1: \documentclass[12pt,preprint]{aastex}
2: \shortauthors{Carciofi et al. }
3: \shorttitle{}
4:
5: \newcommand{\teff}{T_{\rm eff}}
6: \newcommand{\um}{\;\mu\rm m}
7: \newcommand{\bw}{\bar{\varpi}}
8:
9: \begin{document}
10:
11: \title{On the Determination of the Rotational Oblateness of Achernar}
12:
13: \author{
14: A.~C.~Carciofi \altaffilmark{1},
15: A. Domiciano de Souza\altaffilmark{2},
16: A. M. Magalh\~aes\altaffilmark{1},
17: J. E. Bjorkman\altaffilmark{3},
18: F.~Vakili\altaffilmark{2}
19: }
20:
21: \altaffiltext{1}{Instituto de Astronomia, Geof{\'\i}sica e Ci{\^e}ncias
22: Atmosf{\'e}ricas, Universidade de S\~ao Paulo, Rua do Mat\~ao 1226,
23: Cidade Universit\'aria, S\~ao Paulo, SP 05508-900, Brazil}
24: \altaffiltext{2}{LUAN, Universit{\'e} de Nice-Sophia Antipolis (UNSA),
25: CNRS, Observatoire de la C{\^o}te d'Azur (OCA), 28 avenue de Valrose,
26: Parc Valrose, F-06108 Nice - France} \altaffiltext{3}{University
27: of Toledo, Department of Physics \& Astronomy, MS111 2801 W.
28: Bancroft Street Toledo, OH 43606 USA}
29: \begin{abstract}
30:
31: The recent interferometric study of Achernar, leading to the
32: conclusion that its geometrical oblateness cannot be explained in
33: the Roche approximation, has stirred substantial interest in the community, in view
34: of its potential impact in many fields of stellar astrophysics.
35: It is the purpose of this paper to reinterpret the interferometric
36: observations with a fast rotating, gravity
37: darkened central star surrounded by a small equatorial disk, whose
38: presence is consistent with contemporaneous spectroscopic data.
39: We find that we can only fit the available data assuming a critically rotating central star.
40: We identified two different disk models that
41: simultaneously fit the spectroscopic, polarimetric, and
42: interferometric observational constraints: a tenuous disk in
43: hydrostatic equilibrium (i.e., with small scaleheight) and a smaller, scaleheight enhanced disk.
44: We believe that these relatively small disks correspond to the
45: transition region between the photosphere and the circumstellar
46: environment, and that they are probably perturbed by some photospheric mechanism.
47: The study of this interface between photosphere and circumstellar
48: disk for near-critical rotators is crucial to our understanding of
49: the Be phenomenon, and the mass and angular momentum loss of stars
50: in general. This work shows that it is nowadays possible to
51: directly study this transition region from simultaneous
52: multi-technique observations.
53:
54: \end{abstract}
55:
56: \keywords{polarization --- interferometry --- spectroscopy --- stars: emission
57: line, Be --- stars: individual (Achernar)}
58:
59: \section{Introduction \label{introduction}}
60:
61: Interferometry greatly increased our knowledge of the Be Star
62: Achernar ($\alpha$ Eri, HD10144). \citet[][hereafter D03]{dom03}
63: observed Achernar during fall 2002 with the ESO-VLTI/VINCI instrument \citep{ker03}
64: in the $K$ band and found that the star is highly
65: oblate. By converting the individual visibilities into equivalent
66: uniform disk angular diameters they derived a ratio between the
67: maximum and minor elongation axis $a/b=1.56\pm0.05$. Later,
68: \citet{ker06} detected a tenuous polar wind and very recently
69: \citet{ker07} found that Achernar has a main-sequence, lower mass
70: companion.
71:
72: The available interferometric data, together with the abundance of
73: data from the literature, forms a body of information that is
74: probably unrivaled by that of any other Be star. There is,
75: however, much that is still unknown about the star and its
76: circumstellar environment. As pointed out by D03, the
77: determination of the actual stellar oblateness from the
78: interferometric data is not a simple task, since there are so many
79: unknowns involved. In their original work D03, based on
80: simultaneous spectroscopic data, assumed that there was no
81: circumstellar material at the time of the observations, and that
82: the observed shape was purely photospheric in origin. Their
83: modeling of the observations with a rotationally deformed, gravity darkened star,
84: assuming the Roche approximation of uniform rotation and centrally condensed
85: mass,
86: led to the conclusion that the stellar flattening required to explain the
87: observations exceeds that of a critical rigid rotator .
88:
89: Recent theoretical developments in the theory of differentially
90: rotating stars (e.g., \citealt{jac04}) show that they can have
91: equatorial radii larger than the Roche limit, which is physically
92: compatible with the interferometric results. However,
93: \citet{vin06} demonstrated the existence of a tiny, yet non
94: negligible emission in H$\alpha$ at the time of the VLTI observations,
95: an indication that there was some circumstellar material around
96: Achernar, probably associated with a small rotating disk.
97: Furthermore, \citet{vin06} showed, from temporal analysis of
98: profile variations of the line He I $\lambda6678$ \AA, that orbiting gas clouds
99: are a very frequent feature of Achernar, even during its quiescent
100: phase.
101:
102: The presence of this
103: circumstellar material raises the question of how much it can
104: contribute to the observed visibilities. In this paper we
105: investigate how the presence of a small disk around Achernar may
106: alter to the interferometric signal, and the effects this may
107: have on determination of fundamental parameters such as stellar
108: rotation rate and stellar flattening.
109:
110: \section{Models \label{models}}
111:
112: To compute the emergent spectrum of the system we use the computer
113: code HDUST \citep{car06a,car06b}. This code solves the coupled
114: problem of the non-local thermodynamic equilibrium (NLTE) and
115: radiative equilibrium for arbitrary three-dimensional envelope geometries,
116: using the Monte Carlo method. Given a prescription for the star
117: (rotation rate, shape, flux distribution, effective temperature,
118: luminosity, etc.) and for the circumstellar gas (density and
119: velocity distribution) the code calculates the NLTE Hydrogen level
120: populations and the electron kinetic temperature. Once those
121: quantities are determined, the emergent spectrum (spectral energy
122: distribution, line profiles, polarization, synthetic images, etc.)
123: is calculated.
124:
125: For the present work, we modified HDUST to treat non-spherical,
126: gravity darkened stars. We assume the stellar shape to be an
127: spheroid, which is a reasonable approximation for the shape of a
128: rigidly rotating star \citep{fre05}. Given the stellar rotation
129: rate, $\Omega/\Omega_\mathrm{crit}$, the ratio between the
130: equatorial and polar radii is determined from the Roche
131: approximation for the stellar surface equipotentials
132: \citep{fre05}. For the gravity darkening we use the standard von
133: Zeipel flux distribution \citep{von24}, according to which
134: $F(\theta) \propto g_\mathrm{eff}(\theta) \propto
135: T_\mathrm{eff}^4(\theta)$, where $g_\mathrm{eff}$ and
136: $T_\mathrm{eff}$ are the effective gravity and temperature at
137: stellar latitude $\theta$. In HDUST, the star is divided into a
138: number of latitude bins (typically 50) which have an associated
139: $T_\mathrm{eff}$ and $g_\mathrm{eff}$ and emit with an spectral
140: shape given by the corresponding Kurucz model atmosphere
141: \citep{kur94}.
142:
143: For the disk density distribution we assume the following expression
144: \begin{equation}
145: \rho(\varpi,z)=\frac{\rho_0R_\mathrm{eq}}{\sqrt{2\pi}H}
146: \left(\frac{R_\mathrm{eq}}{\varpi}\right)^{2} \exp\left(-z^2/H^2\right),
147: \end{equation}
148: which is similar to the density distribution of an isothermal
149: viscous decretion disk in vertical hydrostatic equilibrium (e.g.
150: \citealt{car06b}). In the above equation, $\varpi$ is the radial
151: distance in cylindrical coordinates, $R_\mathrm{eq}$ is the
152: equatorial radius of the star and $\rho_0$ is the disk density
153: scale. We write the disk vertical scaleheight, $H$, as
154: \begin{equation}
155: H=H_0 \left(\varpi/R_\mathrm{eq}\right)^{1.5},
156: \label{eq:H}
157: \end{equation}
158: where $H_0$ is the scaleheight at the base of the disk. For
159: isothermal disks in vertical hydrostatic equilibrium $H_0=aV_{\rm
160: crit}^{-1}R_\mathrm{eq}$, where $a$ is the sound speed and the
161: critical velocity, $V_{\rm crit} \equiv \left(GM/R_{\rm eq}\right)^{1/2}$, is the Keplerian orbital speed at
162: the stellar surface.
163:
164: We chose a set of fixed stellar parameters, listed in
165: Table~\ref{tab:star}. The value for the polar radius came directly
166: from the interferometry, once we fixed the stellar inclination
167: angle to be $65\deg$ after the recent work of \citet{car07}. It is
168: worthy noting, however, that the results we show below are little
169: affected if we allow the inclination angle to vary within a
170: reasonable range (say, $\pm 5\deg$).
171: The value for the stellar luminosity was obtained by fitting the available photometric data (see caption of Figure~\ref{fig:sed}).
172:
173: Each model of the star plus
174: disk system has 4 free parameters: (1) the stellar angular
175: rotation rate, $\Omega/\Omega_\mathrm{crit}$, (2) the disk density
176: scale, $\rho_0$, (3) the disk outer radius, $R_\mathrm{d}$, and
177: (4) the disk scaleheight, $H_0$.
178:
179: We have two well-known observational constraints, viz., the 2002
180: interferometric observations (D03) and the 2002 H$\alpha$ line
181: profile (\citealt{vin06}, fig. 12).
182: Unfortunately, no
183: contemporaneous measurement of the linear polarization exists, but
184: we can impose, with a reasonable level of confidence, an upper
185: limit for the polarization based on the recent results of
186: \citet{car07}. In that paper, the results of a
187: polarization monitoring of Achernar, carried out between 2006 July and
188: November, is reported. Those measurements were taken during a period in
189: which Achernar was active, with a tenuous circumstellar disk, as
190: indicated by the weak H$\alpha$ emission (\citealt{car07}, fig. 3). We adopt as an upper limit for the
191: polarization in our modeling the lowest value reported in
192: \citet{car07}, which was 0.12\% in the $B$ band. As we shall see
193: below, fixing this upper limit for the polarization level has
194: important consequences for the modeling.
195:
196: \section{Results \label{results}}
197:
198: Our modeling procedure is as follows. For a given set of
199: $\Omega/\Omega_\mathrm{crit}$, $\rho_0$, $R_\mathrm{d}$ and $H_0$
200: we calculate synthetic images in the $K$ band, centered at
201: $\lambda=2.15\;\mu\rm m$, the H$\alpha$ line profile, the
202: continuum polarization and the spectral energy distribution (SED).
203: The model amplitude visibilities where obtained from the Fourier
204: transform moduli of the model images at $2.15\;\mu\rm m$,
205: normalized by %the total flux $F^\lambda$.
206: the total $K$ band flux of the synthetic image.
207:
208: Since the H$\alpha$ emission from October 2002 is very small and
209: the true photospheric profile of Achernar is not known, we model,
210: instead of the observed profile, the residual emission profile of
211: \citet{vin06}, fig. 12. This profile was obtained by subtracting
212: from the October 2002 observations the average profile of the 1999
213: period, which is believed to be purely photospheric. The
214: equivalent width (hereafter EW) of this emission profile is
215: $-0.29$ \AA.
216:
217: In Figure~\ref{fig:models} we show the visibility curves along the
218: polar and equatorial directions together with the corresponding
219: $K$ band images for five representative models that were chosen to
220: illustrate different aspects of our solution.
221: In Table~\ref{tab:models} we list the model parameters, along with
222: some model results, such as the $B$ band polarization level, the
223: H$\alpha$ EW and the $K$ band flux excess, $E_K$, which is defined
224: as $F^\lambda/F^\lambda_\star-1$, where $F^\lambda_\star$ is the
225: stellar flux without the disk at wavelength $\lambda$.
226:
227: Let us begin discussing models 1 to 3, for which the star was
228: assumed to be rotating critically. We have used
229: $\Omega/\Omega_\mathrm{crit} = 0.999999$
230: ($V_\mathrm{eq}/V_\mathrm{crit}=0.9993$) instead of 1 to avoid the
231: unphysical situation of having $T_{\rm eff} = 0$ in the stellar
232: equator. Model 1 corresponds to a small and relatively dense disk,
233: with $H_0$ given by Eq.~\ref{eq:H}, i.e., model 1 corresponds to a
234: disk in vertical hydrostatic equilibrium. For this model the
235: visibility curves matches very well the observations and the
236: polarization is within the adopted limit, but the H$\alpha$
237: emission is very weak, as a result of the small disk size.
238:
239: One way to increase the H$\alpha$ emission is to make the disk larger.
240: For model 2 we adopted a larger value for $R_d$, but the density had to be decreased in order to keep the polarization within the adopted limit.
241: This model reproduces well both the visibility curves and the H$\alpha$ EW. %Note, however, that the model line profile differs substantially from the observed one (see \S~\ref{conclusions}, below).
242:
243: Another way of increasing the line emission is to raise the disk scaleheight.
244: If one raises $H$ by a factor of, say, $k$, the disk density scale must be decreased by a factor $k^{1/2}$ to keep the polarization approximately constant.
245: Model 3 corresponds to our best scaleheight enhanced model for a critically rotating star.
246: The visibility curves and the H$\alpha$ EW are well reproduced by the model, and the polarization is within the adopted upper limit.
247:
248: We also studied models with lower values of $\Omega/\Omega_\mathrm{crit}$. In this case, our attempts to fit simultaneously the visibility curves, polarization and H$\alpha$ EW were unsuccessful.
249:
250: Models 4 and 5 are examples of the results we have obtained for
251: sub-critical stars. Model 4 corresponds to a small and dense disk
252: in vertical hydrostatic equilibrium. This model reproduces quite
253: well the visibilities, but have too large $B$ band polarization and no H$\alpha$ emission. A
254: scaleheight enhanced model with lower density, such as model 5,
255: have better values for polarization and H$\alpha$ EW, but does not
256: have enough IR flux in the equator to account for the observed
257: size.
258:
259: Our lack of success on fitting the observations with sub-critical
260: models is a direct result of two physical effects, both related to
261: the presence of gravity darkening. For the critical models 1 to 3,
262: the bolometric stellar flux at the equator is about 1100 times
263: lower than the flux at the pole. Since the disks are relatively
264: small and the star oblate, light from the pole cannot reach the
265: disk and, as a result, the net scattered flux is small. As
266: $\Omega/\Omega_\mathrm{crit}$ decreases, the flux at the equator
267: rises, and the star becomes less oblate. Both mechanisms cause a
268: significant increase in the scattered flux and, thus, in the
269: polarization level.
270:
271: \section{Discussion \label{conclusions}}
272:
273: The presence of a small disk around Achernar is a realistic
274: possibility to explain the interferometric observations. As shown
275: by D03 (see also Figure~\ref{fig:sed}), Roche models without a disk cannot
276: account for the observed aspect ratio.
277:
278: The combined constraints imposed by the interferometry,
279: spectroscopy and the adopted upper limit for the polarization have
280: allowed us to impose very narrow limits on the model parameters,
281: within our assumption of a rigidly rotating star in the Roche
282: approximation surrounded by a small disk. The main result is that the star must be rotating
283: very close to critical, since all of our models with
284: $\Omega/\Omega_\mathrm{crit} < 0.992$ that successfully reproduced
285: both the visibilities and the $H\alpha$ EW (e.g. model 5) had too
286: large polarization levels, for the reasons discussed above.
287:
288: We have found two critical models that reproduce equally well the observations: a large and dense disk in hydrostatic equilibrium (i.e., geometrically thin; model 2) and a small and more tenuous disk with enhanced scaleheight (i.e., geometrically thick; model 3).
289: Because the lack of contemporaneous polarization measurements, the model parameters shown in Table~\ref{tab:models} have a degree of uncertainty that is
290: difficult to estimate. This stresses the importance of having simultaneous multi-technique data.
291:
292: The current paradigm for Be star disks is that of a geometrically thin viscous
293: Keplerian disk in vertical hydrostatic equilibrium. This paradigm
294: has been corroborated by several recent studies
295: \citep[e.g.][]{car06b,mei07}.
296:
297: When a star is close to critical rotation, the stellar photospheric material near the
298: equator, which is weakly bound to the star because of the strong
299: centrifugal force, can eventually escape the star provided it is
300: given an extra energy by some other mechanism (e.g. stellar
301: pulsation, interaction in a binary system, photospheric activity).
302: This material is likely to have, initially, a very complicated
303: density and velocity distribution. As the gas diffuses outwards as
304: a result of viscosity, the density and velocity distributions will
305: tend to relax and become in hydrostatic equilibrium.
306:
307: Thus, it is reasonable to assume that there is a \emph{transition
308: region} between the photosphere (which is characterized by large
309: densities of the order of $10^{-10}\;\rm g\; cm^3$) and the disk
310: itself, which has densities at least 10 times lower. Many unknowns
311: exist concerning the size and physical properties
312: (e.g. density, temperature, velocity field) of this smooth
313: transition region.
314:
315: With the available data we were able to narrow down substantially the range of possible values for Achernar's model parameters, but we cannot distinguish yet between model 2 and 3, i.e., between a model in hydrostatic equilibrium and a model with enhanced scaleheights that might result from the perturbation of the gas by some photospheric mechanism.
316:
317: As shown in Figure~\ref{fig:sed}, high-precision spectropolarimetry might add invaluable information, since models 2 and 3 have different values for the size of Balmer and Pashen jumps in polarization.
318: Also, with simultaneous spectro-interferometry and spectropolarimetry one can
319: spatially resolve the velocity field, thus establishing whether the material is in equilibrium or not. The observed residual emission profile (Figure~\ref{fig:sed}) is indicative of velocities fields more complex than the simple Keplerian rotation that was assumed in this work.
320:
321: In any case, one important consequence of this work is that we have shown that with current observing techniques and state-of-the-art modeling it is already possible to study the properties of the very inner layers of the disk in Achernar and possibly other nearby Be stars. Future simultaneous spectropolarimetry, spectro-interferometry and photometry will allow us to study the properties of this region and determine, for instance, the
322: size of the transition region and the point at which the disk becomes in vertical hydrostatic equilibrium.
323:
324: \acknowledgements{
325: This work was supported by FAPESP grant 04/07707-3 to ACC, FAPESP grant 01/12598-1 to AMM and NSF grant AST-0307686 to the University of Toledo (JEB). AMM also acknowledges partial support by CNPq.
326: }
327:
328: \begin{thebibliography}{}
329:
330: \bibitem[Carciofi \& Bjorkman(2006)]{car06a} Carciofi, A.~C.,
331: \& Bjorkman, J.~E.\ 2006, \apj, 639, 1081
332:
333: \bibitem[Carciofi et al.(2006)]{car06b} Carciofi, A.~C., Miroshnitchenko, A. S., Kusakin, A. V., et
334: al.\ 2006, \apj, 652, 1617
335:
336: \bibitem[Carciofi et al.(2007)]{car07} Carciofi, A.~C.,
337: Magalh{\~a}es, A.~M., Leister, N.~V., Bjorkman, J.~E., \& Levenhagen,
338: R.~S.\ 2007, \apjl, in press
339:
340: \bibitem[Cutri et al.(2003)]{2mass} Cutri, R.~M., et al.\
341: 2003, The IRSA 2MASS All-Sky Point Source Catalog, NASA/IPAC Infrared
342: Science Archive.~http://irsa.ipac.caltech.edu/applications/Gator/
343:
344: \bibitem[Domiciano de Souza et al.(2003)]{dom03} Domiciano de Souza, A., Kervella,
345: P., Jankov, S., Abe, L., Vakili, F., di Folco, E., \& Paresce, F.\ 2003, \aap, 407, L47 (D03)
346:
347: \bibitem[Fr{\'e}mat et al.(2005)]{fre05} Fr{\'e}mat, Y.,
348: Zorec, J., Hubert, A.-M., \& Floquet, M.\ 2005, \aap, 440, 305
349:
350: \bibitem[Jackson et al.(2004)]{jac04} Jackson, S., MacGregor,
351: K.~B., \& Skumanich, A.\ 2004, \apj, 606, 1196
352:
353: \bibitem[Kervella et al.(2003)]{ker03} Kervella, P., et al.\
354: 2003, \procspie, 4838, 858
355:
356: \bibitem[Kervella \& Domiciano de Souza(2006)]{ker06}
357: Kervella, P., \& Domiciano de Souza, A.\ 2006, \aap, 453, 1059
358:
359: \bibitem[Kervella \& Domiciano de Souza(2007)]{ker07}
360: Kervella, P., \& Domiciano de Souza, A.\ 2007, \aap, 474, L49
361:
362: \bibitem[Kurucz (1994)]{kur94} Kurucz, R. L. 1994, Kurucz CD-ROM No. 19, 20, 21, Cambridge Mass.: Smithsonian Astrophysical Observatory
363:
364: \bibitem[Meilland et al.(2007)]{mei07} Meilland, A., Stee, P., Vannier, M. et al. \ 2007, \aap, 464, 59
365:
366: \bibitem[Perryman et al.(1997)]{per97} Perryman, M. A. C., Lindegren, L., Kovalevsky, J., et al. \ 1997,
367: \aap, 323, L49
368:
369: \bibitem[Vinicius et al.(2006)]{vin06} Vinicius, M.~M.~F.,
370: Zorec, J., Leister, N.~V., \& Levenhagen, R.~S.\ 2006, \aap, 446, 643
371:
372: \bibitem[Zacharias et al.(2005)]{nomad} Zacharias, N., Monet,
373: D.~G., Levine, S.~E., Urban, S.~E., Gaume, R., \& Wycoff, G.~L.\ 2005,
374: VizieR Online Data Catalog, 1297, 0
375:
376: \bibitem[von Zeipel(1924)]{von24} von Zeipel, H.\ 1924,
377: \mnras, 84, 665
378:
379: \end{thebibliography}
380:
381: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
382: \clearpage
383:
384: \begin{deluxetable}{lll}
385: \tablecaption{Fixed Stellar Parameters. \label{tab:star}}
386: \tablewidth{0pt}
387: \tablehead{
388: \colhead{Parameter} &
389: \colhead{Value} &
390: \colhead{Reference}}
391: \startdata
392: $R_\mathrm{pole}$ & $7.3\; R_\sun$ & This work \\
393: Luminosity & $3150\; L_\sun$ & This work \\
394: $V_\mathrm{crit}$ & $350\;\rm km\;s^{-1}$ & \citet{vin06} \\
395: inclination & $65\deg$ & \citet{car07} \\
396: distance & $44.1\;\rm pc$ & \citet{per97}\\
397: \enddata
398: \end{deluxetable}
399:
400: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
401: %\clearpage
402:
403: \begin{deluxetable}{cccccccccc}
404: %\tablecaption{Parameters and results of some Succesful Models. \label{tab:models}}
405: \tablecaption{Model Parameters. \label{tab:models}}
406: \tablewidth{0pt}
407: \tablehead{
408: \colhead{Model} &
409: \colhead{$\Omega/\Omega_\mathrm{crit}$} &
410: \colhead{$\rho_0\;(\rm g\;cm{-3})$} &
411: \colhead{$R_\mathrm{d}$} &
412: \colhead{$H\;(R_\sun)$} &
413: %\colhead{$R_\mathrm{eq}/R_\mathrm{p}$} &
414: \colhead{$\frac{R_\mathrm{eq}}{R_\mathrm{p}}$} &
415: %\colhead{$T_\mathrm{p}/T_\mathrm{eq}$} &
416: \colhead{$\frac{T_\mathrm{p}}{T_\mathrm{eq}}$} &
417: \colhead{H$\alpha$ EW (\AA)} &
418: \colhead{$P_B$ (\%)} &
419: %\colhead{$\Delta V$} &
420: \colhead{$E_K$}
421: }
422: \startdata
423: 1 & 0.999999 & $1.0 \times 10^{-11}$ & 13.7 & 0.45 & 1.5 & 5.8 &
424: 0.10 & 0.02 & 0.17 \\ %mod220aa
425: %
426: 2 & 0.999999 & $3.8 \times 10^{-12}$ & 19.1 & 0.45 & 1.5 & 5.8 &
427: -0.34 & 0.10 & 0.14 \\ %mod220t
428: %
429: 3 & 0.999999 & $2.6 \times 10^{-12}$ & 16.1 & 1.1 & 1.5 & 5.8 &
430: -0.27 & 0.12 & 0.12 \\ %mod220u
431: %
432: 4 & 0.992 & $1.0 \times 10^{-11}$ & 13.7 & 0.45 & 1.4 &1.8 &
433: 0.20 & 0.31 & 0.15 \\ %mod222aa
434: %
435: 5 & 0.992 & $1.7 \times 10^{-12}$ & 13.7 & 0.91 & 1.4 &1.8 &
436: -0.05 & 0.15 & 0.03 \\ %mod222d
437: %
438: \enddata
439: \end{deluxetable}
440:
441:
442: \clearpage
443:
444: \begin{figure}[bp]
445: \plotone{f1.eps} \figcaption[]{
446: Emergent spectrum for models 1 to 5.
447: Top left: visible SED. The red triangles correspond to BVR photometry from the NOMAD catalog \citep{nomad}, for which no observational errors were available.
448: Top right: IR SED. The JHK photometry is from the 2MASS catalog \citep{2mass} and the mid-IR photometry is from \citet{ker07}. The observational errors for the mid-IR data is of the order of 2\%. %, smaller than the symbols used.
449: Bottom left: continuum polarization.
450: Bottom right: H$\alpha$ emission profile. The red curve corresponds to the residual emission profile of \citet{vin06}.
451: \label{fig:sed}}
452: \end{figure}
453:
454: \begin{figure}[bp]
455: %\plottwo{mod1_220aaa.eps}{mod2_220n.eps}
456: %\plottwo{mod3_220i.eps}{mod4_222aa.eps}
457: %\plottwo{mod5_222d.eps}{modsem_220sem.eps}
458: \plottwo{f2a.eps}{f2b.eps}
459: \plottwo{f2c.eps}{f2d.eps}
460: \plottwo{f2e.eps}{f2f.eps}
461: \figcaption[]{Squared visibilities for the models of Table~\ref{tab:models} (solid lines), along the polar (upper curves) and equatorial (lower curves) directions. The dashed lines represent the $K$ band visibilities for uniform disk (UD) angular diameters of 1.62~mas and 2.53~mas derived by D03; the corresponding $\pm1$-sigma uncertainties are shown as light-color bands. These UD diameters indicate the maximum (equatorial direction) and minimum (polar direction) sizes measured on Achernar in the $K$ band with VLTI/VINCI. The vertical dotted lines indicate the maximum baseline available from the VLTI data. The insets show the model image in the $K$ band, in logarithmic scale. The bottom right plot shows the squared visibilities for a gravity darkened Roche star at near-critical velocity (models 1 to 3) without the disk, and indicates, as shown by D03, that such model does not reproduce the observations. \label{fig:models}}
462: \end{figure}
463:
464: %com fator de escala 0.93:
465: % ref_file='K_mod220aaagrande_R20.0_T70.0_P0.0_L2.00000' & model='Model 1' & out_filename=dir+'mod1_220aaa'
466: % ref_file='K_mod220igrande_R20.0_T70.0_P0.0_L2.00000' & model='Model 2' & out_filename=dir+'mod2_220i'
467: % ref_file='K_mod220ngrande_R20.0_T70.0_P0.0_L2.00000' & model='Model 3' & out_filename=dir+'mod3_220n'
468: % ref_file='K_mod222aagrande_R20.0_T70.0_P0.0_L2.00000' & model='Model 4' & out_filename=dir+'mod4_222aa'
469: % ref_file='K_mod222dgrande_R20.0_T70.0_P0.0_L2.00000' & model='Model 5' & out_filename=dir+'mod5_222d'
470: % ref_file='bin_K_mod220sem_R18.0_T70.0_P0.0_L2.00000' & model='Models 1-3 w/o disk' & out_filename=dir+'modsem_220sem'
471:
472:
473: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
474: %\clearpage
475:
476:
477: \end{document}
478: