1: \documentclass[12pt,preprint]{aastex}
2:
3: %% manuscript produces a one-column, double-spaced document:
4: % \documentclass[manuscript]{aastex}
5:
6: %% preprint2 produces a double-column, single-spaced document:
7: % \documentclass[preprint2]{aastex}
8:
9: \newcommand{\vdag}{(v)^\dagger}
10: \def\etal {et al.~}
11: \def\ie {{i.e.},\ }
12: \def\eg {{e.g.},\ }
13: \def\onesig {$1\sigma$}
14: \def\twosig {$2\sigma$}
15: \def\threesig {$3\sigma$}
16:
17: \def\AIPS {{\small AIPS}}
18: \def\IMAGR {{\small IMAGR}}
19: \def\CLEAN {{\small CLEAN}}
20: \def\uvplane {{\it(u,v)}-plane}
21: \def\uvdist {{\it(u,v)}-distance}
22:
23: \def\hho {H$_2$O}
24: \def\Htwo {H$_2$}
25: \def\sioone {SiO v$=1$, J=1--0}
26: \def\siotwo {SiO v$=1$, J=2--1}
27: \def\rstar {\ifmmode {R_*} \else {$R_*$} \fi}
28: \def\oceti {$o$~Ceti}
29: \def\rleo {R~Leo}
30: \def\whya {W~Hya}
31: \def\Msun {M$_\odot$}
32: \def\Lsun {L$_\odot$}
33: \def\Hminus {H$^-$ free-free}
34: \def\pe {${\rm p}^+/{\rm e}^-$}
35: \def\mjyb {\ifmmode {{\rm mJy~beam}^{-1}} \else{mJy~beam$^{-1}$}\fi}
36: \def\jyb {\ifmmode {{\rm Jy~beam}^{-1}} \else{Jy~beam$^{-1}$}\fi}
37: \def\degrees {$^\circ$}
38: \def\Vorb {\ifmmode {V_{orb}} \else ${V_{orb}}$ \fi}
39: \def\Vdot {\ifmmode {\.V} \else ${\.V}$ \fi}
40: \def\ls{\vskip 12.045pt}
41: \def\hs{\hskip 2.5em}
42: \def\et{{\it et\thinspace al.}\ }
43: \def\arcm{\ifmmode {' }\else $' $\fi}
44: \def\arcs{\ifmmode {'' }\else $'' $\fi}
45: \def\arcmper{\ifmmode \rlap.{'} \else $\rlap{.}' $\fi}
46: \def\arcsper{\ifmmode \rlap.{''} \else $\rlap{.}'' $\fi}
47: \def\porm {\ifmmode\pm\else$\pm$\fi}
48: \def\kms {\ifmmode{{\rm ~km~s}^{-1}}\else{~km~s$^{-1}$}\fi}
49: \def\peryr {\ifmmode{{\rm ~yr}^{-1}}\else{~yr$^{-1}$}\fi}
50: \def\gmpercc {\ifmmode{{\rm gm~cm}^{-3}}\else{gm~cm$^{-3}$}\fi}
51: \def\percmcube {\ifmmode{{\rm ~cm}^{-3}}\else{~cm$^{-3}$}\fi}
52: \def\percmsquare {\ifmmode{{\rm ~cm}^{-2}}\else{~cm$^{-2}$}\fi}
53: \def\masy {\ifmmode{{\rm mas~y}^{-1}}\else{mas~y$^{-1}$}\fi}
54: \def\micron {\ifmmode{\mu{\rm m}}\else{$\mu$m}\fi}
55: \def\a {\ifmmode {\rlap.}^{''}\! \else ${\rlap.}^{''}\!$\fi}
56: \def\p {\phantom{0}}
57: \def\q {\phantom{00}}
58: \def\r {\phantom{000}}
59: \def\s {\phantom{0.0000}}
60:
61: \def\blankline {\par\vskip 12 pt}
62: \def\bl {\vskip 0.3truecm}
63: %
64: \newbox\grsign \setbox\grsign=\hbox{$>$} \newdimen\grdimen \grdimen=\ht\grsign
65: \newbox\laxbox \newbox\gaxbox
66: \setbox\gaxbox=\hbox{\raise.5ex\hbox{$>$}\llap
67: {\lower.5ex\hbox{$\sim$}}}\ht1=\grdimen\dp1=0pt
68: \setbox\laxbox=\hbox{\raise.5ex\hbox{$<$}\llap
69: {\lower.5ex\hbox{$\sim$}}}\ht2=\grdimen\dp2=0pt
70: \def\gax{\mathrel{\copy\gaxbox}}
71: \def\lax{\mathrel{\copy\laxbox}}
72:
73: \slugcomment{2007 July 20}
74:
75: \shorttitle{Imaging Mira Radio Photospheres}
76:
77: \shortauthors{Reid \& Menten}
78:
79: \begin{document}
80:
81: \title{Imaging the Radio Photospheres of Mira Variables}
82:
83: \author{M. J. Reid}
84: \affil{Harvard--Smithsonian Center for Astrophysics,
85: 60 Garden Street, Cambridge, MA 02138}
86: \email{reid@cfa.harvard.edu}
87:
88: \author{K. M. Menten}
89: \affil{Max-Planck-Institut f\"ur Radioastronomie,
90: Auf dem H\"ugel 69, D-53121 Bonn, Germany}
91: \email{kmenten@mpifr-bonn.mpg.de}
92:
93:
94: \begin{abstract}
95: We have used the VLA at 43 GHz to image the radio continuum
96: emission from \oceti, \rleo, and \whya\ and to precisely locate their SiO
97: maser emission with respect to the star. The radio continuum emission
98: region for all three stars has a diameter close to 5.6~AU. These diameters
99: are similar to those measured at infrared wavelengths in bands
100: containing strong molecular opacity and about twice those measured
101: in line-free regions of the infrared spectrum. Thus, the radio photosphere
102: and the infrared molecular layer appear to be coextensive.
103: The 43~GHz continuum emission is consistent with temperatures near 1600~K
104: and opacity from \Hminus\ interactions. While the continuum image of
105: \oceti\ appears nearly circular, both \rleo\ and \whya\ display
106: significant elongations. The SiO masers for all three stars show
107: partial rings with diameters close to 8~AU.
108: \end{abstract}
109:
110: \keywords{Stars: Long-Period Variables, Circumstellar Matter,
111: Atmospheres -- Radio Continuum: Stars -- Masers}
112:
113:
114: \section{Introduction}
115:
116: A typical oxygen-rich Mira variable has a radius, \rstar, defined by
117: line-free regions of the spectrum, of about 1--2 AU. Surrounding
118: the star is an extensive ``molecular layer,'' which extends
119: to nearly 2\rstar\ and can
120: be opaque across most of the visible and infrared spectrum.
121: The time-variable formation of metallic oxides in the molecular
122: photosphere can explain the thousand-fold (8 mag) variations in
123: the optical light curves of some Miras \citep{RG02}.
124: Also at about 2\rstar\ one detects a
125: ``radio photosphere,'' surrounded by an SiO maser emitting region.
126: Beyond the SiO region dust grains form, radiation pressure
127: accelerates the dust, transferring
128: outward momentum to the gas and leading to significant mass loss.
129:
130: \citet{RM97a}
131: published measurements made with the Very Large Array
132: (VLA) of a sample of Mira-like variables
133: (\oceti, \rleo, \whya, R~Aql, $\chi$~Cyg, and R~Cas).
134: Their results revealed that the these stars exhibit a cm-wavelength
135: spectral index of $\approx 1.9$ and nearly constant radio light curves
136: (variation $\lax\pm15$\% about the mean).
137: Also, they imaged the radio continuum of \whya\ at 22 GHz and found
138: the star precisely at the center of a ring of
139: \hho\ masers with a radius of 0\a15 \citep{RM90}.
140:
141: The detection of a ``radio photosphere" near $2R_*$ provides a new
142: observational tool to increase our understanding of the region
143: between the star and the dust formation zone.
144: For the physical conditions expected in the radio photosphere,
145: free electrons, obtained mostly from the ionization of potassium
146: and sodium, provide the dominant opacity through free-free
147: interactions with neutral H and \Htwo. \citet{RM97a} proposed
148: a simple physical model which can explain the radio frequency observations.
149: At cm-wavelengths, unity optical depth is achieved at a radius of
150: about 3 to 4~AU, where the density and temperature
151: are $\sim10^{12}$~\percmcube\ and $\approx1600$~K, respectively.
152: This produces a flux density which is roughly twice that expected for
153: black-body emission from the underlying star, as is observed.
154:
155: In this paper, we report observations at 7~mm wavelength (43 GHz)
156: in which we simultaneously imaged the radio photospheres and SiO maser emission
157: from two Miras (\oceti\ and \rleo) and one semi-regular variable (\whya).
158: These observations and supporting models constrain the
159: physical and dynamical conditions between $2$ and $3~\rstar$
160:
161:
162:
163: \section{Observations \& Data Analysis}
164:
165: The observations were made with the NRAO
166: \footnote{The National Radio Astronomy Observatory (NRAO) is operated
167: by Associated Universities, Inc., under a cooperative
168: agreement with the National Science Foundation.}
169: Very Large Array (VLA) under program AR446.
170: Our goal was to measure directly the angular size of radio photospheres
171: with the highest angular resolution afforded by the VLA.
172: Our synthesized beam at 43 GHz in the A-configuration was
173: $\approx 0\arcsper04$ full-width at half-maximum (FWHM),
174: which well resolves a typical Mira to a distance of $\approx200$~pc.
175: Since radio frequency ``seeing" is limited primarily by fluctuations of
176: water vapor in the atmosphere and is rarely better than $0\arcsper1$,
177: we used a calibration scheme developed by us \citep{RM90,RM97a}
178: in which circumstellar masers provide a phase reference for the
179: continuum data. This allows diffraction limited imaging to be achieved.
180:
181: Between 2000 October 25 and November 6, we observed three stars:
182: \oceti, \rleo, and \whya. Optical stellar phases for the three stars
183: were approximately 0.05, 0.55, and 0.25, based on examination of
184: AAVSO visual magnitudes for several cycles before and after the observations.
185:
186: The observations to measure stellar sizes
187: were conducted in 4-IF continuum mode. The A-C IF pair
188: (50 MHz bandwidth in right and left circular polarization) was centered
189: at 43164.9 MHz, which is 43~MHz above the rest frequency
190: of 43122.08 MHz of the \sioone\ line.
191: The B-D IF pair used a 1.563~MHz bandwidth centered on the SiO
192: line at local standard-of-rest (LSR) velocities of
193: 47.0~\kms\ for \oceti, 1.0~\kms\ for \rleo, and 42.0~\kms\ for \whya.
194:
195: The interferometer data were calibrated using the Astronomical Image
196: Processing System (\AIPS).
197: Absolute flux density calibration was obtained from an observation of
198: 3C286, assuming a flux density extrapolated from the formula
199: given by \citet{B77}.
200: Following the procedures described in detail in \citet{RM97a},
201: we used roughly hourly observations of a strong continuum source to
202: update pointing, to remove the electronic phase differences among the bands
203: and to monitor the slowly-varying amplitude response of the antennas.
204: The narrow-band data, containing very intense SiO maser emission,
205: were phase-only self-calibrated on each 10~sec integration.
206: The resulting antenna--dependent
207: phase corrections were then applied to the broad-band data
208: containing the continuum emission from the star.
209: Since the masers are observed simultaneously with the continuum data
210: and come from essentially the same position on the sky,
211: this calibration removes all significant atmospheric phase-fluctuations.
212:
213: Once the data were ``cross-self-calibrated,'' both the narrow-band maser
214: and broad-band continuum signals were imaged with the \AIPS\ task \IMAGR.
215: Map resolutions and noise levels are given in
216: Table~\ref{table:map_parameters}.
217: In order to simplify map analysis, \CLEAN\ maps used a round restoring
218: beam equal approximately to the geometric mean of the dirty beam's
219: major and minor axes. These maps are shown in left column of panels in
220: Fig.~\ref{fig:star_visibilities}.
221:
222: In addition to images, we also analyzed the interferometer visibility
223: data in the following manner.
224: Firstly, we measured the position of peak continuum emission by fitting
225: an elliptical Gaussian brightness distribution to the map.
226: Next, we ``shifted'' the data to place the continuum
227: peak at the interferometer phase-center. After time averaging
228: the visibilities for 1~minute, we fitted the data in the \uvplane\
229: with a uniform-brightness, elliptical-disk model.
230: This fitting was done using MIRIAD analysis software.
231: Parameters of these fits are given in Table~\ref{table:fit_parameters}.
232:
233: Displaying the 2-dimensional \uvplane\ data and model fits for
234: weak sources is difficult. In order to obtain sufficient
235: signal-to-noise ratio, the data need to be vector averaged
236: over large portions of the irregularly sampled \uvplane.
237: This precludes, for example, plotting visibility amplitude
238: versus projected baseline length along a narrow strip at
239: given position angle in the \uvplane.
240: We have chosen to bin the data and plot
241: the real part of the visibility versus baseline length.
242: These plots appear in the right-hand panels of
243: Fig.~\ref{fig:star_visibilities}.
244: For an elliptical source, this mixes data with differing
245: visibilities, owing to the greater resolution of a source
246: for a given baseline length if it is aligned with the
247: source major axis than the minor axis.
248: In order to indicate the magnitude of this effect, we have
249: plotted the theoretical visibilities for circular sources
250: with diameters equal to the major and minor axes of the
251: fitted elliptical model.
252:
253: We also observed the SiO maser emission at high spectral resolution
254: with several scans in spectral-line mode interspersed among the
255: dual-band continuum observations. We covered the entire SiO line
256: emission with a bandwidth of 3.125 MHz and used 64 spectral channels,
257: which provided a velocity resolution of 0.34 \kms. The line data
258: were self-calibrated by choosing a channel with strong emission
259: as a reference, and the resulting phase and amplitude corrections
260: were applied to the other channels. Because the maser emission was
261: very strong ($\sim10^3$ Jy), we did not require bandpass calibration.
262: From the line data we produced a spectral-line data cube, which we
263: restored with a 20~mas FWHM Gaussian beam, taking advantage of the high
264: signal-to-noise ratio and nearly point-like emission in each spectral
265: channel to ``super-resolve'' by a factor of about two.
266:
267: The continuum emission (in the 50 MHz A-C IF pair) and maser emission
268: (in the 1.563 MHz B-D IF pair), which were observed simultaneiously,
269: are automatically aligned by the calibration procedures to an accuracy
270: of better than 1~mas. In order to align the maser emission from the
271: spectral-line data, not observed simultaneously, with the continuum,
272: we performed the following steps:
273: 1) we produced a pseudo-continuum map from
274: the line data, using spectral channels covering the velocities that were
275: within the 1.563 MHz passband of the dual-band continuum setup;
276: 2) we overlaid the pseduo-continuum map and on the map from the 1.563 MHz
277: passband; and
278: 3) we translated the pseduo-continuum map to achieve the best alignment
279: by eye. Since the maps were nearly identical and contained unresolved
280: spots with high signal-to-noise ratios, alignment of the continuum and
281: maser emission to better than $\approx2$~mas (or 10\% of the restoring
282: beam) was achieved.
283: Based on this cross-registration, we show velocity-integrated SiO maser
284: emission superposed on the continuum emission of the stars in
285: Fig.~\ref{fig:star_maser_overlays_grey}.
286: In order to avoid introducing excess noise in the velocity-integrated
287: SiO maser maps, we set the flux density of all pixels with less
288: than 1~\jyb\ to zero before summing the emission of the
289: individual spectral channels.
290:
291:
292: \section {Radio Photospheric Characteristics}
293:
294:
295: \subsection{Diameters}
296:
297: We clearly resolve the radio photospheres of all three stars.
298: For \oceti, we find a nearly circular image with a uniform-disk
299: diameter of about 52~mas (see Table~\ref{table:fit_parameters}).
300: At a distance of 110~pc, based on the Period-Luminosity (PL)
301: relation of \citet{F89} and the photometry of \citet{H95},
302: this corresponds to a radio diameter of 5.7~AU.
303: Throughout this paper, we adopt the PL-based distances,
304: which appear to be more accurate than {\it Hipparcos} parallaxes
305: for these large, variable and irregularly bright stars \citep{RM97a}.
306: We measure a mean diameter of 50~mas for \rleo,
307: corresponding to 5.5~AU at a distance of 110~pc.
308: \whya\ displays a mean angular diameter
309: of 58~mas, which at 95~pc distance implies a physical diameter
310: of 5.5~AU. Thus, the radio-frequency mean diameters of all three
311: stars are remarkably similar, averaging close to 5.6 AU.
312:
313: The diameters of the radio photospheres are similar to
314: those measured in the visible and infrared in narrow bands
315: that contain strong molecular opacity. For example, in
316: a 10-\micron-wide band centered at 710~\micron\ containing
317: strong TiO absorption, \citet{H95} find (uniform-disk) angular
318: diameters of 53, 45 and 66~mas for \oceti, \rleo\ and \whya,
319: respectively. Similarly, \citet{P04} infer a ``molecular layer''
320: diameter of 52 and 50~mas for \oceti\ and \rleo, respectively.
321: Generally it is difficult to ``see'' through the molecular
322: photosphere and down to \rstar. Using narrow bands placed
323: in spectral regions with minimal molecular opacity, \citet{P04}
324: find stellar diameters of 26~mas for \oceti\ and 22~mas for \rleo.
325: Model-fit Rosseland mean diameters from observations using broader
326: bands yield diameters that are $\approx20$\% larger \citep{W04,F05}.
327: Thus, the radio and molecular layers occur at radii of
328: $\approx2\rstar$, adopting the narrow-band stellar diameters.
329:
330: \subsection {Mira Binary}
331:
332: Mira A is \oceti, an M-type giant, and Mira B is possibly a low-mass
333: main-sequence star \citep{I06} or a white dwarf. \cite{MK06} detected
334: both stars in the Mira binary with the VLA. They found the 43~GHz flux
335: density of Mira B is weaker than that of \oceti\ by a factor
336: of $\approx3$. We well resolve the binary, which has a projected
337: separation of $\approx500$~mas, and fail to detect Mira B.
338: Our \threesig\ limits are 0.4 and 0.7~\mjyb\ for untapered and
339: tapered maps with (geometric mean) interferometer beams of 50 and 250~mas,
340: respectively.
341: Were the radio emission from Mira B unresolved and at the
342: $\approx0.9$~mJy level seen by Matthews \& Karovska, we would have
343: detected it. Since the Matthews \& Karovska B-configuration
344: observations at 43~GHz had a beam size comparable to our tapered
345: map, it appears that angular resolution alone cannot explain
346: our failure to detect Mira B and that it must be time variable.
347: Since we well resolve ($\approx10$ beams) the binary and Mira B
348: is not detected, our \oceti\ results should not be significantly
349: affected by the secondary.
350:
351:
352: \subsection{Asymmetries}
353:
354: While \oceti's image is circular to within $\pm4$\%, we
355: detect non-circular radio photospheres for \rleo\ and \whya.
356: The differences between the major and minor axes are significant at
357: about the $2\sigma$ level.
358: %major-to-minor axis ratios that are $1.56\pm0.35$
359: %and $1.50\pm0.32$, respectively.
360: We are reasonably confident that the non-circular structures
361: do not come from instrumental limitations for several reasons:
362: 1) for \rleo\ the intrinsic interferometer beam is nearly round
363: and for \whya\ the image is elongated nearly perpendicular to the
364: intrinsic interferometer beam;
365: 2) analysis of the data both in the image and \uvplane\ yield similar
366: results; and
367: 3) the plots of visibility versus \uvdist\ for the elliptically shaped
368: stars show scatter larger than expected for a circular source, but
369: consistent with an elliptical source.
370:
371:
372: \subsection{Physical Conditions}
373:
374: Both \oceti\ and \rleo\ have (uniform-disk) brightness temperatures
375: near 1650~K. These temperatures are close to those expected for
376: the radio photospheres \citep{RM97a}.
377: We measure a somewhat higher brightness temperature of $2380\pm550$~K
378: for \whya.
379: Since the collisional ionization of metals, and hence the \Hminus\ opacity,
380: is strongly temperature dependent below 2500~K, the higher measured
381: brightness temperature of \whya\ might suggest that the total density in
382: the \whya\ photosphere is lower than in \oceti\ or \rleo.
383: However, the uncertainties in brightness temperature are fairly large,
384: owing predominantly to the squared-dependence on the uncertainty in the
385: size measurement, and all three stars could have similar temperatures.
386: Our temperature measurements are consistent, within uncertainties,
387: with our earlier measurement of brightness temperature of \whya\ of
388: $1570\pm570$~K at a lower frequency of 22 GHz \citep{RM97a}.
389:
390: The brightness temperatures of the radio photospheres of \oceti\ and
391: \rleo\ are also similar to those inferred by \citet{P04} of about 2000
392: and 1600~K for the molecular layers of these stars.
393: Spectroscopic observations of \hho\ in the infrared
394: also indicate temperatures and densities close to those we previously
395: measured and modeled.
396: \citet{M02} model their 2.5 to 4.0~\micron\ ISO telescope data
397: with a hot (2000~K) layer close to the stellar photosphere
398: (between 1 and 2\rstar) and a cool (1400 to 1000~K) layer
399: between 2.5 and \hbox{4\rstar.} The total hydrogen densities in these
400: layers are estimated to be $\sim10^{12-13}$~\percmcube.
401: These layers roughly straddle the radio photosphere where
402: we have previously found temperatures of about 1600~K and
403: hydrogen densities of $\sim10^{12}$~\percmcube\
404: \citep{RM97a}. Since both the radii, temperatures and densities
405: measured in the radio and infrared agree, this provides added confidence
406: in both measurements.
407:
408: \section{SiO Masers}
409:
410: The SiO masers in \oceti, \rleo\ and \whya\ form partial rings
411: with radii of about 40, 35 and 41~mas, respectively
412: (Fig.~\ref{fig:star_maser_overlays_grey}).
413: These radii were measured by visually overlaying circles of varying
414: radii centered on the stellar continuum position and are probably
415: accurate to $\pm10$\%, limited mostly by the lack of complete maser rings.
416: These partial rings are consistent with, but perhaps slightly larger than,
417: the Very Long Baseline Array (VLBA) results by \citet{C04},
418: who find radii at different epochs ranging from 30--38, 26--30, and 42~mas
419: for \oceti, \rleo\ and \whya, respectively, for the \sioone\ transition .
420: The differences in ring radii between our VLA images, which are sensitive
421: to lower brightness emission, and the VLBA results, which have
422: higher angular resolution and use a different approach to define
423: the ring radii, are small. We note that \citet{DK03}, in their
424: extensive VLBA monitoring of TX Cam, find lower brightness SiO
425: masers radially outside of the brightest masers.
426:
427: When we convert our ring diameters from angular to linear units,
428: we find radii of 4.4, 3.8 and 3.9~AU, respectively.
429: Thus, these three stars show very similar SiO masing radii
430: close to 4~AU.
431: Previous observations have also detected partial rings of SiO masers
432: in Mira-like stars \citep{M94,G95,B00,DK03} and have usually
433: {\it assumed} that the stars were near the center of the rings.
434: However, for the first time, we are able to demonstrate that the
435: maser rings are precisely centered on the stars.
436: These results nicely complement our previous result that the
437: \hho\ masers in \whya\ also exhibit a partial ring and are
438: precisely centered on the stellar continuum emission \citep{RM90}.
439: We plot the integrated \hho\ maser emission from \whya\ in
440: Fig.~\ref{fig:whya_h2o}, superposed on its SiO
441: maser and radio continuum emission.
442: The \hho\ maser ring had a radius of $\approx150$~mas, which is
443: about 3.7-times that of the SiO masers.
444:
445: The simplest interpretation for the SiO masers is that they form
446: in a spherical shell near radii of about 3\rstar. If one adopts
447: the somewhat smaller SiO radii from VLBA maps, which sample the higher
448: brightness masers, and the larger stellar radii, from Rosseland
449: mean opacities, the maser shells would have radii closer to 2\rstar.
450: The inner boundary of the SiO masers is determined either by
451: densities $\gax10^{10}$~\percmcube, leading to thermalization of the
452: level populations \citep{LE92}, or by high radio-continuum
453: opacity from \Hminus\ interactions \citep{RM97a}.
454: The outer boundary occurs when SiO densities become too low for
455: strong maser amplification. Note that in addition to the rapid
456: decrease in total density with increasing radius in the extended
457: atmosphere, the SiO density can fall even faster
458: owing to silicate grain formation.
459:
460: Proper motions of SiO masers in TX~Cam by \citet{DK03} indicate
461: complex dynamics.
462: Expansion at a speed of $\approx7~(D/390~{\rm pc})$~\kms, where
463: $D$ is the star's distance, is the dominant motion.
464: However, some SiO maser spots are also observed to move inwards
465: at a few \kms, suggesting a possible large-scale convective structure.
466: \citet{DK03} adopt a distance based on the period-luminosity and
467: photometry compiled by \citet{OWM01}. An alternative distance estimate
468: could come from the radius of the SiO maser ring, since we find
469: all three of our stars have a ring radius near 4 AU. Adopting this
470: ring radius for TX Cam, the measured mean angular radius of 16~mas
471: gives a distance of 250~pc. At this distance the dominant SiO
472: expansion speed would only be 4.5 \kms.
473: In any event, there are no indications of pervasive, fast-moving
474: ($\gax10$~\kms) shocks in the SiO masing region.
475:
476:
477:
478: \section{Discussion}
479:
480:
481: A simple, consistent picture of the inner envelopes of oxygen-rich
482: Mira variables is emerging from radio and infrared observations.
483: Radio observations of stars like
484: \oceti, \rleo\ and \whya\ provide data which
485: constrains the kinematics and physical conditions between
486: $2$ and $3\rstar$. Radio continuum emission
487: comes predominantly from \Hminus\ interactions.
488: An optical depth of unity is observed at $\approx2\rstar$
489: and requires a temperature and density of $\approx 1600$~K and
490: $\sim10^{12}$~\percmcube, respectively \citep{RM97a}.
491:
492: The radio photosphere corresponds very closely to the molecular layer,
493: identified at infrared wavelengths. The molecular layer can
494: be directly observed in lines of \hho\ and CO \citep{H95,M02,P04}.
495: As the star cools cyclically,
496: this region can become nearly opaque at visible wavelengths,
497: owing to metallic oxide formation. This leads to dramatic
498: changes in visible light as the temperature in the molecular
499: photosphere at $\approx2\rstar$ can be low enough for visible
500: light to ``fall off'' the Wien side of the black body spectrum \citep{RG02}.
501:
502: SiO maser emission surrounds the radio/molecular layer and
503: appears as partial rings with radii of $2$ to $3\rstar$. These rings
504: sometimes display significant ellipticity as, for example, shown
505: in TX Cam by \citet{DK03}.
506: SiO masers trace both outwardly and inwardly moving material, possibly indicative
507: of large scale convection. Temperatures and densities inferred
508: from SiO masers are $\approx 1000$~K and $\lax10^{10}$~\percmcube.
509: SiO maser proper motions suggest that outside of $3~\rstar$
510: shocks are mostly damped and propagate at $\lax5-7$~\kms.
511: This complements a
512: similar conclusion, that strong, fast shocks are not found in the
513: underlying radio photosphere, based on the observed nearly constant
514: radio light curves \citep{RM97a,RM97b}.
515: Thus, while radio data indicate a dynamic region between $2$ and $3~\rstar$,
516: there are no signs of strong shocks heating large portions of the atmosphere
517: to $\sim10,000$~K near $2\rstar$ as in some models of these stars
518: \citep{W00}.
519:
520: The ``radio photospheres'' of \oceti\ and \whya\ appear somewhat
521: elongated. The non-circular structure we detect in the radio
522: photospheres of two AGB stars may be of significance in understanding
523: the highly elongated structures seen in planetary nebulae,
524: following a post-AGB phase.
525: Clearly second epoch images of the radio photospheres,
526: after several pulsation cycles, would be very interesting.
527: If the non-circular structures change significantly, either in
528: shape or orientation, this would suggest a (non-radial) pulsation
529: origin. Alternatively, should the structures be stable over time,
530: this would suggest a possible rotational or magnetic field origin,
531: which may play a role in creating asymmetric structures in
532: planetary nebulae.
533:
534:
535: \acknowledgments
536: We thank L. Matthews for comments on an early version of the manuscript.
537: We acknowledge with thanks the variable star observations from the AAVSO
538: International Database contributed by observers worldwide and used in this
539: research.
540:
541:
542: \begin{thebibliography}{}
543: \bibitem[Baars \etal (1977)]{B77}
544: Baars, J. W. M., Genzel, R., Pauliny-Toth, I. I. K \&
545: Witzel, A. 1977, \aap, 61, 99
546: \bibitem[Boboltz \& Marvel (2000)]{B00}
547: Boboltz, D. A. \& Marvel, K. B. 2000, \apj, 545, L149
548: \bibitem[Cotton \etal (2004)]{C04}
549: Cotton, W. D. \etal 2004, \aap, 414, 275
550: \bibitem[Dalgarno \& Lane (1966)]{DL66}
551: Dalgarno, A. \& Lane, N. F. 1966, \apj, 145, 623
552: \bibitem[Diamond \& Kemball (2003)]{DK03}
553: Diamond, P. J. \& Kemball, A. J. 2003, \apj, 599, 1372
554: \bibitem[Feast \etal (1989)]{F89}
555: Feast, M. W., Glass, I. S., Whitelock, P. A. \& Catchpole, R. M.
556: 1989, \mnras, 241, 375
557: \bibitem[Fedele \etal (2005)]{F05}
558: Fedele, D., Wittkowski, M., Paresce, F., Scholz, M., Wood, P. R.
559: \& Ciroi, S. 2005, \aap, 431, 1019
560: \bibitem[Greenhill \etal (1995)]{G95}
561: Greenhill, L. J., Colomer, F., Moran, J. M., Backer, D. C.,
562: Danchi, W. C. \& Bester, M. 1995, \apj, 449, 365
563: \bibitem[Haniff, Scholz \& Tuthill (1995)]{H95}
564: Haniff, C. A., Scholz, M. \& Tuthill, P. G. 1995, \mnras, 276, 640
565: \bibitem[Ireland \etal (2007)]{I06}
566: Ireland, M. J. \etal 2007, \apj, 662, 651
567: \bibitem[Lockett \& Elitzur (1992)]{LE92}
568: Lockett, P. \& Elitzur, M. 1992, \apj, 399, 704
569: \bibitem[Matthews \& Karovska (2006)]{MK06}
570: Matthews, L. D. \& Karovska, M. 2006, \apj, 637, L49
571: \bibitem[Matsuura \etal (2002)]{M02}
572: Matsuura, M., Yamamura, I., Cami, J., Onaka, T. \& Murakami, H.
573: 2002, \aap, 383, 972
574: \bibitem[Miyoshi \etal (1994)]{M94}
575: Miyoshi, M., Matsumoto, K., Kameno, S., Takaba, H. \& Iwata, T.
576: 1994, \nat, 371, 395
577: \bibitem[Olivier, Whitelock \& Marang (2001)]{OWM01}
578: Olivier, E. A., Whitelock, P. \& Marang, F. 2001, \mnras, 326, 490
579: \bibitem[Perrin \etal (2004)]{P04}
580: Perrin, G. \etal 2004, \aap, 426, 279
581: \bibitem[Reid \& Goldston (2002)]{RG02}
582: Reid, M. J. \& Goldston, J. E. 2002, \apj, 568, 931
583: \bibitem[Reid \& Menten (1990)]{RM90}
584: Reid, M. J. \& Menten, K. M. 1990, \apj, 360, L51
585: \bibitem[Reid \& Menten (1997a)]{RM97a}
586: Reid, M. J. \& Menten, K. M. 1997a, \apj, 476, 327
587: \bibitem[Reid \& Menten (1997b)]{RM97b}
588: Reid, M. J. \& Menten, K. M. 1997b, {\it Ap\&SS}, 251, 41
589: \bibitem[Reid \& Menten (2003)]{RM03}
590: Reid, M. J. \& Menten, K. M. 2003, in
591: Mass-losing pulsating stars and their circumstellar matter,
592: Astrophysics and Space Science Library, Vol. 283,
593: eds. Y. Nakada, M. Honma and M. Seki.
594: (Kluwer Academic Publishers, Dordrecht), p. 283
595: \bibitem[Willson (2000)]{W00}
596: Willson, L. A. 2000, ARA\&A, 38, 573
597: \bibitem[Woodruff \etal (2004)]{W04}
598: Woodruff \etal 2004, \aap, 421, 703
599: \end{thebibliography}
600:
601: \clearpage
602:
603: \begin{deluxetable}{lccccc}
604: \tabletypesize{\scriptsize}
605: \tablecaption {Interferometer Parameters \label{table:map_parameters}
606: }
607: \tablehead{ \colhead{Star} &\colhead{$\theta_B(maj)$} &\colhead{$\theta_B(min)$}
608: &\colhead{P.A.} &\colhead{Continuum RMS Noise} &\colhead{Line RMS Noise}
609: \\ &\colhead{(mas)} &\colhead{(mas)} &\colhead{(deg.)}
610: &\colhead{(mJy)} &\colhead{(mJy)}
611: }
612: \tablecolumns{6}
613: \startdata
614: \\
615: \oceti~~......... &53 &39 &$-13$ &0.13 &10 \\
616: \rleo~~......... &61 &39 &$-20$ &0.14 &15 \\
617: \whya~~......... &81 &37 &$-\p7$ &0.30 &20 \\
618: \enddata
619: \tablecomments{Beam major, $\theta_B(maj)$, and minor, $\theta_B(min)$,
620: axes and East-of-North position angles, P.A.,
621: are from the best fit elliptical Gaussians
622: to the ``dirty beams.'' The maps displayed in
623: Fig.~\ref{fig:star_visibilities} and Fig.~\ref{fig:star_maser_overlays_grey}
624: used round restoring beams as discussed in the text.
625: }
626: \end{deluxetable}
627:
628: \begin{deluxetable}{lcccccccc}
629: \tabletypesize{\scriptsize}
630: \tablecaption {Measured Stellar Parameters \label{table:fit_parameters}
631: }
632: \tablehead{ \colhead{Star} &\colhead{Distance} &\colhead{Opt. Phase}
633: &\colhead{$\theta_{maj}$} &\colhead{$\theta_{min}$}
634: &\colhead{P.A.} &\colhead{$S_\nu$} &\colhead{$T_b$}
635: &\colhead{SiO Radius}
636: \\ &\colhead{(pc)} & &\colhead{(mas)} &\colhead{(mas)}
637: &\colhead{(deg.)}
638: &\colhead{(mJy)} &\colhead{(K)} &\colhead{(mas)}
639: }
640: \tablecolumns{9}
641: \startdata
642: \\
643: \oceti~~.........&$110\pm9$ &0.05 &$54\pm\p5$ &$50\pm5$ &$+39\pm50$
644: &$4.8\pm0.2$ &$1680\pm250$ &$40\pm4$ \\
645: \rleo~~.........&$110\pm9$ &0.55 &$61\pm10$ &$39\pm6$ &$-20\pm12$
646: &$4.1\pm0.2$ &$1630\pm410$ &$35\pm4$ \\
647: \whya~~.........&$\p95\pm8$ &0.25 &$69\pm10$ &$46\pm7$ &$+83\pm18$
648: &$8.0\pm0.4$ &$2380\pm550$ &$41\pm6$ \\
649: \enddata
650: \tablecomments{Distances are from \citet{H95} based on the PL
651: relation of \citet{F89}.
652: Stellar major, $\theta_{maj}$, and minor, $\theta_{min}$,
653: diameters and East-of-North position angles, P.A., are from best-fit,
654: uniform-brightness ($T_b$), elliptical-disk models.
655: Flux density, $S_\nu$, uncertainties do not include
656: a roughly $\pm10$\% uncertainty in the absolute flux density scale.
657: SiO maser shell radii are estimated from the brightness peaks in
658: Fig.~\ref{fig:star_maser_overlays_grey}.
659: }
660: \end{deluxetable}
661:
662:
663: \clearpage
664:
665: \begin{figure}
666: \epsscale{0.65}
667: \plotone{f1.eps}
668: \caption{Images shown as contours ({\it left panels}) and
669: visibility vs. interferometer baseline length ({\it right panels})
670: of the 43 GHz continuum emission from \oceti\ ({\it top panels}),
671: \rleo\ ({\it middle panels}) and \whya\ ({\it bottom panels}).
672: Contour levels are integer multiples of 0.5 mJy for \oceti\
673: and \rleo\ and multiples of 1.0 mJy for \whya.
674: In the image panels, the interferometer FWHM ``dirty'' beams are shown as
675: solid ellipses and the round restoring beams are shown as dotted
676: circles in the lower left.
677: The dashed ellipses superposed on the images are the sizes of
678: uniformly bright elliptical disks that best fit the data.
679: In the visibility plots, the two model curves correspond to
680: expected visibilities for circular disks with diameters
681: corresponding to the major and minor axes listed in
682: Table~\ref{table:fit_parameters}.
683: \label{fig:star_visibilities}
684: }
685: \end{figure}
686:
687:
688: \clearpage
689:
690: \begin{figure}
691: \epsscale{0.40}
692: \plotone{f2.eps}
693: \caption{SiO maser emission ({\it contours and gray scale)}
694: superposed on the fitted uniformly bright elliptical disk models
695: of the radio continuum emission ({\it green filled ellipses}) from
696: \oceti\ ({\it top panel}), \rleo\ ({\it middle panel}) and \whya\ ({\it bottom panel}).
697: SiO contour levels are 1, 2, 4, 8, 16, and 32 times 100, 40,
698: and 200 \jyb\ for the three stars, respectively.
699: The relative positions of the SiO masers and the radio continuum
700: have been measured with accuracies of about 2~mas. The restoring
701: beams for the radio continuum (larger dotted circle) and SiO
702: maser data (smaller dotted circle) are indicated in the lower
703: left corner of each panel.
704: The true stellar sizes ({\it red filled circles}) are from
705: circular, uniformly bright, disk models of IR interferometer data
706: \citep{P04}.
707: \label{fig:star_maser_overlays_grey}
708: }
709: \end{figure}
710:
711: \clearpage
712:
713: \begin{figure}
714: \epsscale{0.80}
715: \plotone{f3.eps}
716: \caption{\whya\ \hho\ maser emission ({\it gray scale)} from \citet{RM90},
717: superposed on the SiO maser emission ({\it yellow contours)} and the
718: radio continuum emission (elliptical disk model)
719: ({\it green filled ellipse}) from Fig.~\ref{fig:star_maser_overlays_grey}.
720: The \hho\ velocity-integrated flux scale in \jyb\
721: is shown along the right side of the figure.
722: The relative positions of the \hho\ masers and the radio continuum
723: is accurate to about 10~mas. The restoring
724: beam for the \hho\ emission is indicated in the lower left corner.
725: \label{fig:whya_h2o}
726: }
727: \end{figure}
728:
729:
730: \end{document}
731:
732: