1: \documentclass[12pt,preprint]{aastex}
2: %\usepackage[T1]{fontenc}
3: %\setcounter{tocdepth}{3}
4: %\documentclass{aastex}
5: %\usepackage{emulateapj5}
6: %\usepackage{apjfonts}
7:
8: \begin{document}
9:
10: \title{Line and Mean Opacities for Ultracool Dwarfs and Extrasolar
11: Planets}
12:
13:
14: \author{Richard S. Freedman\altaffilmark{1} \& Mark S. Marley}
15:
16: \affil{NASA Ames Research Center}
17:
18: \affil{Mail Stop 245-3}
19:
20: \affil{Moffett Field, CA 94035-1000}
21:
22:
23: \email{freedman@darkstar.arc.nasa.gov}
24:
25:
26: \author{Katharina Lodders}
27:
28:
29: \affil{Planetary Chemistry Laboratory, Department of Earth and Planetary
30: Sciences, Washington University, Campus Box 1169, Saint Louis, MO
31: 63130-4899}
32:
33:
34: \email{lodders@wustl.edu}
35:
36: \altaffiltext{1}{SETI Institute, Mountain View, CA}
37:
38: \begin{abstract}
39: Opacities and chemical abundance data are crucial ingredients of
40: ultracool dwarf and extrasolar giant planet atmosphere models. We
41: report here on the detailed sources of molecular opacity data employed by our
42: group for this application. We also present tables
43: of Rosseland and Planck mean opacities which are of use in some studies
44: of the atmospheres, interiors, and evolution of
45: planets and brown dwarfs.
46: For the tables presented here we have
47: included the opacities of important atomic and molecular species, including the alkali elements,
48: pressure induced absorption by hydrogen, and other significant opacity
49: sources but neglect opacity from condensates. We report for each species how we have assembled
50: molecular line data from a combination of public databases, laboratory data
51: that is not yet in the public databases, and our own numerical calculations.
52: We combine these opacities with abundances
53: computed from a chemical equilibrium model using recently revised solar abundances
54: to compute mean opacities. The chemical equilibrium calculation
55: accounts for the settling of condensates in a gravitational field, and
56: is applicable to ultracool dwarf and extrasolar planetary atmospheres, but not
57: circumstellar disks. We find that the inclusion of alkali atomic opacity
58: substantially increases the mean opacities over those currently in the literature
59: at densities relevant to the atmospheres
60: and interiors of giant planets and brown dwarfs. We provide our opacity
61: tables for public use and discuss their limitations.
62:
63: \end{abstract}
64:
65: \keywords{line profiles, molecular data, stars: atmospheres, stars: low-mass, brown dwarfs}
66:
67:
68: \section{Introduction}
69:
70: The thermal structure and radiative transfer within the atmospheres of
71: ultracool dwarfs and extrasolar planets ultimately depends upon an
72: entire suite of molecular and atomic opacities relevant to the the
73: temperatures and pressures found in these objects. Our group utilizes
74: these opacities in models of ultracool dwarfs and extrasolar planets
75: \cite[e.g.,][]{Marley et al. 2002, Fortney et al. 2006, Saumon et al. 2007} and provides them to other members
76: of the community. Here we briefly summarize our current suite of
77: molecular opacities that we depend upon in our modeling and consider
78: some relevant issues in their construction. We also report on the
79: chemical equilibrium calculation used to compute molecular abundances.
80:
81: Although less commonly used than in the past, some calculations,
82: particularly of planetary evolution, utilize Rosseland ($\kappa_{\rm R}$) or Planck ($\kappa_{\rm P}$) mean
83: opacities \citep[e.g.,][]{Hubickyj et al. 2005}. Canonically, Rosseland
84: mean opacities are the appropriate choice for radiative transfer models
85: of optically thick atmospheric regions where radiation propagates by
86: diffusion, whereas Planck mean opacities are the appropriate choice in
87: optically thin regions \cite{Mihalas 1978}.
88: By their nature, mean opacities are sensitive to the sum of the relevant opacity sources involved in their computation.
89: Changes both in molecular and atomic abundances and in the individual opacity sources themselves can impact the final product. Here
90: we present improved calculations for both the abundance of atoms used in our equation of state and new calculations
91: for the relevant opacity sources.
92:
93: The best available data defining the `solar abundance' of the atoms has evolved
94: over time as understanding of the sun's atmospheric composition has improved. Here we employ
95: solar and scaled solar metallicities that use the recently revised
96: abundances presented in \cite{Lodders 2003}. The chemical equilibrium gas
97: compositions computed as a function of temperature and total pressure
98: take condensate formation and condensate settling into clouds into
99: account in the calculation of the chemical equilibrium as is appropriate for the conditions in brown dwarfs,
100: extrasolar giant planets and cool stars. However we purposefully neglect grain opacity in the calculation
101: of the Rosseland and Planck means. We emphasize that the results here
102: do not apply to lower-gravity environments such as proto-planetary
103: disks because there the condensation chemistry, and therefore the gas
104: chemistry, is significantly different than that in gravitationally
105: bound atmospheres of substellar objects \cite{Lodders 1999a, Lodders and Fegley 2002, Lodders and Fegley 2006}.
106:
107: To compute mean opacities, we calculated a new set of molecular and atomic
108: opacities on a grid of 324 pressure and temperature points
109: ranging from 75 to 4000K and $3\times 10^{2}$ to $3\times 10^{8}\,\rm dyne\,cm^{-2}$
110: ($3\times 10^{-4}$ to 300 bar). Every atomic and molecular line opacity in our database
111: of 10 different gaseous species was included, and not simply a sampling. These opacities
112: were computed on a fixed wavenumber grid that completely resolved individual line profiles.
113: These opacities were then combined with
114: collision induced absorption due to interactions of hydrogen and helium
115: as well as several other opacity sources. This new set of mean
116: opacities uses the latest physical information including new terms due
117: to alkali atoms that were not included in previous investigations. Condensate opacity is not
118: included in the mean opacities presented here because condensate opacities
119: depend on the particulars of cloud models chosen to model substellar
120: atmospheres \citep[e.g.,][]{Ackerman and Marley 2001} and may very
121: greatly, depending upon the actual thermal structure in a given
122: situation. Because we neglect grain opacity, the opacities presented here should
123: be regarded as a lower limit to the true opacity, which may be several orders of magnitude higher.
124: The grain free results are of
125: interest as they highlight the important role of the alkali elements and isolate
126: assumptions regarding the grain opacity.
127:
128: Recent tabulations of mean opacities include the work by Lenzuni et al. (1991)
129: for a zero metallicity gas and Ferguson et al. (2005). Tabulations from the latter work
130: generally include opacity of solid condensates but a few cases without grains
131: are also presented. We compare our results to the latter work
132: over the relatively limited region of overlap. No other pure gaseous opacity tables
133: are readily available for solar metallicity over our temperature and pressure regime.
134:
135: We discuss the opacity sources and methods for treating line broadening
136: in Section 2. In Section 3 we
137: discuss our chemical equilibrium calculation. Mean opacities are
138: presented and discussed in Section 4.
139:
140: \section{Molecular Opacity Sources and Data}
141:
142: We maintain a large and constantly updated database of molecular and
143: atomic opacities. Some of the opacities
144: are from standard sources, such as the HITRAN database, while others
145: are a mixture of standard and other sources. In this section we discuss
146: the opacity sources we employ for each molecule of interest. Sharp and Burrows (2007)
147: recently published a thorough discussion of molecular opacities for
148: ultracool dwarfs and discuss which species are of greatest interest for modeling
149: these objects. We refer the reader to that work for more extensive background discussions than
150: are included here. Furthermore, several of the molecular and atomic opacity
151: tabulations that they review are also employed by our group. Thus for a number
152: of opacity sources we simply defer to their discussion. In other
153: cases, particularly for $\rm CH_4$, our opacity line list is unique and
154: we discuss it in some detail.
155:
156: Given a list of atomic transitions, it is further necessary to
157: compute a line shape for each atomic or molecular line. Thus in
158: Section 2.2 we discuss our choices for molecular line shapes and pressure
159: broadening.
160:
161: \subsection{Molecular opacities}
162:
163: In this section we discuss the molecules for which we compute opacities
164: (although many more species are included in the
165: chemical equilibrium calculation). The selection of molecules is
166: dictated by the chemistry in solar-composition brown dwarf and
167: extrasolar giant planet atmospheres, and, to some extent, the availability
168: of line data. Our list includes the most important opacity sources at
169: the temperatures appropriate to our calculations as validated by
170: observations of cool stars, brown dwarfs, and giant planets. We do not include certain
171: sources that affect only the far UV portion of the spectra as our high
172: temperature cutoff is currently 4000 K.
173:
174: Most spectroscopic databases are built upon measurements taken at or
175: near room temperature, and theoretical calculations supply the missing
176: transitions that can become important at high temperatures. If only
177: room temperature databases are used, the transitions from highly
178: excited energy levels (usually referred to as hot bands in the
179: literature) would be missing and the true opacity at elevated
180: temperatures would be substantially underestimated. Whenever possible,
181: we use expanded databases here.
182:
183: Note that all references to HITRAN are to the HITRAN database. The HITRAN
184: website\footnote{ \url{http://cfa-www.harvard.edu/hitran}} provides the
185: latest, updated data and copies of all the papers
186: that give details on each molecule included.
187:
188: \begin{itemize}
189:
190: \item $\rm H_{2}O$: The two most extensive line lists for water are
191: those of Partridge and Schwenke (1997) and Barber et al. (2006). Both
192: are computational lists with hundreds of millions of lines. We have
193: made extensive comparisons between the two lists and find that over the
194: temperature range relevant to brown dwarf and extrasolar planet
195: atmospheres, the differences are slight. We utilize the former list
196: here, supplemented with lines for minor isotopes ($\rm HD^{18}O$ and
197: $\rm HD^{17}O$) from HITRAN\footnote{In calculations specifically for
198: brown dwarfs we set the deuterium abundance equal to zero, see \S 3.1.}. The
199: entire database includes about $2.9\times10^8$ lines. Line widths are
200: computed using $\rm H_2$ broadening data from Gamache \citep{Gamache et
201: al. 1998, Gamache 2001}; a recent paper
202: \cite{Ma et al. 2007} discusses possible problems with the current
203: implementation of this theory for some other molecules. However, the few examples shown in the Ma et al. paper only cover HF
204: broadened by HF or $\rm N_2$. The data for $\rm N_2 - HF$ broadening showed little differences
205: from the earlier theory, leading us to have some confidence in our approach until data is available
206: for $\rm H_2$ broadening of $\rm H_2O$.
207:
208: We note that Allard et al. \cite{Allard et al. 2000} considered the completeness
209: of the Schwenke H$_{\rm{2}}$O (and TiO) databases at high temperature
210: and
211: concluded that the water database still lacked transitions from high
212: vibrational energy levels needed for calculating models of M stars and proposed
213: that a then preliminary (and never, to the best of our knowledge, publicly released) database by the Tennyson group was superior for these purposes.
214: However, as discussed above the latest release of the
215: H$_{\rm{2}}$O database from the Tennyson
216: group \cite{Barber et al. 2006} no longer shows a significant difference with the earlier Schwenke
217: result at these temperatures.
218:
219: \item $\rm CH_4$: The laboratory analysis of the methane spectrum is
220: incomplete and is unlikely to be completed from laboratory measurements alone. The
221: difficulties arise from the high
222: degree of symmetry of the molecule, which causes a great number of the bands
223: to overlap, and the fact that the average line separation is generally less than the
224: doppler width at room temperature.
225:
226: The HITRAN methane database is not a true high temperature
227: database because many bands originating from higher vibrational levels
228: are not included. Instead our line lists for
229: $\rm ^{12}CH_{4}$ and $\rm ^{13}CH_4$ were generated using the latest Spherical Top Data System (STDS)
230: software \cite{Wenger and Champion 1998} from the group at the University of
231: Bourgogne\footnote{\url{http://icb.u-bourgogne.fr/OMR/SMA/SHTDS/STDS.html}} which allows
232: us to calculate methane spectra to much higher total angular momentum
233: value, $J$, than tabulated in HITRAN. For the calculations reported
234: here we computed to a maximum rotational level of
235: $J_{\rm max}=60$, which likely covers
236: lines of importance in cool objects. Homeier et al. \cite{Homeier et al. 2003} also used
237: the STDS software, but only calculated up to $J_{\rm max}=40$.
238: For our calculation one band of $\rm ^{13}CH_4$ that is missing in the
239: Bourgogne list was added from HITRAN and $\rm CH_3D$ was included from
240: HITRAN. Extensive line width data for all bands is not available
241: so we separated the available information by
242: reference to the vibrational symmetry of the transitions and applied it
243: to all bands \citep{Brown 1996}. There
244: are about $2\times10^8$ total $\rm CH_4$ lines in the database. Figure 1
245: compares our
246: computed hot methane opacity at 1000 K and $10^6\,\rm dyne\,cm^{-2}$ to an opacity computed solely from HITRAN data,
247: both without any $\rm CH_3D$.
248: There is generally good agreement near band centers and the computational database clearly provides
249: important additions in regions where the HITRAN data are lacking.
250:
251: Both our database generated from the STDS software and HITRAN
252: do not extend to wavenumbers much higher than about 6400 $\rm cm^{-1}$ (although
253: many weak lines due to the $\rm 2\nu_{3}$ band extend up to 10000 $\rm cm^{-1}$).
254: We thus supplement the numerical line database with a continuum opacity derived from
255: laboratory data by Strong \citep{Strong et al. 1993} even though the experimental results are not available at
256: elevated temperatures. This situation is unlikely to improve until new
257: theoretical predictions are available for $\rm CH_4$. The currently available
258: laboratory data for bands above 6400 $\rm cm^{-1}$ remained unanalyzed, in part because
259: of lack of knowledge of the detailed energy levels of the transitions.
260:
261: Recently Leggett et al. (2007) compared our model spectra produced with the methane line list described above
262: to the spectra of late type T dwarfs. They found that the models still provide a poor match to observed spectra near $1.67\,\rm \mu m$, where the models show too much flux.
263: This is just the region where the calculated hot methane line list effectively
264: ends except for the few very high $J$ lines that extend to shorter wavelengths. We note that
265: the CIA opacity of $\rm H_2-H_2$ (see below) is predicted to vary rapidly with wavelength (by about a factor of 2)
266: in this region. Thus, the inability to match the observations in this region could
267: be ascribed either to the lack of good methane data, problems in the CIA simulations,
268: or some combination of the two factors. It is also conceivable, but unlikely, that some other opacity source could also
269: be missing from the models in this region. Further improvements to the methane opacity
270: data will eventually resolve this problem.
271:
272:
273: \item $\rm NH_{3}$: For ammonia we rely on the line list from HITRAN
274: supplemented with additional measurements, not yet fully analyzed, made at room temperature.
275: These additional lines are in the
276: 6600 - 7000 $\rm cm^{-1}$ region and are from the $\nu_3 + 2\nu_4$,
277: $\nu_1 + \nu_3$, and the $2\nu_3$ bands \citep{McBride and Nicholls 1972, Brown 2000}.
278: Only line strengths at $296\,\rm K$ and estimated values
279: for the lower energy level are available for these bands.
280: The final line list has about
281: 34,000 lines. Line widths are computed as arising 90\% from collisions
282: with $\rm H_2$ and 10\% with He (from work by Nemchikov as reported in Brown (2000)).
283:
284: In their comparison of our spectral models for late T dwarfs to data, Leggett et al. (2007) also considered exploratory
285: models that employed laboratory measurements of ammonia opacity by Irwin et al. (1999) over the spectral range
286: of 0.91 to $1.9\,\rm\mu m$. They found that even when atmospheric depletion of ammonia by
287: non-equilibrium chemistry was accounted for, the predicted near-infrared ammonia features were not seen
288: in the spectra of two T8 dwarfs. We conclude that the Irwin et al. data overestimate the ammonia opacity and
289: we do not employ that dataset here.
290:
291:
292: \item CO: As with methane, for carbon monoxide we favor a high
293: temperature line list over that available from HITRAN. We utilize the
294: list from Goorvitch \cite{Goorvitch 1994} which includes bands that originate
295: from highly excited energy levels. We supplement this list with data
296: on $\Delta V=4$ transitions from Tipping (1993) and somewhat
297: fragmentary information on $\rm H_2$ and He line widths from the
298: literature (Bulanin et al. 1984, Le Moal and Severin 1986, Mannucci
299: 1991). Minor isotopes missing from the Goorvitch list were added from HITRAN.
300:
301: \item $\rm H_2S$: Data for the main isotope of $\rm H_{2}\,^{32}S$ are
302: from a calculated list by Richard
303: Wattson \cite{Wattson 1996} plus minor isotopes from HITRAN. $\rm H_{2}$
304: broadening was included from data in the literature \cite{Kissel et
305: al. 2002}. There are about 188,000
306: lines in the list. This list is not a true high temperature list but
307: does include
308: many weak lines below the intensity cutoff of the current HITRAN line
309: list.
310: The wavenumber coverage is thus much greater than HITRAN, extending to
311: 19500 $\rm cm^{-1}$.
312:
313: \item $\rm PH_{3}$: For this molecule we use the latest HITRAN list,
314: including new bands and broadening information
315: from Linda Brown \cite{Brown 2000}. The list includes about 20,000
316: lines.
317:
318: \item TiO: We include five isotopes in our TiO tabulation from
319: David Schwenke \citep{Schwenke 1998} with modifications to the
320: strengths of the $\delta$ and $\phi$ bands
321: based on a comparison \cite{Allard et al. 2000} of models with
322: stellar spectra.
323: The list of about $1.7 \times 10^8$ lines includes transitions from
324: higher energy levels. There is no data on broadening from $\rm H_2$,
325: instead we compare data from other species such as $\rm H_2S$ to try to
326: set some reasonable limits on the broadening as discussed in \S 2.2.
327: The TiO molecular opacities should be reasonably complete
328: for all temperatures considered except perhaps for the very highest
329: values considered in these tables \citep{Allard et al.
330: 2000}
331:
332: As reported in Sharp and Burrows (2007), one of us (RF) discovered an error in line
333: strengths in the program to convert the predictions of Schwenke
334: \cite{Schwenke 1998} from atomic units to HITRAN units,
335: with the strengths being off by a multiplicative factor of $2J^{\prime
336: \prime} +1$. This error is corrected here.
337:
338: \item {VO:} For this molecule we rely on the line list, consisting of
339: about 3.1 million lines, from Plez (1998) which is briefly reviewed in
340: Sharp and Burrows (2007). No information on line broadening is
341: available, so it is treated as was $\rm TiO$.
342:
343: \item {FeH:} The line list for FeH is known to be incomplete in the near infrared \cite{Cushing
344: et al. 2005}. We rely on lists tabulated by
345: Dulick et al. (2003) and discussed in more detail by Sharp and Burrows (2007). Only the most abundant $\rm ^{56}Fe$ isotope is
346: used. No width data are available so it was estimated by using data for similar molecules from the literature.
347:
348: \item {CrH:} For this molecule, primarily of interest in M and L
349: dwarfs, we rely on a list from Burrows et al. (2002) which is further reviewed by \cite{Sharp and Burrows 2007}. The list
350: includes 55,300 lines, but again no width data are available and it was
351: estimated.
352:
353: \end{itemize}
354:
355: \subsection{Line By Line Calculations}
356:
357: As noted previously, we compute the opacity on a fine, fixed wavenumber grid. We add the opacity arising from each individual molecular line (hence, `line by line') using
358: a program that takes information for each molecular absorber from a database that contains line
359: strengths and positions, the lower energy level and line broadening
360: information. The line profiles are generated from the line database
361: with a Voigt profile algorithm; at higher pressures the profile
362: are essentially Lorentzian. We neglect the problem of how to treat the shape
363: of the far line wings, where it is known \cite{Levy et al. 1992} that the line shape
364: should eventually become sub-Lorentzian. In most cases actual measurements, particularly at the higher pressures,
365: are lacking and in many cases only
366: a few theoretical predictions are available for selected bands.
367: We make no attempt to simulate the
368: specialized line shapes that are appropriate in the far IR and microwave
369: regions \cite[see][for a discussion in the context of Jupiter's deep atmosphere]{de Pater et al. 2005}. The line shapes in these regions are expected to be asymmetric,
370: but because of the large overlap of the low and high frequency wings
371: due to the high density of lines, the effects due to the deviation
372: of the line shape from a Lorentzian will tend to average out and should
373: not cause a significant change in the integrated values of the mean
374: opacities.
375:
376: Likewise we neglect $\rm{\chi}$ factors \cite{Levy et al. 1992} which describe the deviation of the line wings
377: from a pure Lorenztian form. This includes effects such
378: as line mixing and line narrowing, for example. Some information
379: is available for selected species on the deviation of line shapes
380: from a Lorentzian, but in many cases these studies have covered only
381: a single band of a given molecular species, and the broadening agent
382: was usually some mixture of $\rm N_2$ and $\rm O_2$ instead of $\rm H_2$
383: and He which would be the appropriate choice for brown dwarfs and
384: giant planets. After conducting a number of tests, we found
385: that the inclusion of a $\chi$ factor for $\rm H_2O$, a major source
386: of opacity, had no substantive effect on the Rosseland mean results.
387: Considering all the other sources of uncertainty, we did
388: not include $\rm{\chi}$ factors in this study.
389:
390: The collisional (pressure broadened) line widths for several of the species
391: used in our opacity calculations are not well known. In general, there have
392: been no measurements reported in the literature and in many cases the main
393: information about the line positions and intensities
394: that we use comes from theoretical predictions. One could estimate, in principle,
395: the line widths in comparison to other molecules by examining the relative values
396: of the molecular polarizabilities \cite{Hirschfelder et al. 1954, CRC Handbook 2000-2001}.
397: Unfortunately, even this information is not
398: available to us for many species of interest. It might be possible to again estimate
399: some of these polarizabilities by quantum mechanical calculations but again this data
400: is not currently available. For the one case that can be compared to measurements,
401: namely $\rm H_2S$, the large relative value of the static polarizability compared to other molecules
402: in our list does correlate well
403: with the rather large line widths actually measured \cite{Kissel et al. 2002}.
404: Currently, we simply use estimates for the line widths of $\rm TiO$ and the metal
405: hydrides that are $\sim 25 - 50\%$ larger than the line widths for other molecules that are
406: earlier in the periodic table and generally have smaller effective radii when they are
407: formed into molecules.
408:
409: As a check on the effect of uncertainties in the line broadening for the various
410: metal compounds, we calculated a set of $\rm TiO$ opacities with twice the
411: assumed value of the pressure broadening widths. A direct comparison shows
412: no significant effect on the overall mean opacity. The total value of the
413: absorption is conserved over the line profile as it should be, assuming that
414: the line wings are allowed to extend to larger values at higher pressures.
415: The overall effect of broader lines is to smooth out the central peaks of the absorption
416: lines and to fill in the low points in the far wings. Since the doubled
417: widths approach the largest measured values for any molecule in our study ($\rm H_2S$),
418: it is apparent that line width uncertainties of a factor of 2 are not a significant
419: source of error in our calculations.
420:
421: The remaining line broadening parameters were taken from the literature when
422: available. Available information can include actual measurements or
423: theoretical predictions.
424: In a few cases, no experimental or theoretical data of any kind for
425: $\rm H_2$ and He broadening was available so estimates of the line
426: width were made. In several cases data was available for broadening
427: by $\rm H_2$ and He and this was used. The assumed line broadening
428: parameters were scaled by pressure and an assumed temperature dependence
429: for each $(T, P)$ combination. This temperature dependence could come
430: from actual measurements or from theory. In practice, the laboratory measurements
431: do not cover a very large range in $T$, but this is usually all that
432: is available \cite{Homeier 2005}. In contrast, the Van der Waals theory of broadening
433: by foreign gases, commonly used in stellar atmospheres calculations,
434: predicts the same temperature dependence for all lines irrespective
435: of their angular momentum quantum numbers and the identity of the foreign broadener is usually
436: assumed to be hydrogen atoms.
437:
438: The scaling of the line width linearly with pressure ignores the problem
439: of what happens to the line width at very high pressures, as a hard
440: sphere cutoff to the pressure scaling should exist. This is related
441: to the value of the second and higher virial coefficients in the
442: equation
443: of state for the gas. Since reliable experimental data on collision
444: cross sections as derived from viscosity and diffusion data are only
445: available for a few combinations of species and broadener, it is
446: difficult
447: to validate the theory under the physical conditions that apply to
448: the astronomical case for more than a few molecule-broadener
449: combinations.
450: In particular, in the astronomical case the highest pressures are
451: associated with the highest temperatures which is just the region
452: where the parameters that determine the equation of state are the
453: most uncertain.
454:
455:
456: \subsection{Collision Induced Absorption due to $\rm{H_{2}-H_{2}}$,
457: $\rm{H_{2}-He}$ and $\rm{H_{2}-H}$}
458:
459: Collision induced absorption produces a broad continuum that sculpts
460: the foundation of ultracool dwarf spectra.
461: Our source for the subroutines to calculate the collision induced
462: absorption comes from the recent work of A. Borysow and her
463: collaborators
464: \cite{Borysow 2002, Borysow et al. 2000, Borysow et al. 1997, Birnbaum et al. 1996, Zheng & Borysow 1995, Borysow 1992, Borysow & Frommhold 1990, Borysow et al. 1985}. We
465: have used the latest available
466: versions of all programs to compute the CIA absorption due to $\rm
467: H_2-H_2,$ $\rm H_2-He$ and $\rm H_2-H$
468: collisions. The FORTRAN programs and opacity tables available on
469: Borysow's web page\footnote{http://www.astro.ku.dk/\textasciitilde{}aborysow/} were used
470: to construct tables that
471: represent the absorption by a mixture of {}``normal'' (3:1) $\rm H_2$.
472: This can be contrasted with an ``equilibrium'' mixture where the
473: ratio of the ortho and para forms of $\rm H_2$ is 1:1. At the high
474: temperatures of these objects a normal distribution would be expected.
475: This topic is discussed more thoroughly in Massie and Hunten (1982)
476: and Carlson et al. (1992). In any case, the difference in
477: the results between the two cases is only significant at the lowest
478: temperatures and at the low frequency end of the spectrum. This would
479: lead to small changes in the results of a few percent in these cases,
480: well within the other sources of uncertainty in this problem.
481:
482: \subsection{Opacity from Alkali Atoms: Na, K, Cs, Rb and Li}
483:
484: The importance of alkali atoms to atmospheric opacity in cool
485: substellar objects was first
486: recognized from the influence of these atoms of the far red spectra of
487: T dwarfs \cite{Burrows et al. 2000}. These pressure broadened lines,
488: particularly of Na and K, are major opacity sources over certain spectral ranges, temperatures, and
489: densities, particularly
490: above the Na and K condensation temperatures (about 600 K, Lodders (1999a)).
491: Although the exact form of the line shapes of the alkali atoms due
492: to pressure broadening by $\rm{H_{2}}$ is still not completely
493: understood, it is important to include it. We use a computer
494: code \cite{Burrows et al. 2000} kindly provided to us by A. Burrows
495: to generate line profiles for atomic lines of neutral alkali atoms
496: using a line width parameter setting of 1.0 (as defined by those
497: authors).
498:
499: Burrows and Volobuyev (2003) and Allard et al. (2003) have further modeled
500: the pressure broadened alkali lines. Future improvements in understanding
501: of the alkali line widths will certainly impact spectral modeling of ultracool dwarfs
502: and may impact the mean opacities. The importance of the alkali elements to
503: the mean opacities is discussed in Section 4.
504:
505: \subsection {Other Opacity Sources}
506:
507: Several other opacity sources are included in our calculation.
508: Bound-free absorption by H and $\rm H^{-}$ and free-free absorption by
509: H, $\rm H_2$, $\rm H_{2}^{-}$ and $\rm H^{-}$ (see Lenzuni et al.
510: 1991) were added using algorithms provided by Tristan
511: Guillot \cite{Guillot 1999}.
512: Rayleigh scattering from $\rm H_{2}$ and Thompson scattering are also
513: included in the Rosseland mean following Lenzuni et al.
514: Opacity from electrons and H atoms does not provide a large
515: contribution to the overall opacity below $\sim 2500\,\rm K$ but
516: becomes important at higher temperatures relevant to M dwarf
517: atmospheres.
518:
519: The opacity from the more abundant atomic species are not important in our case because our high temperature cases also involve high pressure. Since we are concerned with high gravity objects the fractional abundances of atoms like Fe, Mg, Si, and Al never exceed $8\times10^{-5}$ and are usually much
520: lower. Furthermore the most important lines for these atoms are in the UV where there is little flux over our temperature regime.
521:
522: \section{Chemical Equilibrium Calculations}
523:
524: The thermochemical equilibrium abundances used in the opacity models
525: were computed with the CONDOR code described and applied to brown dwarf
526: studies in several papers (e.g., Fegley \& Lodders 1994, 1996, Lodders
527: 1999a, 2002, Lodders \& Fegley 2002). Brief overviews about the gas
528: and cloud chemistry in substellar atmosphere can be found in Lodders
529: (2004) and Lodders \& Fegley (2006). Here we summarize this work and
530: highlight important issues for the problem at hand.
531:
532: \subsection{Gas and Condensate Chemistry Computations}
533:
534: The CONDOR code simultaneously computes the chemical equilibrium compositions for
535: more than 2200 gases (including ions) and more than 1700 solids and liquids of all naturally
536: occurring elements by considering dual constraints of mass balance and chemical equilibrium.
537: The major thermodynamic data sources are given in Fegley \& Lodders (1994), and data are
538: frequently updated (see e.g., Lodders 1999b, 2004). Note that some frequently used compilations of thermodynamic properties and/or polynomial fits contain errors that also affect compounds important for brown dwarf chemistry (see Appendix 1 in Lodders 2002). Therefore chemical equilibrium species and abundances computed by other groups could be different than our results if erroneous thermodynamic data were used by them, which can also introduce differences in calculated opacities.
539:
540: The CONDOR code uses elemental abundances, total pressure and
541: temperature as inputs. We use the solar system abundances in Lodders
542: (2003), uniformly enhanced or depleted to model metallicity effects
543: where appropriate. Equilibrium compositions considering cloud
544: condensate formation for solar elemental abundances were computed for
545: 324 pressure-temperature sample points in a grid ranging from 50 to
546: 4000 K and $\log P({\rm dyne\, cm^{-2}}) = 2$ to +8.5 which spans characteristic
547: conditions in the atmospheres of low mass objects. Similar calculations
548: were done for other metallicities; here we only include results for
549: metallicities of 2 and 1/2 times solar ($\rm [M/H] = \pm 0.3$).
550:
551: Despite the plethora of gas species present in the thermochemical
552: calculations, only a few compounds are abundant or major opacity
553: sources, and the subset of the gases selected here for constructing
554: the opacity tables are discussed in Section 2.
555: Compared to previous solar elemental abundance compilations, the more
556: recent data include significant downward revisions in the C, N, and O
557: abundances around 20\% to 40\%,
558: which have consequences for the abundances of important opacity sources
559: such as methane, CO, water, and ammonia in substellar atmospheres. The
560: new lower C, N, and O abundances resemble a decrease in CNO
561: metallicity, and a detailed discussion of how metallicity affects the
562: $\rm CH_4/CO$ and $\rm NH_3/N_2$ equilibria is given in Lodders \& Fegley
563: (2002). For example, one important consequence of the lower C, N, and O
564: abundances is that the methane to CO as well as the $\rm NH_3$ to $\rm
565: N_2$ conversions are shifted towards higher temperatures (at constant
566: total pressure or gravity), and are shifted towards lower total
567: pressure (at constant temperature).
568:
569: The thermochemical calculations also include results for deuterium.
570: However, in the opacity modeling for brown dwarfs, the deuterium
571: abundance was set to zero because it is assumed that all D is destroyed
572: in objects more massive than 13 Jupiter masses. Thus, the opacities due from
573: HDO and $\rm CH_{3}D$ are not included. However, in lower mass objects,
574: the HDO and $\rm CH_{3}D$ opacities must be included and these were
575: calculated using $\rm D/H = 1.94 \times 10^{-5}$ \cite{Lodders 2003}. However, the additional
576: opacity from HDO and $\rm CH_{3}D$ generally has only a very marginal
577: effect on the overall results.
578:
579: \subsection{Condensate treatment}
580: The CONDOR code has two principle pathways to treat condensate
581: formation, depending on the desired application. Under ideal
582: equilibrium conditions in protoplanetary disks or stellar winds,
583: gas-solid equilibria are maintained within a cooling gas and therefore
584: high temperature condensates that formed first from a cooling gas
585: (primary condensates)
586: still can react with the gas to form secondary condensates at lower
587: temperatures. A well-known gas-solid reaction is the reaction of
588: primary iron metal with $\rm H_2S$ gas to secondary troilite (FeS) at
589: low temperatures (see below). However, such reactions may not apply to
590: substellar
591: atmospheres where cloud layer formation prevents the formation of
592: secondary condensates (see Lodders \& Fegley 2002). If a primary
593: condensate settles from the gas into a cloud layer (sometimes called
594: rainout),
595: the primary condensate becomes depleted in the atmosphere above its
596: cloud and cannot participate in reactions that are thermochemically
597: favorable at the cooler temperatures above the cloud. For example, if
598: iron metal condenses and settles into a cloud, the reaction of iron
599: metal with $\rm H_2S$ to secondary troilite cannot happen. This has
600: important consequences for the gas chemistry because troilite formation
601: would remove all $\rm H_2S$ (essentially all the atmospheric sulfur
602: inventory) at the low temperatures where troilite would be stable.
603: However, if the secondary troilite does not form, $\rm H_2S$ remains in
604: the atmosphere until $\rm NH_4SH$ condenses, which only happens at $T
605: <<300\,\rm K$. The {\it Galileo} entry probe mass spectrometer detection of
606: $\rm H_2S$ at about three times the solar $\rm S/H_2$ ratio in
607: Jupiter's atmosphere below the $\rm NH_4SH$ cloud level (Niemann et al. 1998)
608: shows that the cloud layer condensation approach works well in giant
609: planet atmospheres such as in Jupiter (see also Lodders \& Fegley 2002,
610: Visscher et al. 2006).
611:
612: The computations with cloud settling thus require the knowledge of
613: which condensates are primary (formed directly from the gas) and which
614: ones are secondary (formed by gas-solid reactions). This is easily
615: gauged from the equilibrium calculations by checking the distribution
616: of an element between condensates and gas. For example, under typical
617: total pressure conditions (e.g., $>10^5\,\rm dyne\,cm^{-2}$) in substellar atmospheres,
618: iron starts to condense as metal at temperatures above 1600 K, and the
619: metal settles into a cloud layer with a base at the temperature level
620: where metal condensation starts. At the cloud base, iron metal is in
621: evaporation-condensation equilibrium with the hotter atmosphere below.
622: Above the cloud base, the iron gas abundance is determined by the iron
623: vapor pressure of the metal cloud. The iron vapor pressure drops
624: exponentially with the decreasing temperatures above the cloud, and
625: mass balance dictates that more iron is in the condensates, and less
626: iron gas remains in the atmospheric gas. Typically it takes
627: about a $100 - 200\,\rm K$ drop from the condensation temperature
628: (i.e., the temperature where the condensate appears first from a
629: cooling gas) to the temperature where $>99\%$ of an element is removed by
630: its condensate. This is the normal vapor-pressure driven condensation
631: process. The only special consideration here is that the condensate
632: settles so that the amount of condensate becomes depleted at
633: atmospheric levels above the condensate cloud that forms from the
634: settled condensate particles. However, this does not affect the fact
635: that the gas abundances remain established by the vapor pressure over
636: the condensing solid or liquid because the vapor pressure over a
637: substance is independent of the absolute amount of the substance
638: present.
639:
640: Troilite formation would only occur at temperatures below about 700 K,
641: which is significantly lower than the temperatures at which iron vapor
642: pressures are so low that essentially all iron is condensed. The
643: gas-sold reaction for troilite formation requires the presence of iron
644: metal, but at the temperature level where FeS becomes stable, the iron
645: metal condensate has settled out at greater depth in low mass object
646: atmospheres, and thus, the secondary troilite cannot form.
647:
648: Note that the thermochemical equilibrium calculations of the gas
649: abundances do not require information about the cloud particle sizes,
650: their settling efficiencies, or the vertical extent of the cloud layer;
651: nor can these calculations provide this information without further
652: model assumptions. These parameters are not needed if we are only
653: interested in gas opacities, however, these parameters are needed if
654: condensate opacities are to be considered, which requires inclusion of
655: a detailed cloud model (see e.g., Ackerman and Marley 2001, Marley et
656: al. 2002).
657:
658:
659: Other observations also demonstrate that other refractory elements are
660: depleted by condensate cloud formation at high temperatures deep in the
661: atmospheres of low-mass objects. For example, the gradual disappearance
662: of TiO and VO bands in the optical spectra of L dwarfs requires
663: condensation of Ca-titanates into which VO dissolves (Fegley \& Lodders
664: 1994, 1996, Lodders 1999a, 2002). The absence of silane ($\rm SiH_4$)
665: indicates deeply seated silicate clouds and the presence of volatile
666: germane ($\rm GeH_4$) in the atmospheres of Jupiter and Saturn shows
667: that germanium did not condense into iron metal at low temperatures (as
668: it did in the solar nebula, see Lodders 2003) because the metal settled
669: (Fegley \& Lodders 1994). The presence of monatomic K gas in the
670: atmospheres of T dwarfs requires that the refractory
671: feldspar anorthite has settled into a deep cloud because otherwise, the
672: more volatile K could dissolve in it at lower temperatures, (e.g.,
673: Lodders 1999a, Burrows et al. 2000, Geballe et al. 2001, Marley et al. 2002).
674:
675: In summary, the major difference for {\it gas} chemistry between the ``nebula" and ``cloudy" cases is that the gas abundances of elements forming condensates by gas-solid reactions will be lower in the ``nebula" models than in the ``cloudy" models at the temperatures and total pressures where such secondary condensation is expected. In the cloudy models, the secondary condensates cannot form because the solid required for the gas-solid reaction is not present; e.g., anorthite is required for Na and K condensation in solid-solution. Therefore, the {\it gas opacities} will be different for the nebula and cloudy models. As a practical matter, the biggest impact on the opacities arises from the timing of removal of gaseous alkali species. In the ``cloudy'' chemisty, Na and K persist to lower temperatures than would otherwise be the case. The spectra of T dwarfs confirms that in fact the ``cloudy'' chemistry is the correct choice for atmospheres. Since we neglect grain opacity, the differences in the timing and arrival of the condensates themselves does not impact our opacity calculation.
676:
677: \section{Results}
678:
679: \subsection{Mean Opacity Tables}
680:
681: Given the opacities described in Section 2 and the elemental and
682: molecular abundances described in Section 3, we computed Rosseland and
683: Planck mean opacities for three metallicities ($\rm [M/H]= -0.3, 0.0,
684: +0.3$).
685: The integration
686: over frequency was carried out using a grid of equally spaced wavenumber
687: points. The frequency spacing was based on the temperature and pressure
688: of each layer, such that the spectral resolution was always fine (1/4 of a line width
689: or $0.5\,\rm cm^{-1}$, whichever is less) compared to a
690: typical
691: line profile under those conditions. Spacings are as small as $4\times10^{-3}\,\rm cm^{-1}$ in some
692: instances. Such a fine grid is usually not
693: required for mean opacity calculations, but we also use the same grid
694: for high resolution spectral modeling.
695:
696: Our grid of 324
697: $(P,T)$ pairs ranges from 75 to $4000\,\rm K$ and $3\times 10^{2}$ to $3\times 10^8\,\rm dyne\,cm^{-2}$ (or $3\times 10^{-4}$ to 300 bar). We do not compute opacities at high pressure and low
698: temperature or high temperature and low pressure, as such points are not reached by
699: brown dwarf and giant planet atmosphere models. The table spacing is not uniform in order
700: to better sample important processes, including water condensation.
701: Figure 2 illustrates the pressure-temperature domain over
702: which we computed opacities as well as the magnitude of the Rosseland
703: mean opacities for a solar composition gas. This figure also shows
704: temperature-pressure profiles computed for Jupiter and a
705: variety of other substellar objects.
706: Tables I, II, and III provide Rosseland and Planck mean opacities as a function
707: of pressure and temperature for the three metallicities we consider on this grid.
708: The complete tables may be found in the on line
709: supplement to this paper. In the tables opacities have been converted to units
710: of $\rm cm^2/g $ by calculating
711: the mean molecular weight (for gaseous species only) at each temperature-pressure level, for each
712: chemical composition.
713:
714: Because of our choice of a uniform pressure-temperature $(P, T)$ grid, our data are not
715: on a uniform mass density ($\rho$) grid.
716: For easier comparison with earlier work, we also interpolated to a set of
717: constant densities for graphical purposes. Figures 3 illustrates the
718: Rosseland mean opacities along profiles of constant density. We note that
719: such an interpolation on occasions crosses chemical equilibria and condensation
720: boundaries, which produces some kinks in the interpolated data shown on the figure.
721:
722: \subsection{Opacity from Alkali Atoms: Na, K, Cs, Rb, and Li}
723:
724: Perhaps the greatest difference from previous tabulations of mean
725: opacities arises
726: from the inclusion of the pressure-broadened lines of the alkali
727: elements, particularly
728: sodium and potassium. These molecules, with their large absorption
729: cross sections
730: in the far red and very near infrared (Burrows et al. 2000) fill what
731: would otherwise
732: be a region of fairly low opacity. Figure 4 compares the summed
733: opacity as a function of wavenumber at 1400 K and $10^6\,\rm dyne\,cm^{-2}$ in our baseline case with
734: a calculation that neglects the alkali opacity. The substantial role of
735: the alkali opacity to the total summed opacity is unmistakable above about
736: $10,000\,\rm cm^{-1}$. The influence
737: of the alkali opacity on the total gaseous mean opacity is illustrated in Figure 5
738: which compares the Rosseland
739: mean opacity with and without the contribution of alkali metals
740: at several densities. It is clear that at the higher
741: densities the alkali opacity substantially fills in the opacity
742: minimum from about 1000 to 1500 K. Differences at lower densities are
743: slight since the pressure-broadened lines play a much smaller role.
744:
745: Guillot et al. (1994a, b) predicted that low opacity around 1,000 to 2,000 K in
746: Jupiter's deep atmosphere would lead to the formation of a radiative region
747: within what was then expected to be the fully convective deep interior of Jupiter.
748: In Figure 2 it can be seen by extrapolation that Jupiter's deep adiabat indeed passes through
749: a trough in opacity in this temperature range. We now understand that the opacity in this local
750: minimum region--owing to the contribution of pressure-broadened alkali opacity--is much
751: larger than was considered by Guillot et al. (1994a,b) using the opacities available at that time.
752: In a reevaluation of their earlier work, Guillot et al. (2004) indeed found that the alkali opacity is sufficient to prevent substantial
753: energy transport by radiation in Jupiter's interior. The Rosseland opacities we report here are consistent with
754: those used in the latest work by Guillot and collaborators. Thus, despite the intrinsic uncertainty
755: in the alkali line widths at high pressure, a thin radiative shell within the interior
756: of Jupiter need not be considered in the construction of interior models. This removes one source
757: of uncertainty in the construction of evolution and interior models of Jupiter.
758:
759: As noted in \S2.4, the pressure broadened line shape for the alkali metals
760: remains uncertain, particularly
761: at high pressure. In the future, it may be possible to obtain
762: experimental data on the line absorption coefficients of pressure
763: broadened alkali lines so that the effects of their shapes on the total
764: opacity can be quantified to replace the semi empirical profiles \cite{Johnas et al. 2006}.
765:
766: \section{Discussion}
767:
768: The opacities and chemical equilibrium calculations described here are examples of
769: the current state of the art for understanding giant planet and brown dwarf atmospheres.
770: Nevertheless, a number of uncertainties remain, particularly in the treatment of
771: the opacities at high temperatures. Because it is a dominant source
772: of opacity, of course water is of special concern,
773: but comparisons between the most
774: recent theoretical line lists suggest that, at least the the temperatures considered here,
775: the line list is in reasonably good shape.
776: Further work on calculating an updated
777: version of the the water spectrum
778: is currently underway. Missing opacity at high temperatures is certainly
779: a problem in even greater
780: measure for most of the other molecular opacities, particularly methane, but these
781: other molecules are generally less important than water to the mean opacity. In fact, only water, CO, TiO, and to a lesser
782: extent VO have a significant contribution from hot bands while $\rm
783: CH_4$ and, to a lesser extent, $\rm H_2S$ contain lines originating from
784: higher rotational quantum values which have been predicted from a mixture of theory
785: and available experimental data.
786:
787: For comparisons with spectra of ultracool dwarfs, however, the lack of adequate
788: high temperature opacities for methane and ammonia is an important
789: limitation (e.g., Leggett et al. 2007).
790: Improving this situation will require in the case of methane a
791: substantial theoretical and computational effort. Newer data are under
792: development for water by David Schwenke and for CO by one of us (R.F.) that
793: would allow a recomputation of the line lists. In the case of water
794: this could lead to a better representation of the opacity at the highest
795: temperatures, while any changes in the CO database will probably be
796: less substantial.
797:
798: Because the calculations presented here do not include condensates as an opacity source
799: a direct comparison with earlier work that includes grains is difficult.
800: In our brown dwarf and extrasolar planet modeling calculations
801: cloud opacity is computed from the local description of the atmospheric
802: structure, rather than relying on a pre-computed table.
803: Because these additional
804: sources of opacity may appear at different pressures and temperatures
805: in a series of models depending on the assumptions built into the
806: calculation, it is not possible to give a general set of results that
807: include solid material. Any such tables must be regarded with some
808: caution as condensate size and abundance depends on other parameters,
809: including the convective velocity, and no single prescription as a function of only
810: density and temperature can be given. In fact the grain opacity plays
811: an important role in the gaseous accretion of giant planets by the core accretion
812: mechanism \cite{Hubickyj et al. 2005}.
813:
814: We show in Figure 6 a
815: comparison with recent grain free calculations \cite{Ferguson et al. 2005}.
816: This work uses updated atomic abundances to compute mean opacities at
817: relatively low densities relevant to circumstellar disks. As such the region
818: of overlap in density and temperature is relatively small. In the overlapping region,
819: however, the correspondence is reasonably good. Since this is a low density
820: region, the effect of the alkali opacity is negligible and is not a factor.
821:
822: \section{Conclusion}
823:
824: The tables presented here and in the on line supplement to this paper provide Rosseland and Planck mean opacities for three elemental compositions relevant to the study of ultracool dwarfs and extrasolar giant planets. We have also described the databases for line transitions and our approach to computing the line broadening as well as the chemical equilibrium calculation. Future improvements
825: in the molecular opacities--particularly at high temperature--will certainly improve the quality of model spectra for the comparison with astronomical data. Barring the addition of substantial new
826: opacity sources or further updates to the solar abundance of the elements, we do not expect to see
827: significant changes to the Rosseland and Planck mean opacities reported here.
828:
829:
830: \acknowledgements
831: We thank Tristan Guillot for providing his hydrogen opacity routine and for
832: helpful advice, Jonathan Fortney for help with
833: figures and formatting, Didier Saumon for helpful conversations,
834: Adam Burrows for use of his alkali atom opacity code, and the referee for suggestions which improved the manuscript.
835: We also received generous support from David Schwenke and his collaborators.
836: He has provided extensive calculations and information concerning the quantum
837: mechanical modeling of various spectra.
838: R.F. acknowledges support from NASA grant NAG5-4970, M.M. acknowledges
839: support from the NASA Office of Space Sciences. Work by K.L. is supported by NSF grant AST0406963 and NASA grant NNG06GC26G.
840:
841:
842:
843:
844: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
845:
846: \begin{thebibliography}{99}
847: \bibitem[Ackerman and Marley 2001]{Ackerman and Marley 2001} Ackerman, A.~S., \&
848: Marley, M.~S.\ 2001, \apj, 556, 872
849: \bibitem[Alexander and Ferguson 1994]{Alexander and Ferguson
850: 1994}Alexander, D.R., and Ferguson, J.W. 1994, ApJ,437,879
851: \bibitem[Allard et al. 2000]{Allard et al. 2000}Allard, F.,
852: Hauschildt, P. H., and Schwenke, D. 2000, \apj, 540, 1005
853: \bibitem[Allard et al. 2003]{2003A&A...411L.473A} Allard, N.~F., Allard,
854: F., Hauschildt, P.~H., Kielkopf, J.~F., \& Machin, L.\ 2003, \aap, 411, L473
855: \bibitem[Allende Prieto and Lambert 2002]{Allende Prieto and Lambert
856: 2002}Allende Prieto, C.A. and Lambert, D. 2002,ApJ,573,L137
857: \bibitem[Barber et al. 2006]{Barber et al. 2006}Barber, R.J., Tennyson,
858: J., Harris, G.J., and Tolchenov, R.N. 2006 \mnras, 368, 1087
859: \bibitem[Birnbaum et al. 1996]{Birnbaum et al. 1996} Birnbaum, G., Borysow,
860: A., \& Orton, G.~S.\ 1996, Icarus, 123, 4
861: \bibitem[Borysow 2002]{Borysow 2002} Borysow, A.\ 2002, \aap, 390,
862: 779
863: \bibitem[Borysow et al. 2000]{Borysow et al. 2000} Borysow, A., Borysow,
864: J., \& Fu, Y.\ 2000, Icarus, 145, 601
865: \bibitem[Borysow et al. 1997]{Borysow et al. 1997} Borysow, A., Jorgensen,
866: U.~G., \& Zheng, C.\ 1997, \aap, 324, 185
867: \bibitem[Borysow 1992]{Borysow 1992} Borysow, A.\ 1992, Icarus, 96,
868: 169
869: \bibitem[Borysow 1991]{Borysow 1991} Borysow, A.\ 1991, Icarus, 92,
870: 273
871: \bibitem[Borysow \& Frommhold 1990]{Borysow & Frommhold 1990} Borysow, A., \&
872: Frommhold, L.\ 1990, \apjl, 348, L41
873: \bibitem[Borysow et al. 1985]{Borysow et al. 1985} Borysow, J., Trafton,
874: L., Frommhold, L., \& Birnbaum, G.\ 1985, \apj, 296, 644
875: \bibitem[Boudon et al. 2006]{Boudon et al. 2006}Boudon, V., Rey, M.,
876: and L\"{o}ete, M., 2006 \jqsrt, 98, 394
877: \bibitem[Brown 1996]{Brown 1996}Brown, L., 1996 Private Communication
878: \bibitem[Brown 2000]{Brown 2000}Brown, L., 2000 Private Communication
879: \bibitem[Bulanin et al. 1984]{Bulanin et al. 1984}Bulanin, M.O., Dokuchaev, A.B., Tonkov, M.V., \& Filippov, N.N. 1984 \jqsrt, 31, 521
880: \bibitem[Burrows 1999]{Burrows 1999}Burrows, A. 1999, Private
881: Communication
882: \bibitem[Burrows et al. 2000]{Burrows et al. 2000}Burrows, A., Marley,
883: M., and Sharp, C. 2000 \apj, 531, 438
884: \bibitem[Burrows et al. 2002]{Burrows et al. 2002}Burrows, Adam, Ram,
885: R.S., Bernath, Peter, Sharp, C.M., and Milsom, J.A., 2002 ApJ,577,986
886: \bibitem[Burrows and Volobuyev 2003]{Burrows and Volobuyev 2003}
887: Burrows, A., and Volobuyev, M. 2003 \apj, 583, 985
888: \bibitem[Carlson et al. 1992]{Carlson et al. 1992}Carlson, Barbara E.,
889: Lacis, Andrew A., Rossow, and William B., 1992, \apj, 393, 357
890: \bibitem[CRC Handbook 2000-2001]{CRC Handbook 2000-2001}Lide, David R., ed.,
891: Handbook of Chemistry and Physics 81st ed. 2000-2001, (Boca Raton, FL: CRC Press)
892: \bibitem[Cushing et al. 2005]{Cushing et al. 2005} Cushing, M.~C., Rayner,
893: J.~T., \& Vacca, W.~D.\ 2005, \apj, 623, 1115
894: \bibitem[de Pater et al. 2005]{de Pater et al. 2005} de Pater, I., Deboer,
895: D., Marley, M., Freedman, R., \& Young, R.\ 2005, Icarus, 173, 425
896: \bibitem[Dulick 2003]{Dulick et al. 2003}Dulick, M., Bauschlicher,
897: C.W., Jr., Burrows, A, Sharp, C.M., Ram, R.S., Bernath, Peter 2003
898: \apj, 594, 651
899: \bibitem[Fortney et al. 2006]{Fortney et al. 2006}Fortney,J.J., Saumon,
900: D., Marley, M.S., Lodders, K., and Freedman, R. 2006 \apj, 642,1,495
901: \bibitem[Fegley and Lodders 2005]{Fegley and Lodders 2005}Fegley, B.
902: and Lodders, K. 2005, Astrophysics Update,2,1
903: \bibitem[Ferguson et al. 2005]{Ferguson et al. 2005}Ferguson, Jason W., Alexander, David R., Allard,
904: France, Barman, Ravis, Bodnarik, Julia G., Hauschildt, Peter H., Heffner-Wong, Amanda,
905: and Tamanai, Akemi, 2005, \apj, 623, 585
906: \bibitem[Fortney et al.(2006)]{2006ApJ...642..495F} Fortney, J.~J., Saumon,
907: D., Marley, M.~S., Lodders, K., \& Freedman, R.~S.\ 2006, \apj, 642, 495
908: \bibitem[Gamache et al. 1998]{Gamache et al. 1998}Gamache, R.~R., Lynch,
909: R., Neshbya, S.~P. 1998, \jqsrt, 59, 319
910: \bibitem[Gamache 2001]{Gamache 2001}Gamache, R.R. 2001, Private
911: Communication
912: \bibitem[Goorvitch 1994]{Goorvitch 1994}Goorvitch, D. 1994, \apjs, 95, 535
913: \bibitem[Guillot 1999]{Guillot 1999}Guillot,T. 1999, Private Communication
914: \bibitem[Guillot et al. 1994a]{Guillot et al. 1994a} Guillot, T., Gautier,
915: D., Chabrier, G., \& Mosser, B.\ 1994a, Icarus, 112, 337
916: \bibitem[Guillot et al. 1994b]{Guillot et al. 1994b} Guillot, T., Chabrier,
917: G., Morel, P., \& Gautier, D.\ 1994b, Icarus, 112, 354
918: \bibitem[Guillot et al. 2004]{2004jpsm.book...35G} Guillot, T., Stevenson,
919: D.~J., Hubbard, W.~B., \& Saumon, D.\ 2004, Jupiter.~The Planet, Satellites
920: and Magnetosphere, 35
921: \bibitem[Harris et al. 2004]{Harris et al. 2004}Harris, G.~J.,
922: Lynas-Gray, A.E., Miller, S., and Tennyson, J. 2004, ApJ,600,1025
923: \bibitem[Hirschfelder et al. 1954]{Hirschfelder et al. 1954}Hirschfelder, Joseph O., Curtiss,
924: Charles F., and Bird, R. Byron 1954, Molecular Theory of Gases and Liquids, (New York, NY: John
925: Wiley and Sons)
926: \bibitem[Homeier et al. 2003]{Homeier et al. 2003}Homeier, Derek, Hauschildt, Peter H. and,
927: Allard, France 2003, ASP Conf. Series,288,357
928: \bibitem[Homeier 2005]{Homeier 2005}Homeier, Derek, 2005, Mem. S.A.It,7,157
929: \bibitem[Hubickyj et al. 2005]{Hubickyj et al. 2005} Hubickyj, O.,
930: Bodenheimer, P., \& Lissauer, J.~J.\ 2005, Icarus, 179, 415
931: \bibitem[Irwin et al. 1999]{Irwin et al. 1999}Irwin, P.G.J., Calcutt, S.,B., Sihra, K., Taylor, F.W., Weir, A.L., Ballard, J. \& Johnston, W.B. 1999, \jqsrt, 62, 193
932: \bibitem[Johnas et al. 2006]{Johnas et al. 2006}Johnas, C.~M.~S., Allard, N.~F., Homeier, D.,
933: Allard, F. and Hauschildt, P.~H., 2006, AIP Conf. Proc. 874: Spectral Line Shapes: XVIII,354
934: \bibitem[Jones et al. 2005]{Jones et al. 2005}Jones, H.R.A, Pavlenko,
935: Y., Viti, S., Barber, R.J., Yakovina, L.A., Pinfield, D., and
936: Tennyson,J. 2005 \mnras, 358, 105
937: \bibitem[Kissel 2002]{Kissel et al. 2002}Kissel, A., Sumpf, B.,
938: Kronfeldt, H.-D., Tikhomirov, B. A., and Ponomarev, Yu. N. 2002
939: JMolSpec, 216, 345
940: \bibitem[Lellouch et al. 2001]{Lellouch et al. 2001}Lellouch, E.,
941: Bezard, B., Fouchet, T., Feuchtgruber, H., Encrenaz, T., and de Graauw,
942: T. 2001, \aap, 670,610
943: \bibitem[Leggett et al. 2007]{Leggett et al. 2007}Leggett, S., Marley, M.S., Freedman, R., Saumon, D., Liu, M.C., Geballe, T.R., Golimowski, D., \& Stephens, D. 2007, \apj, submitted.
944: \bibitem[Le Moal and Severin 1986]{Le Moal and Severin 1986}Le Moal,
945: M.F., and Severin, F. 1986, \jqsrt, 35, 145
946: \bibitem[{Lenzuni et al. }{1991}]{Lenzuni et al. 1991}Lenzuni, P.,
947: Chernoff, D.F., and Salpeter, E.E. 1991, ApJS, 76, 759
948: \bibitem[Levy et al. 1992]{Levy et al. 1992}Levy, Armand, Lacome,
949: Nelly, and Chackerian Jr.,
950: Charles 1992, Spectroscopy of the Earth's Atmosphere and Interstellar
951: Medium, (Academic Press)
952: \bibitem[Lodders 1999a]{Lodders 1999a}Lodders, K 1999a, \apj, 519, 791
953: \bibitem[Lodders 1999b]{Lodders 1999b}Lodders, K J. 1999b, Chem. Phys. Ref. Data 28, 1705
954: \bibitem[Lodders 2002]{Lodders 2002}Lodders, K., 2002, \apj, 577, 974
955: \bibitem[Lodders 2003]{Lodders 2003}Lodders, K., 2003, \apj, 591, 1220
956: \bibitem[Lodders and Fegley 1996]{Lodders and Fegley 1996}Lodders, K.
957: and Fegley, B. 1996,\apj, 472, L37
958: \bibitem[Lodders and Fegley 2002]{Lodders and Fegley 2002}Lodders, K.
959: and Fegley, B. 2002, Icarus, 155, 393
960: \bibitem[Lodders 2004]{Lodders 2004}Lodders, K. 2004, Science, 303, 323
961: \bibitem[Lodders and Fegley 2006]{Lodders and Fegley 2006}
962: Lodders, K., and Fegley, B. 2006, In Astrophysics Update 2, J.W. Mason, ed., Springer/Praxis Publ. Ltd., Chichester, UK, pp 1
963: \bibitem[Ma et al. 2007]{Ma et al. 2007}Ma, Q., Tipping,
964: R.H., and Boulet, C. 2007, \jqsrt, 103, 588
965: \bibitem[Mannucci 1991]{Mannucci 1991}Mannucci, Anthony J. 1991,
966: JChemPhys,95,7795
967: \bibitem[Marley et al. 2002]{Marley et al. 2002}Marley, M.~S., Seager,
968: S., Saumon, D., Lodders, K., Ackerman, A.~S., Freedman, R.~S., \& Fan,
969: X.\ 2002, \apj, 568, 335
970: \bibitem[Massie and Hunten 1982]{Massie and Hunten 1982}Massie, S.~T.,
971: and Hunten, D.M. 1982, Icarus, 49, 214
972: \bibitem[McBride and Nicholls 1972]{McBride and Nicholls 1972}McBride,
973: J.O.P. and Nicholls, R.W. 1972, J.Phys. B, 5, 408
974: \bibitem[Mihalas 1978]{Mihalas 1978}Mihalas,D. 1978, Stellar
975: Atmospheres, (2nd ed; San Francisco, CA: W. H. Freeman)
976: \bibitem[Nassar and Bernath 2003]{Nassar and Bernath 2003}Nassar, Ray
977: and Bernath, Peter 2003 \jqsrt, 82, 279
978: \bibitem[Partridge and Schwenke 1997]{Partridge and Schwenke
979: 1997}Partridge, H., and Schwenke, D. 1997, JChemPhys, 106, 4618
980: \bibitem[Saumon et al. 2007]{Saumon et al. 2007} Saumon, D., et al.\
981: 2007, \apj, 656, 1136
982: \bibitem[Schwenke 1998]{Schwenke 1998}Schwenke, D. 1998, Faraday
983: Discussions, 109, 321
984: \bibitem[Sharp and Burrows 2007]{Sharp and Burrows 2007}Sharp, C.~M.,
985: and Burrows, A. 2007, \apjs, 168, 140
986: \bibitem[Strong et al. 1993]{Strong et al. 1993}Strong, K., Taylor,
987: F.~W., Calcutt, S.~B., Remedios, J.~J., and Ballard, J. 1993, \jqsrt, 50, 363
988: \bibitem[Tipping 1993]{Tipping 1993}Tipping, R. 1993 Private
989: Communication
990: \bibitem[Visscher et al.(2006)]{2006ApJ...648.1181V} Visscher, C., Lodders,
991: K., \& Fegley, B.~J.\ 2006, \apj, 648, 1181
992: \bibitem[Wattson 1996]{Wattson 1996}Wattson, R., 1996 Private
993: Communication
994: \bibitem[Wenger and Champion 1998]{Wenger and Champion 1998}Wenger,
995: Ch., and Champion, J.-P. 1998, \jqsrt, 59, 471
996: \bibitem[Zheng \& Borysow 1995]{Zheng & Borysow 1995} Zheng, C., \&
997: Borysow, A.\ 1995, \apj, 441, 960
998:
999: \end{thebibliography}
1000:
1001: \clearpage
1002: \begin{deluxetable}{ccccc}
1003: %\rotate
1004: \center
1005: \tablecolumns{5}
1006: \tablewidth{0pc}
1007: \tablecaption{Mean Opacities for $\rm [M/H]=0.0$}
1008: \tablehead{
1009: \colhead{T (K)} &\colhead{$\rm P(dyne\,cm^{-2}$)} & \colhead{$\rho(\rm g\, cm^{-3})$} & \colhead{$\kappa_{\rm R}(\rm cm^2\,g^{-1})$} & \colhead{$\kappa_{\rm P}(\rm cm^2\,g^{-1})$} }
1010: \startdata
1011: 75 & 3E+02 & 1.1277E-07 & 2.5619E-06 & 7.1083E-06 \\
1012: 75 & 3E+03 & 1.1277E-06 & 2.5589E-05 & 6.4309E-05 \\
1013: 75 & 1E+04 & 3.7591E-06 & 8.5261E-05 & 2.1238E-04 \\
1014: 75 & 3E+04 & 1.1277E-05 & 2.5571E-04 & 6.3555E-04 \\
1015: 75 & 1E+05 & 3.7591E-05 & 8.5211E-04 & 2.1167E-03 \\
1016: 75 & 3E+05 & 1.1277E-04 & 2.5557E-03 & 6.3485E-03 \\
1017: 75 & 1E+06 & 3.7591E-04 & 8.5180E-03 & 2.1160E-02 \\
1018: 75 & 3E+06 & 1.1277E-03 & 2.5553E-02 & 6.3478E-02 \\
1019: 75 & 1E+07 & 3.7591E-03 & 8.5176E-02 & 2.1159E-01 \\
1020: 100 & 3E+02 & 8.4584E-08 & 4.5393E-06 & 2.4757E-02 \\
1021: 100 & 3E+03 & 8.4583E-07 & 3.9962E-05 & 2.5407E-03 \\
1022: 100 & 1E+04 & 2.8193E-06 & 1.2854E-04 & 1.0837E-03 \\
1023: 100 & 3E+04 & 8.4582E-06 & 3.7709E-04 & 1.0589E-03 \\
1024: 100 & 1E+05 & 2.8193E-05 & 1.2345E-03 & 2.5780E-03 \\
1025: 100 & 3E+05 & 8.4582E-05 & 3.6583E-03 & 7.3903E-03 \\
1026: 100 & 1E+06 & 2.8193E-04 & 1.2104E-02 & 2.4401E-02 \\
1027: 100 & 3E+06 & 8.4582E-04 & 3.6260E-02 & 7.3044E-02 \\
1028: 100 & 1E+07 & 2.8193E-03 & 1.2088E-01 & 2.4334E-01 \\
1029: 100 & 3E+07 & 8.4582E-03 & 3.6261E-01 & 7.2982E-01 \\
1030: \enddata
1031: \tablecomments{Complete table on line.}
1032: \end{deluxetable}
1033: \clearpage
1034: \begin{deluxetable}{ccccc}
1035: %\rotate
1036: \center
1037: \tablecolumns{5}
1038: \tablewidth{0pc}
1039: \tablecaption{Mean Opacities for $\rm [M/H]=+0.3$}
1040: \tablehead{
1041: \colhead{T (K)} &\colhead{$\rm P(dyne\,cm^{-2}$)} & \colhead{$\rho(\rm g\, cm^{-3})$} & \colhead{$\kappa_{\rm R}(\rm cm^2\,g^{-1})$} & \colhead{$\kappa_{\rm P}(\rm cm^2\,g^{-1})$} }
1042: \startdata
1043: 75 & 3E+02 & 1.1313E-07 & 2.5527E-06 & 7.6946E-06\\
1044: 75 & 3E+03 & 1.1313E-06 & 2.5496E-05 & 6.4814E-05\\
1045: 75 & 1E+04 & 3.7710E-06 & 8.4951E-05 & 2.1216E-04\\
1046: 75 & 3E+04 & 1.1313E-05 & 2.5478E-04 & 6.3330E-04\\
1047: 75 & 1E+05 & 3.7710E-05 & 8.4894E-04 & 2.1074E-03\\
1048: 75 & 3E+05 & 1.1313E-04 & 2.5460E-03 & 6.3191E-03\\
1049: 75 & 1E+06 & 3.7710E-04 & 8.4845E-03 & 2.1060E-02\\
1050: 75 & 3E+06 & 1.1313E-03 & 2.5452E-02 & 6.3178E-02\\
1051: 75 & 1E+07 & 3.7710E-03 & 8.4837E-02 & 2.1059E-01\\
1052: 100 & 3E+02 & 8.4848E-08 & 4.5234E-06 & 2.4853E-02\\
1053: 100 & 3E+03 & 8.4846E-07 & 3.9824E-05 & 2.6108E-03\\
1054: 100 & 1E+04 & 2.8282E-06 & 1.2808E-04 & 1.1500E-03\\
1055: 100 & 3E+04 & 8.4846E-06 & 3.7575E-04 & 1.1219E-03\\
1056: 100 & 1E+05 & 2.8282E-05 & 1.2300E-03 & 2.6327E-03\\
1057: 100 & 3E+05 & 8.4846E-05 & 3.6443E-03 & 7.4225E-03\\
1058: 100 & 1E+06 & 2.8282E-04 & 1.2056E-02 & 2.4353E-02\\
1059: 100 & 3E+06 & 8.4846E-04 & 3.6115E-02 & 7.2772E-02\\
1060: 100 & 1E+07 & 2.8282E-03 & 1.2042E-01 & 2.4226E-01\\
1061: 100 & 3E+07 & 8.4846E-03 & 3.6123E-01 & 7.2650E-01\\
1062: \enddata
1063: \tablecomments{Complete table on line.}
1064: \end{deluxetable}
1065:
1066: \begin{deluxetable}{ccccc}
1067: %\rotate
1068: \center
1069: \tablecolumns{5}
1070: \tablewidth{0pc}
1071: \tablecaption{Mean Opacities for $\rm [M/H]=-0.3$}
1072: \tablehead{
1073: \colhead{T (K)} &\colhead{$\rm P(dyne\,cm^{-2}$)} & \colhead{$\rho(\rm g\, cm^{-3})$} & \colhead{$\kappa_{\rm R}(\rm cm^2\,g^{-1})$} & \colhead{$\kappa_{\rm P}(\rm cm^2\,g^{-1})$} }
1074: \startdata
1075: 75 & 3e+02 & 1.1260E-07 & 2.5662E-06 & 6.8130E-06\\
1076: 75 & 3e+03 & 1.1260E-06 & 2.5632E-05 & 6.4041E-05\\
1077: 75 & 1e+04 & 3.7535E-06 & 8.5404E-05 & 2.1245E-04\\
1078: 75 & 3e+04 & 1.1260E-05 & 2.5615E-04 & 6.3653E-04\\
1079: 75 & 1e+05 & 3.7535E-05 & 8.5363E-04 & 2.1208E-03\\
1080: 75 & 3e+05 & 1.1260E-04 & 2.5605E-03 & 6.3618E-03\\
1081: 75 & 1e+06 & 3.7535E-04 & 8.5345E-03 & 2.1205E-02\\
1082: 75 & 3e+06 & 1.1260E-03 & 2.5603E-02 & 6.3614E-02\\
1083: 75 & 1e+07 & 3.7535E-03 & 8.5343E-02 & 2.1205E-01\\
1084: 100 & 3e+02 & 8.4455E-08 & 4.5476E-06 & 2.4781E-02\\
1085: 100 & 3e+03 & 8.4453E-07 & 4.0027E-05 & 2.5126E-03\\
1086: 100 & 1e+04 & 2.8151E-06 & 1.2874E-04 & 1.0526E-03\\
1087: 100 & 3e+04 & 8.4453E-06 & 3.7770E-04 & 1.0280E-03\\
1088: 100 & 1e+05 & 2.8151E-05 & 1.2366E-03 & 2.5503E-03\\
1089: 100 & 3e+05 & 8.4453E-05 & 3.6650E-03 & 7.3736E-03\\
1090: 100 & 1e+06 & 2.8151E-04 & 1.2127E-02 & 2.4421E-02\\
1091: 100 & 3e+06 & 8.4453E-04 & 3.6329E-02 & 7.3175E-02\\
1092: 100 & 1e+07 & 2.8151E-03 & 1.2109E-01 & 2.4383E-01\\
1093: 100 & 3e+07 & 8.4453E-03 & 3.6325E-01 & 7.3143E-01\\
1094: \enddata
1095: \tablecomments{Complete table on line.}
1096: \end{deluxetable}
1097:
1098: \clearpage
1099:
1100: \begin{figure}
1101: \includegraphics[angle=270,scale=0.6]{f1}
1102: \caption{Opacity $\kappa$ at $10^6\,\rm dyne\,cm^{-2}$ (1 bar) and 1000 K as a function of wavenumber for two opacity databases
1103: of the $\rm CH_4$ molecule. Black data are computed from the standard HITRAN database. Red
1104: curve is derived from our computational database, as described in the text. }
1105: \end{figure}
1106:
1107:
1108: \begin{figure}
1109: \includegraphics[angle=90,scale=0.7]{f2}
1110: \caption{Contour plot showing Rosseland mean opacity over the temperature-pressure domain
1111: included in this study. Model temperature-pressure profiles from several planets and a brown dwarf
1112: are shown for comparison. Temperatures along profile refer to the effective temperature;
1113: brown dwarf profile is for an object with surface gravity $\log g \rm (cm^2/sec) = 5$.}
1114: \end{figure}
1115:
1116: \begin{figure}
1117: \includegraphics[angle=-90,scale=0.5]{f3}
1118: \caption{Rosseland mean opacity at seven densities as a function of temperature. Opacity data
1119: at each $(\rho, T)$ pair is computed by interpolation within our standard grid. Some discontinuities in this
1120: figure arise from interpolation over chemical equilibria and condensation boundaries. Densities range from $1\times10^{-8}$ (bottom) to
1121: $1\times10^{-2}\rm g/cm^3$. }
1122: \end{figure}
1123:
1124: \begin{figure}
1125: \includegraphics[angle=270,scale=0.6]{f4}
1126: \caption{Total opacity as a function of wavenumber at 1400 K and 1 bar. Thin curve gives opacity without the
1127: contribution of alkali atoms. Thicker curve shows summed opacity including alkalis, computed using the
1128: theory of Burrows, Marley, and Sharp (1999). Under these conditions the alkali opacity dominates at wavenumbers above
1129: about $10,000\,\rm cm^{-1}$. The derivative of the Planck function, $dB/dT$, which weights the
1130: opacity in the computation of the Rosseland mean, is shown as as a dotted line on an arbitrary log scale. }
1131: \end{figure}
1132:
1133: \begin{figure}
1134: \includegraphics[angle=270,scale=0.6]{f5}
1135: \caption{Rosseland mean opacity at three densities as a function of temperature. Opacity data
1136: at each $(\rho, T)$ pair is computed by interpolation within our standard grid. Some discontinuities in this
1137: figure arise from interpolation noise. Densities range from $1\times10^{-6}$ (bottom) to
1138: $1\times10^{-2}\rm g/cm^3$. Solid line denotes our standard calculation, the dotted line is for a case
1139: with the alkali opacity set equal to zero. The alkali opacity substantially increases the mean opacity at
1140: high densities and temperatures above about 1000 K. Differences in the lowest density case plotted
1141: are negligible. }
1142: \end{figure}
1143:
1144: \begin{figure}
1145: \includegraphics[angle=270,scale=0.6]{f6}
1146: \caption{Rosseland mean opacity as a function of density for three temperatures. Solid and broken
1147: lines are the ``no grain'' opacity from
1148: Ferguson et al. (2005) for three temperatures (specifically their case with $X=0.7$ and $Z=0.02$, filename=``cunha06.nog.7.02.tron''). This opacity database is optimized for protostellar disks and thus there is relatively little overlap in density space with the cool atmospheric conditions we consider here. The closest densities from our calculation (including alkali opacity) are shown as isolated symbols. There is excellent agreement at 2000 K. The trends for 1000 and 3000 K suggest reasonably good agreement between the two calculations.}
1149: \end{figure}
1150:
1151: \clearpage
1152:
1153:
1154:
1155: \end{document}
1156:
1157: