0705.3922/ms.tex
1: 
2: \documentclass{emulateapj}
3: %\documentclass[preprint]{aastex}
4: 
5: %\documentclass[12pt,preprint]{aastex}
6: 
7: 
8: \slugcomment{Accepted to Ap.J.}
9: 
10: \shortauthors{Hubeny, \& Burrows}
11: \shorttitle{Non-chemical equilibrium models}
12: 
13: 
14: \begin{document}
15: 
16: \title{A Systematic Study of Departures from Chemical Equilibrium 
17: in the Atmospheres of Substellar Mass Objects
18: }
19: 
20: \author{Ivan Hubeny\altaffilmark{1} \& Adam Burrows\altaffilmark{1}}
21: %\author{Adam Burrows\altaffilmark{1}}
22: 
23: \altaffiltext{1}{Department of Astronomy and Steward Observatory, The University
24: of Arizona, Tucson, AZ 85721}
25: 
26: \begin{abstract}
27: We present a systematic study of the spectral consequences
28: of departures from chemical equilibrium in the atmospheres of L and T 
29: dwarfs, and for even cooler dwarfs.  The temperature/pressure profiles of the non-equilibrium
30: models are 
31: fully consistent with the non-equilibrium chemistry. Our grid of non-equilibrium models 
32: includes spectra for effective temperatures from 200 K to 1800 K, three surface gravities,  
33: four possible values of the coefficient of eddy diffusion in the radiative zone,
34: and three different CO/CH$_4$ chemical reaction prescriptions. We also provide
35: clear and cloudy model variants. We find, in keeping with previous studies,
36: that there are essentially only two spectral regions where the effects of departures
37: from chemical equilibrium can influence the predicted spectrum.  
38: These are in the M ($\sim$4$-$5 $\mu$m) and N (8$-$14 $\mu$m)
39: bands due to CO and NH$_3$, respectively.  The overabundance of CO
40: translates into flux suppressions of at most $\sim$40\% between
41: effective temperatures of 600 and 1800 K.  The effect is largest around
42: $T_{\rm eff} \approx 1100$ K.  The underabundance of ammonia 
43: translates into flux enhancements of no more than
44: $\sim$20\% for the $T_{\rm eff}$ range from 300 to 1800 K, with the largest
45: effects at the lowest values of $T_{\rm eff}$.  The magnitude
46: of the departure from chemical equilibrium increases with decreasing gravity,
47: with increasing eddy diffusion coefficient, and with decreasing speed of
48: the CO/CH$_4$ reaction.  Though these effects are modest, they lead to
49: better fits with the measured T dwarf spectra.  Furthermore, the suppression in 
50: the M band due to non-equilibrium enhancements in the CO abundance disappears
51: below $\sim$500 K, and is only partial above $\sim$500 K, preserving   
52: the M band flux as a useful diagnostic of cool atmospheres and maintaining
53: its importance for searches for the cooler brown dwarfs beyond the Ts.  
54: \end{abstract}
55: 
56: \keywords{stars: abundances---stars: atmospheres---stars: individual (Gliese 570D)
57: ---stars: low-mass, brown dwrafs}
58: 
59: \section{Introduction}
60: \label{intro}
61: 
62: It has been long realized that departures from the local chemical equilibrium (LCE)
63: may play a significant role in the atmospheres of the substellar-mass objects
64: (SMO), that is of giant planets and brown dwarfs. Essentially, some chemical 
65: reactions occurring in these atmospheres may be very slow, so that vertical 
66: transport via convective motions or eddy diffusion
67: can lead to departures from LCE. The mechanism was first suggested to operate in
68: the Jovian planets of the solar system (Prinn \& Barshay 1977, Barshay \& Lewis 1978,
69: Fegley \& Prinn 1985, Yung et al. 1988, and Fegley \& Lodders 1994).
70: 
71: Fegley \& Lodders (1996) suggested that the same mechanism can also operate in the 
72: atmospheres of brown dwarfs. Noll, Geballe, \& Marley (1997) studied in more
73: detail the non-equilibrium enhancement of CO in the atmosphere of Gl 229B.
74: Subsequently, Griffith \& Yelle (1999) and Saumon et al. (2000) provided a more
75: detailed analysis. The latter paper also suggested that a non-equilibrium
76: depletion of NH$_3$ may be important. Saumon et al. (2003) demonstrated the 
77: effects of non-equilibrium chemistry for a range of effective temperatures 
78: ($T_{\rm eff}$ between 800 and 1600 K). Recently, Saumon et al. (2006, 2007)
79: showed that the non-equilibrium effects in the carbon and nitrogen
80: chemistry have significant effects on the predicted spectrum of brown dwarfs, 
81: and in turn fit the observed spectrum of the T7.5 dwarf Gliese 570D (Saumon 
82: et al. 2006) and 2MASS 0415 (Saumon et al. 2007) much better than
83: models based on local chemical equilibrium.  Finally, Leggett et al. (2007)
84: extended the Saumon et al. (2006) models to a larger range of parameters, and
85: presented a comparison of model predictions with observed infrared colors.
86: 
87: 
88: In this paper, we extend and generalize the Saumon et al. and Leggett et al.
89: treatments by computing a large grid of solar-metallicity
90: models for effective temperatures $T_{\rm eff}$
91: covering the whole domain of T and L dwarfs (700 - 1800 K), as well as
92: for cooler dwarfs (200 $-$ 700 K), for three values of surface gravity
93: ($\log g = 4.5,\, 5.0,\, 5.5$ cm s$^{-2}$), for several values of the coefficient
94: of eddy diffusion (see \S 3.1), and for three different reaction rate
95: timescales for the carbon reactions (see \S 3.1). We 
96: compute self-consistent models in the sense that the atmospheric structure
97: ($T/P$ profile) is computed taking into account departures from chemical
98: equilibrium. We have computed two grids of models, one cloudless, and one
99: including clouds composed of silicate condensates.
100: 
101: In \S\ref{model}, we outline our procedures for computing model atmospheres, while
102: allowing for departures from chemical equilibrium. In \S\ref{results}, we present  
103: and analyze the predicted spectra computed with and without chemical
104: equilibrium for L and T dwarfs.  We explore the dependence on $T_{\rm eff}$,
105: gravity, and eddy mixing coefficient in the radiative zone and compare clear and 
106: representative cloudy models.  We also discuss the effect of non-equilibrium chemistry
107: for objects cooler than the coolest T dwarfs down to effective temperatures of 200 K.  
108: Then, in \S\ref{gl570d}, we briefly present a comparison
109: between theory and observation in the mid-IR for Gliese 570D. 
110: Finally, in \S\ref{conclusions}, we summarize our conclusions.
111: 
112: 
113: 
114: \section{Modeling Procedures}
115: \label{model}
116: 
117: \subsection{Non-equilibrium Carbon and Nitrogen Chemistry}
118: 
119: The chemistry of carbon and nitrogen in the atmospheres of SMOs is
120: essentially described by the net reactions
121: %
122: \begin{equation}
123: \label{co}
124: {\rm CO + 3H_2 \longleftrightarrow CH_4 + H_2O}\, ,
125: \end{equation}
126: %
127: for carbon, and
128: %
129: \begin{equation}
130: \label{n2}
131: {\rm N_2 + 3H_2 \longleftrightarrow 2 NH_3}\, ,
132: \end{equation}
133: %
134: for nitrogen.
135: %
136: Because of the strong C=O and N$\equiv$N bonds, the reactions (\ref{co}) and
137: (\ref{n2}) proceed much faster from right to left than from left
138: to right. For instance, for carbon the reaction
139: in which CO is converted to CH$_4$ is very slow, and, therefore, CO
140: can be vertically transported in the atmosphere to the upper and cooler
141: regions by convective motions or eddy diffusion. The net result is an
142: overabundance of CO and N$_2$ and an underabundance of NH$_3$ and CH$_4$ 
143: in the upper atmosphere. Since more oxygen atoms are now tied in CO,
144: the mixing ratio of water is also reduced because of conservation of 
145: the total number of oxygen atoms.
146: 
147: The mixing (vertical
148: transport) timescale, $t_{\rm mix}$, was suggested by Griffith \& Yelle (1999) 
149: to be parameterized by
150: %
151: \begin{equation}
152: \label{tmix1}
153: t_{\rm mix} = H^2/K_{zz}\, ,
154: \end{equation}
155: %
156: where $H$ is the pressure scale height, and $K_{zz}$ is the coefficient of
157: eddy diffusion. Equation (\ref{tmix1}) applies in the radiative (convectively
158: stable) zone. In the convection zone, the mixing time is given by
159: (Saumon et al. 2006):
160: %
161: \begin{equation}
162: t_{\rm mix} = 3 H_c/v_c\, ,
163: \end{equation}
164: %
165: where $H_c$ is the convective mixing length (adopted here to be equal to $H$),
166: and $v_c$ is the convective velocity.
167: We stress that while the mixing time is well defined in the convection zone,
168: it can only be roughly parameterized in the radiative zone. As discussed
169: e.g. by Saumon et al. (2006, 2007), the reasonable values of $K_{zz}$ are in the
170: range $10^2$ -- $10^6$ cm$^2$ s$^{-1}$.
171: 
172: The chemical reaction timescale, $t_{\rm chem}$, generally decreases rapidly 
173: with increasing pressure and temperature.  
174: In the atmospheric regions where $t_{\rm mix} \geq t_{\rm chem}$, the species
175: acquire their equilibrium mixing ratios. This occurs at temperatures and
176: pressures larger than that at the depth where $t_{\rm mix} = t_{\rm chem}$.
177: At smaller pressures, the species acquire mixing ratios equal to those at
178: the $t_{\rm mix} = t_{\rm chem}$ layer, as first suggested for the case of
179: the CO/CH$_4$ reaction by Prinn \& Barshay (1977). 
180: 
181: The non-equilibrium mixing ratios thus depend sensitively on the
182: chemical reaction timescales, which unfortunately are poorly known.
183: For the N$_2$ $\rightarrow$ NH$_3$ conversion, we adopt the timescale
184: given by Lodders \& Fegley (2002):
185: %
186: \begin{eqnarray}
187: t_{\rm chem}\equiv t_{\rm N_2} = \frac{1}{\kappa_{\rm N_2} N({\rm H}_2)}\, \nonumber\\
188: \kappa_{\rm N_2} = 8.54\times 10^{-8}\, \exp\left(-\frac{81515}{T}\right)\, ,
189: \end{eqnarray}
190: %
191: where $\kappa_{\rm N_2}$ is the rate constant for the N$_2$ $\rightarrow$ NH$_3$
192: reaction, and $N({\rm H}_2)$ is the number density of H$_2$.
193: 
194: The reaction timescales for the CO $\rightarrow$ CH$_4$ conversion are
195: uncertain because the correct chemical pathway is still unknown.
196: We will, therefore, use three different timescales, suggested earlier in
197: the literature:
198: 
199: \noindent i) the ``slow'' chemical timescale after Prinn \& Barshay (1977):
200: %
201: \begin{eqnarray}
202: t_{\rm chem}\equiv t_{\rm CO} = \frac{N({\rm CO})}{\kappa_{\rm CO} N({\rm H}_2)
203: \, N({\rm H_2CO})}\, \nonumber\\
204: \kappa_{\rm CO} = 2.3\times 10^{-10}\, \exp\left(-\frac{36200}{T}\right)\, ,
205: \end{eqnarray}
206: %
207: where $N({\rm M})$ is the number density of species M.
208: 
209: \noindent ii) and iii) the ``fast'' chemical timescale after Yung et al. (1988)
210: or Dunning et al. (1984). In both cases the timescale is given by 
211: %
212: \begin{equation}
213: t_{\rm CO} = \frac{N({\rm CO})}{\kappa_7 N({\rm H})
214: \, N({\rm H_2CO})}\,  ,
215: \end{equation}
216: %
217: where the rate coefficient differs in each reference. We use tabulated
218: values presented in Yung et al. (1988 - their Table III), and interpolate
219: to current temperatures. In this paper, we call the Dunning et al.
220: timescale ``fast1,'' and the Yung et al. timescale ``fast2.'' As we shall
221: show in \S 3, these two fast timescales produce very similar results.
222: 
223: 
224: \subsection{Atmosphere and Spectrum Modeling}
225: 
226: We use the updated code {\sc CoolTLUSTY}, described
227: in Sudarsky, Burrows, \& Hubeny (2003), Hubeny, Burrows, \& Sudarsky (2003), 
228: and Burrows, Sudarsky, \& Hubeny (2006 $-$ BSH), which is a variant of the
229: universal atmospheric code {\sc TLUSTY} (Hubeny 1988; Hubeny \& Lanz 1995). 
230: {\sc CoolTLUSTY} solves self-consistently a set of radiative transfer
231: equations for selected frequency points (typically 5000 points
232: logarithmically spaced between $7\times 10^{14}$ Hz and $10^{12}$ Hz),
233: and the equation of radiative+convective equilibrium. The atmosphere is
234: assumed to be in hydrostatic equilibrium. The opacities are not computed on the
235: fly; instead, they are interpolated from pre-calculated opacity tables for the
236: current values of temperature and density. Our opacity database is described
237: in detail by Sharp \& Burrows (2007). The atmospheric structure is
238: computed iteratively, applying the Newton-Raphson method
239: (first applied in the case of stellar atmospheres by Auer \& Mihalas 1969
240: with the name ``complete linearization'').
241: The discretized and linearized equations are re-organized in the so-called
242: Rybicki scheme (Mihalas 1978), which renders the computer time
243: linearly proportional to the number of frequencies (and not proportional
244: to the cube of the number of frequencies as in the case of the original complete
245: linearization). This allows us to avoid using the accelerated lambda iteration
246: (ALI) scheme to treat the radiative transfer equation, which leads to 
247: faster and more stable convergence, without compromising on computer time.
248: 
249: We have found that the most stable procedure to compute non-LCE models
250: is to perform several iterations (typically 6) of the linearization
251: method assuming chemical equilibrium (using the algorithm of
252: Burrows \& Sharp 1999 to determine the equilibrium abundances of all
253: species), and then to switch the non-LCE algorithm. 
254: In each subsequent iteration, one calculates the intersection point 
255: where the mixing time for the current $T/P$ profile equals the reaction time
256: (computed again for the current $T/P$ profile); the number densities
257: of CO, CH$_4$, H$_2$O, and NH$_3$ are set to constant values equal to
258: those found at the intersection points for pressures lower than the
259: pressure at the intersection. The opacities in these regions are modified 
260: accordingly. The next iteration of the linearization scheme is then performed
261: with the new opacities. The overall process usually converges in a few
262: additional iterations.
263: 
264: When clouds are taken into account, we employ the scheme described in
265: BSH. In the case of L and T dwarfs, we
266: represent individual cloud decks of many condenstate species by a single 
267: extended cloud deck of one representative condenstate. As in BSH, we take
268: forsterite as a representative condensate, and assume the cloud deck
269: to extend between the intersection of the condensation curve of
270: forsterite with the current $T/P$ profile, and the pressure where the local
271: temperature equals 2300 K (which schematically represents the condensation
272: curve of the most refractory condenstate), and with exponential decreases
273: of particle density on
274: both sides of the cloud. In the terminology of BSH, this is an E-type
275: cloud. In all the present models, we assume the modal particle size 100 $\mu$m. 
276: 
277: 
278: % \subsection{Treatment of Clouds}
279: 
280: \section{Results}
281: \label{results}
282: 
283: \subsection{L and T Dwarfs}
284: \label{res_lt}
285: 
286: 
287: Our effort has generated an extensive grid of
288: non-equilibrium model spectra for effective temperatures
289: from 700 K to 1800 K (in steps of 100 K), for three values of surface
290: gravity, $g = 10^{4.5},\, 10^{5.0}$, and $10^{5.5}$ cm s$^{-2}$, for four
291: values of the coefficient of eddy diffusion in the radiative zone
292: $K_{zz}= 10^2, 10^4, 10^6$, and $10^8$ cm$^2$ s$^{-1}$, and for three
293: different CO/CH$_4$ reaction time prescriptions\footnote{Electronic versions
294: of
295: these spectral models can be found online at {\tt
296: http://zenith.as.arizona.edu/}$\widetilde{\,\,\,}${\tt burrows}.}.
297: We have also provided clear and cloudy model variants. The latter assumes
298: an E-type forsterite cloud with a modal particle size of 100 $\mu$m.
299: In the present paper, we consider only solar-composition models,
300: where the solar elemental abundances are from Asplund, Grevesse, \&
301: Sauval (2006).
302: 
303: We will first discuss composition profiles and their trends with
304: effective temperature, surface gravity, diffusion coefficient,
305: and the CO/CH$_4$ reaction speed.
306: %
307: Figure \ref{fig1} displays the composition profiles of the
308: non-equilibrium species, CO, CH$_4$, NH$_3$, and H$_2$O, as a function
309: of the local temperature in the atmosphere, for a representative model
310: with $T_{\rm eff} = 900$ K, $g=10^{5.5}$ cm s$^{-2}$, and $K_{zz}= 10^4$
311: cm$^2$ s$^{-1}$.  The full lines show the non-equilibrium mole fractions or,
312: equivalently, the number fractions (expressed as a fraction of the total 
313: number of particles), while the dashed lines show the equilibrium
314: number fractions. For this model, the non-equilibrium effects in 
315: water and methane are negligible. The non-equilibrium number fraction of ammonia
316: is smaller than the equilibrium one for $T<1200$ K, as first shown by 
317: Saumon et al. (2006). The non-equilibrium mixing ratio of CO is significantly
318: larger than the equilibrium one, which in turn leads to a significant
319: strengthening of the 4.7-$\mu$m feature (and, to a lesser extent, to other
320: CO features at 3.4 and 2.8 $\mu$m) in non-equilibrium models.
321: Moreover, the non-equilibrium mixing ratio of CO is larger for the slower
322: CO/CH$_4$ reaction, because the intersection of the $t_{\rm mix}(T)$
323: curve with the $t_{\rm CO}(T)$ curve occurs deeper in the atmosphere
324: for higher temperature, at which the equilibrium CO mixing ratio is higher
325: (see also Fig. \ref{fig5}). Figure \ref{fig1} is analogous to Fig. 3 of
326: Saumon et al. (2006).
327: 
328: To depict the trend of composition profiles with effective temperature, we
329: display in Fig. \ref{fig2} composition profiles as in Fig. \ref{fig1}, but
330: for $T_{\rm eff}/100\, {\rm K} = 8,10,12,14,16$, and 18. To avoid cluttering 
331: the figure, the models are not labeled by the values of $T_{\rm eff}$. 
332: Instead, the models are distinguished by the fact that 
333: the higher the effective temperature is, the longer the curves 
334: continue toward high temperatures (i.e., to the right side of the plot).
335: First, we stress that the equilibrium abundance patterns follow from
336: the basic chemistry (Burrows \& Sharp 1999; Lodders \& Fegley 2002),
337: and are well understood.
338: Briefly, the same local temperature is reached at lower pressure for
339: models with higher effective temperatures (see Fig. \ref{fig7} for an
340: explicit demonstration). Due to Le Chatelier's principle,  
341: at a given local temperature the abundance of CH$_4$ 
342: decreases and the abundance of CO increases with increasing $T_{\rm eff}$. 
343: Analogously, the abundance of NH$_3$ decreases, and the abundance of N$_2$
344: (not shown here) increases with increasing $T_{\rm eff}$. These effects
345: are clearly shown in Fig. \ref{fig2}.
346: 
347: For all models displayed here, the non-equilibrium effects in water 
348: and methane are negligible. The non-equilibrium abundance of CO is 
349: always larger than the equilibrium abundance for 
350: $T \stackrel{\sim}{<} 1000$ K. For very low $T_{\rm eff}$, although 
351: the non-equilibrium abundance of CO is significantly larger than 
352: the equilibrium abundance, it is still small in absolute value, and, 
353: therefore, does not have much of an effect on the model structure 
354: and predicted spectra. For the highest $T_{\rm eff}$, the equilibrium
355: abundance profile of CO becomes flatter, and, consequently,
356: the differences between the non-equilibrium and equilibrium abundance 
357: profiles of CO decrease, and so do the differences in the predicted 
358: spectra. This is demonstrated in Fig. \ref{fig8}. For ammonia, departures 
359: from equilibrium become smaller for higher effective temperatures, because 
360: the equilibrium ammonia composition profile becomes flat. Therefore, fixing 
361: the ammonia number fraction at the intersection of the $t_{\rm mix}$ 
362: and the $t_{\rm NH_3}$ curves does not lead to a significant difference 
363: in the composition. Therefore, the non-equilibrium effects in ammonia, and, 
364: consequently, in the predicted flux in the 8 - 14 $\mu$m region, decrease 
365: with increasing $T_{\rm eff}$. This also is demonstrated in Fig. \ref{fig8}. 
366: Moreover, because of its diminished number fraction, ammonia becomes
367: a less important source of opacity as the temperature increases.
368: 
369: In Fig. \ref{fig3}, we show the trend of the composition
370: profiles with the surface gravity, for $T_{\rm eff} = 900$~K, 
371: $K_{zz} = 10^4$ cm$^2$ s$^{-1}$, and the fast CO/CH$_4$ reaction. 
372: The intersection of the 
373: $t_{\rm mix}$ and the $t_{\rm CO}$ curves occurs at essentially the same
374: temperature for all three surface gravities (see Fig. \ref{fig4}).
375: However, at a given temperature, the model with the lowest gravity exhibits
376: the lowest pressure, and, therefore, the CO mixing ratio at the intersection is
377: higher for the lower-gravity models. Consequently, the effects of departures
378: from chemical equilibrium are greater for the low-gravity models.
379: 
380: Figure \ref{fig4} displays the mixing time and the reaction times for the
381: models displayed in Fig.~\ref{fig3}. Notice the abrupt change of the mixing
382: time at $T\approx 1800$ K; this corresponds to the onset of convection,
383: where the mixing times are about five orders of magnitude faster.
384: %
385: To explain the behavior of models with different coefficients of eddy
386: diffusion, and with different reaction timescales for the CO/CH$_4$ reaction,
387: we display in Fig. \ref{fig5} the mixing times for four different diffusion
388: coefficients, $K_{zz} = 10^2, 10^4, 10^6$, and $10^8$ cm$^2$ s$^{-1}$.
389: By increasing the coefficient of eddy diffusion, $K_{zz}$, the mixing time 
390: decreases and, consequently, the intersection of the chemical reaction timescale
391: curves and the mixing time curve occurs at higher temperatures. Therefore,
392: the non-equilibrium mixing ratio of CO is higher. For the N$_2$/NH$_3$ reaction,
393: and for $K_{zz} > 10^4$ cm$^2$ s$^{-1}$, the chemical time curve 
394: intersects the mixing time curve in the convection zone at very high
395: temperature ($T \approx 2100$ K).
396: 
397: In Fig. \ref{fig6}, we display trends of the chemical reaction time with
398: the effective temperature. The N$_2$/NH$_3$ reaction timescale is relatively
399: insensitive to the model effective temperature, while for the CO/CH$_4$ reaction,
400: the intersection of the curves moves to higher temperatures and pressures for
401: higher effective temperatures. It would seem at first sight that the departures
402: from chemical equilibrium would increase with increasing $T_{\rm eff}$.
403: However, as we have shown in Fig. \ref{fig2}, the abundance profile of CO becomes 
404: flatter and, thus, the departures from chemical equilibrium decrease instead.
405: 
406: In Fig. \ref{fig7} we display the temperature/pressure profiles
407: for selected models of our grid. The differences in the local temperature between
408: the equilibrium and non-equilibrium models are not large, but nevertheless
409: appreciable.
410: 
411: 
412: We now turn to emergent spectra.
413: %
414: In Fig. \ref{fig8}, we display the emergent spectra for all effective temperatures,
415: and for one representative value of $g = 10^{5.5}$ cm s$^{-2}$, for $K_{zz}=10^4$
416: cm$^2$ s$^{-1}$, and for the ``slow'' reaction rate of the CO/CH$_4$ reaction
417: of Prinn \& Barshay (1977). 
418: There are essentially only two wavelength regions where
419: the departures from chemical equilibrium influence the predicted spectrum,
420: the region around 4.7 microns (due to non-equilibrium abundances of CO), 
421: and a wide region between 8 and 14 microns, due to non-equilibrium
422: abundances of NH$_3$. The CO feature has its peak for effective
423: temperatures between 800 K and 1200 K; the effect generally decreases
424: for hotter models and essentially disappears at 1800 K. 
425: The non-equilibrium effect of ammonia already disappears for 
426: $T_{\rm eff} \stackrel{>}{\sim} 1200$ K. The behavior of the associated spectra
427: is readily understood upon examination of the composition 
428: profiles displayed in Figs. \ref{fig1} - \ref{fig3} (see also the related
429: discussion).
430: %
431: Figure \ref{fig8} is a synoptic display of the overall model trends,
432: but is not meant to communicate details of the predicted spectra. Therefore,
433: we later present in Figs. \ref{fig11}, \ref{fig12}, and \ref{fig13} enlarged plots
434: of the two non-equilibrium regions, the CO region between 3 and 5 $\mu$m,
435: and the ammonia region between 8 and 14 $\mu$m.
436: In these and subsequent figures that display the
437: predicted spectra, we smooth the the computed spectrum by a 10-point boxcar
438: average, which corresponds to a resolution of about 1/100.
439: The corresponding models with clouds are displayed in Fig. \ref{fig9}
440: and behave analogously, thus demonstrating that the presence of clouds
441: does not lead to any qualitative differences in the magnitude of departures
442: from chemical equilibrium. 
443: 
444: Figure \ref{fig10} shows the sensitivity of the predicted spectrum to
445: the surface gravity. The non-equilibrium 
446: effects generally increase with decreasing gravity, as was explained in
447: the discussion of Fig. \ref{fig3}. 
448: Figure \ref{fig11} presents a more detailed view of the CO non-equilibrium
449: region between 3 and 5 $\mu$m, for one particular model with $T_{\rm eff}=900$ K
450: and $g=10^{5.5}$ g cm$^{-2}$. We display both the LCE model and
451: four non-equilibrium models with $K_{zz}= 10^2, \, 10^4, \, 10^6$, and
452: $10^8$ cm$^2$ s$^{-1}$. All models are for the ``slow'' reaction rate
453: of Prinn \& Barshay (1977) for the CO/CH$_4$ reaction. An analogous plot
454: for the models with the ``fast2'' reaction of Yung et al. (1988) is displayed
455: in Fig.~\ref{fig12}. The magnitude of departures from LCE increases significantly
456: with increasing $K_{zz}$. This is because the mixing time decreases with
457: increasing $K_{zz}$ and, thus, the intersection of the $t_{\rm mix}$ and the
458: $t_{\rm CO}$ curves occurs deeper in the atmosphere, where the mixing ratio
459: of CO is larger. This was already demonstrated in Fig. \ref{fig5}.
460: %
461: As Fig. \ref{fig12} shows, the departures from LCE are smaller for the
462: faster CO/CH$_4$ reaction rate, as demonstrated previously in Fig. \ref{fig1}.
463: 
464: In Fig. \ref{fig13}, we display a more detailed view of the
465: ammonia region between 8 and 14 $\mu$m, for the same models as displayed in
466: Fig. \ref{fig12}. The departures from chemical equilibrium are essentially
467: the same as predicted by Saumon et al. (2006, 2007).
468: Unlike the case of CO, the abundance of ammonia and,
469: therefore, the flux in the region between 8 and 14 $\mu$m, is relatively
470: insensitive to the value of $K_{zz}$. This is because the abundance profile
471: of ammonia is rather flat (see Figs. \ref{fig1} - \ref{fig3}), and, therefore,
472: the depth of intersection of the $t_{\rm N_2}$ and the $t_{\rm mix}$ curves
473: does not matter very much. The departures of 
474: ammonia from LCE are largely insensitive to the speed of the CO/CH$_4$
475: reaction, as expected. We do not show this in the figure because
476: the corresponding plots are essentially identical. 
477: The predicted departures of NH$_3$ from LCE seem to be quite
478: robust. 
479: 
480: To better display departures from chemical equilibrium, we plot
481: in Fig. \ref{fig14} the ratio of the integrated flux between
482: 4.4 $\mu$m and 5 $\mu$m of the non-equilibrium and equilibrium models, 
483: as a function of effective temperature. The upper panel displays the
484: dependence of the non-equilibrium ratio for $g=10^5$ cm s$^{-2}$, for
485: the ``slow'' CO/CH$_4$ reaction, and the ``fast2'' CO/CH$_4$ reaction. 
486: Each set of models is computed for three values of $K_{zz}$. The curves 
487: are labelled by the value of $\log_{10} K_{zz}$ in cm$^2$ s$^{-1}$.
488: As discussed above, assuming the ``slow'' CO/CH$_4$ reaction increases
489: the departures from LCE significantly, reaching about 30\% at 
490: $T_{\rm eff}=1100$ K (at $K_{zz}=10^4$ cm$^2$ s$^{-1}$), 
491: and decreasing toward higher temperatures.
492: For the ``slow2'' reaction, the maximum effect is around $T_{\rm eff}=1300$ K,
493: where it reaches about  10\% (again, at $K_{zz}=10^4$ cm$^2$ s$^{-1}$).
494: %
495: The lower panel displays the dependence of the non-equilibrium ratio
496: on surface gravity for two sets of models. In all models, $K_{zz}=10^4$
497: cm$^2$ s$^{-1}$. The upper set of models (diamonds connected by dashed
498: lines) are for the ``fast2'' CO/CH$_4$ reaction, while the lower set
499: of models (stars connected by dotted lines) are for the ``slow''
500: CO/CH$_4$ reaction. The curves are labelled by the value of $\log_{10} g$  
501: in cm s$^{-2}$. As discussed above, the lower-gravity models exhibit
502: larger departures from equilibrium for temperature below 1300 K. However,
503: the trend is reversed for $T_{\rm eff} > 1300$ K, where the lower gravity
504: models are closer to equilibrium. 
505: 
506: \subsection{Beyond the T Dwarfs}
507: \label{beyond}
508: 
509: In this section, we briefly explore the effects on spectra of departures 
510: from chemical equilibrium for brown dwarfs cooler than the coolest known
511: T dwarfs (see also Burrows, Sudarsky, \& Lunine 2003).  To this end,
512: we extend the grid of models described in \S\ref{res_lt}, and
513: compute models for $T_{\rm eff} = 200,\, 300,\, 400,\, 500,$
514: and 600 K.  For the models below 500 K, we include the effect 
515: of a water cloud with a modal particle size of 100 $\mu$m
516: and a super-saturation parameter of 1.0\%.
517: The other basic parameters are the same as in \S \ref{res_lt}.
518: %
519: In Fig. \ref{fig16}, we display the mixing time and the reaction times
520: for three models with $T_{\rm eff} = 200,\, 400,$ and 600 K,
521: for $g=10^5$ cm s$^{-2}$, $K_{zz}=10^4$ cm$^2$ s$^{-1}$, and the
522: ``fast2'' CO/CH$_4$ reaction. In all models, the intersection 
523: of the mixing time and the reaction times occurs in the convection zone.
524: Therefore, the actual value of the diffusion coefficient $K_{zz}$, which 
525: influences the mixing time only at temperatures and pressures lower
526: than those at the intersection, is inconsequential. 
527: (In other words, $K_{zz}$ influences the $T_{\rm mix}$ curve only on the
528: left of the intersection point, which is already determined by the local
529: value of the convective velocity.)
530: Note that the intersections move to lower temperature with decreasing $T_{\rm eff}$.
531: The somewhat wiggly character of the $t_{\rm mix}$ curve in convection zone
532: follows form the fact the the convective velocity is given by (e.g., Mihalas 1978)
533: %
534: \begin{equation}
535: v = (gQH/8)^{1/2} (\nabla-\nabla_E)(l/H)^2\, ,
536: \end{equation}
537: %
538: where $g$ is the surface gravity, $H$ the pressure scale height,
539: $Q=(\partial\ln\rho/\partial\ln T)_P$, $l$ is the mixing length, 
540: and $(\nabla-\nabla_E)$ is a difference between the true temperature
541: gradient, $\nabla=\partial\ln T/\partial\ln P$, and the gradient
542: of the convective elements. This is computed by considering the
543: efficiency of convective transport. In the standard mixing-length
544: approach, the efficiencies are computed differently for Rosseland optical
545: depth smaller and larger than unity. The wiggly behavior corresponds to 
546: a transition between Rosseland optical depth larger than and smaller than 1.
547: In any case, the uncertainty in the exact value of the convective
548: velocity does not influence the non-equilibrium abundances of CO, CH$_4$,
549: and NH$_3$ in a significant way.
550: 
551: Figure \ref{fig17} depicts the abundance profiles for the models displayed
552: in Fig. \ref{fig16}. The non-equilibrium number fraction of CO sharply decreases mixing time
553: with decreasing $T_{\rm eff}$.  This is a consequence of the fact that
554: while the intersection points of the $t_{\rm mix}$ and the
555: $t_{\rm CO}$ curves occur at similar temperatures around 1100 K,
556: these intersection points are reached at much higher pressures for lower $T_{\rm eff}$.
557: Consequently, the equilibrium number fraction of CO at the intersection point
558: is lower. For ammonia, the equilibrium abundance profiles become flatter for  
559: deceasing $T_{\rm eff}$ and, therefore, the differences between the
560: equilibrium and non-equilibrium number fractions are smaller.
561: 
562: In Fig. \ref{fig18}, we depict the predicted flux for the three models
563: displayed in Figs. \ref{fig16} and \ref{fig17}. The hottest model, at
564: $T_{\rm eff}=600$ K, exhibits the same features as late T dwarfs (see \S\ref{res_lt}),
565: namely the non-equilibrium flux is higher in the 8$-$14 $\mu$m region (by about
566: 15-20\%) and smaller in the the 4$-$5 $\mu$m region (on average, by about 10\%).
567: In the cooler model at $T_{\rm eff}=400$ K, the departures in the 8$-$14 $\mu$m region
568: due to non-equilibrium chemistry of ammonia are slightly smaller than in the 600-K model.
569: However, the differences in the 4$-$5 $\mu$m region essentially disappear because
570: the non-equilibrium mixing ratio of CO, though much higher than the
571: equilibrium one, is already quite small.  Therefore, CO contributes but 
572: little to the opacity in this region. Finally, for the coolest model 
573: at $T_{\rm eff}=200$ K, the non-equilibrium effects almost disappear 
574: in both the 4$-$5 $\mu$m and 8$-$14 $\mu$m regions.
575: 
576: To diagnose the trends described above, we display in Fig. \ref{fig19} the 
577: temperature/pressure profiles for the same models (both 
578: non-equilibrium and equilibrium) depicted in Figs. \ref{fig16} 
579: - \ref{fig18}. In Fig. \ref{fig20}, we display the ``radiation formation
580: temperature'' as a function of wavelength for the two extreme models with 
581: $T_{\rm eff} = 200$ and 600 K. The radiation formation temperature is defined as a local
582: temperature in the atmosphere where the monochromatic optical depth is 2/3.
583: This quantity indicates the position/layer from which a photon of a given wavelength
584: effectively decouples from the atmosphere. 
585: 
586: 
587: \section{Model Comparison with Gliese 570D Observations}
588: \label{gl570d}
589: 
590: In Fig. \ref{fig21}, we display the observed spectrum of Gliese 570D, 
591: kindly supplied to us in digital form by M. Cushing (Cushing et al. 2006). 
592: This is one of the latest known T dwarfs, and is the object studied by Saumon et al.
593: (2006). We also display two theoretical models computed for  
594: $T_{\rm eff}=800$ K and $g=10^{5}$ cm s$^{-2}$, namely an LCE and a non-LCE model.
595: The non-equilibrium model is computed with $K_{zz}=10^4$ cm$^2$ s$^{-1}$,
596: and for the ``fast2'' CO/CH$_4$ reaction rate, but, as demonstrated in Fig. \ref{fig13},
597: this predicted flux essentially coincides with the flux computed for all other
598: values of $K_{zz}$ and CO/CH$_4$ reaction time prescriptions. The agreement between
599: the observed and predicted spectrum for non-LCE model is excellent in the
600: region between 8 and 14 $\mu$m, while the fit is much worse for the LCE model.
601: We thus confirm the results of Saumon et al.\footnote{Note that the predicted 
602: fluxes for both our models and those of Saumon et al. (2006) are too 
603: low between 7 and 8 $\mu$m, for both LCE and
604: non-LCE models. The reason for this is unclear, but may be related to
605: deficiencies in the extant water opacity databases.}.
606: In a future work, we will show comparisons between our non-equilibrium
607: spectral models and observations in the near- and mid-IR of other late T dwarfs.
608: 
609: 
610: 
611: 
612: \section{Discussion and Conclusions}
613: \label{conclusions}
614: 
615: In this paper, we have performed a systematic study of the spectral consequences
616: of departures from chemical equilibrium in carbon and nitrogen chemistry
617: in the atmospheres of L dwarfs, T dwarfs, and substellar-mass 
618: objects down to 200 K. Unlike previous investigations, the
619: temperature/pressure profiles of the non-equilibrium models are fully consistent
620: with the non-equilibrium chemistry.  Such models reveal that failure to 
621: calculate the non-equilibrium atmospheric profiles in a self-consistent fashion 
622: can result in errors of as much as $\sim$50$-$100 K in the atmospheric temperatures
623: at a given pressure.  Our effort has generated an extensive grid of  
624: non-equilibrium model spectra for effective temperatures
625: from 700 K to 1800 K (in steps of 100 K), for three values of surface gravity,
626: $g = 10^{4.5},\, 10^{5.0}$, and $10^{5.5}$ cm s$^{-2}$, for four values of the 
627: coefficient of eddy diffusion in the radiative zone
628: $K_{zz}= 10^2, 10^4, 10^6$, and $10^8$ cm$^2$ s$^{-1}$, and for three different
629: CO/CH$_4$ reaction time prescriptions. We have also provided 
630: clear and cloudy model variants.  We find, in keeping with previous studies, 
631: that there are essentially only two regions where the effects of departures 
632: from chemical equilibrium significantly influence the
633: predicted spectrum.  These are in the M ($\sim$4$-$5 $\mu$m) and N (8$-$14 $\mu$m)
634: bands due to CO and NH$_3$, respectively.  The overabundance of CO can be very large,
635: but translates into flux suppressions in the M band of at most $\sim$40\% between 
636: effective temperatures of 800 and 1800 K.  The effect is largest around
637: $T_{\rm eff} \approx 1100$ K.  The underabundance of ammonia can be 
638: significant, but translates into flux enhancements in the N band of no more than
639: $\sim$20\% for the $T_{\rm eff}$ range from 700 to 1800 K, with the largest
640: effects at the lowest values of $T_{\rm eff}$.  The magnitude
641: of the departure from chemical equilibrium increases with decreasing gravity,
642: with increasing eddy diffusion coefficient, and with decreasing speed of
643: the CO/CH$_4$ reaction.  Though these effects are modest, they lead to
644: better fits with the measured T dwarf spectra (see also Saumon et al. 2006,2007).
645: 
646: In addition, we have calculated an exploratory sequence of non-equilibrium models 
647: down to 200 K, including the possible effects of water cloud formation. 
648: These models show that in the region ``beyond the T dwarfs" (Burrows, 
649: Sudarsky, \& Lunine 2003) the suppression effect of non-equilibrium CO 
650: in the M band diminishes, then disappears, with decreasing $T_{\rm eff}$.
651: 
652: Hence, despite the modest suppression in the M band due to 
653: non-equilibrium CO chemistry for $T_{\rm eff}$s above $\sim$500 K,  
654: non-equilibrium effects do not compromise the use of the M band flux 
655: as a useful diagnostic of cool atmospheres and as one of their 
656: most prominent features.  At the higher $T_{\rm eff}$s above $\sim$500 K, this is 
657: due in part to the saturation of the augmented CO absorption, but importantly 
658: it is due to the fact that only about half of the M band overlaps with the 4.7-$\mu$m CO 
659: absorption feature.  At lower $T_{\rm eff}$s below $\sim$400 K, non-equilibrium effects in the  
660: M and N bands all but disappear. Overall, the prominence and importance of 
661: the M band flux for searches for the coolest dwarfs and for brown 
662: dwarfs (``Y" dwarfs ?) beyond the T dwarf range is not substantially affected
663: by non-equilibrium chemistry. 
664:   
665: 
666: 
667: \acknowledgments
668: 
669: We thank M. Cushing, J. Budaj, and W. Hubbard for fruitful discussions 
670: and help during the course of this work. We also thank the referee, 
671: K. Lodders, for her detailed comments. This study was supported in part 
672: by NASA grants NAG5-10760, NNG05GG05G, NNG04GL22G and NNX07AG80G,
673: and through the NASA Astrobiology Institute under Cooperative
674: Agreement No. CAN-02-OSS-02 issued through the Office of Space
675: Science.
676: 
677: 
678: 
679: \begin{thebibliography}{}
680: 
681: \bibitem[Asplund et al. 2006]{asp06} Asplund, M., Grevesse, N., \& Sauval, A. J.
682: 2006, Commun. in Asteroseismology, 147, 76
683: 
684: \bibitem[Auer \& Mihalas 1969]{aumi69} Auer, L. H., \& Mihalas, D. 1969, \apj,
685: 158, 641
686: 
687: \bibitem[Barshay \& Lewis 1978]{bl78} Barshay, S. S., \& Lewis, J. S. 1978, Icarus, 
688: 33, 593
689: 
690: \bibitem[Burrows \& Sharp 1999]{bs99} Burrows, A., \& Sharp, C. M. 1999, 
691: \apj, 512, 843
692: 
693: \bibitem[Burrows et al. 2006]{bsh06} Burrows, A., Sudarsky, D., \&  Hubeny, I.
694: 2006, \apj, 640, 1063
695: 
696: \bibitem[Burrows, Sudarsky, \& Lunine 2003]{bsl} Burrows, A., 
697: Sudarsky, D., \& Lunine, J.I. 2003, \apj, 596, 587
698: 
699: \bibitem[Cushing et al. 2007]{cush07} Cushing, M. C. et al. 2006, \apj, 648, 614
700: 
701: \bibitem[Dunning et al. 1984]{dun84} Dunning, T. H., Harding, L. B., \& Wagner, A. F.
702: 1984, in Theory and Modeling for Material Designs, Texas A \& M University
703: 
704: \bibitem[Fegley \& Prinn 1985]{fegpri78} Fegley, B., Jr., \& Prinn, R. G. 1985,
705: \apj, 299, 1067
706: 
707: \bibitem[Fegley \& Lodders 1996]{feglod96} Fegley, B., Jr., \& Lodders, K. 1996,
708: \apj, 472, L37
709: 
710: \bibitem[Griffith \& Yelle 1999]{grye99} Griffith, C. A., \& Yelle, R. V. 1999,
711: \apj, 519, L85
712: 
713: \bibitem[Hubeny 1988]{hub88} Hubeny, I. 1988, Comput. Phys. Commun., 52, 103
714: 
715: \bibitem[Hubeny et al. 2003]{hbs03} Hubeny, I., Burrows, A., \& Sudarsky, D. 2003,
716: \apj, 594, 1011
717: 
718: \bibitem[Hubeny \& Lanz 1995]{hl95} Hubeny, I., \& Lanz, T. 1995, \apj, 439, 875
719: 
720: \bibitem[Leggett et al. 2007]{leg07} Leggett, S. K., Saumon, D., Marley, M. S.,
721: Geballe, T. R., Golimowski, D. A., Stephens, D., \& Fan, X. 2007, \apj, 655, 1079
722: 
723: \bibitem[Lodders \& Fegley 2002]{lf02} Lodders, K., \& Fegley, B., Jr. 2002,
724: Icarus, 155, 393
725: 
726: \bibitem[Mihalas 1978]{mi78} Mihalas, D. 1978, Stellar Atmospheres, 2nd ed.,
727: Freeman, San Francisco
728: 
729: \bibitem[Noll t.al. 1997]{ngm97} Noll, K. S., Geballe, T. R., \& Marley, M.S. 1997,
730: \apj, 489, L87
731: 
732: \bibitem[Prinn \& Barshay 1977]{pb77} Prinn, G. G., \& Barshay, S. S. 1977,
733: Science, 198, 1031
734: 
735: \bibitem[Saumon et al. 2000]{saum00} Saumon, D., Geballe, T. R., Leggett, S. K.,
736: Marley, M. S., Freedman, R. S., Lodders, K., Fegley, B. Jr., \& Sengupta, S. K. 2000,
737: \apj, 541, 374
738: 
739: \bibitem[Saumon et al. 2003]{saum03} Saumon, D.,  Marley, M. S., Lodders, K., \& 
740: Freedman, R. S. 2003, in ``Brown Dwarfs,'' IAU Symposium 211, 345
741: 
742: 
743: \bibitem[Saumon et al. 2006]{saum06} Saumon, D., Marley, M. S., Cushing, M. C.,
744: Leggett, S. K., Roellig, T. L., Lodders, K., \& Freedman, R. S. 2006,
745: \apj, 647, 552
746: 
747: \bibitem[Saumon et al. 2007]{saum07} Saumon, D., Marley, M. S., Leggett, S.K.,
748: Geballe, T.R., Stephens, D., Golimowski, D.A., Cushing, M. C., Fan, X., Rayner, J.T.,
749: Lodders, K., \& Freedman, R.S. 2007, \apj, 656, 1136
750: 
751: \bibitem[Sharp \& Burrows]{sb07} Sharp, C. M., \& Burrows, A. 2007, \apjs, 168, 140
752: 
753: \bibitem[Sudarsky et al. 2003]{sbh03}  Sudarsky, D., Burrows, A., \&  Hubeny, I.
754: 2003, \apj, 588, 1121
755: 
756: \bibitem[Yung et al. 1988]{yun88} Yung, Y. L., Drew, W. A., Pinto, J. P., \&
757: Friedl, R. R. 1988, Icarus, 73, 516
758: 
759: \end{thebibliography}
760: 
761: \newpage
762: 
763: \begin{figure}
764: \plotone{f1.eps}
765: \caption{
766: A comparison of the equilibrium (dashed lines) and non-equilibrium
767: (solid lines) abundances of water (black), methane (yellow),
768: ammonia (red), and CO (blue), for a representative model
769: with $T_{\rm eff} = 900$ K, $g=10^{5.5}$ cm s$^{-2}$, and 
770: $K_{zz}= 10^4$ cm$^2$ s$^{-1}$.
771: In this case, water and methane remain essentially at their equilibrium
772: abundances, while ammonia becomes underabundant at the upper layers
773: (leading to a lower opacity and, thus, higher predicted flux in the
774: 8-14 $\mu$m region for non-equilibrium models). The plots also show that
775: CO exhibits larger departures from equilibrium with the slow reaction.
776: See Fig. \ref{fig5} and the text for a discussion.
777: }
778: \label{fig1}
779: \end{figure}
780: 
781: 
782: \clearpage 
783: 
784: \begin{figure}
785: \plotone{f2.eps}
786: \caption{
787: Similar to Fig. \ref{fig1}, but with a range of effective temperatures to display
788: the trends with $T_{\rm eff}$. The color scheme is
789: the same as in Fig. \ref{fig1}. The association of the individual lines with
790: effective temperatures is made by noting that in the plot the lines corresponding 
791: to progressively lower effective temperatures end at progressively lower
792: local temperatures. (In other words, the longer the lines go to
793: the right, the higher the effective temperature.) In all cases displayed
794: here, water and methane remain at their equilibrium abundances.
795: The degree of departure from equilibrium for ammonia decreases with increasing
796: effective temperature. 
797: }
798: \label{fig2}
799: \end{figure}
800: 
801: 
802: 
803: \clearpage 
804: 
805: 
806: \begin{figure}
807: \plotone{f3.eps}
808: \caption{
809: Similar to Fig. \ref{fig2}, but showing the trend with changing surface gravity.
810: This figure demonstrates that the non-equilibrium abundance of CO increases 
811: with decreasing gravity. See explanation in the text.
812: }
813: \label{fig3}
814: \end{figure}
815: 
816: 
817: \clearpage 
818: 
819: 
820: 
821: \begin{figure}
822: \plotone{f4.eps}
823: \caption{
824: A plot of the mixing time (solid lines), the reaction time
825: for the CO/CH$_4$ (dashed lines), and the reaction time for
826: the NH$_3$/N$_2$ reaction (dash-dotted lines), for models with
827: $T_{\rm eff} = 900$~K, $K_{zz}=10^4$ cm$^2$ s$^{-1}$, and $\log_{10} g$ (cm s$^{-2}$)
828: = 5.5 (black), 5.0 (blue), and 4.5 (red). The abrupt change of the 
829: mixing-time curves represents the boundary of the convection zone.
830: }
831: \label{fig4}
832: \end{figure}
833: 
834: 
835: \clearpage 
836: 
837: \begin{figure}
838: \plotone{f5.eps}
839: \caption{
840: Similar to Fig. \ref{fig4}, but depicting
841: the mixing times (solid lines) corresponding to (from top to bottom)
842: the four values of the diffusion coefficient, 
843: $K_{zz}=10^2, \, 10^4,\, 10^6,\, 10^8$ cm$^2$ s$^{-1}$,
844: for models with $T_{\rm eff}=900$ K, and $g=10^{5.5}$ cm s$^{-2}$.
845: Also shown are the reaction times
846: for CO/CH$_4$ (dashed lines) chemistry, for the three different treatments
847: of the reaction speed, ``slow,'' ``fast1,'' and ``fast2'' 
848: (the curves are labeled accordingly), and the reaction 
849: times for the NH$_3$/N$_2$ reaction (dot-dashed lines).
850: }
851: \label{fig5}
852: \end{figure}
853: 
854: 
855: \clearpage 
856: 
857: 
858: %Fig.9---
859: \begin{figure}
860: \plotone{f6.eps}
861: \caption{
862: Similar to Fig. \ref{fig4}, but depicting
863: the mixing times (solid lines), and the reaction times
864: for CO/CH$_4$ (dashed lines) for the ``fast2'' reaction rate, and the reaction 
865: time for the NH$_3$/N$_2$ reaction (dot-dashed line),
866: for a set of models with 
867: $g=10^{5.5}$ cm s$^{-2}$,  $K_{zz}=10^4$ cm$^2$ s$^{-1}$, and
868: $T_{\rm eff}/100$ K = 9, 12, 15, 18. The reaction times for
869: CO/CH$_4$ are labeled accordingly.
870: }
871: \label{fig6}
872: \end{figure}
873: 
874: 
875: \clearpage 
876: 
877: 
878: %Fig.10---
879: \begin{figure}
880: \plotone{f7.eps}
881: \caption{
882: A comparison of temperature-pressure profiles for equilibrium (dashed lines)
883: and non-equilibrium (solid lines) models of $T_{\rm eff}=800,\, 1000,\, 1200,\, 1400, \, 1600$, 
884: and 1800 K (from bottom to top). $g$ is equal to $10^{5.5}$ cm s$^{-2}$.
885: }
886: \label{fig7}
887: \end{figure}
888: 
889: 
890: 
891: \clearpage 
892: 
893: 
894: 
895: \begin{figure}
896: %\plotone{f8.eps}
897: \includegraphics[width=15cm,height=17cm,keepaspectratio=false]{f8.eps}
898: \caption{
899: Synoptic comparison of model atmosphere spectra computed with 
900: chemical equilibrium (black lines), and with departures from 
901: equilibrium (red lines), for a representative value of $g = 10^{5.5}$ cm s$^{-2}$,
902: and $K_{zz} = 10^4$ cm$^2$ s$^{-1}$,
903: for a ``slow'' CO/CH$_4$ reaction, and for effective
904: temperatures between 700 and 1800 K, covering most of the T and 
905: L spectral types. There are essentially two wavelength regions where
906: the departure from chemical equilibrium influences the predicted spectrum:
907: the region around 4.7 $\mu$m (due to non-equilibrium abundances of CO), 
908: and a wide region between 8 and 14 $\mu$m, due to non-equilibrium
909: abundances of NH$_3$. The CO feature has its peak for effective
910: temperatures between 800 K and 1200 K; the effect generally decreases
911: for hotter models and essentially disappears at 1800 K.
912: The figure is similar to that presented by Saumon et al. (2003).
913: }
914: \label{fig8}
915: \end{figure}
916: 
917: 
918: \clearpage 
919: 
920: 
921: \begin{figure}
922: %\plotone{f9.eps}
923: \includegraphics[width=15cm,height=17cm,keepaspectratio=false]{f9.eps}
924: \caption{
925: The same as Fig. \ref{fig8}; the only difference is that the models are computed
926: with clouds. We assume a cloud composed of a collection of magnesium 
927: silicate componds represented by a single species, in this case
928: forsterite, with an extension of the cloud layer toward
929: higher pressures, as suggested by BSH (E-type cloud model in
930: their notation). We assume a modal particle size of 100 $\mu$m.
931: The presence of clouds leads to differences in the predicted spectrum,
932: in particular for wavelengths below 2.5 $\mu$m, but otherwise the
933: effect of departures from chemical equilibrium is essentially 
934: the same as for the cloudless models.
935: }
936: \label{fig9}
937: \end{figure}
938: 
939: \clearpage 
940: 
941: 
942: %Fig.3---
943: \begin{figure}
944: \plotone{f10.eps}
945: \caption{
946: A comparison of predicted spectra for a model with $T_{\rm eff}=900$ K,
947: $K_{zz} = 10^4$ cm$^2$ s$^{-1}$, and three different surface gravities. 
948: We display the equilibrium models (black lines), and non-equilibrium models
949: (red lines) assuming the ``slow'' CO/CH$_4$ reaction rate of Prinn \&
950: Barshay (1977).
951: The plot also clearly shows that the effects of departures from
952: chemical equilibrium are more pronounced for lower gravities.
953: See Fig. \ref{fig1} and the text for explanations.
954: }
955: \label{fig10}
956: \end{figure}
957: 
958: 
959: \clearpage 
960: 
961: 
962: 
963: \begin{figure}
964: \plotone{f11.eps}
965: \caption{A more detailed view of the CO non-equilibrium
966: region between 3 and 5 $\mu$m, for models with $T_{\rm eff}=900$ K
967: and $g=10^{5.5}$ cm s$^{-2}$. Displayed are the equilibrium model (black), as well
968: four non-equilibrium models with $K_{zz}= 10^2$ (blue), $10^4$ (red), $10^6$
969: (yellow), and $10^8$ (green) cm$^2$ s$^{-1}$. 
970: All models are for the ``slow'' reaction rate for the CO/CH$_4$ reaction 
971: of Prinn \& Barshay (1977). 
972: }
973: \label{fig11}
974: \end{figure}
975: 
976: \clearpage 
977: 
978: 
979: \begin{figure}
980: \plotone{f12.eps}
981: \caption{Similar to Fig. \ref{fig11}, but for the ``fast2'' CO/CH$_4$ 
982: reaction rate of Yung et al. (1988). Departures from equilibrium are now
983: smaller, as explained in the text. 
984: }
985: \label{fig12}
986: \end{figure}
987: 
988: 
989: 
990: \begin{figure}
991: \plotone{f13.eps}
992: \caption{A more detailed view of the
993: ammonia region between 8 and 14 $\mu$m, for the same models as displayed in
994: Fig. \ref{fig12}. The curves for all non-equilibrium models
995: essentially coincide, except for the highest value of $K_{zz}=10^8$
996: cm$^2$ s$^{-1}$.
997: The flux in the region between 8 and 14 $\mu$m is relatively
998: insensitive to the value of $K_{zz}$.
999: }
1000: \label{fig13}
1001: \end{figure}
1002: 
1003: 
1004: \clearpage 
1005: 
1006: 
1007: 
1008: \begin{figure}
1009: \plotone{f14.eps}
1010: \caption{Departures from chemical equilibrium in the 4.7-$\mu$m CO region.
1011: The plot displays the ratio of the integrated flux in the region between
1012: 4.4 $\mu$m and 5 $\mu$m of the non-equilibrium and equilibrium models, 
1013: as a function of effective temperature. The upper panel displays the
1014: dependence of the non-equilibrium ratio for $g=10^5$ cm s$^{-2}$, for
1015: the ``slow'' CO/CH$_4$ reaction (the upper set of diamonds connected
1016: by dashed lines) and the ``fast2'' CO/CH$_4$ reaction (the lower set
1017: of stars connected by dotted lines); each set is shown for three values of
1018: $K_{zz}$. The curves are labelled by the value of $\log_{10} K_{zz}$
1019: in cm$^2$ s$^{-1}$. The curves for $K_{zz}=10^6$ cm$^2$ s$^{-1}$ for
1020: the ``fast2'' CO/CH$_4$ reaction and for $K_{zz}=10^2$ cm$^2$ s$^{-1}$ for
1021: the ``slow'' CO/CH$_4$ reaction essentially coincide.
1022: The lower panel displays the dependence of the non-equilibrium ratio
1023: on surface gravity for two sets of models. In all models, $K_{zz}=10^4$
1024: cm$^2$ s$^{-1}$. The upper set of models (diamonds connected by dashed
1025: lines) are for the ``fast2'' CO/CH$_4$ reaction, while the lower set
1026: of models (stars connected by dotted lines) are for the ``slow''
1027: CO/CH$_4$ reaction. The curves are labelled by the value of $\log_{10} g$  
1028: in cm s$^{-2}$.
1029: }
1030: \label{fig14}
1031: \end{figure}
1032: 
1033: 
1034: \clearpage 
1035: 
1036: 
1037: 
1038: \begin{figure}
1039: \plotone{f15.eps}
1040: \caption{Similar to Fig. \ref{fig14}, but for the ammonia region between
1041: 9 $\mu$m and 14 $\mu$m. Since the predicted spectra are insensitive to the
1042: speed of the CO/CH$_4$ reaction, and only weakly sensitive to $K_{zz}$,
1043: we display only models with $K_{zz}=10^4$, and the ``fast2'' CO/CH$_4$
1044: reaction. The curves are labelled by the value of $\log_{10} g$ in
1045: cm s$^{-2}$.
1046: }
1047: \label{fig15}
1048: \end{figure}
1049: 
1050: 
1051: 
1052: \clearpage 
1053: 
1054: 
1055: 
1056: \begin{figure}
1057: \plotone{f16.eps}
1058: \caption{A plot of the mixing time (solid lines), the reaction time
1059: for the CO/CH$_4$ (dashed lines), and the reaction time for
1060: the NH$_3$/N$_2$ reaction (dash-dotted lines), for models with
1061: $g=10^5$ cm s$^{-2}$
1062: $K_{zz}=10^4$ cm$^2$ s$^{-1}$, and $T_{\rm eff} = 600$ K (black lines),
1063: 400 K (blue lines), and 200 K (red lines).
1064: }
1065: \label{fig16}
1066: \end{figure}
1067: 
1068: 
1069: \clearpage 
1070: 
1071: 
1072: 
1073: \begin{figure}
1074: \plotone{f17.eps}
1075: \caption{A comparison of the equilibrium (dashed lines) and non-equilibrium
1076: (solid lines) abundances of water (black), methane (yellow),
1077: ammonia (red), and CO (blue), for the same models as displayed in
1078: Fig. \ref{fig16}. Thin lines depict models for $T_{\rm eff} = 600$ K,
1079: thicker lines for for $T_{\rm eff} = 400$ K, and the thickest lines for  
1080: $T_{\rm eff} = 200$ K. The curves for CO are labeled by the value of $T_{\rm eff}$
1081: in K.
1082: }
1083: \label{fig17}
1084: \end{figure}
1085: 
1086: 
1087: \clearpage 
1088: 
1089: 
1090: 
1091: \begin{figure}
1092: \plotone{f18.eps}
1093: \caption{Predicted flux for models with equilibrium (black) and
1094: non-equilibrium (red) models with $g=10^5$ cm s$^{-2}$
1095: $K_{zz}=10^4$ cm$^2$ s$^{-1}$, and $T_{\rm eff} = 600$ K (upper panel),
1096: 400 K (middle panel), and 200 K (lower panel).
1097: }
1098: \label{fig18}
1099: \end{figure}
1100: 
1101: 
1102: \clearpage 
1103: 
1104: 
1105: \begin{figure}
1106: \plotone{f19.eps}
1107: \caption{A comparison of temperature-pressure profiles for equilibrium (dashed lines)
1108: and non-equilibrium (solid lines) models of $T_{\rm eff}=600,\, 400$,
1109: and 200 K (from top to bottom). $g$ is equal to $10^{5}$ cm s$^{-2}$.
1110: }
1111: \label{fig19}
1112: \end{figure}
1113: 
1114: 
1115: \clearpage 
1116: 
1117: 
1118: 
1119: \begin{figure}
1120: \plotone{f20.eps}
1121: \caption{Radiation formation temperature as a function of wavelength for models
1122: with $T_{\rm eff}=600$ K (upper curve) and 200 K (lower curve).
1123: $g$ is equal to $10^{5}$ cm s$^{-2}$.
1124: }
1125: \label{fig20}
1126: \end{figure}
1127: 
1128: 
1129: 
1130: \clearpage 
1131: 
1132: 
1133: 
1134: 
1135: \begin{figure}
1136: \plotone{f21.eps}
1137: \caption{Comparison of observed and predicted spectra of Gl 570D.
1138: Thick black histogram displays the observed flux, the red and blue
1139: lines are the predicted spectra for non-equilibrium (red) and equilibrium
1140: (blue) models, computed for $T_{\rm eff}=800$ K and $g=10^{5}$ cm s$^{-2}$.
1141: The non-equilibrium model displayed is computed with $K_{zz}=10^4$ cm$^2$ s$^{-1}$,
1142: and for fast CO/CH$_4$ reaction rate, but, as demonstrated in Fig. \ref{fig13},
1143: this predicted flux essentially coincides with the flux computed for all other
1144: values of $K_{zz}$ and CO/CH$_4$ reaction time prescriptions.
1145: }
1146: \label{fig21}
1147: \end{figure}
1148: 
1149: 
1150: \end{document}
1151: