398: magnitudes for both dwarfs, with individual band discrepancies as large as 0.3 magnitudes. Hence 
399: non-equilibrium chemistry must be included in studies of the mid-infrared fluxes of the late T dwarfs, as 
400: also found by Golimowski et al. (2004), Patten et al. (2006) and Leggett et al. (2007).
401: 
402: 
403: 
404: \subsection{Adopted Parameters for HD~3651B}
405: 
406: Our best-fit parameters for HD~3651B are summarised in Table 2. 
407: Comparison to the well-studied dwarf Gl~570D, and synthetic spectra, show that 
408: HD~3651B has $T_{\rm eff}=$820 -- 830~K, log~$g=$5.4 -- 5.5 and [m/H]$\approx +0.2$.
409: The corresponding age and mass is 8 -- 12~Gyr and 60 -- 70~M$_{\rm Jupiter}$
410: (Figure 3 of LLC).
411: Our derived gravity agrees with the high end of the 
412: range determined by LLC (Table 1), which is consistent with an age at the high end of their allowed 
413: range. The latter is in turn consistent with the relative ages of the Gl~570 and HD~3651 systems (\S 2.1).  
414: The low end of our range of values for $T_{\rm eff}$,  log~$g$, age and mass agree with the
415: values derived by Burgasser (2007) when the BBK flux indices are recalibrated for a metallicity
416: of [m/H]$=+$0.24, consistent with the value implied by our model fit.  The Burgasser values for
417: lower metallicity require the HD~3651 system to be younger than 4.7~Gyr (see his Table 1), 
418: which is inconsistent with the various age indicators described in \S 2.1.   
419: 
420: 
421: \section{Re-examination of 2MASS0939-24 and 2MASS1114-26}
422: 
423: The distances to 2MASS0939-24 and 2MASS1114-26 have not yet been published and so an absolute comparison of synthetic to 
424: observed spectra, i.e. with a known scale factor similar to Figures 4 and 6, cannot be made for these dwarfs.  Neither can 
425: their luminosities be measured and $T_{\rm eff}$ derived from structural models, as has been done for 2MASS0415-09, Gl~570D 
426: and HD~3651B. Nevertheless, a comparison of the model spectra shown in Figure 3 to the observed spectra shown in Figure 2 
427: allows us to constrain $T_{\rm eff}$, log~$g$ and [m/H] for them, relative to the well-studied 
428: dwarf 2MASS0415-09. As described in \S 2, Saumon et al. (2007) shows that the age of 2MASS0415-09 is 3 -- 10~Gyr, with 
429: $T_{\rm eff}=$ 725 -- 775~K and log~$g=$ 5.00 -- 5.37; metallicity could 
430: not be constrained. We suggest that 2MASS0415-09 has solar or only slightly higher metallicity, as Figure 
431: 4 shows that the quality of the fit, especially in the metallicity-sensitive $K$-band, is very similar to that of Gl~570D, 
432: for which a solar metallicity has been measured (\S 2.1).
433: 
434: Figure 3 shows that the ratio of the $H$-band flux peak to the depth of the CH$_4$ absorption around 1.7~$\mu$m, is 
435: sensitive to $T_{\rm eff}$.  Although the model opacities are incomplete in this region, we assume the strength of this 
436: absorption can still be used for relative estimates of $T_{\rm eff}$.  Hence $T_{\rm eff}$ for 2MASS0939-24 and 2MASS1114-26 
437: appears to be essentially identical to that for 2MASS0415-09, as would also be surmised from the strengths of the H$_2$O 
438: absorptions at 1.15 and 1.45~$\mu$m.
439: 
440: The suppressed $K$-band flux for 2MASS0939-24 and 2MASS1114-26, relative to 2MASS0415-09,
441:  can be explained by either higher gravity or lower metallicity. These effects however can be distinguished by their effects 
442: on the $Y$-band, as shown in Figure 3. (Again, although the models do not perfectly reproduce the $Y$-band observation, we 
443: believe that comparisons in a relative sense are valid.) The relative heights of the $K$-bands and the width of the $Y$-band 
444: flux peaks suggest that both 2MASS0939-24 and 2MASS1114-26 are quite metal-poor compared to 2MASS0415-09. Adopting solar 
445: metallicity for 2MASS0415-09 implies [m/H]$\approx-0.3$ for these two dwarfs.  The additional suppression at $K$ for 
446: 2MASS0939-24 suggests that it also has a higher gravity than 2MASS0415-09.
447: 
448: Our adopted parameter ranges for these dwarfs are given in Table 2 and are: $T_{\rm eff}=$ 725 -- 775~K, [m/H]$\approx-0.3$, and 
449: log~$g\approx$5.0 -- 5.3 for 2MASS1114-26 and 5.3 -- 5.45 for 2MASS0939-24.  The upper gravity limit for 2MASS0939-24 is 
450: determined by evolutionary arguments, assuming an upper age limit of 12~Gyr for the dwarf. We do not confirm the lower 
451: temperature of $\lesssim$700~K determined for these dwarfs by BBK (Table 1); instead we find that these dwarfs each have 
452: $T_{\rm eff}$ essentially identical to that of 2MASS0415-09, and suggest that the method of BBK needs to be revised to
453: take metallicity into account (as done for HD~3651B by Burgasser 2007).
454: 
455: Saumon et al. (2007) and LLC show that $T_{\rm eff}$ and gravity give mass and age for a brown dwarf, and 
456: that changing metallicity by 0.3 dex impacts derived mass and age by $<$5\%.  We derive ages and 
457: masses for these metal-poor late-T dwarfs of 10 -- 12~Gyr and 50 -- 65~M$_{\rm Jupiter}$ for 
458: 2MASS0939-24, and 3 -- 8~Gyr and 30 -- 50~M$_{\rm Jupiter}$ for 2MASS1114-26.  When available, parallaxes 
459: will constrain these parameters further.
460: 
461: 
462: \section{Dwarfs Cooler than 700~K and the Next Spectral Type}
463: 
464: \subsection{Expected Spectral Trends and the Opacity Problem}
465: 
466: Many researchers are currently searching for objects cooler than the coldest known brown dwarfs with $T_{\rm 
467: eff}\approx$700~K, and in particular for the next, cooler, spectral type, for which the letter ``Y'' has been suggested 
468: (e.g. Kirkpatrick 2005).  Until examples of cooler dwarfs are found one cannot be certain what spectral change will
469: trigger the use of a new letter; if no obvious spectral changes are seen, the T classification will need to 
470: be extended beyond T8.  
471: 
472: For practical reasons, the spectral indicators for classification of the extreme-T or Y dwarfs should 
473: remain in the far-red or near-infrared, and should be identifiable
474: based on low spectral-resolution data, as these dwarfs will be faint. 
475: As the temperature decreases
476: more flux emerges at mid-infrared wavelengths, relative to the red and near-infrared; 
477: the calculated ratios of the 4.5~$\mu$m and 10.0~$\mu$m
478: fluxes relative to the $J$-band 1.27~$\mu$m flux, in Jy, are 2.3 and 1.2 
479: at 600~K, and 3.4 and 1.7 at 500~K.  
480: However, detector sensitivity in the mid-infrared is worse than in the 
481: near-infrared, such that, 
482: for example, it would take 2--5 times longer to obtain a  5~$\mu$m $Spitzer$ spectrum than a 
483: $J$-band spectrum at the same S/N ratio with current 
484: instrumentation on 8~m-class telescopes.  Although $JWST$ is expected to be very sensitive in the 
485: mid-infrared, it will be $\sim$6 magnitudes more sensitive in the near-infrared. 
486: 
487: In this Section we use planetary and laboratory data,  and new model calculations, to 
488: investigate spectral trends in the near-infrared as $T_{\rm eff}$ drops 
489: below 700~K. If the changes are too subtle to trigger a new spectral type, it will still be important 
490: to find and understand indicators of lower temperature in the 1.0--2.2~$\mu$m region,
491: as very little flux remains in the H$_2$O and CH$_4$ absorption
492: features currently used to classify the T dwarfs (see Figures 2 and 3).
493: 
494: A summary of the expected spectral changes at $T_{\rm eff}<$~700~K is given by Burrows, Sudarsky \& Lunine (2003).  
495: They describe how, around $T_{\rm eff}=$600~K, the strong red and near-infrared absorption features of K~I weaken 
496: as this element forms grains of KCl, and NH$_3$ features become apparent in the blue wings of the $H$ and $K$ flux 
497: peaks, around 1.5 and 1.95~$\mu$m.  Below 400~K, water clouds are expected to form causing 
498: the colors of dwarfs
499: to change dramatically. It is possible that the NH$_3$ features
500: in the $H$- and $K$-bands may trigger the next spectral type. 
501: (NH$_3$ is already seen at 10~$\mu$m in T dwarfs (Roellig et al. 2004),
502: however its appearance in the more easily measured near-infrared could be used as a trigger 
503: in the same way that CH$_4$ at $H$ and $K$ is used to trigger the T type, 
504: despite CH$_4$ absorption being prominent at 3~$\mu$m in mid- and late-L dwarfs (Noll et al. 2000)).
505: However, although our 700~K and 600~K model spectra in Figure 3 (and our 500~K  spectra, not shown)
506: show a change in the shape of the $Y$-band flux peak, and a steady 
507: decline of the $K$-band flux, there is no obvious spectral change that would indicate the necessity of a new 
508: spectral type. 
509: 
510: Unfortunately, both our analysis and that of Burrows, Sudarsky \& Lunine, suffer from incomplete or non-existent
511: line lists for the important opacity sources CH$_4$ and NH$_3$. 
512: The models shown here use line lists for CH$_4$ at $\lambda > 1.5 \mu$m, and CH$_4$ laboratory and 
513: astrophysical data (Strong et al. 1993; Karkoschka 1994) for $\lambda \leq 1.5 \mu$m. For NH$_3$, a line list is used for 
514: $\lambda > 1.4 \mu$m, but it is known to be  incomplete at 1.4 -- 1.9~$\mu$m and no 
515: NH$_3$ opacity is included at shorter wavelengths. Discrepancies between synthetic spectra and 
516: the observations can be seen in the $Y$-, $J$- and $H$-bands in Figure 4, most likely due to these incomplete opacities.  
517: Due to these  problems, we revisit the predictions of the SEDs of dwarfs cooler 
518: than 700~K, first from the observational and then from the theoretical point of view.
519: 
520: \subsection{Observational Indicators}
521: 
522: 
523: Figure 9 shows observed 0.96 -- 2.2~$\mu$m spectra of the 750~K dwarf 2MASS0415-09, the $\sim$160~K planet 
524: Jupiter{\footnote {from the IRTF Spectral Library, 
525: accessible at http://irtfweb.ifa.hawaii.edu/~spex/spexlibrary/IRTFlibrary.html, 
526: see Rayner, Cushing \& Vacca (2008)}},
527: and the $\sim$95~K planet Saturn.  Also shown are room-temperature laboratory data for CH$_4$ (Cruikshank \& 
528: Binder 1968) and NH$_3$ (Irwin et al. 1999). The physical environment of the brown dwarf atmosphere will be quite 
529: different from either the laboratory gas or the planets' atmospheres. Also, there is no H$_2$O in 
530: the planets' spectrum as all H$_2$O is condensed in deep cloud decks. Moreover, the planetary spectra are 
531: reflection spectra.  However,  both planets have NH$_3$ features in their spectra, with the warmer
532: Jupiter having less condensed NH$_3$ and stronger NH$_3$ gas absorption.  Given the 
533: absence of more appropriate data, we use these spectra to investigate what changes might occur at cold temperatures.
534: 
535: Starting at the longest wavelength, it can be seen that there is very little flux from Saturn or Jupiter in the 
536: $K$-band, due to strong pressure-induced H$_2$ opacity. Figure 3 shows that brown dwarfs also become very 
537: faint at $K$ with decreasing temperature. Figure 9 shows that the observed T8 $K$-band spectrum is very complex, due to the 
538: wealth of H$_2$O and CH$_4$, and possibly NH$_3$, absorption features 
539: (see also the higher resolution spectra of the T7.5 dwarfs in Figures 5 and 7, and the discussion of the Gl~229B 
540: $K$-band spectrum in Saumon et al. 2000).  The feature seen in 2MASS0415-09 at 2.02~$\mu$m is as strong 
541: in hotter 800~K spectra (Figures 1 and 3) and so is not a useful spectral indicator despite the 
542: apparent match to an NH$_3$ feature in the laboratory spectrum. Although the blue wing of this band 
543: was suggested by Burrows et al. (2003) as a region where increasing NH$_3$ absorption would become 
544: apparent for the temperature range of interest, Figures 3 and 9 show that 
545: identifying the increasing NH$_3$ opacity at 1.95 -- 2.05~$\mu$m amongst the existing features, 
546: and with low flux levels, is not practical. Hence the $K$-band is not an appropriate region to 
547: search for features that may indicate the next spectral type.
548: 
549: The $H$-band offers only a very narrow region of peak flux due to the strong H$_2$O features to the blue, 
550: and the strong CH$_4$ to the red of this band.  Identifying the increasing NH$_3$ opacity at 1.48 -- 
551: 1.55~$\mu$m, as suggested by Burrows et al. (2003), again is problematical due to the already strong
552: H$_2$O absorption at these wavelengths.  
553: A steepening of the blue wing of the $H$-band peak may become visible, and we discuss this further 
554: below.   If the dip seen in the laboratory spectrum of NH$_3$ 
555: at 1.58~$\mu$m, which appears to carry through in the Jupiter
556: spectrum, is also seen in brown dwarf spectra, it may be a useful Y dwarf indicator, as it is in a region 
557: relatively free of H$_2$O and CH$_4$ absorption.
558: 
559: It appears that the $Y$ and $J$ regions may provide the best Y dwarf indicators.  This 
560: would be fortuitous, not only for the coincidence of the letter ``Y'', but because these 
561: extreme dwarfs emit more energy at 1.0--1.3~$\mu$m than at 1.5--2.2~$\mu$m.   
562: Weak absorption features in the 2MASS0415-09 spectrum at 1.21--1.22~$\mu$m and 1.29~$\mu$m  
563: apparently align with NH$_3$ features, and are weaker or non-existent in the warmer Gl~570D spectrum.
564: Even if these are due to CH$_4$ and not NH$_3$, if they become dominant features of colder dwarfs 
565: they could be used as spectral type indicators. The feature at 
566: 1.27~$\mu$m is seen in the warmer spectrum and coincides with a strong sky emission feature, 
567: and so is not a useful spectral indicator. 
568: 
569: The blue wing of the $Y$-band peak may 
570: be a useful Y dwarf indicator due to the loss of gaseous K~I from the photosphere, as the strong 0.77~$\mu$m K~I 
571: feature impacts the blue edge of this band.
572: Also, the Jovian spectrum suggests that the 1.02--1.035~$\mu$m region may show NH$_3$ features.
573: The feature seen in 2MASS0415-09 at 1.02~$\mu$m is weak or non-existent in the 
574: Gl~570D spectrum, and so, whether due to NH$_3$ or CH$_4$, may be a useful later-T or Y dwarf indicator.
575:  
576: 
577: \subsection{Theoretical Indicators}
578: 
579: To further explore the appearance of NH$_3$ in very-late T dwarf spectra, we have complemented our NH$_3$ line opacity in 
580: the models with the Irwin et al. (1999) laboratory measurements, which were conducted at $T=230$ -- 296$\,$K at a resolution 
581: of 5$\,$cm$^{-1}$.  We applied their opacities over the range 0.91 -- 1.9$\,\mu$m by extrapolating the tabulated opacities 
582: to higher temperatures using their estimate for the energy of a representative lower level for the transitions in each 
583: frequency range.  As a test of the temperature extrapolation, we computed spectra with the measured opacities at a fixed 
584: temperature of 300$\,$K and obtained very similar synthetic spectra.
585: 
586: Figure 10 shows the resulting spectra for  brown dwarfs with $T_{\rm eff}$ of 600 and 700$\,$K.
587: A comparison of the 700$\,$K spectrum with one computed without the Irwin et al. (1999)
588: NH$_3$ opacity shows that a very strong feature would be expected on the blue side of the $H$ band
589: peak ($\sim 1.54\,\mu$m), that the $J$ band peak would be shaved on both sides, and that the blue
590: side of the $Y$ band peak would be suppressed by an NH$_3$ band.  While this would seem a promising way 
591: of identifying NH$_3$ features to define the spectral characteristics of Y dwarfs, none of these features
592: are apparent in the spectra of the coolest known T dwarfs with $T_{\rm eff} \sim 700$ -- 800$\,$K
593: (Figure 2).  The shoulders seen in the red wing of the $J$-band and the blue wing of the $H$-band are
594: in particular discord with the observational data (e.g. for 2MASS0415-09 in Figures 2, 4 and 9).
595: 
596: On the other hand, detailed analyses of cool T dwarfs have revealed a depletion of NH$_3$ in the atmosphere by a factor of 
597: $\sim 10$ due to non-equilibrium chemistry arising from vertical transport in the atmosphere (Saumon et al. 2006, 2007). 
598: Figure 10 shows a 600$\,$K spectrum with the Irwin et al. (1999) NH$_3$ opacity computed out of chemical equilibrium with an 
599: eddy diffusion coefficient of $K_{zz}=10^4\,$cm$^2\,$s$^{-1}$ (see e.g. Saumon et al. 2006). The NH$_3$ features from the 
600: Irwin et al. (1999) opacity are significantly weaker but remain stronger than seen in the 
601: observational data.  It is highly
602: unlikely that non-equilibrium chemistry can further decrease the NH$_3$ abundance by the required factor.  Additional 
603: laboratory work on the opacity of NH$_3$ well above room temperature, as well as {\it ab initio} line list calculations,
604: are urgently needed to understand the spectra of the new cooler brown dwarfs that will soon be uncovered by modern large 
605: surveys.
606: 
607: 
608: 
609: \section{Conclusions}
610: 
611: We have used the well-studied T7.5 and T8 dwarfs Gl~570D and 2MASS0415-09 as templates to determine 
612: physical parameters for the very similar dwarfs HD~3651B, 2MASS09393-24 and 2MASS1114-26.  Synthetic 
613: spectra and photometry, together with new and published near-infrared spectra and mid-infrared photometry, 
614: show that HD~3651B, a distant companion to an exoplanet host star, has both higher gravity and higher 
615: metallicity than Gl~570D.  The derived parameters for this dwarf are: $T_{\rm eff}=$820 -- 830~K, 
616: log~$g=$5.4 -- 5.5, [m/H]$=+$0.2 and $K_{zz}=10^4$~cm$^2$s$^{-1}$.  The corresponding age is around 10~Gyr 
617: and mass 65~M$_{\rm Jupiter}$. Synthetic spectra, together with published near-infrared spectra, show that 
618: 2MASS0939-24 and 2MASS1114-26 are both metal-poor compared to 2MASS0415-09, and have similar or slightly 
619: higher gravities.  The derived parameters for these dwarfs are: $T_{\rm eff}=$725 -- 775~K and 
620: [m/H]$=-$0.3; log~$g=$5.3 -- 5.45 for 2MASS0939-24 and log~$g=$5.0 -- 5.3 for 2MASS1114-26. The 
621: corresponding ages and masses are $\sim$10~Gyr and 60~M$_{\rm Jupiter}$ for 2MASS0939-24, and $\sim$5~Gyr 
622: and 40~M$_{\rm Jupiter}$ for 2MASS1114-26. These temperatures are significantly higher than those derived 
623: by BBK, and hence these objects are not the coolest known T dwarfs. For these two dwarfs (which do not have
624: published parallax values) the atmospheric 
625: properties are mostly derived from the ratio of the $J$ and $K$ flux peaks, which is sensitive to both 
626: gravity and metallicity, and the shape of the $Y$ flux peak, which is sensitive primarily to metallicity.
627: 
628: The intrinsic accuracy of the inferred atmospheric parameters depends, of course, on the veracity of the 
629: atmosphere models.  Until other model sets are similarly applied to comparable datasets, the systematic 
630: uncertainties are difficult to evaluate.  However, given the apparent simplicity of late T dwarfs 
631: (relative to the cloudy L dwarfs, for example), large systematic 
632: errors are unlikely.  Errors larger than 5 to 10\% in effective temperature and 0.3 dex in $\log g$ and 
633: metallicity would be surprising.  The uncertainties here in $\log g$ and metallicity are smaller than this 
634: and are estimated to be 0.1~dex (or 26\%), as we  have carried out relative comparisons to the well-studied 
635: dwarfs Gl~570D and 2MASS0415-09. 
636: 
637: When the metallicities and gravities of more T dwarfs have been determined (following 
638: improvements to current models), the community will need to agree on metallicity and gravity labels in
639: spectral classification schemes.  We suggest that these labels should follow existing stellar schemes.
640: Metal-poor main-sequence M dwarfs are labelled as ``subdwarfs'' due to their location in optical
641: color-magnitude diagrams (e.g. Figure 10 of Monet et al. 1992).  Although the location of brown dwarfs in 
642: such diagrams will be a strong function of gravity, the ``sd'' prefix could be maintained as a 
643: metal-poor indicator, as already being done for some L dwarfs (e.g. Burgasser 2004).
644: Similarly, gravities should be described numerically, using the log of the value in 
645: c.g.s. units, following the spectral type. It would be desirable to specify gravity to at least 
646: one-half a dex, as this parameter has a large impact on brown dwarf SEDs. Also, for a given 
647: $T_{\rm eff}$, a range of 1.0~dex in log~$g$ implies a large range in age and mass
648: (for example if $T_{\rm eff}= 800$~K, log~$g=4.5$ corresponds to mass and age around 15~M$_{\rm Jupiter}$
649: and 0.5~Gyr, while log~$g=5.5$ corresponds to 70~M$_{\rm Jupiter}$ and 15~Gyr).  
650: Given this requirement,
651: and the existing association of roman numerals with luminosity class, we suggest that arabic numerals are
652: used for brown dwarf gravities, following a ``g'' to separate it from the spectral type.  
653: Thus a metal-poor T7.5 dwarf with log~$g=5.5$ would be described as sdT7.5g5.5. For L and early-T dwarfs
654: an additional classification dimension will be required to describe the ``dustiness'' or ``redness'' of
655: the spectrum caused by the condensate cloud decks, perhaps using the letter ``r'' 
656: (as ``d'' would be confused with ``dwarf''). 
657: 
658: Existing models would  greatly benefit from improved line lists for 
659: transitions of CH$_4$ and NH$_3$ at $\lambda \leq 1.7 \mu$m.  Such lists would  allow substantially better 
660: analyses of the SEDs of the T dwarfs, and would yield better predictions for the 
661: spectra and colors of the next, cooler, spectral type. 
662: We use observed near-infrared spectra of 2MASS0415-09, Jupiter and Saturn, and laboratory observations of 
663: CH$_4$ and NH$_3$ gas, to investigate the appearance of this spectral type.  The expectation was 
664: that the appearance of NH$_3$ features in the blue wings of the $H$ and $K$ flux peaks would be significant
665: enough to require a new (``Y'') spectral type, and that this would occur at $T_{\rm eff}\approx 600$~K.
666: However, we find that the lack of flux in the $H$ and $K$ wings,
667: combined with the complexity of already strong H$_2$O and CH$_4$ absorption features, make it unlikely that these 
668: NH$_3$ features will be useful.  Instead we find that previously overlooked NH$_3$ features 
669: that occur in the brighter $Y$- and $J$-bands
670: may  be the indicator of the next spectral type, combined with the changing shape of the $Y$-band 
671: as K~I solidifies into KCl. Large uncertainties remain still, as 
672: an attempt to include NH$_3$ laboratory data into our models produced spectra that do not agree 
673: with observations, and as it is now known from mid-infrared analyses that 
674: the abundance of NH$_3$ in late-T dwarfs is around 10 times lower than expected from chemical equilibrium models.
675: It is possible that the next spectral type may not be triggered until much cooler temperatures are reached. 
676: Progress will be made as observers
677: continue to search to fainter limits to find cooler, and possibly spectrally different, brown dwarfs,
678: while theoreticians continue  improving
679: the opacity line lists and models.
680: 
681: 
682:  
683: 
684: 
685: \acknowledgments
686: 
687: We are grateful for the award of Director's Discretionary time at the Gemini
688: Observatory, which is operated by the
689: Association of Universities for Research in Astronomy, Inc., under a cooperative agreement
690: with the NSF on behalf of the Gemini partnership: the National Science Foundation (United
691: States), the Particle Physics and Astronomy Research Council (United Kingdom), the
692: National Research Council (Canada), CONICYT (Chile), the Australian Research Council
693: (Australia), CNPq (Brazil) and CONICET (Argentina).
694: TRG and SKL are supported by the Gemini Observatory, and acknowledge the above partnership
695: for this support. This work was supported in part under the auspices of the U.S. Department of
696: Energy at Los Alamos National Laboratory under Contract W-7405-ENG-36.
697: MSM acknowledges the support of the NASA Office of Space Sciences. 
698: MCL acknowledges support for this work from NSF grants AST-0407441 and
699:  AST-0507833 and an Alfred P. Sloan Research Fellowship.
700: Dale Cruikshank gave 
701: invaluable assistance with the CH$_4$ and NH$_3$ laboratory data. Katarina Lodders contributed
702: significantly in many helpful discussions. John Rayner generously provided the Jovian spectrum pre-publication.
703: Finally, we are grateful to the referee, who significantly improved the manuscript.
704: 
705: 
706: 
707: 
708: \begin{thebibliography}{}
709: 
710: \bibitem[Ackerman \& Marley (2001)] {ack01} Ackerman, A. S. \& Marley, M. S., 2001, \apj, 5556, 872.
711: \bibitem[Allende Prieto (2007)]{allende07} Allende Prieto, C. 2007, in Proc. of the 14$^{\rm th}$ Cambridge Workshop on Cool Stars, Stellar Systems and the Sun, G. van Belle, ed. in press.
712: %\bibitem[Anders \& Grevesse (1989)]{ag89} Anders, E. \& Grevesse, N. 1989, Geoch. Cosmo. Act. 53, 197
713: \bibitem[Asplund et al. (2004)]{asp04} Asplund, M., Grevesse, N., Sauval, A.~J., Allende Prieto, C., \& Kiselman, D., 2004, A\&A, 417,  751
714: \bibitem[Ayres, Plymate \& Keller (2006)]{ayr06} Ayres, T. R., Plymate, C. \& Keller, C. U. 2006, \apjs, 165, 618
715: \bibitem[Beichman et al. (1998)] {bei98} Beichman, C. A., Chester, T. J., Skrutskie, M., Low, F. J.  \& Gillett, F. 1998, \pasp, 110, 480
716: \bibitem[Burgasser (2004)]{bur04} Burgasser, A. J., 2004, \apj, 614, L73
717: \bibitem[Burgasser (2007)]{bur07} Burgasser, A. J., 2007, \apj, 658, 617
718: \bibitem[Burgasser et al. (1999]{bur99} Burgasser, A. J. et al. 1999, \apj, 522, L65
719: \bibitem[Burgasser et al. (2000)]{bur00} Burgasser, A. J. et al. 2000, \apj, 531, L57
720: \bibitem[Burgasser et al. (2002)]{bur02} Burgasser, A. J. et al. 2002, \apj, 564, 421
721: %\bibitem[Burgasser et al. (2003)]{bur03} Burgasser, A. J., Kirkpatrick, J. D., Liebert, J. \& Burrows, A. 2003, \apj, 594, 510
722: \bibitem[Burgasser et al. (2006a)]{bur06} Burgasser, A. J., Geballe, T. R., Leggett, S. K., Kirkpatrick, J. Davy \& Golimowski, D. A. 2006a, \apj, 637, 1067
723: \bibitem[Burgasser, Burrows \& Kirkpatrick (2006b)]{bur06b} Burgasser, A. J., Burrows, A. \& Kirkpatrick, J. D. 2006b, \apj, 639, 1095 (BBK) 
724: %\bibitem[Burrows et al. (1997)] {bu97} Burrows A. et al. 1997, \apj, 491, 856
725: %\bibitem[Burrows et al. (2001)] {bu01} Burrows A., Hubbard W. B., Lunine, J. I. \& Liebert, J. 2001, Reviews
726: %of Modern Physics, 73, 719
727: \bibitem[Burrows, Sudarsky \& Lunine (2003)]{bu03}Burrows, A., Sudarsky, D. \& Lunine, J. I. 2003, \apj, 596, 587
728: \bibitem[Cruikshank \& Binder (1968)]{cru68} Cruikshank, D. P. \& Binder, A. B. 1968,  
729: Communications of the Lunar and Planetary Laboratory 6, 275
730: %\bibitem[Cruikshank \& Binder (1969)]{cru69} Cruikshank, D. P. \& Binder, A. B. 1969,
731: % Astrophys. Space Sci. 3, 347
732: \bibitem[Cushing et al. (2006)] {cu06} Cushing, M. C. et al. 2006, \apj, 648, 614
733: \bibitem[Eggen (1998)] {eg98} Eggen O.J. 1998, \aj, 115, 2397
734: \bibitem[Elias et al. (2006)]{gnirs}
735: Elias, J. H., Rodgers, B., Joyce, R. R., Lazo, M., Doppmann, G., Winge, C., Rodriguez-Ardila, A., 2006,
736: in Ground-based and Airborne Instrumentation for Astronomy, eds. I. S. McLean and I. Masanori, SPIE 6269, 36
737: \bibitem[ESA (1997)]{hip} The Hipparcos and Tycho Catalogues, ESA 1997 (ESA SP-1200) (Noordwijk: ESA)
738: \bibitem[Fazio et al. (2004)] {faz04} Fazio, G. et al., 2004, \apjs, 154, 10
739: \bibitem[Feltzing \& Gustaffson (1998)] {fg98} Feltzing, S.  \& Gustaffson, B. 1998, \aaps, 129, 237
740: \bibitem[Fischer et al. (2003)] {fis03} Fischer, D. A., Butler, R. P., Marcy, G. W., Vogt, S. S. \&
741: Henry, G. W., 2003, ApJ, 590, 1081
742: \bibitem[Freedman  et al. (2007)] {fre07}  Freedman, R.F., Marley, M.S. \& 
743: Lodders, K., 2007, \apj, submitted
744: \bibitem[Geballe et al. (2001)] {geb01} Geballe, T. R., Saumon, D., Leggett, S. K., Knapp, G. R., Marley, M. S. \& Lodders, K. 2001, \apj, 556, 373
745: %\bibitem[Geballe et al. (2002)] {geb02} Geballe, T. R. et al. 2002, \apj, 564, 466
746: \bibitem[Golimowski et al. (2004)] {gol04} Golimowski D. A. et al. 2004, \aj, 127, 3516
747: \bibitem[Gray et al. (2003)] {gr03}
748: Gray, R. O., Corbally, C. J., Garrison, R. F., McFadden, M. T., \& Robinson, P. E. 2003, \aj, 126, 2048
749: \bibitem[Grevesse \& Sauval (1998)]{grev98} Grevesse, N., \& Sauval, A.~J., 1998, Space Science Reviews, 85, 161
750: \bibitem[Hewett et al. (2006)] {he06} Hewett, P. C., Warren, S. J., Leggett, S. K. \& Hodgkin, S. T. 2006, \mnras, 367, 454
751: \bibitem[Houck et al. (2004)] {hou04} Houck, J. et al. 2004, \apjs, 154, 18
752: \bibitem[Irwin et al. (1999)]{irw99} Irwin, P. G. J., Calcutt, S. B., Sihra, K., Taylor, F. W., Weir, A. L., Ballard, J. \& Johnston, W. B., 1999, JQSRT, 62, 193
753: \bibitem[Karkoschka (1994)]{kar94} Karkoschka, E. 1994, Icarus, 111, 174
754: \bibitem[Kim \& Geballe (2005)] {kg05} Kim, S. J. and Geballe, T. R. 2005, Icarus, 179, 449, 
755: \bibitem[Kirkpatrick (2005)]{kir05} Kirkpatrick, J. D. 2005, ARA\&A, 43, 195
756: \bibitem[Knapp et al. (2004)] {kna04}  Knapp, G. R. et al. 2004, \aj, 127, 3553
757: \bibitem[Lawrence et al. (2007)] {law07} Lawrence, A. et al. 2007, \mnras, in press
758: %\bibitem[Leggett et al. (2002)] {leg02} Leggett, S. K. et al. 2002, \apj, 564, 452
759: \bibitem[Leggett et al. (2007)] {leg07} Leggett, S. K., Saumon, D., Marley, M. S, Geballe, T. R., Golimowski, D. A., Stephens, D. S. \& Fan, X. 2007, \apj, 655, 1079 
760: \bibitem[Liebert \& Burgasser (2007)]{lb07} Liebert, J. \& Burgasser, A. J.  2007, \apj, 655, 522
761: \bibitem[Liu et al. (2007)]{liu07} Liu, M. C., Leggett, S. K., Chiu, K. C., 2007, \apj, 660, 1507 (LLC)
762: \bibitem[Lodders (1999)] {lodders99} Lodders, K. 1999, \apj, 519, 793
763: \bibitem[Lodders (2003)] {lodders03} Lodders, K. 2003, \apj, 591, 1220
764: \bibitem[Lodders \& Fegley (2002)] {lf02} Lodders, K. \& Fegley, B. Jr. 2002, Icarus, 155, 393
765: \bibitem[Lodieu et al. (2007)]{lod07} Lodieu, N. et al., 2007, \mnras, submitted 
766: \bibitem[Luhman et al. (2007)]{lu07} Luhman, K. L. et al., 2007, \apj, 654, 570 
767: \bibitem[Marley et al. (2002)]{mar02} Marley, M. S., Seager, S., Saumon, D., Lodders, K., 
768: Ackerman, A. S., Freedman, R. S. \& Fan, X., 2002, \apj, 568, 335
769: \bibitem[McLean et al. (2003)]{mcl03}
770: McLean, I. S., McGovern, M. R., Burgasser, A. J., Kirkpatrick, J. D., Prato, L. \& Kim, S. S. 2003, \apj, 596, 561
771: \bibitem[Monet et al. (1992)]{mon92} Monet, D. G., Dahn, C. C., Vrba, F. J., Harris, H. C., 
772: Pier, J. R., Luginbuhl, C. B. \& Ables, H. D., 1992, \aj, 103, 638
773: \bibitem[Mugrauer et al. (2006)] {mu06} Mugrauer, M., Seifahrt, A., Neuhaeuser, R., \& Mazeh, T., 2006,
774: \mnras, 373, L31
775: \bibitem[Nakajima et al. (1995)]{gl229} 
776: Nakajima, T., Oppenheimer, B. R., Kulkarni, S. R., Golimowski, D. A., Matthews, K. \& 
777: Durrance, S. T., 1995, Nature, 378, 463
778: \bibitem[Noll et al. (2000)]{nol00} Noll, K. S., Geballe, T. R., Leggett, S. K. \& Marley, M. S., 2000,
779: \apj, 541, L75
780: \bibitem[Patten et al. (2006)] {pat00} Patten, B. M., et al. 2006, \apj, 651, 502
781: \bibitem[Rayner, Cushing \& Vacca (2008)]{irtf} Rayner, J. T., Cushing, M. C. \& Vacca, W. D., 2008,
782: in preparation
783: \bibitem[Roellig et al. (2004)] {roe04} Roellig, T. L. et al., 2004, \apjs, 154, 418
784: \bibitem[Santos, Israelian \& Mayor (2004)]{san04}
785: Santos, N. C., Israelian, G., \& Mayor, M., 2004, \aap, 415, 1153
786: \bibitem[Santos et al. (2005)]{san05}
787: Santos, N. C., Israelian, G., \& Mayor, M., Bento, J. P., Almeida, P. C., Sousa, S. G., \& Ecuvillon, A., 
788: 2005, \aap, 437, 1127
789: \bibitem[Saumon et al. (2000)] {sau00} Saumon, D., Geballe, T. R., Leggett, S. K., Marley, M. S., 
790: Freedman, R. S., Lodders, K., Fegley, B. Jr. \& Sengupta, S. K., 2000, \apj, 541, 374
791: \bibitem[Saumon et al. (2003)] {sau03} Saumon, D., Marley, M. S., Lodders, K. \& Freedman, R. S., 2003,
792: Brown Dwarfs, Proceedings of IAU Symposium 211, ed. E. Martin, 
793: San Francisco: Astronomical Society of the Pacific, 345
794: \bibitem[Saumon et al. (2006)] {sau06} Saumon, D., Marley, M. S., Cushing, M. C., Leggett, S. K., Roellig, T. L., Lodders, K. \& Freedman, R. S. 2006, \apj, 647, 552
795: \bibitem[Saumon et al. (2007)] {sau07} Saumon, D. et al. 2007, \apj, 656, 1136
796: \bibitem[Sharp \& Burrows (2007)] {sb07} Sharp, C. M. \& Burrows, A. 2007, \apjs, 168, 140
797: \bibitem[Skrutskie et al. (2006)] {sk06} Skrutskie, M. F. et al. 2006, \aj, 131, 1163
798: \bibitem[Strong et al. (1993)]{str93} Strong, K., Taylor, F. W., Calcutt, S. B., Remedios, J. J.
799: \& Ballard, J. 1993, JQSRT, 50, 363
800: \bibitem[Tinney et al. (2003)] {ti03} Tinney, C. G.,Burgasser, A. J. \& Kirkpatrick, J. D. 2003, \aj, 126, 975
801: \bibitem[Tinney et al. (2005)] {ti05} Tinney, C. G., Burgasser, A. J., Kirkpatrick, J. D. 
802: \& McElwain, M. W. 2005, \aj, 130, 2326
803: %\bibitem[Tokunaga, Simons \& Vacca (2002)] {tol02} Tokunaga, A. T., Simons, D. A. \& Vacca, W. D. 2002, \pasp, 114, 180
804: \bibitem[Toon et al. (1989)]{too89} Toon, O. B., McKay, C. P., Ackerman, T. P. \& Santhanam, K., 1989, 
805: JGR, 9416287
806: \bibitem[Valenti \& Fischer (2005)]{vf05} Valenti, J. A. \& Fischer, D. A. 2005, \apjs, 159, 141
807: \bibitem[van Altena, Lee \& Hoffleit (1995)] {ypc}
808: van Altena, W. F., Lee, J. T. \& Hoffleit, E. D. 1995, The General Catalogue of Trigonometric Stellar Parallaxes (4th ed.; New Haven; Yale Univ. Obs.)
809: \bibitem[Vrba et al. (2004)] {vrb04} Vrba, F. J. et al. 2004, \aj, 127, 2948
810: \bibitem[Warren et al. (2007a)]{dr1} Warren, S. J.et al. 2007a, \mnras, 375,213
811: \bibitem[Warren et al. (2007b)]{0034} Warren, S. J.et al. 2007b, \mnras, submitted
812: %\bibitem[Werner et al. (2004)]{wer04} Werner, M., et al. 2004, ApJS, 154, 1 
813: \bibitem[York et al. (2000)]{yo00} York, D. G. et al. 2000, \aj, 120, 1579
814: \end{thebibliography}
815: %\clearpage
816: 
817: \begin{figure} \includegraphics[angle=-90,scale=.65]{f1.pdf} 
818: \caption{Observed R$\approx$400 CGS4 near-infrared spectra of Gl~570D 
819: (Geballe et al. 2001) and  R$\approx$150 Spex spectra of HD~3651B (Liu et al. 2007).  
820: Spectra are scaled to the peak flux at 1.27~$\mu$m.}
821: \end{figure}
822: \clearpage
823: 
824: \begin{figure} \includegraphics[angle=-90,scale=.65]{f2.pdf} 
825: \caption{Observed R$\approx$400 CGS4 near-infrared spectra of 2MASS0415-09
826: (Knapp et al. 2004) and 
827: R$\approx$150 Spex spectra of 2MASS0939-24 and 2MASS1114-26
828: (Burgasser, Burrows \& Kirkpatrick 2006b).  
829: Spectra are scaled to the peak flux at 1.27~$\mu$m. }
830: \end{figure}
831: \clearpage
832: 
833: \begin{figure} \includegraphics[angle=0,scale=.70]{f3.pdf} 
834: \caption{Synthetic spectra smoothed to R$=$400 and scaled to the peak flux at 1.27~$\mu$m.
835: %Note opacity sources are incomplete at $\lambda < 1.4 \mu$m; CH$_4$ is 
836: %known to be important at 1--1.5~$\mu$m, see Figure 2.
837: }
838: \end{figure}
839: \clearpage
840: 
841: 
842: \begin{figure} \includegraphics[angle=0,scale=.70]{f4.pdf} 
843: \caption{Observed R$\approx$400 near-infrared spectra of Gl~570D and 2MASS0415-09 (black),
844: and synthetic spectra scaled for the observed flux at Earth (red).
845: }
846: \end{figure}
847: \clearpage
848: 
849: \begin{figure} \includegraphics[angle=0,scale=.70]{f5.pdf} 
850: \caption{Observed medium-resolution R$=$1700--2000 near-infrared spectra of Gl~570D and HD~3651B, 
851: from the NIRSPEC brown dwarf spectroscopic survey (McLean et al. 2003) and this work. 
852: Spectra are scaled to the peak flux in each panel. 
853: See Figure 1 for line identifications.}
854: \end{figure}
855: \clearpage
856: 
857: 
858: \begin{figure} \includegraphics[angle=-90,scale=.60]{f6.pdf}
859: \caption{Observed low-resolution near-infrared spectrum of HD~3651B (black curve), compared to synthetic 
860: spectra scaled for the observed flux at Earth for various temperatures, gravities and metallicities.}
861: \end{figure}
862: \clearpage
863: 
864: 
865: \begin{figure} \includegraphics[angle=-90,scale=.60]{f7.pdf}
866: \caption{Medium-resolution spectra of Gl~570D (McLean et al. 2003, black) and HD~3651B (this work, red)
867: compared to synthetic spectra for [$T_{\rm eff}$, log~$g$, [m/H]] models [800, 5.0, 0.0] (green) and [800, 5.5, +0.3] (blue). The spectra are scaled to the peak flux and offset for clarity.
868: }
869: \end{figure}
870: \clearpage
871: 
872: 
873: \begin{figure} \includegraphics[angle=-90,scale=.60]{f8.pdf}
874: \caption{Modelled (for $T_{\rm eff}=825$~K)
875: and observed absolute IRAC magnitudes, symbols are described in the legend. 
876: Observations are taken from Patten et al. (2006) and Luhman et al. (2007); errors in the photometry
877: are 2--5\% for both dwarfs for [3.55] and [4.49], and 7--14\% at [5.73] and [7.87].
878: Observational datapoints are shifted to shorter wavelengths for clarity.}
879: \end{figure}
880: \clearpage
881: 
882: 
883: 
884: \begin{figure} \includegraphics[angle=-90,scale=.60]{f9.pdf} 
885: \caption{ Spectra of the T8 dwarf 2MASS0415-09 (black, Knapp et al. 2004), Jupiter (magenta, 
886: Rayner, Cushing \& Vacca 2008) and Saturn (cyan, Kim 
887: \& Geballe 2005), together with room temperature transmission laboratory data for CH$_4$ (yellow, Cruikshank 
888: \& Binder 1968) and NH$_3$ (green, Irwin et al. 1999). } \end{figure} \clearpage
889: 
890: 
891: \begin{figure} \includegraphics[angle=-90,scale=.60]{f10.pdf}
892: \caption{
893: Ammonia features in the near-infrared spectra of very cool brown dwarfs.
894:          The black curves show cloudless synthetic spectra with $T_{\rm eff}=700\,$K 
895:          (upper curve) and 600$\,$K (lower curve), $\log g=5$, and including the
896:          NH$_3$ opacity of Irwin et al. (1999) for $\lambda < 1.9\,\mu$m and the
897:          NH$_3$ opacity computed from a line list for $\lambda > 1.9\,\mu$m.  The upper 
898:          curve (red) shows the 700$\,$K model computed without the Irwin et al. opacity.
899:          A 600$\,$K model computed with the Irwin et al. 
900:          opacity but with NH$_3$ depleted by non-equilibrium chemistry 
901:          (with the eddy diffusion coefficient 
902:          $K_{zz}=10^4\,$cm$^2$/s) is shown by the lower curve (green). All spectra for a
903:          given $T_{\rm eff}$ are computed with the same $(T,P)$ structure. The spectra 
904:          have been smoothed to a resolving power of R=400.}
905: \end{figure}
906: \clearpage
907: 
908: 
909: \voffset 1.5truein
910: \begin{deluxetable}{llrrrccccccccc}
911: \rotate
912: \tabletypesize{\scriptsize}
913: \tablecaption{Published Physical Parameters for  T7.5--T8 Dwarfs}
914: \tablewidth{625pt}
915: \tablehead{
916: \colhead{Name} &  \colhead{Spectral} & \colhead{Parallax}  & \colhead{$V_{\rm tan}$(error)} &   
917: \colhead{log($L_{\rm bol}/$} & &\multicolumn{4}{c}{Parameters from Luminosity} & &
918: \multicolumn{2}{c}{BBK Parameters\tablenotemark{b}} 
919: &  \colhead{References\tablenotemark{c}} \\
920: \cline{7-10}\cline{12-13} 
921: &\colhead{Type\tablenotemark{a}} & \colhead{(error)(mas)}& \colhead{(km~s$^{-1}$)} & 
922:  \colhead{$L_{\odot}$)(error)} & &
923: \colhead{$T_{\rm eff}$~K} & \colhead{log~$g$} & \colhead{Age Gyr} & \colhead{Mass M$_J$} & &
924: \colhead{$T_{\rm eff}$~K} & \colhead{log~$g$} & 
925: \\
926: }
927: \startdata
928: 2MASS J04151954-0935066         & T8   & 174.34(2.76) & 61.4(1.0) & -5.67(0.02) & &
929: 725--775 & 5.00--5.37 & 3--10 & 33--58 && 740--760  & 4.9--5.0 & 1, 2, 3  \\
930: 2MASS J09393548-2448279         & T8   & \nodata      & 46(8)   & \nodata & &
931: \nodata & \nodata &   \nodata & \nodata & & $\lesssim$ 700 &   \nodata &  4 \\
932: 2MASS J11145133-2618235         & T7.5 & \nodata      & 109(20)   & \nodata & &
933: \nodata & \nodata &   \nodata & \nodata & & $\lesssim$ 700 &   \nodata & 4 \\
934: 2MASS J12171110-0311131         & T7.5 & 93.2(2.06)   & 53.5(1.2) & -5.31(0.03) & &
935: 850-950 & 4.80--5.42 & 1--10 & 25--66 & & 860--880 & 4.7--4.9 & 2, 3, 5, 6  \\
936: Gl 570 D & T7.5 & 170.16(1.45) & 56.4(0.9)   & -5.54(0.01) & &
937: 800-820 & 5.09--5.23 & 3--5 & 38--47 && \nodata & \nodata & 7 , 8, 9, 10\\
938: HD 3651B & T7.5 & 90.03(0.72) & 31.0(1.2)    & -5.60(0.05) & &
939: 780--840 & 5.1--5.5 & 3--12 & 63--72 & & 760--820 &  4.7--5.3 & 11, 12, 13, 14 \\
940: \enddata
941: 
942: 
943: \tablenotetext{a}{~Spectral types are based on the near-infrared classification scheme for T dwarfs 
944: by Burgasser et al.\ (2006a) and have an uncertainty of 0.5 of a subclass. }
945: \tablenotetext{b}{~Derived from the near-infrared spectral ratio technique of 
946: Burgasser, Burrows \& Kirkpatrick (2006b)}
947: \tablenotetext{c}{
948: (1) Burgasser et al. (2002); (2) Vrba et al. (2004); (3) Saumon et al. (2007);
949: (4) Tinney et al. (2005),  
950: tangential velocities are estimated from distances derived from $M_J$:spectral type;
951: (5) Burgasser et al. (1999); (6) Tinney et al. (2003); 
952: (7) Burgasser et al. (2000);  (8) ESA (1997) 
953: (9) van Altena, Lee \& Hoffleit (1995) (10) Saumon et al. (2006) 
954: (11) Mugrauer et al. (2006) (12) Luhman et al. (2007) (13) Burgasser (2007) (14) Liu et al. (2007)
955: }
956: 
957: \end{deluxetable}
958: \clearpage
959: 
960: \begin{deluxetable}{llccccc}
961: \tabletypesize{\scriptsize}
962: \tablecaption{Derived Properties of HD~3651B, 2MASS J09393548-2448279 and 2MASS J11145133-2618235}
963: \tablewidth{400pt}
964: \tablehead{
965: \colhead{Name} &  \colhead{Spectral} & \colhead{Metallicity}  & 
966: \colhead{$T_{\rm eff}$~K} & \colhead{log~$g$\tablenotemark{c}} & \colhead{Age Gyr\tablenotemark{c}} & \colhead{Mass M$_J$\tablenotemark{c}} \\
967: &\colhead{Type\tablenotemark{a}} & \colhead{[m/H]\tablenotemark{b}}&  &  & & \\ 
968: }
969: \startdata
970: HD~3651B &   T7.5 & $+$0.2 & 820--830 & 5.4--5.5 & 8--12 & 60--70 \\
971: 2MASS J09393548-2448279 & T8 & $-$0.3 & 725-775 & 5.3--5.45 & 10--12 & 50--65 \\
972: 2MASS J11145133-2618235 & T7.5 & $-$0.3 & 725-775 & 5.0--5.3 & 3--8 & 30--50 \\
973: \enddata
974: \tablenotetext{a}{Uncertainty in spectral type is 0.5 of a subclass.}
975: \tablenotetext{b}{Uncertainty in metallicity is estimated to be 0.1 dex.}
976: \tablenotetext{c}{Gravity, age and mass are correlated such that lower gravity 
977: implies younger age and lower mass, and higher gravity implies an older and more massive dwarf.}
978: \end{deluxetable}
979: 
980: 
981: 
982: 
983: \end{document}
984: