1: % aa.dem
2: % AA vers. 6.1, LaTeX class for Astronomy & Astrophysics
3: % demonstration file
4: % (c) Springer-Verlag HD
5: % revised by EDP Sciences
6: %-----------------------------------------------------------------------
7: %
8: %\documentclass[referee]{aa} % for a referee version
9: %\documentclass[onecolumn]{aa} % for a paper on 1 column
10: %\documentclass[longauth]{aa} % for the long lists of affiliations
11: %\documentclass[rnote]{aa} % for the research notes
12: %\documentclass[letter]{aa} % for the letters
13: %
14: \documentclass[structabstract]{aa}
15: %\documentclass[traditabstract]{aa} % for the abstract without structuration
16: % (traditional abstract)
17: %
18: \usepackage{graphicx}
19: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
20: \usepackage{txfonts}
21: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
22:
23: \newcommand{\TP}{\emph{$P$-$T$} profile}
24: \newcommand{\Tz}{\emph{T-z} profile}
25: \newcommand{\hd}{\object{HD~209458b}}
26:
27:
28:
29: %
30: \begin{document}
31: %
32: \title{TiO and VO broad band absorption features in the optical spectrum of the atmosphere of the hot-Jupiter \hd}
33: \subtitle{}
34: \titlerunning{TiO and VO broad band absorption features in the atmosphere of \hd} %
35: \authorrunning{J.-M. D\'esert et al.}
36:
37: \author{J.-M. D\'esert\inst{1} \and A. Vidal-Madjar\inst{1} \and A. Lecavelier des Etangs\inst{1} \and D.
38: Sing\inst{1} \and D. Ehrenreich\inst{2} \and G. H\'ebrard\inst{1} \and R. Ferlet\inst{1}}
39:
40: \institute{
41: Institut d'Astrophysique de Paris, CNRS (UMR~7095) -- Universit\'e Pierre \& Marie
42: Curie;
43: 98 bis, boulevard Arago 75014 Paris, France, \email{desert@iap.fr}
44: \and
45: Laboratoire d'astrophysique de l'observatoire de Grenoble, Universit\'e Joseph Fourier, CNRS (UMR 5571), BP53 38041 Grenoble cedex 9, France}
46:
47: \offprints{desert@iap.fr}
48:
49: \date{Received June xx, 2008; accepted September xx, 2008}
50:
51: % \abstract{}{}{}{}{}
52: % 5 {} token are mandatory
53:
54:
55: \abstract
56: % context heading (optional)
57: %{} %leave it empty if necessary
58: % aims heading (mandatory)
59: {The presence of titanium oxide (TiO) and vanadium oxide (VO) gas phase species is searched for in the atmosphere of the hot Jupiter \hd.}
60: % methods heading (mandatory)
61: {We compared a model for the planet's transmitted spectrum to
62: multi-wavelength eclipse-depth measurements (from 3\,000 to
63: 10\,000~\AA), obtained by Sing et al. (2008a) using archived
64: HST-STIS time series spectra. We make use of these observations to
65: search for spectral signatures from extra absorbers in the planet
66: atmosphere between 6\,000 and 8\,000~\AA.}
67: % results heading (mandatory)
68: {Along with sodium depletion and Rayleigh scattering recently
69: published for this exoplanet atmosphere, an extra absorber of
70: uncertain origin, redward of the sodium lines, resides in the
71: atmosphere of the planet. Furthermore, this planet has a
72: stratosphere experiencing a thermal inversion caused by the
73: capture of optical stellar flux by absorbers that resides at
74: altitude. Recent models have predicted that the presence of TiO
75: and VO in the atmosphere of \hd\ may be responsible for this
76: temperature inversion. Although no specific TiO and VO spectral band head signatures have
77: been identified unambiguously in the observed spectrum, we suggest
78: here that the opacities of those molecules are possible candidates
79: to explain the remaining continuous broad band absorption observed
80: between 6\,200 and 8\,000~\AA. To match reasonably well the data, the
81: abundances of TiO and VO molecules are evaluated from ten to one
82: thousand times below solar. This upper limit result is in
83: agreement with expected variations with altitude due to depletion
84: effects such as condensation.}
85:
86: % conclusions heading (optional), leave it empty if necessary
87: {}
88: \keywords{planetary atmospheres -- extrasolar planets -- HD209458b}
89: \maketitle
90: %
91:
92:
93: \section{Introduction}
94: \label{sec:intro}
95:
96: The discovery of transiting extrasolar giant planets (EGP) has
97: opened the window to the direct detections and characterization of
98: their atmospheres. Because of the wavelength-dependent opacities
99: of absorbing species, measurement of relative changes in eclipse
100: depth as a function of wavelength during primary transit has the
101: potential to reveal the presence (or absence) of specific chemical
102: species (\cite{Seager2000a}; \cite{Hubbard01}; \cite{Brown01}). In
103: the case of \hd's atmosphere, the transmission spectroscopy method
104: led to the detection of sodium (Charbonneau et~al. 2002). The
105: small atmospheric sodium signature has made its detection from
106: ground-based telescopes difficult. Only recently, Redfield et al.
107: (2008) and Snellen et al. (2008) reported the detection of
108: Na\,{\sc i} in the atmosphere of HD189733b and HD209458b,
109: respectively. Absorptions of several percents for H\,{\sc i}
110: Lyman-$\alpha$, O\,{\sc i} and C\,{\sc ii} have been measured in
111: the hydrodynamically escaping upper atmosphere (\cite{Madjar03},
112: 2004, 2008). More recently, Rayleigh scattering by H$_2$ molecules
113: has been identified (Lecavelier et~al. 2008b) and a
114: temperature-pressure (TP) profile with inversion was derived (Sing
115: et~al. 2008b). This temperature inversion leads to high
116: temperature at both low and high pressure. Note that this
117: temperature bifurcation was very well predicted by atmospheric
118: models of strongly irradiated planets (\cite{Hubeny03}). In the
119: lower part of the atmosphere ($\sim$30 mbar), the temperature is
120: found to be in the range $1\,900 -2\,400$~K (Lecavelier et~al.
121: 2008b), which corresponds to the M/L/T brown dwarf regime, as
122: expected for a hot Jupiter such as \hd\ (\cite{Kirkpatrick05}).
123:
124:
125: In the cool atmosphere of sub-stellar objects like brown dwarfs
126: and close-in EGPs, the strongest absorption features are expected
127: to be those of alkali metals like sodium and potassium
128: (\cite{Seager2000a}; \cite{Sudarsky00}; \cite{Burrows00}). This is
129: due to the condensation and the rain out of elements in the
130: atmosphere which clears the atmosphere of most of the metals and
131: keeps the less refractory alkali metals (\cite{Burrows99};
132: \cite{Lodders99}; \cite{Burrows02}; \cite{Sudarsky03}). Thus, in
133: contrary to the spectrum of HD189733b (Pont et al. 2007;
134: Lecavelier des Etangs et al. 2008a), the spectrum of \hd\ should
135: be dominated by absorption from sodium Na\,{\sc i} and potassium
136: K\,{\sc i}.
137:
138:
139: However, depending on the effective temperature, a large number of
140: diatomic and polyatomic molecules are predicted to be present
141: according to various models of brown dwarf and hot Jupiter
142: (\cite{Burrows99}; \cite{Lodders99}; \cite{Allard01};
143: \cite{Lodders02}; \cite{Hubeny03}; \cite{Burrows06}). Among those
144: molecules and at a temperature above 1\,800 K, titanium oxide
145: (TiO) and vanadium oxide (VO) in gas phase equilibrium are most
146: probably present with a high abundance in strongly irradiated
147: planet atmospheres (Seager et al. 1998; \cite{Hubeny03};
148: \cite{Fortney07}). Furthermore, the low albedo measurements of
149: \hd\ (\cite{Rowe06}) rule out most of the absorbents, but TiO and
150: VO. Recently, the interpretation of visible observations obtained
151: during the transit of this planet (Sing et al. 2008a), as well as
152: the interpretation of near infrared observations taken during its
153: secondary eclipse (Knutson et al. 2008) require that HD 209458b
154: experience an inversion in the atmospheric \TP\ and a have
155: stratosphere (Sing et al. 2008b; Burrows et al. 2007c; Burrows et
156: al. 2008). Such an inversion could be due to the absorption of the
157: incident visible light by TiO and VO molecules (Fortney et al.
158: 2007). Ongoing models (Burrows et al. 2008) claim that this
159: inversion must be caused by the capture of the incident optical
160: stellar flux by a stratospheric absorber that could potentially be
161: TiO and VO. Similarly other studies (\cite{Hubeny03};
162: \cite{Fortney07}) have more particularly investigated the effects
163: of TiO and VO which absorb much of the incoming stellar radiation
164: high in the atmospheres of the hottest CEGPs.
165:
166: In this work, we use {\it STIS} spectra obtained during planetary
167: transit at two spectral resolutions (low and medium) to search for
168: the direct evidence of the presence of TiO and VO in the planet's
169: atmosphere. Both datasets are combined to extend the measurements
170: over the entire optical regime to quantify other possible
171: absorbers appearing in the transmission spectrum. Rayleigh
172: scattering and sodium absorption have been proposed to explain the
173: spectral features observed between 3\,000~\AA\ and 6\,000~\AA\
174: (Lecavelier des Etangs et al. 2008b; Sing et al. 2008a; 2008b)
175: yet, additional absorptions remain unattributed from 6\,100~\AA\
176: up to 8\,000~\AA. These absorptions cannot be due to optically
177: thick high altitude clouds, which would otherwise mask the
178: detected signature of Rayleigh scattering (Lecavelier des Etangs
179: et al. 2008b). Since TiO and VO molecules are expected to be
180: present in high abundance for the range of temperatures and
181: pressures for strongly irradiated hot Jupiter, their opacities
182: could largely contribute to the observed absorption in this
183: wavelength domain (Seager et al.1998; \cite{Burrows99};
184: \cite{Sharp07}).
185:
186:
187: After a brief description of the observation and interpretation
188: status (Sect.~\ref{sec:obs}), we present, in
189: Sect.~\ref{sec:model}, the model together with the method used to
190: calculate the atomic and molecular opacities. Finally, we estimate
191: the contribution of TiO and VO to the observed transmission
192: spectrum for several abundance scenarios (Sect.
193: ~\ref{sec:analysis}) and discuss the consequences on the
194: atmospheric chemistry and aeronomy of HD209458b
195: (Sect.~\ref{sec:discuss}).
196:
197:
198:
199: %__________________________________________________________________
200:
201: \section{Observations}
202: \label{sec:obs}
203:
204:
205:
206: The \emph{HST-STIS} G750L, and G430L low resolution grating
207: observations of \hd\ analyzed here are also detailed in
208: \cite{Knutson07a}, \cite{Ballester07} and Sing et al. (2008a). For both
209: the G750L, and G430L gratings, two visits were done for each
210: grating, of five consecutive orbits each. Each visit span one
211: completed transit of \hd\ in 2003. Together they cover the
212: combined range 2\,900-10\,300~\AA, with some overlap around 5\,300~\AA\
213: with a resolving power $R=500$. While the broad-band
214: spectrophotometric analysis of \cite{Knutson07a} was used to
215: identify water absorption features in the atmosphere of \hd\ at
216: wavelength greater than 9\,000~\AA\ (\cite{Barman07}), here we
217: analyze the spectrum at full resolution.
218:
219:
220: We directly use \hd\ absorption depths (AD, see definition in
221: Sect.~\ref{sec:AD}) between 4\,000~\AA\ and 8\,000~\AA\ provided
222: by \cite{Sing08a}. The low resolution transit spectrum ratio, limb
223: darkening corrected, is plotted in Figure ~\ref{fig_atoms}. The
224: spectrum below 8\,000~\AA\ is considered here where the absorption
225: due to water molecules is negligible. Three prominent broadband
226: absorption features are revealed. The near ultraviolet (NUV)
227: absorption, at wavelength $\lambda$ $\le$ 5\,000~\AA, was first
228: reported by \cite{Ballester07} and explained by the absorption of
229: a hot hydrogen layer within the atmosphere. \cite{Lecavelier08b}
230: propose an alternative explanation invoking the Rayleigh
231: scattering by H$_2$ molecules. Within the same datasets,
232: \cite{Sing08b} found that the Na spectral line profile is
233: characterized by a wide absorption with a sharp transition to a
234: narrow absorption profile at higher altitudes values. This sharp
235: transition is interpreted by condensation or ionization which
236: deplete Na atoms in the upper atmosphere. Using a global fit to
237: these data, from 3\,000~\AA\ to 6\,200~\AA, \cite{Sing08b}
238: determine the average pressure-temperature profile ($P$-$T$) at
239: the planetary terminator. Two types of \TP\ were derived which
240: both can lead to the depletion of Na\,{\sc i} atoms. One profile
241: implies the condensation of Na\,{\sc i} and the other the
242: ionization of Na\,{\sc i}. We use this former one
243: (Fig.~\ref{fig_TP}). Both profiles experience hot temperatures,
244: above 1\,800~K at 50~mbar, due to Rayleigh scattering (Lecavelier
245: et al. 2008b). At such a temperature, TiO and VO are in gas phase.
246:
247: A third broad absorption feature in the range 6\,200 - 8\,000~\AA\
248: appears on the AD spectrum and is still unexplained. This
249: absorption is composed of two remarkable features visible in
250: Fig.~\ref{fig_atoms}:
251:
252:
253: \begin{enumerate}
254:
255: \item A strong and broad spectral absorption feature centered at
256: about 6\,250~\AA\ which is slightly smaller than for the Na\,{\sc i}
257: ($AD=0.0148\%$ over ~100~\AA). Other spectral signatures appearing
258: at higher wavelengths are weaker than the 6\,250~\AA\ one.
259:
260: \item A broad and continuous absorption level above 7\,000~\AA\
261: systematically higher than the lowest observational AD ($0.0146\%$
262: over 1500~\AA). This broad continuous flat absorption excess is
263: weaker than the spectral feature at 6\,250\AA\ but is equally as important
264: since it covers a large spectral domain, and has a much larger
265: absorption level than predicted by extrapolating Rayleigh
266: scattering and Na\,{\sc i} or K\,{\sc i} broad wings.
267:
268: \end{enumerate}
269:
270: We overplotted in Fig.~\ref{fig_atoms} the radii presented by
271: Knutson et al. (2008) that were derived using spectrophotometric
272: bandpass across roughly thousand angstrom bins. These radii are
273: found to be slightly different than those presented in this paper,
274: especially in the region of interest of this study (over
275: 6\,000~\AA), though the differences are either close to or within
276: the error bars of the two studies. This differences could be
277: explained by the supplementary corrections in the data treatement
278: employed by Sing et al. (2008) who corrected the transmission
279: spectrum for telluric contamination and various systematics,
280: quantified the level of systematic ‘red noise’ to take it into
281: account in the computed errors and finally applied a spectral limb
282: darkening correction before binning. Note that the systematic
283: errors have a weak wavelength dependance, affecting the absorption
284: depth values over large wavelength regions in a similar manner,
285: largely preserving the wavelength structure of narrow bands such
286: as the 6\,250~\AA\ feature.
287:
288:
289: Looking at the low resolution spectrum, we find the minimum AD of
290: 0.01453\% around 5\,000~\AA, which indicates a base pressure of
291: $33$ mbar in the \TP (Lecavelier des Etangs et al. 2008b).
292: Conversely, the maximum AD of 0.0149\% is obtained for the central
293: pixel of the Na\,{\sc i} line which corresponds to a pressure of
294: $1$ mb (Sing et al. 2008b).
295:
296: Below we present possible identifications of absorbers that can
297: simultaneously explain the spectral features between 6\,200 and
298: 7\,000 \AA\ and the broad relatively flat absorption excess from
299: 7\,000 to 8\,000~\AA.
300:
301:
302: %______________________________________________________________
303:
304: \section{Model description}
305: \label{sec:model}
306:
307: In order to interpret the transit observations, we developed a
308: model that calculates the absorptions of atomic and molecular
309: lines in the atmosphere of a transiting planet.
310:
311: Transit spectroscopy probes the transition region between the day
312: and night sides, called {\em limb}. We used the geometry of a
313: transiting system (\cite{Brown01}) and a model originally
314: developed by Ehrenreich et al.\ (2006) and adapted for this work.
315: Along a single cord, stellar photons cross several levels of the
316: spherically stratified atmosphere of the planetary limb.
317:
318: The total opacity, $\tau_{\lambda}$, along a cord (parallel to the
319: line of sight) as a function of its impact parameter $b$ and the
320: altitude $h$, is given as:
321:
322: \begin{equation}
323: \label{eq:opacity} \tau_{\lambda,i}(b) = 2 \int_{0}^{+\infty}
324: \sigma_{\lambda,i} n_i(h) \mathrm{d} l,
325: \end{equation}
326:
327: \noindent where $\sigma_{\lambda,i}$ is the cross section
328: coefficient for the species $i$ at the wavelength $\lambda$, and
329: $n_i(h)$ is the density of the species $i$ at an altitude $h$ in
330: the atmosphere (see Fig.~1 in Ehrenreich et al. 2006).
331:
332:
333: %______________________________________________________________
334:
335:
336: \subsection{Atomic and molecular line opacities}
337: \label{sec:opacities}
338:
339:
340: textbf{In the case of alkali atoms, we calculate the opacity using
341: the oscillator strength, the radiative lifetime and the
342: collisional broadening taken from Morton et al. (1991) and Iro et
343: al. (2005) which used Burrows et al. (2000) for determining the
344: line shapes. Further progresses have been achieved in the theory
345: of these profiles (Burrows et al. 2003; Allard et al. 2005). The
346: difference between these new line shapes and the ones used in this
347: study resides mainly in the far wings. However, the contribution
348: of other absorbents dominates the spectrum in the far wings of the
349: alkali lines. Thus, we found that the calculations of Morton et
350: al. (1991) and Iro et al. (2005) were precise enough for our
351: study.
352:
353:
354: In the case of molecules, we calculate the line strengths for each
355: line of each species using the available data. We derive the
356: strength at local thermal equilibrium (LTE), noted $S$, of the
357: molecular spectral line from the general equation
358: (\cite{Sharp07}):
359:
360: \begin{equation}
361: S = \frac{\pi e^2 g_if_{ij}}{m_ec} \frac{e^{-hcF_i/kT}}{Q(T)}
362: \left[1 - e^{-hc(F_j-F_i)/kT}\right]\, ,
363: \label{oscil}
364: \end{equation}
365:
366: \noindent where $g_i$ and $f_{ij}$ are the statistical weight of
367: the $i^{th}$ energy level and the oscillator strength for a
368: transition from that level to a higher level $j$, respectively.
369: The excitation energies $F_i$, and $F_j$ are the term values in
370: cm$^{-1}$ of the $i^{th}$ and $j^{th}$ levels participating in the
371: transition. In addition, $Q(T)$ is the partition function of the
372: species at some temperature $T$. We compute the partition function
373: as described by \cite{Sauval84}, considering a fourth degree
374: polynomial function of $\log(\frac{5040}{T})$. The molecular lines
375: are broadened using a Voigt profile. At pressure $P$, the
376: collisional broadening is calculated using (Sharp \& Burrows
377: 2007):
378:
379: \begin{equation}
380: \Delta\nu_{Lorentz}=w_oP,
381: \label{dnu_lorentz}
382: \end{equation}
383: with $w_o=0.1$ cm$^{-1}$~bar$^{-1}$
384:
385: Finally, we obtain the total monochromatic opacity of the
386: atmosphere by summing the individual contributions for each atom
387: and molecule weighted by their respective abundances at each
388: altitude level.
389:
390:
391: \subsection{Grid of opacities}
392: \label{sec:grid}
393:
394: Computing the opacities directly at each altitude with the
395: corresponding temperature and pressure would require intense
396: calculation. Instead, we precalculate a grid of opacities for each
397: molecule. Our grid is composed of 42 cells with a range of
398: temperature from 100 to 5\,000~K and a range of pressure from 1 to
399: $10^{-6}$ bar. Opacities have been computed from 2\,900 to
400: 10\,000~\AA\ with a resolution of 0.005~\AA\, which corresponds to
401: 5 million elements per cell. For a given \TP\, we then calculate
402: the corresponding opacities by applying a linear interpolation on
403: the grid cells.
404:
405:
406: \subsection{Absorption Depth (AD)}
407: \label{sec:AD}
408:
409:
410: The output of the model is a spectrum ratio defined by Brown
411: (2001) as the ratio of the flux received during the transit with
412: the flux received when the planet is not occulting the star of
413: radius $R_\star$. Otherwise indicated by the absorption depth
414: (AD), that is the occultation of the surface of the star by the
415: surface of the planet at each wavelength. The surface of the
416: planet is twofold: the optically thick disk of radius $R_P$ and
417: the wavelength-dependent occultation by the atmosphere that
418: surrounds the planetary disk expressed as an equivalent surface.
419: We used $R_{P}=1.32~R_{jup}$ and $R_{\star}=1.125~R_{\odot}$ from
420: \cite{Knutson07a}. \cite{Lecavelier08b} show that the observed AD
421: is well approximated by:
422:
423: \begin{equation}
424: AD_{\lambda}=AD_0(1+\frac{2H}{R_P}\ln\frac{\sigma_{\lambda}}{\sigma_{\lambda_0}})
425: \label{ADbin}
426: \end{equation}
427:
428: \noindent where AD$_0$ is AD at $\lambda=\lambda_0$ and H the
429: scale height. Thus the observed mean AD over a wavelength range of
430: a given spectral element (typically 50~\AA), is proportional to
431: the mean of the logarithm of the cross section. Therefore, we
432: calculate the effective cross section in a given spectral element
433: by averaging the logarithm of the cross section calculated at a
434: much larger resolution:
435:
436:
437: \begin{equation}
438: \sigma_{bin}[\lambda_i ; \lambda_j]=
439: exp \left( \int_{\lambda_i}^{\lambda_j} \frac{\ln
440: \sigma_{\lambda}}{\lambda_i - \lambda_j}d_{\lambda}\right)
441: \end{equation}
442:
443:
444:
445:
446: %_____________________________________________________________
447:
448: \begin{table}
449: \caption{List of the most probable species with their
450: corresponding solar abundances which have significant absorptions
451: in the 6\,000-8\,000~\AA\ wavelength domain (from \cite{Lodders02} and
452: \cite{Sharp07}). This list is ordered by the product of an average
453: of the cross section between 4\,000 and 8\,000~\AA\ and its
454: corresponding solar abundance. This ordered list gives a rough
455: estimate of the plausible detectability for each atomic and
456: molecular species.}
457:
458:
459:
460: \label{table:solar_abund} % is used to refer this table in the text
461: \centering % used for centering table
462: \begin{tabular}{c c c c} % centered columns (4 columns)
463: \hline\hline % inserts double horizontal lines
464: Element & $\chi_{\odot}$ & $<\sigma_{\lambda}>$ ($cm^{2}$)& $\chi_{\odot}<\sigma_{\lambda}>$ ($cm^{2}$) \\ % table heading
465: \hline % inserts single horizontal line
466: Na\,{\sc i} & 3.e-6 & 1.e-19 & 1.e-25 \\
467: K\,{\sc i} & 2.e-7 & 1.e-20 & 1.e-27 \\
468: Li\,{\sc i} & 1.e-9 & 1.e-19 & 1.e-28 \\
469: \hline % inserts single horizontal line
470: SiO & 1.e-5 & 1.e-17 & 1.e-22 \\
471: TiO & 1.e-7 & 1.e-16 & 1.e-23 \\
472: VO & 1.e-8 & 1.e-16 & 1.e-24 \\
473: MgH & 1.e-9 & 1.e-17 & 1.e-26 \\
474: H$_2$O & 1.e-3 & 1.e-24 & 1.e-27 \\
475: FeH & 1.e-9 & 1.e-19 & 1.e-28 \\
476: CaH & 1.e-11 & 1.e-17 & 1.e-28 \\
477: CrH & 1.e-10 & 1.e-19 & 1.e-29 \\
478:
479: \hline %inserts single line
480: \end{tabular}
481: \end{table}
482:
483: %______________________________________________
484:
485:
486:
487:
488: \section{Analysis}
489: \label{sec:analysis}
490:
491:
492: Spectral signatures, line profiles and abundances vary depending
493: on the temperature and pressure vertical profile in the
494: atmosphere. Here we use the vertical \TP\ presented in Figure
495: ~\ref{fig_TP} and two sodium mixing ratios from \cite{Sing08b} as
496: well as the Rayleigh scattering derived by \cite{Lecavelier08b}.
497: The AD of 0.01453\%, corresponding to a pressure of 33 mbar, is
498: taken as the reference at 1.32 Jupiter radius (Knutson {et~al.}
499: 2007). Using those hypothesis, we calculate the absorption
500: spectrum with only the abundances as free parameter. We vary the
501: abundances to find the best fit to the data using a $\chi^2$
502: minimization between the data and the model.
503:
504: \subsection{Atomic line:}
505: \label{sec:atoms}
506:
507: The most important absorptions by atomic lines are those of alkali
508: metals, Na\,{\sc i}, K\,{\sc i} and Li\,{\sc i} (see
509: Table~\ref{table:solar_abund}). This table provides the lines with
510: the strongest abundance weighted cross-section in the considered
511: wavelength range. None of this atomic lines, nor H$\alpha$, can
512: explain these observed spectral feature at 6\,250~\AA.
513:
514: In the low resolution transit spectrum, we do not detect
515: significant absorption in the K\,{\sc i} line at 7\,698~\AA\ nor in
516: the Li\,{\sc i} doublet at 6\,708~\AA. These non detections exclude
517: the possibility of solar abundance over the whole atmosphere for
518: these two species (see Fig.~\ref{fig_atoms}). Assuming a constant
519: mixing ratio, we constrain the K\,{\sc i} and Li\,{\sc i}
520: abundances to be lower than $2\times 10^{-3}$ and $2\times
521: 10^{-1}$ solar, respectively (1$\sigma$). Of course, these upper
522: limits does not exclude larger abundances below the altitude
523: corresponding to the observed absorption depth of $\sim$0.0145\%.
524:
525: We conclude that the absorption depth measured between 6\,200 and
526: 8\,000~\AA\ cannot be explained by the strong absorption lines
527: from abundant atomic species. The line profile of Na\,{\sc i} does
528: not affect the results on TiO/VO abundances since the remaining
529: broad band absorption begin at 200~\AA\ away from the center of
530: the line on the red side. In the following, we consider possible
531: absorptions by molecules.
532:
533:
534: \subsection{Molecular line:}
535: \label{sec:molecules}
536:
537: A list of plausible molecular species absorbing in the 6\,000~\AA\
538: region is presented in Table~\ref{table:solar_abund}.
539:
540: Although the molecules CH$_4$, NH$_3$ and CO are the most
541: important source of opacities in the infrared, their effects are
542: negligible in the visible wavelength domain; they are not
543: considered in the following analysis.
544:
545: Using the abundance weighted mean cross-section list, we found
546: that most plausible species are SiO, TiO and VO respectively.
547: Silicium oxide (SiO) absorption mainly occurs below 2\,000~\AA,
548: outside our wavelength domain. As a result, TiO and VO are
549: predicted to be the most abundant and the most important sources
550: of opacity in the present domain (\cite{Fortney06};
551: \cite{Lodders02}). However, since both species condense, their
552: abundances decrease below 1500K.
553:
554: Amongst all the other molecular absorbers that can be used to
555: explain the observation in this wavelength range, the hydride
556: metals CrH, FeH, MgH, and CaH were considered as potential
557: candidates to explain the shape of the spectrum. However, the
558: abundance of those elements drops with decreasing temperature
559: below 2\,000K, due to the formation of condensates
560: (\cite{Lodders06}; \cite{Kirkpatrick05}). Furthermore, their
561: absorption domain does not match the one considered here, like for
562: H$_2$O.
563:
564: Among plausible molecules in the atmosphere of \hd, TiO and VO
565: have the largest cross-section in the visible domain. We explore
566: the extent to which various abundances of TiO and VO can be used
567: to reproduce both the medium and low resolution spectra. We
568: derived the LTE strength $S$ (Eq.~\ref{oscil}) of the molecular
569: spectral lines from \cite{Plez98}. These data list about 3 million
570: lines from 4\,000 to nearly 30,000 cm$^{-1}$, covering the whole
571: visible part of the spectrum for the transitions of the most
572: abundant isotopes $^{48}$TiO, together with $^{46}$TiO,
573: $^{47}$TiO, $^{49}$TiO, and $^{50}$TiO, for nine different
574: electronic band systems. Using the grid previously described in
575: Sect.~\ref{sec:grid}, we compute theoretical AD including TiO and
576: VO absorption at each wavelength (see Fig.~\ref{fig_sigma}). At a
577: temperature of 2\,000~K and a pressure of 0.1 bar, the lines are
578: significantly broadened, thus the rapid fluctuations in
579: cross-section over short wavelength intervals are suppressed
580: revealing the main band features. At significantly lower
581: pressures, the broadening of the lines is much smaller, the
582: cross-section shows rapid variations as a function of wavelength
583: and the broad main features do not show up so clearly.
584:
585:
586: In the following sections, we explore different scenarios for the
587: atmospheric abundance of TiO/VO in order to adjust our model to
588: the observations. We fit the data with an atmospheric model
589: including Na I, Rayleigh scattering, and TiO and/or VO in
590: different abundances. In a first step, we will consider a unique
591: abundance of TiO/VO over the whole atmosphere. In a second step,
592: we will assume two levels of abundance of TiO/VO. We finally
593: discuss the fits obtained using those different abundances.
594:
595:
596:
597:
598: \subsection{TiO :}
599: \label{sec:TiO}
600:
601:
602:
603: As a first step, we consider only TiO and constrain the abundance
604: by fitting the observational low resolution spectrum ratio.
605: Starting with the simplest assumption, we consider a solar
606: abundance over the whole atmosphere of the planet.
607:
608: Our model gives an AD average of 0.155\% over the full bandpass
609: 4\,000-8\,000~\AA\ (see upper dashed-dotted line in
610: Fig.~\ref{fig_molecules}a). Such a value is significantly above
611: what is observed. Therefore, TiO abundance must be much lower than
612: solar above altitudes corresponding to the observed AD level.
613:
614: Assuming a constant mixing ratio, the best fit of the low
615: resolution data set over the region 4\,000-5\,700~\AA\ gives a TiO
616: abundance of $(8.\pm0.6)\times 10^{-4}$ solar. With this TiO
617: abundance, the $\chi^2$ is reduced from 1754 to 1207 for $n = 942$
618: degrees of freedom (using non rebinned spectra). However, this
619: abundance cannot reproduce the 6\,250~\AA\ spectral feature and the
620: flat absorption of the observations between 6\,200-8\,000~\AA. We
621: conclude that other absorbers must be considered to explain this
622: part of the AD spectral profile.
623:
624:
625:
626:
627: \subsection{TiO and VO:}
628: \label{sec:TiO_VO} Since VO is the most important source of
629: opacity after TiO (see Table~\ref{table:solar_abund}), we add this
630: molecule to our model. The chemistry of the vanadium is quite
631: similar to the one of titanium, where the monoxide is in the gas
632: phase at a temperature higher than the temperature of
633: condensation.
634:
635: In a first step, we consider a solar abundance over the whole
636: atmosphere. As for TiO, VO with solar abundance cannot reproduce
637: the observations, especially around 5\,000~\AA. The abundance of
638: VO must be lower. Assuming constant mixing ratios for both
639: molecules, the best fit gives for TiO $(6\pm0.6)\times 10^{-4}$
640: and for VO $(3\pm0.5)\times 10^{-2}$ as shown in
641: Fig.~\ref{fig_molecules}b. We obtain a satisfactory fit with a
642: $\chi^2$ of 817, or a $\chi^2/n \approx 0.86$ in the low
643: resolution data set. The continuous flat part of the AD curve can
644: be well reproduce by absorption of VO molecules. The region around
645: 5\,000~\AA\, already well reproduced with Rayleigh scattering and
646: with sodium lines absorptions, is not affected by the presence of
647: TiO and VO molecules. However, the spectral features around
648: 6\,250~\AA\ remains unexplained assuming constant mixing ratios
649: for both species. Note that sligthly lower abundances are
650: necessary for TiO and VO if we consider the Knutson et al.(2007)
651: results for the radii over 6\,000\AA.
652:
653:
654:
655:
656:
657:
658: \subsection{TiO, VO, and condensation:}
659: \label{sec:TiO_VO_cond}
660:
661:
662: With a large temperature gradient, TiO and VO abundances are
663: expected to vary along the vertical atmospheric profile.
664: Furthermore, the \TP\ derived by \cite{Sing08b} crosses the TiO
665: and VO condensation curve (See Fig.~\ref{fig_TP}). To check if
666: signature of condensation can be found in the present data set, we
667: assume in this section that TiO/VO molecules can be depleted above
668: a given altitude. Indeed, TiO and VO are expected to condense
669: into many Ti and V-bearing compounds as described in details by
670: Lodders (2002). To implement this assumption, we introduce a
671: maximum altitude, corresponding to an absorption depth of
672: condensation ($AD_{cond}$), above which the absorption by the
673: corresponding molecules in gas phase vanishes.
674:
675:
676: In the model, we derive the altitude of condensation from the
677: intersection between the \TP\ and the condensation curve. The
678: condensation curves of TiO and VO depend mainly on the
679: temperature (see Fig.~\ref{fig_TP} and \cite{Sharp07}). The TiO
680: condensation point occurs at a temperature of 1\,850~K and a
681: pressure of 0.02 bar (see Fig.~\ref{fig_TP}). This altitude
682: corresponds to an AD of condensation $AD_{cond}=0.0146\%$. In the
683: case of VO, the condensation point is at 1\,600~K and 0.01 bar,
684: leading to $AD_{cond}=0.0147\%$ (see condensation limits in
685: Fig.~\ref{fig_molecules}c).
686:
687: Using the \TP\ presented in Fig.~\ref{fig_TP}, TiO and VO
688: $AD_{cond}$ levels are lower than the 6\,250~\AA\ spectral feature
689: AD peak. However, both $AD_{cond}$ levels are found to be above
690: the observed average AD, above the continuous flat part between
691: 6\,500~\AA\ $-$ 7\,500~\AA, and significantly above the lowest
692: $AD=0.0144\%$ around the left wing of the Na\,{\sc i} at 5\,000~\AA.
693: Thus, these molecules cannot be in solar abundances below the
694: altitude of condensation, otherwise the resulting AD curve would
695: be at the level of the condensation limits. We fit the AD curve
696: between 4\,000-8\,000~\AA\ and obtain an abundance for TiO and VO
697: below their altitudes of condensation of $(8\pm0.5)\times 10^{-4}$
698: and $(6\pm0.4)\times 10^{-2}$ solar, respectively. We obtain a fit
699: with a similar $\chi^2$ as before (Sect.~\ref{sec:TiO_VO}) with
700: one level of abundance for TiO/VO over the whole atmosphere.
701:
702:
703: Although the "cold-trap" can explain the lowest observational ADs,
704: our assumption of null abundances above this altitude cannot
705: describe the observed 6\,250~\AA\ spectral signature. Thus TiO/VO
706: molecules should remain in the upper part of the atmosphere if the
707: feature is due to these species.
708:
709:
710:
711: \subsection{TiO and VO in two separated layers:}
712: \label{sec:TiO_VO_cond2}
713:
714: In this section, we explore the possibility of the presence of TiO
715: and VO molecules in the upper part of the atmosphere above the
716: condensations altitudes. In the lower part of the atmosphere, we
717: keep the abundances found in Sect.~\ref{sec:TiO_VO_cond}. Then a
718: level of null abundance is imposed, corresponding to the
719: condensation level. Finally another abundance level is chosen for
720: the upper part of the atmosphere, where the temperature rise up
721: and cross again the condensation curves of TiO/VO. In the upper
722: part of the atmosphere, where the pressure is lower, the TiO lines
723: are less pressure broadened, numerous narrow peaks appear in the
724: TiO opacity curve, in particular, a strong peak appears at 6\,200
725: \AA\ (Fig.~\ref{fig_molecules}d).
726:
727:
728: The best fit gives a TiO and VO respectively $(1\pm0.2)\times
729: 10^{-4}$ and $(1\pm0.2)\times 10^{-2}$ solar abundances in the
730: upper layer, above the altitudes of condensation. However the
731: $\chi^2$ does not significantly decrease (802 for 944 degrees of
732: freedom). The spectral features of the AD curve is now better
733: fitted with TiO, at least to first order, whereas VO molecules
734: reproduce mainly the continuous flat part. In that way, We can
735: reproduce more successfully the spectral feature, the continuous
736: flat part and match the Na\,{\sc i} blue wing simultaneously. Note
737: that the flat part of the spectrum ratio cannot be reproduced if
738: we choose a null abundance of TiO/VO in the lower part of the
739: atmosphere.
740:
741:
742: \begin{table}
743: \caption{Bests $\chi_{2}$ fits performed for various model and the derived TiO and VO abundances.}
744:
745: \label{table:Chi2} % is used to refer this table in the text
746: \centering % used for centering table
747: \begin{tabular}{c c c c} % centered columns (4 columns)
748: \hline\hline % inserts double horizontal lines
749: Model & $\chi_{2}$ & TiO $(\times 10^{-4}$ & V0 $(\times 10^{-2})$ \\ % heading
750: \hline % inserts single horizontal line
751: Only TiO & 1207 & $(8.\pm0.6)$ & $0$ \\
752: TiO and VO & 817 & $(6.\pm0.6)$ & $(3.\pm0.5)$ \\
753: TiO and VO with condensation & 816 & $(8.\pm0.5)$ & $(6.\pm0.4)$ \\
754: 2 levels of TiO and VO & 802 & $(1.\pm0.2)$ & $(1.\pm0.2)$ \\
755:
756: \hline %inserts single line
757: \end{tabular}
758: \end{table}
759:
760: %______________________________________________
761:
762:
763:
764:
765:
766: \section{Discussion}
767: \label{sec:discuss}
768:
769: We have highlighted that the \hd\ atmosphere is optically thick at
770: low pressures and this requires new absorbers between
771: 6\,000-8\,000~\AA. The observation of the Na\,{\sc i} in the
772: atmosphere exclude the presence of clouds at pressure under $30$
773: mbar in the observed limb of the planet. We have shown that
774: Li\,{\sc i} and K\,{\sc i} absorption cannot reproduce the broad
775: band absorptions observed. Nonetheless, we derive upper limits for
776: their abundances at $2\times 10^{-3}$ solar for K\,{\sc i} and
777: $2\times 10^{-1}$ for Li\,{\sc i}.
778:
779:
780: The Barman (2007) analysis of the Knuston et al. (2007) results
781: and Sing et al. (2008b) analysis both agree that remaining
782: absorbents are required to explain the continuous broad band
783: absorption observed between 6\,200 and 8\,000~\AA, since the radii
784: derived in both reductions are significantly above the model with
785: no TiO nor VO molecules (see Fig.~\ref{fig_atoms}). As a
786: consequence, the presence of remaining absorbents, possibly TiO
787: and VO, is mandatory to explain the observations. This result does
788: not depend on the \TP\ adopted. The abundances of TiO and VO are
789: only poorly constrained by any \TP. We decided to use the \TP\
790: derived by Sing et al. (2008) as an example. In that case, the
791: lower part of the \TP, i.e below the condensation curves, is the
792: most critical part for the determination of the TiO and VO
793: abundances. The abundances of TiO and VO molecules are mainly
794: constrained by the level of the observed AD curve between 6\,500
795: and 8\,000~\AA\ and by the pressure of the lower point. Indeed,
796: for this lower point, the TiO and VO lines are mostly pressure
797: broadened. Thus, the abundances of these species are mainly
798: sensitive to the pressure and not to the temperature in the lower
799: part of this profile. The \TP\ above the condensation curves
800: constrains mostly the spectral features, especially the
801: 6\,250~\AA\ one, as seen for the case two level of TiO and VO.
802: However, we do not detect any TiO/VO narrow spectrale features,
803: thus any \TP\ above the condensation curve can be adopted. With
804: the \TP\ used in this study, we have shown that no clear solutions
805: emerge wether condensation is taken into account, or if two levels
806: of TiO and VO molecules are considered (see
807: Table~\ref{table:Chi2}). Thus, other profiles could led to the
808: same conclusion for the presence of TiO and VO and/or of potential
809: additional absorbers.
810:
811:
812: Although no typical TiO and VO spectral signatures have been
813: identified unambiguously in the observed spectrum, we suggest that
814: the opacities of those molecules are the best candidates to
815: explain the remaining continuous broad band absorption observed
816: between 6\,200 and 8\,000~\AA. Theoretical absorption with models
817: including TiO/VO with abundances below solar were evaluated to
818: match the data reasonably well. Using the \TP\ from
819: \cite{Sing08b}, we derived upper limits for the TiO and VO
820: abundances. The model without TiO or with TiO but no VO molecules
821: give the worst $\chi_{2}$ solutions. We found that the abundance
822: of TiO should be around $10^{-4}$ to $10^{-3}$ solar, and the
823: abundance VO around $10^{-3}$ to $10^{-2}$ solar. The fits become
824: marginally worse on the blue side of the sodium wing when adding
825: TiO and VO (see Fig.~\ref{fig_molecules}). This is because the
826: \TP\ used in our work was obtained by fitting the observation on
827: the blue side of the sodium line with a model which contains only
828: Rayleigh scattering and sodium opactity (Sing et al. 2008b) which
829: helped limit the number of free parameters in the fit, such that a
830: convergent solution could be found easilly. This model fits
831: precisely the observed data for this part of the spectrum (see the
832: dotted line in Figs.~\ref{fig_atoms} and ~\ref{fig_molecules}).
833: The addition of the TiO and VO opacities in this model,
834: contributes to slightly decrease the quality of the fit on the
835: blue side of the sodium line (see continous line in
836: Fig.~\ref{fig_molecules}). This is, however, of marginal
837: consequence on the TiO and VO evaluations in our study since these
838: are mostly controlled by the main part of the observations in the
839: redder range of the spectrum. Considering the quality of the
840: observations this does not affect the TiO and VO evaluations.
841: Hence, the use of the \TP\ with only Rayleigh and sodium
842: calculated by Sing et al.2008 is sufficient to derive reasonably
843: good estimates of the TiO and VO abundances.
844:
845:
846: Two layers of TiO/VO in the atmosphere are better suited to
847: explain the 6\,250~\AA\ spectral feature, however the $\chi^2$ is
848: not significantly improved (Table~\ref{table:Chi2}). Note that,
849: the 6\,250~\AA\ spectral feature is seen in the two independent
850: spectra obtained at both low resolution, as mentioned here, and at
851: medium resolution as observed by \cite{Sing08a} in the G750M STIS
852: spectra.
853:
854:
855: The tholins, polyacetylenes, or various non-equilibrium compounds
856: at high-altitude could also be responsible for the temperature
857: inversion. However, the study of these molecules requires a full
858: non-equilibrium chemical model which is not the purpose of that
859: work. TiO and VO molecules are very good high-altitude absorber
860: candidates (\cite{Hubeny03}; \cite{Fortney07}). Fortney et al.\
861: (2008) highlighted the importance of gaseous TiO and VO opacity in
862: their model of highly irradiated close-in giant planets. They
863: define two classes of irradiated atmospheres. Those which are warm
864: enough to have a strong opacity due to TiO and VO gases (``pM
865: Class'' planets), and those that are cooler (``pL Class'' planets)
866: dominated by Na\,{\sc i} and K\,{\sc i}. Our possible detection of
867: TiO/VO and Na\,{\sc i} in the hot atmosphere of \hd\ confirm that
868: this planet is located in the transition region between the two
869: classes defined by these authors. As observed here, due to the
870: presence of TiO/VO, \hd's atmosphere absorbs incident energy
871: between 6\,500~\AA\ and 8\,000~\AA. Consequently, this planet has a
872: hot stratosphere (around 2\,500~K) with a temperature inversion
873: (\cite{Burrows07c}). As another consequence, because of the
874: thermal emission of the energy trapped by the TiO/VO absorption,
875: the planet appears very bright during the mid infrared secondary
876: eclipse (\cite{Knutson08}).
877:
878: The first titanium condensates appearing at high temperature and
879: high pressure are Ti$_3$O$_5$ and CaTiO$_3$ (\cite{Burrows99};
880: \cite{Lodders02}). The condensation of vanadium starts at lower
881: temperature (1\,600 K) than for titanium (1\,800 K). The vanadium
882: condenses into solid VO and then into V$_2$O$_3$. Thus, depletion
883: of TiO should start at lower altitude than for VO. Below 1\,600K,
884: TiO, VO and major refractory elements are absent, leaving
885: monoatomic Na\,{\sc i} and K\,{\sc i} to dominate the spectrum.
886: Nevertheless, sodium and potassium chlorides become increasingly
887: abundant with decreasing temperature, especially for KCl which is
888: the dominant K-bearing compound. Condensed potassium depletes the
889: atmosphere of atomic K\,{\sc i}, as seen with Na\,{\sc i}, leading
890: to reduced signatures. This could explain why Na\,{\sc i} is seen
891: in abundance in the low resolution data, but not K\,{\sc i}. In
892: the lower warmer atmosphere, where wide atomic K line wings would
893: be observable even at low resolution, TiO and VO likely make up
894: the surrounding continuum further masking the signature.
895:
896: An alternative \TP\ with Na ionization has been also proposed
897: (Sing et al. 2008b). A model with Na ionization leaves a wide
898: range of temperatures possible in the middle atmosphere. Within
899: the framework of equilibrium chemistry, this profile presents a
900: lower temperature gradient \textbf{(Fortney et al. 2003)}. The
901: minimum temperature is below the TiO/VO condensation curves, and
902: above the sodium condensation curve. We did not enter into a
903: detailed study of TiO/VO abundances with these \TP; however, we
904: note that these profiles would have the effect to increase the
905: abundance of TiO and VO molecules in the middle atmosphere. See
906: Sing et al. (2008b) for more details about ionization.
907:
908:
909:
910: If present in the upper part of the atmosphere, TiO/VO molecules
911: are not fully depleted by condensation. Thermochemical equilibrium
912: calculations with rainout (\cite{Burrows99}, \cite{Lodders02},
913: \cite{Hubeny03}) have shown that TiO and VO can exist at
914: high-$T$/low-$P$ points such as in the upper part of our \TP. The
915: presence of TiO and VO at high-$T$/high-$P$ points leads to a
916: situation in which there is two levels of TiO/VO. This has been
917: already proposed by Hubeny et~al. (2003) to explain the presence
918: of a temperature inversion in strongly irradiated planets
919: atmosphere's. Since the \TP\ used here crosses twice the Ti and V
920: condensation curves, two levels containing TiO/VO molecules in gas
921: phase could be separated by a middle level that is free of those
922: molecules, as proposed by Hubeny et al. 2003. Nonetheless, a cold
923: trap region is usually expected to deplete the upper low-$P$
924: region. Flushing out the upper atmosphere of TiO/VO would rule out
925: 6\,250~\AA\ feature as being due to TiO. However, for now, it has
926: not been proven that cold-trap effect would deplete those
927: molecules, especially for the hydrodynamically escaping type of
928: upper atmosphere (\cite{Madjar03}, 2004, 2008; Lecavelier 2004;
929: Lecavelier et al. 2007; \cite{Ehrenreich08}).
930:
931: New observations with higher signal-to-noise ratio and better resolution,
932: together with improved chemical models, are required to address
933: TiO and VO spectral absorption features and detect specific band
934: heads.
935:
936:
937:
938:
939:
940: %_____________________________________________________________
941: % BEGIN figures
942: %-------------------------------------------------------------
943:
944:
945:
946: \newpage
947: \begin{figure*}
948: % \begin{figure}[!thb]
949: %\centering
950: %\includegraphics[angle=90,width=15cm]{fig_molecules.eps}
951: %\includegraphics[angle=90,width=17cm]{fig_atoms_modif.eps}
952: \includegraphics[width=12cm]{fig_Atoms.ps}
953: %\resizebox{\hsize}{!}{\includegraphics[angle=90]{fig_molecules.eps}}
954:
955:
956: \caption{The low resolution STIS measurements of the planetary
957: transit absorption depth (AD) corrected from limb-darkenning
958: effects and binned by 60 pixels (histogram). The observed
959: 1$\sigma$ errorbar is plotted above the spectrum. The radii
960: derived by Knutson et al.(2007) with the corresponding error bars
961: and spectrale bins are plotted (squares). The dashed line
962: corresponds to the best fit model assuming Rayleigh scattering and
963: sodium absorption with the physical T-P profile plotted in
964: Fig.~\ref{fig_TP}. The difference between the dashed curve and the
965: observed AD spectrum for wavelength over 6\,000~\AA\ indicates
966: that remaning absorbents should be presents in this spectral
967: domain. Overplotted in continuous line, is the same model, with
968: atomic lithium and potassium with solar abundances.This model
969: cannot reproduce the observations above 6\,200~\AA.}
970: \label{fig_atoms}
971: \end{figure*}
972: \newpage
973:
974:
975:
976: \newpage
977: \begin{figure*}
978: \centering
979: \includegraphics[angle=90,width=10cm]{TPprofile_escape.ps}
980: \caption{The atmospheric Temperature-Pressure profile (\TP) with
981: error bars derived by \cite{Sing08b} used in this study. The
982: dashed lines correspond to the condensation curves for titanium,
983: vanadium and sodium. The three points of this T-P profile are
984: derived from the fit of the observed absorption depth curve. The
985: hot point at a pressure of 0.05~Bar is imposed by the Rayleigh
986: scattering (See Lecavelier et al. 2008).}
987: \label{fig_TP}%
988: \end{figure*}
989:
990:
991: \newpage
992: \begin{figure*}
993: \centering
994: %\includegraphics[angle=90,width=10cm]{sigma_TiO.ps}
995: \includegraphics[angle=90,width=10cm]{sigma_TiO_VO.ps}
996:
997: \caption{The $\log_{10}$ of the monochromatic cross-section
998: $\sigma$ (cm$^2$) as a function of wavelength for the
999: vibration-rotation transitions of TiO and VO. The contribution due
1000: to different isotopes is included. TiO has a
1001: strong absorption feature shortward of 7\,500~\AA, and has a strong
1002: peak near of 6\,200~\AA\ as observed in Fig.~\ref{fig_atoms}.}
1003: \label{fig_sigma}%
1004: \end{figure*}
1005:
1006:
1007: \newpage
1008: \begin{figure*}
1009: %\includegraphics[width=17cm]{fig_TiO_VO_Na_VariousAbund_style4.eps}
1010: %\includegraphics[width=17cm]{fig_TiO_VO_Na_VariousAbund_style4.eps}
1011: \includegraphics[width=17cm]{fig_TiO_VO_Na_Rayleigh_VariousAbund_style4.eps}
1012: \newline
1013: \caption{Same as Fig.~\ref{fig_atoms} with fits obtain with
1014: various models (continuous thick line). The observed 1$\sigma$
1015: error bars are plotted around 7\,800~\AA\ (see also Fig.10 of Sing
1016: et al. 2008). The dashed line shows the fit obtained using only
1017: the model with Rayleigh scattering and sodium absorption with the
1018: T-P profile (Fig.~\ref{fig_TP}). A remaining broad band
1019: absorption appears in between 6\,200 to 8\,000~\AA\ which could be
1020: attributed to TiO and VO opacities. a: Model with a constant
1021: mixing ratio of TiO. A constant solar abundance of TiO is excluded
1022: (upper dashed-dotted line). The best fit gives an abundance of
1023: $(8\pm0.6)\times 10^{-4}$ solar (red solid line). b: Model with a
1024: constant mixing ratio of TiO and VO. Solar abundance for VO is
1025: also excluded. The red solid line is the best fit which includes
1026: TiO and VO. c: TiO, VO, and condensation. The altitudes of
1027: condensation for titanium and for vanadium are both plotted (dash
1028: dotted lines) together with the best fit (red solid line). d: Best
1029: fit for two distinct levels containing TiO and VO (See
1030: Sect.~\ref{sec:TiO_VO_cond2}). No clear solution emerge in between
1031: the three last models (red solid line).} \label{fig_molecules}
1032: \end{figure*}
1033: \newpage
1034:
1035:
1036:
1037: %_____________________________________________________________
1038: % END figures
1039: %-------------------------------------------------------------
1040:
1041: % custom format ADS pour A&A
1042: % \\bibitem%[\Y]%{\1.1M%\Y}%\5.3l %\Y,%\j,%\V,%\p\n
1043:
1044:
1045:
1046: \begin{acknowledgements}
1047: We thank our anonymous referee and our editor for their comments
1048: that strengthen the presentation of our results. We also thank
1049: Adam Burrows, Jonathan Fortney and Mark Marley for useful
1050: discussions.
1051: \end{acknowledgements}
1052:
1053: \begin{thebibliography}{}
1054:
1055: \bibitem[{{Allard} {et~al.} 2001}]{Allard01}
1056: Allard, F., Hauschildt, P.~H., Alexander, D.~R., Tamanai, A., \& Schweitzer, A.\ 2001, \apj, 556, 357
1057:
1058: \bibitem[2005]{Allard et al.2005}Allard, N.~F., Allard, F., \& Kielkopf, J.~F.\ 2005, \aap, 440, 1195
1059:
1060: \bibitem[{{Ballester} {et~al.} (2007)}]{Ballester07}
1061: {Ballester}, G.~E., {Sing}, D.~K., \& {Herbert}, F. 2007, \nat,
1062: 445, 511
1063:
1064: \bibitem[{{Barman} 2007}]{Barman07}
1065: {Barman}, T. 2007, \apjl, 661, L191
1066:
1067: \bibitem[{{Brown} 2001}]{Brown01}
1068: {Brown}, T.~M. 2001, \apj, 553, 1006
1069:
1070: \bibitem[{{Burrows} \& {Sharp} 1999}]{Burrows99}
1071: {Burrows}, A. \& {Sharp}, C.~M. 1999, \apj, 512, 843
1072:
1073: \bibitem[{{Burrows} {et~al.} 2000}]{Burrows00}
1074: Burrows, A., Marley, M.~S., \& Sharp, C.~M.\ 2000, \apj, 531, 438
1075:
1076: \bibitem[{{Burrows} {et~al.} 2002}]{Burrows02}
1077: Burrows, A., Burgasser, A.~J., Kirkpatrick, J.~D., et al.\ 2002, \apj, 573, 394
1078:
1079: \bibitem[2003]{Burrows et al.2003}Burrows, A., \& Volobuyev, M.\ 2003, \apj, 583, 985
1080:
1081: \bibitem[{{Burrows} {et~al.} 2006}]{Burrows06}
1082: Burrows, A., Sudarsky, D., \& Hubeny, I.\ 2006, \apj, 650, 1140
1083:
1084: \bibitem[{{Burrows} {et~al.} 2007c}]{Burrows07c}
1085: Burrows, A., Hubeny, I., Budaj, J., Knutson, H.~A., \& Charbonneau, D.\ 2007, \apjl, 668, L171
1086:
1087: \bibitem[{{Burrows} {et~al.} 2008}]{Burrows08}
1088: Burrows, A., Budaj, J., \& Hubeny, I.\ 2008, \apj, 678, 1436
1089:
1090: \bibitem[{{Charbonneau} {et~al.} (2002)}]{Charbonneau02}
1091: Charbonneau, D., Brown, T.~M., Noyes, R.~W., \& Gilliland, R.~L.\ 2002, \apj, 568, 377
1092:
1093: \bibitem[{{Charbonneau} {et~al.} 2005}]{Charbonneau05}
1094: Charbonneau, D., Allen, L.~E., Megeath, S.~T., et al.\ 2005, \apj, 626, 523
1095:
1096: \bibitem[2006]{Ehrenreich et al.2006}Ehrenreich, D., Tinetti, G.,
1097: Lecavelier Des Etangs, A., Vidal-Madjar, A.,
1098: \& Selsis, F.\ 2006, \aap, 448, 379
1099:
1100: \bibitem[{{Ehrenreich} {et~al.} 2008}]{Ehrenreich08}
1101: Ehrenreich, D., Lecavelier Des Etangs,
1102: A., H{\'e}brard, G., et al.\ 2008, \aap, 483, 933
1103:
1104: \bibitem[{{Fegley} \& {Lodders} 1996}]{Fegley96}
1105: {Fegley}, B.~J. \& {Lodders}, K. 1996, \apjl, 472, L37+
1106:
1107: \bibitem[{{Fortney} {et~al.} 2003}]{Fortney03}Fortney, J.~J., Sudarsky, D., Hubeny, I.,
1108: et al.\ 2003, \apj, 589, 615
1109:
1110: \bibitem[{{Fortney} {et~al.} 2006b}]{Fortney06}Fortney, J.~J., Cooper, C.~S., Showman,
1111: A.~P., Marley, M.~S., \& Freedman, R.~S.\ 2006, \apj, 652, 746
1112:
1113: \bibitem[{{Fortney} {et~al.} 2007}]{Fortney07}
1114: Fortney, J.~J., Lodders, K., Marley,
1115: M.~S., \& Freedman, R.~S.\ 2008, \apj, 678, 1419
1116:
1117: \bibitem[{{Hubbard} {et~al.} 2001}]{Hubbard01}
1118: Hubbard, W.~B., Fortney, J.~J., Lunine, J.~I., et al.\ 2001, \apj, 560, 413
1119:
1120: \bibitem[{{Hubeny} {et~al.} 2003}]{Hubeny03}
1121: Hubeny, I., Burrows, A., \& Sudarsky, D.\ 2003, \apj, 594, 1011
1122:
1123: \bibitem[{{Iro} {et~al.} 2005}]{Iro05}
1124: Iro, N., B{\'e}zard, B., \& Guillot, T.\ 2005, \aap, 436, 719
1125:
1126: \bibitem[{{Knutson} {et~al.} (2007a)}]{Knutson07a}
1127: Knutson, H.~A., Charbonneau, D., Noyes,
1128: R.~W., Brown, T.~M., \& Gilliland, R.~L.\ 2007, \apj, 655, 564
1129:
1130: \bibitem[{{Knutson} {et~al.} 2008}]{Knutson08}
1131: Knutson, H.~A., Charbonneau, D., Allen,
1132: L.~E., Burrows, A., \& Megeath, S.~T.\ 2008, \apj, 673, 526
1133:
1134: \bibitem[{{Kirkpatrick} 2005}]{Kirkpatrick05}
1135: {Kirkpatrick}, J.~D. 2005, \araa, 43, 195
1136:
1137: \bibitem[2004]{Lecavelier des Etangs et al.2004}Lecavelier des Etangs, A.,
1138: Vidal-Madjar, A., McConnell, J.~C., \& H{\'e}brard, G.\ 2004, \aap, 418, L1
1139:
1140: \bibitem[2007]{Lecavelier Des Etangs2007}Lecavelier Des Etangs, A.\ 2007,
1141: \aap, 461, 1185
1142:
1143: \bibitem[{{Lecavelier Des Etangs} {et~al.} (2008a)}]{Lecavelier08a}Lecavelier Des Etangs, A.,
1144: Pont, F., Vidal-Madjar, A., \& Sing, D.\ 2008, \aap, 481, L83
1145:
1146: \bibitem[{{Lecavelier Des Etangs} {et~al.} (2008b)}]{Lecavelier08b}
1147: Lecavelier des Etangs, A., Vidal-Madjar, A., D\'esert, J.~-.,
1148: \& Sing, D.\ 2008, ArXiv e-prints, 805, arXiv:0805.0595
1149:
1150: \bibitem[{{Lodders} 1999}]{Lodders99}
1151: {Lodders}, K. 1999, \apj, 519, 793
1152:
1153: \bibitem[{{Lodders} 2002}]{Lodders02}
1154: {Lodders}, K. 2002, \apj, 577, 974
1155:
1156: \bibitem[{{Lodders} \& {Fegley} 2006}]{Lodders06}
1157: Lodders, K., \& Fegley, B., Jr.\ 2006, Astrophysics Update 2, 1
1158:
1159: \bibitem[{{Morton} {et~al.} (1991)}]{Morton91}
1160: {Morton}, D.~C. 1991, \apjs, 77, 119
1161:
1162: \bibitem[{{Plez} (1998)}]{Plez98}
1163: {Plez}, B. 1998, \aap, 337, 500
1164:
1165: \bibitem[2008]{Pont et al.2008}Pont, F., Knutson, H., Gilliland, R.~L.,
1166: Moutou, C., \& Charbonneau, D.\ 2008, \mnras, 385, 109
1167:
1168: \bibitem[2008]{Redfield et al.2008}Redfield, S., Endl, M., Cochran, W.~D.,
1169: \& Koesterke, L.\ 2008, \apjl, 673, L87
1170:
1171: \bibitem[{{Rowe} {et~al.} 2006}]{Rowe06}
1172: Rowe, J.~F., Matthews, J.~M., Seager, S., et
1173: al.\ 2006, \apj, 646, 1241
1174:
1175: \bibitem[{{Sauval} \& {Tatum} (1984)}]{Sauval84}
1176: {Sauval}, A.~J. \& {Tatum}, J.~B. \apjs, 56, 193
1177:
1178: \bibitem[{{Sharp} \& {Burrows} 2007}]{Sharp07}
1179: {Sharp}, C.~M. \& {Burrows}, A. 2007, \apjs, 168, 140
1180:
1181: \bibitem[1998]{Seager et al.1998}Seager, S.,
1182: \& Sasselov, D.~D.\ 1998, \apjl, 502, L157
1183:
1184: \bibitem[{{Seager} \& {Sasselov} 2000}]{Seager2000a}
1185: {Seager}, S. \& {Sasselov}, D.~D. 2000, \apj, 537, 916
1186:
1187: \bibitem[2008]{Snellen et al.2008}Snellen, I.~A.~G., Albrecht, S., de
1188: Mooij, E.~J.~W., \& Le Poole, R.~S.\ 2008, \aap, 487, 357
1189:
1190: \bibitem[{{Sudarsky} {et~al.} 2000}]{Sudarsky00}
1191: Sudarsky, D., Burrows, A.,
1192: \& Pinto, P.\ 2000, \apj, 538, 885
1193:
1194: \bibitem[{{Sudarsky} {et~al.} 2003}]{Sudarsky03}
1195: Sudarsky, D., Burrows, A.,
1196: \& Hubeny, I.\ 2003, \apj, 588, 1121
1197:
1198: \bibitem[{{Sing} {et~al.} (2008a)}]{Sing08a}
1199: Sing, D.~K., Vidal-Madjar, A., Desert,
1200: J.~-., Lecavelier des Etangs, A.,
1201: \& Ballester, G.\ 2008, ArXiv e-prints, 802, arXiv:0802.3864
1202:
1203: \bibitem[{{Sing} {et~al.} (2008b)}]{Sing08b}
1204: Sing, D.~K., Vidal-Madjar, A., Lecavelier
1205: des Etangs, A., et al.\ 2008, ArXiv e-prints, 803, arXiv:0803.1054
1206:
1207: \bibitem[{{Vidal-Madjar} {et~al.} 2003}]{Madjar03}Vidal-Madjar, A., Lecavelier des
1208: Etangs, A., D{\'e}sert, J.-M., et al.\ 2003, \nat, 422, 143
1209:
1210: \bibitem[{{Vidal-Madjar} {et~al.} 2004}]{Madjar04}Vidal-Madjar, A., D{\'e}sert, J.-M.,
1211: Lecavelier des Etangs, A., et al.\ 2004, \apjl, 604, L69
1212:
1213: \bibitem[{{Vidal-Madjar} {et~al.} 2008}]{Madjar08}Vidal-Madjar, A., Lecavelier des
1214: Etangs, A., D{\'e}sert, J.-M., et al.\ 2008, \apjl, 676, L57
1215:
1216:
1217: \end{thebibliography}
1218:
1219: \end{document}
1220: