0709.3271/ms.tex
1: %\documentstyle[emulateapj,apjfonts,epsf]{article}
2: %\documentstyle[emulateapj,apjfonts,epsf,versioncopy]{article}
3: %\documentclass[preprint2]{aastex}
4: %\documentclass[preprint]{aastex}
5: \documentclass[12pt,preprint]{aastex}
6: %\usepackage{emulateapj,pstricks,apjfonts}
7: %\documentstyle{emulateapj}
8: %\documentstyle[emulateapj,apjfonts,epsf,versioncopy]{article}
9: %\documentstyle[emulateapj,apjfonts,epsf]{article}
10: %\documentstyle[emulateapj,epsf]{article}
11: %\documentclass[12pt,preprint2]{aastex}
12: %\documentclass[12pt,preprint]{aastex}
13: 
14: \newdimen\minuswidth    %define @ width of minus sign for tables
15: \setbox0=\hbox{$-$}
16: \minuswidth=\wd0
17: \catcode`@=\active
18: \def@{\kern\minuswidth}
19: \newdimen\digitwidth    %define ! a one digit width for tables
20: \setbox0=\hbox{\rm0}
21: \digitwidth=\wd0
22: \catcode`!=\active
23: \def!{\kern\digitwidth}
24: \newcommand{\gras}[1]{\mbox{\boldmath $#1$}}
25: 
26: \def\reference#1\par{\parindent0pt\hangindent20pt\hangafter1 #1\par}
27: \def\btsp{BT-$2^{\rm nd}$P}
28: %\def\msun{\alwaysmath{{M}_{\odot}}}
29: \def\SP{$2^{\rm nd}$P}
30: \def\FP{$1^{\rm st}$P}
31: 
32: \received{}
33: \begin{document}
34: \shorttitle{Empirical Mass Loss Law}
35: \shortauthors{Origlia et al.}
36: 
37: \title{The First Empirical Mass Loss Law for Population II Giants$^1$}
38: \altaffiltext{1}{
39: This work is based on observations made with the Spitzer Space Telescope, 
40: which is operated by the Jet Propulsion Laboratory, 
41: California Institute of Technology under a contract with NASA.
42: Support for this work was provided by NASA through 
43: an award issued by JPL/Caltech.}
44: 
45: \author{Livia Origlia$^2$, Robert T. Rood$^3$, Sara Fabbri$^4$, 
46: Francesco R. Ferraro$^4$, 
47: Flavio Fusi Pecci$^2$, R. Michael Rich$^5$} 
48: \altaffiltext{2}{INAF--Osservatorio Astronomico di Bologna,
49: Via Ranzani 1, I--40127 Bologna, Italy, livia.origlia@oabo.inaf.it,
50: flavio.fusipecci@oabo.inaf.it}
51: \altaffiltext{3}{Astronomy Department, University of Virginia, Charlottesville, VA 22903,
52: rtr@virginia.edu}
53: \altaffiltext{4}
54: {Universit\`a degli Studi di Bologna, Dip. di Astronomia,
55: Via Ranzani 1, I--40127 Bologna, Italy,
56: sara.fabbri@studio.unibo.it,francesco.ferraro3@unibo.it}
57: \altaffiltext{5}{
58: Department of Physics and Astronomy, University of California
59: at Los Angeles, Los Angeles, CA 90095-1547,
60: rmr@astro.ucla.edu}
61: %\medskip
62: 
63: 
64: \begin{abstract}
65: Using the Spitzer IRAC camera we have 
66: obtained mid-IR photometry of the red giant branch stars 
67: in the Galactic globular cluster 47~Tuc. 
68: About  100
69: stars show an excess of mid-infrared light above that expected from
70: their photospheric emission. This is
71: plausibly due to dust formation in
72: mass flowing from these stars. 
73: This mass loss extends down to the level of the horizontal branch
74: and increases with luminosity. The mass loss is episodic, occurring in
75: only a fraction of stars at a given luminosity.  
76: Using a simple model and 
77: our observations we derive mass loss rates for these stars. 
78: Finally, we obtain
79: the first empirical mass loss formula calibrated with observations of
80: Population II stars. The dependence on luminosity of our mass loss rate is considerably shallower 
81: than the widely used Reimers Law. The results presented here are the
82: first from our Spitzer survey of a carefully chosen sample of 17
83: Galactic Globular Clusters, spanning the entire metallicity range from
84: about one hundredth up to almost solar.  
85: \end{abstract}
86: 
87: \keywords{stars: Population II, mass loss -- circumstellar matter -- infrared: stars}
88: 
89: \section{Introduction}
90: \label{intro}
91: 
92: Though the current generation of theoretical models can reasonably
93: reproduce the general framework of stellar evolution, there are still
94: a number of physical phenomena which are poorly understood. Among
95: these the stellar mass loss (hereafter ML) is one of the most
96: vexing. This is particularly the case for cool stars. Indirect
97: evidence shows that ML strongly affects all of their late stages of
98: evolution, yet we have little theoretical or observational guidance of
99: how to incorporate ML into our models.  We have relied on empirical
100: laws like that of \citet{rei75a,rei75b} based on observations of
101: Population~I giants.  While subsequent work
102: \citep{mul78,gol79,jud91,cat00} led to slight refinements, a ``Reimers
103: Law'' or some variant has been the {\em only} basis for stellar
104: evolutionary models of cool stars at all ages and metallicities.
105: Indeed, a ML law directly calibrated on Population II low-mass giants
106: has never been determined. Galactic globular clusters (GGCs) are particularly attractive
107: observational targets because there is so much indirect, {\em but
108: quantitative,} evidence for ML. This includes the observed morphology
109: of the horizontal branch (HB) in the cluster color-magnitude diagrams
110: (CMDs), the pulsational properties of the RR Lyrae stars, and the
111: absence of asymptotic giant branch (AGB) stars brighter than the red
112: giant branch (RGB) tip, and the masses inferred for white dwarfs in
113: GCCs \citep{roo73,ffp75,ffp76,ren77,ffp93,cru96,han05,kal07}. 
114: 
115: \begin{figure*}
116: %\epsscale{1.8}
117: \plotone{fig1.eps}
118: \caption{Three color (3.6$\mu$m (blue), 6$\mu$m (green), 8$\mu$m (red) mosaiced image of 47~Tuc 
119: from IRAC. 
120: \label{ima}}
121: \end{figure*}
122: 
123: \begin{figure*}
124: %\epsscale{2.3}
125: \plottwo{fig2a.eps}{fig2b.eps}
126: \caption{Absolute CMDs in the near IR and Spitzer planes of 47~Tuc. 
127: Left panel: stars detected in the shallow survey down  
128: to $K\approx11$ or $M_{\rm bol}\approx 0$ over a useful
129: $8'\times5'$ field of view.
130: Right panel: stars detected in the deeper survey down to $K\approx14$ which
131: excludes the inner $\approx2'$ (by radius) region.
132: The dashed horizontal line marks the position of the HB level.  
133: Stars with ($K-8$)$_0$ color excess at a 3$\sigma $ level are marked as filled triangles.
134: \label{cmd}}
135: \end{figure*}
136: 
137: Direct evidence for ML in GGC RGB stars has been inferred from blue
138: shifted features of photospheric lines like H$\alpha$
139: \citep{dup86,mau06}.  These are difficult observations and hard to
140: convert to ML rates.  One can also observe the outflowing gas much
141: further from the star after it has formed dust.  We detected this
142: circumstellar (CS) matter using ISOCAM and were able to derive ML
143: rates for a modest sample of RGB stars \citep{ori02}.  Mid-IR
144: observations have the advantage of sampling an outflowing gas fairly
145: far from the star (typically, tens/hundreds stellar radii).  Such gas
146: typically left the star a few decades previously, hence the inferred
147: ML rate is also smoothed over such a timescale.  From a combined
148: physical and statistical analysis, our ISOCAM study provided ML rates and frequency.
149: We found that {\it i)}~the largest ML occurs near the very end of
150: the RGB evolutionary stage and is episodic, {\it ii)} typical rates
151: are in the range $10^{-7} < dM/dt < 10^{-6} M_{\odot}\,{\rm yr^{-1}}$,
152: {\it iii)} ~the modulation timescales must be greater than a few
153: decades and less than a million years, and {\it iv)}~there is evidence
154: for dusty shells at even the lowest metallicities. However, our ISOCAM
155: survey suffered from two significant limitations. The small
156: sample of observed clusters and the consequent low number of detected 
157: giants allowed us to reach only weak conclusions on the ML dependence
158: on metallicity and HB morphology. Further, the modest spatial
159: resolution and sensitivity compromised our ability to measure lower ML
160: rates near the RGB tip and made it impossible to explore ML much below the
161: RGB tip.
162: 
163: \section{The Spitzer IRAC survey}
164: 
165: A sample of 17 massive GGCs, 4--5 per each 0.5 dex bin in
166: metallicity between ${\rm [Fe/H]} =-2.3$ and $-0.5$ and different HB
167: morphologies within each bin has been observed with IRAC onboard
168: Spitzer with 26 hr of observing time allocated to our program (ID
169: \#20298) in Cycle~2.  For all these clusters complementary near-IR and
170: UV photometry are available to properly characterize both the red and
171: the blue sequences.
172: 
173: In this Letter we discuss the first cluster analyzed, 47~Tuc.  This
174: cluster was observed on September 21, 2005.  We used a $1\times3$ grid
175: with small cycling dither pattern to cover the central $9' \times 5'$
176: of the cluster in all the four IRAC filters and the {\it High Dynamic
177: Range} Readout Mode, to avoid saturation of the brightest giants.  A
178: total observing time of 1.2\,hr allowed us to reach $K\le14$ with
179: $S/N\approx20$.  
180: Fig.~\ref{ima} shows the mosaiced three-color image of 47~Tuc.
181: The Post BCD mosaic frames from the Spitzer Pipeline
182: (Software Version: S13.2.0) have been photometrically reduced with
183: ROMAFOT \citep{buo83}, a software package optimized for Point Spread
184: Function fitting in crowded and undersampled stellar fields.  Because
185: of the intrinsic luminosity of 47~Tuc, the crowding in its central
186: region is critical even at 8~$\mu$m.  We thus obtained two photometric
187: catalogs: a shallow one down to $K\approx11$ over a useful
188: $8'\times5'$ field of view and a deeper one down to $K\approx14$ which
189: excludes the inner $\approx2'$ (by radius) region.  The overall
190: photometric uncertainty in all the four IRAC filters is $\le0.1$ mag.
191: Complementary ground-based near-IR photometry of the central region at
192: high spatial resolution has been obtained using IRAC2\@ ESO2.2m, and
193: SOFI@ESO-NTT \citep{fer00,val04} and supplemented with 2MASS data in
194: the external region.  
195: The degree of completeness of the near-IR survey is
196: 100\% over the full magnitude range covered by the Spitzer survey.
197: The shallow Spitzer catalog counts almost 400 stars and it is $\ge 86\%$ complete in the upper two RGB
198: magnitudes, and $\approx66\%$ complete at $0\ge M_{\rm bol}>-2$, while the deeper
199: Spitzer catalog in the outer region counts almost 500 stars and it is $\ge$83\% complete down to $K\le14$.
200: 
201: The two Spitzer catalogs have been combined
202: and placed onto the 2MASS astrometric system by cross-correlating the
203: Spitzer and the ground-based near-IR surveys.  The final catalog
204: contains almost 800 stars with  $J,~K$, and Spitzer photometry in each of the
205: four filters.  
206: 
207: The dereddened $K_0$ magnitudes and $(J-K)_0$ colors have been also
208: used to compute the stellar bolometric magnitudes and effective
209: temperatures, by using the transformation of \citet{mon98},
210: reddening ($E(B-V)=0.04$) and distance modulus ($(m-M)_0=13.32$) from
211: \citet{fer99}.  Fig.~\ref{cmd} shows the absolute color-magnitude
212: diagrams (CMDs) of 47~Tuc in the near and mid IR planes. 
213: By knowing the stellar temperature and bolometric magnitude and using a
214: typical RGB stellar mass of 0.8--0.9 M$_{\odot}$,
215: we finally estimate the stellar radius and gravity 
216: from the following equations:
217: \begin{displaymath}
218: L_{\star} = 4\pi R_{\star}^2 \times \sigma T_{\rm eff}^4 ~~~~~g = G M_{\star}/R_{\star}^2
219: \end{displaymath}
220: 
221: \begin{figure*}
222: %\epsscale{2.0}
223: \plotone{fig3.eps}
224: \caption{ML rates for the Spitzer sources with dust excess, as a
225: function of the observed ($K-8$)$_0$ color (top panel), bolometric
226: magnitude (middle panel) and normalized stellar luminosity (bottom
227: panel).  In the top panel stars with $M_{\rm bol}<-2$ and $M_{\rm bol}\ge-2$ are
228: plotted as filled and open squares, respectively.  In the middle and
229: lower panels
230: filled big triangles mark those giants which are known long period
231: variables, filled small triangles are other AGB stars.  In the middle panel
232: the evolutionary timescale $\Delta t$ in 2 luminosity
233: intervals are also reported.  In the bottom panel
234: all of the AGB stars are excluded
235: from the fit (solid lines). 
236: Typical random error bars are shown in the bottom right corner.
237: The empirical laws by
238: \citet{rei75a,rei75b} with $\eta =0.3$ (short-dashed line) and \citet{cat00}
239: (long-dashed line) calibrated on Population~I giants are also shown
240: for comparison.
241: \label{ml}}
242: \end{figure*}   
243:  
244: 
245: \section{Dust excess and mass loss rates}
246: 
247: As CS dust condenses in an outflowing wind, it can be detected as a
248: mid-IR excess.  In our pilot project using ISOCAM photometry in the 10 $\mu $m
249: window and  assuming a $\nu B_{\nu}$ emissivity, we showed that the
250: bulk of CS dust around the RGB tip giants typically has temperatures
251: in the range 300--500\,K.  The IRAC bands between 3.6 and 8 $\mu $m
252: can also be used to detect this warm dust when coupled with near IR
253: photometry used to properly characterize the stellar counterpart.
254: 
255: In order to select candidate stars with dust excess, we define first
256: the mean ridge lines in each of the $K_0,~(K-IRAC)_0$ CMDs to define
257: the average color of the stars with purely photospheric emission and
258: to determine the overall photometric errors ($\sigma$) at different magnitude
259: bins.  These $(K-IRAC)_0$ colors are $\approx 0.0 \pm 0.1$
260: along the entire RGB range sampled by our survey. This is in good
261: agreement with the prediction of theoretical model atmospheres with
262: $T_{\rm eff}=3500$--5000 K from the Kurucz database.  Since the 8 $\mu
263: $m IRAC band is the most sensitive to warm dust emission, stars are
264: flagged as dusty if they show a $(K-8)_0$ color excess $\ge
265: n\sigma$. These stars are also the reddest in the other IRAC bands.
266: For these stars, we 
267: also compute the pure dust emission at 8 $\mu m$, by subtracting 
268: from the the total observed flux  
269: the photospheric contribution, given by $F_8^{\rm phot} = 
270: F_8^{\rm Vega}\times 10.^{[-0.4\times (K-(K-8)_{\rm phot})]}$, where $(K-8)_{\rm phot}$ 
271: is the mean ridge line color without dust excess.
272: 
273: The number of giants with dust excess in the
274: shallow survey of 47~Tuc is 93 at a 3$\sigma$ level.  In the deeper
275: survey of the cluster outer region, no stars with dust excess have
276: been detected below the HB level.  The seven 47~Tuc stars which showed
277: dust excess in our ISOCAM survey have been also detected by Spitzer
278: and confirmed as dusty stars.  Among the Spitzer dusty giants, four
279: are known long period variables (V1, V4, V6, \& V8 in \citet{cle01})
280: and an other 17 stars are classified as AGB stars by
281: \citet{bec06}.  Hence, in the following we classify the remaining 74
282: giants as true RGB dusty stars.
283: 
284: In order to obtain the ML rates we use our modified version of the
285: DUSTY code \citep{ive99,eli01}, to compute the emerging
286: spectrum and dust emission at the IRAC wavelengths.  We adopt Kurucz
287: model atmospheres for the heating source and for the dust a mixture of
288: warm silicates with an average grain radius $a=0.1\,\mu $m.  Slightly
289: different choices for the latter two parameters have negligible impact
290: in the resulting IRAC colors and mass loss rates.  While radiation
291: pressure acting on dust might plausibly drive winds in luminous, metal
292: rich red giants \citep{wil00}, the GGC stars are generally neither
293: luminous nor metal rich enough for this mechanism to be efficient.
294: Hence we run the DUSTY code under the general assumption of an
295: expanding envelope at constant velocity $v_{\rm exp}$ with a density
296: profile $\eta \propto r^{-2}$, a dust temperature for the inner
297: boundary $r_{\rm in}$ of 1000~K and a shell outer boundary $
298: r_{\rm out}=1000 \times r_{\rm in}$.  We then computed a large grid of DUSTY
299: models with stellar temperatures in the 3500--5000 K range and optical
300: depths at 8 $\mu $m ($\tau_8$) between 10$^{-5}$ and 10$^{-1}$.
301: For each star with dust excess, we enter the grid with its empirical
302: stellar temperature and $(K-IRAC)_0$ colors, and we exit with
303: prediction for the optical depth, emerging flux, dust fractional contribution 
304: and envelope radius.
305: The mass loss rates are computed by using the formula
306: \begin{displaymath}
307: dM/dt = 4 \pi  r_{\rm out}^2 \times \rho_{\rm dust}
308: \times v_{\rm exp} \times \delta
309: \end{displaymath}
310: where 
311: \begin{displaymath}
312: \rho_{\rm dust}\propto
313: \rho_{\rm g} \tau_8 F_8({\rm obs})/F_8({\rm mod}) D^2 
314: \end{displaymath}
315: 
316: \noindent
317: is the dust density, $\rho_{\rm g}=3\,
318: {\rm g\,cm^{-3}}$ is the grain density, 
319: $F_8(obs)$ and $F_8(mod)$ are the observed and model dust emission at 8 micron,
320: respectively,
321: $D$ the distance and $\delta$ the gas to dust ratio.
322: A lower limit to $\delta$ is given by $1/Z$ where $Z$ is the global
323: metallicity.  $v_{\rm exp}$ is a free parameter, which should scale
324: like $\delta ^{-0.5}$ if dust and gas are coupled.  If 
325: the number of grains is increased (by decreasing $\delta$), the momentum per
326: grain is shared with fewer gas molecules, thus implying an enhancement
327: of $v_{\rm exp}$ \citep{hab94,van00}.  
328: For 47~Tuc we adopt
329: $\delta\approx 1/Z\approx 200$ and $v_{\rm exp}=10~{\rm km\,s^{-1}}$. 
330: The latter is the average expansion velocity measured in luminous, low mass giants
331: \citep[see e.g.][]{net93} which ranges between a few and $\approx 20~{\rm km s^{-1}}$. 
332: 
333: \section{Results and Discussion}         
334: 
335: For the dusty stars in 47~Tuc sampled by our Spitzer survey,
336: Fig.~\ref{ml} shows the inferred mass loss rates as a function of {\it
337: i)}~ the observed $(K-8)_0$ color, {\it ii)}~the bolometric magnitude
338: and {\it iii)}~the normalized luminosity.  The mass loss rate
339: increases with increasing color excess and stellar luminosity.  
340: Also,
341: the bulk of the mass loss along the RGB should occur above the HB
342: level.  
343: The provisional empirical law based on such a first set of
344: observations gives:
345: \begin{eqnarray*}
346: dM/dt = 
347: & C \times 4 \times 10^{-10} \times (L/gR)_{\odot}^{0.4}  
348: \end{eqnarray*}
349: %
350: where
351: %
352: \begin{eqnarray*}
353: C = & (\delta/200)^{0.5}\times (v_{\rm
354: exp}/10)\times(\rho_{\rm g}/3) 
355: \end{eqnarray*}
356: and 
357: $L_{\odot}$, $g_{\odot}$, and $R_{\odot}$ are the stellar luminosity, 
358: gravity and radius in solar units. In this study of 47~Tuc $C=1$. 
359: Only true RGB stars are used to derive the fitting formula.
360: Errors are as follows: 
361: $\approx 10$\% for 
362: $L/gR$ and the fit exponent  
363: and $\approx25$\% for the fit zero point. The latter 
364: is also a good estimate of the average random error on the final 
365: mass loss rates.   
366: As shown in Fig.~\ref{ml}, this new law calibrated on Population II
367: RGB stars is significantly flatter than the original Reimers
368: formulation and the one revised by \citet{cat00}, which have slopes of
369: 1 and 1.4, respectively.  
370: 
371: As we found in our ISOCAM survey, only a
372: fraction of stars along the RGB are currently losing mass: $\approx
373: 32$\% in the upper 2 mag, $\approx16$\% down to the HB.  For this
374: reason we conclude that ML is episodic. Basically mass loss is ``on''
375: for only some fraction of the time, $f_{\rm on}$. By using a suitable
376: evolutionary track for a RGB star of $M=0.9\, M_{\odot}$ and $Z=0.004$
377: \citep{pie06}, we can derive the evolutionary timescale in each
378: luminosity interval (see Fig.~\ref{ml}).  Multiplying this by $f_{\rm
379: on}$, we find that the timescale ML is ``on'' is less than a few Myr
380: in each interval.  
381: By using the simple equation 
382: \begin{eqnarray*}
383: \Delta M_{\rm RGB}  = \Sigma_i~ ( dM/dt_i \times \Delta t_i \times f_{{\rm on}_i} )
384: \end{eqnarray*}
385: to integrate the mass loss formula multiplied by
386: $f_{\rm on}$ over the RGB evolution time $\Delta t$, 
387: we find that the total mass
388: lost on the RGB is $\Delta M_{\rm RGB}\approx 0.23\pm 0.07 M_{\odot}$.
389:  
390: In forthcoming papers we will present the results for the other
391: clusters, with the ultimate goal of providing {the first empirical ML
392: law for Population~II stars} calibrated over a large range of
393: metallicity, and investigating whether observed ML in individual stars
394: within a cluster correlates with that cluster's HB morphology and 
395: if ML itself is involved in the second-parameter problem.
396: 
397: \acknowledgments
398: This research was supported by 
399: the Ministero dell'Istru\-zio\-ne, Universit\`a e Ricerca (MIUR), the 
400: PRIN-INAF 2006 and the Italian Space Agency (ASI).
401: RTR and RMR acknowledge support from Spitzer Science Center Grant GO-20298.
402: We warmly thanks Elena Valenti for having provided the SOFI near IR photometric catalog 
403: and the anonymous Referee for his/her useful comments.
404: 
405: \begin{thebibliography}{}
406: 
407: \bibitem[Beccari et al.(2006)]{bec06}    
408: Beccari, G., Ferraro, F. R., Lanzoni, B., Bellazzini, M. 2006, \apj, 652, 121
409: \bibitem[Buonanno et al.(1983)]{buo83}    
410: Buonanno, R., Buscema, G., Corsi C. E., Ferraro, I., \& Iannicola, G.,
411: 1983, \aap, 126, 278
412: \bibitem[Catelan(2000)]{cat00}    
413: Catelan, M. 2000, \apj, 531, 826
414: \bibitem[Clement et al.(2001)]{cle01}    
415: Clement, C. M. 2001, \aj, 122, 2587
416: \bibitem[D'Cruz et al.(1996)]{cru96}    
417: D'Cruz, N. L., Dorman, B., Rood, R.T., O'Connell, R. W. 1996, \apj, 466, 359
418: \bibitem[Dupree(1986)]{dup86}             
419: Dupree, A. K. 1986, ARA\&A, 24, 377
420: \bibitem[Elitzur \& Ivezi\'c(2001)]{eli01}
421: Elitzur, M., \& Ivezi\'c, Z. 2001, \mnras, 327, 403 
422: \bibitem[Ferraro et al.(1999)]{fer99}    
423: Ferraro, F.R., Messineo, M., Fusi Pecci, F., De Palo, M. A., Straniero, O., 
424: Chieffi, A., Limongi, M. 1999, \aj, 118, 1738
425: \bibitem[Ferraro et al.(2000)]{fer00}    
426: Ferraro, F.R., Montegriffo, P., Origlia, L., Fusi Pecci, F. 2000, \aj, 119, 1282
427: \bibitem[Fusi Pecci \& Renzini(1975)]{ffp75}    
428: Fusi Pecci, F., \& Renzini, A. 1975, \aap, 39, 413
429: \bibitem[Fusi Pecci \& Renzini(1976)]{ffp76}    
430: Fusi Pecci, F., \& Renzini, A. 1976, \aap, 46, 447
431: \bibitem[Fusi Pecci et al.(1993)]{ffp93}    
432: Fusi Pecci, F., et al. 1993, \aj, 105, 1145
433: \bibitem[Goldberg(1979)]{gol79}    
434: Goldberg, L. 1979, QJRAS, 20, 361
435: \bibitem[Hansen(2005)]{han05}    
436: Hansen, B. M. S. 2005, \apj, 635, 526
437: \bibitem[Habing, Tignon \& Tielens(1994)]{hab94}    
438: Habing, H.J., Tignon, J., \& Tielens, A.G.G.M. 1994, \aap, 286, 523
439: \bibitem[Kalirai et al.(2007)]{kal07}    
440: Kalirai, J. S. et al. 2007, \apj, in press (astro-ph/07050977)
441: \bibitem[Ivezi\'c, Nenkova \& Elitzur(1999)]{ive99}
442: Ivezi\'c, Z., Nenkova, M., \& Elitzur, M., 1999, User Manual for DUSTY,
443: (Lexington: Univ. Kentucky)
444: \bibitem[Judge(1991)]{jud91}    
445: Judge, P.G., \& Stencel, R. E. 1991, \apj,  371, 357
446: \bibitem[Mauas, Cacciari, Pasquini(2006)]{mau06}
447: Mauas, P.J.D., Cacciari, C., Pasquini, L. 2006, \aap, 454, 609
448: \bibitem[Montegriffo et al.(1998)]{mon98}
449: Montegriffo, P., Ferraro, F.R., Fusi Pecci, F., \& Origlia, L.
450: 1998, \mnras, 297, 872
451: \bibitem[Mullan(1978)]{mul78}    
452: Mullan, D. J. 1978, \apj, 226, 151
453: \bibitem[Netzer \& Elitzur(1993)]{net93}
454: Netzer, N., \& Elitzur, M. 1993, \apj, 410, 701
455: \bibitem[Origlia et al.(2002)]{ori02}   
456: Origlia, L., Ferraro, F.R., Fusi Pecci, F., Rood, R.T. 2002, \apj, 571, 458
457: \bibitem[Reimers(1975a)]{rei75a}   
458: Reimers D. 1975a, in {Problems in Stellar Atmospheres and Envelopes},
459: eds. B. Baschek, W. H. Kegel, \& G. Traving (Berlin: Springer), 229
460: \bibitem[Reimers(1975b)]{rei75b}   
461: Reimers D. 1975b, in {\it Mem. Soc. R. Sci. Li\`ege 6 Ser.}, 8, 369
462: \bibitem[Renzini(1977)]{ren77}   
463: Renzini, A. 1977, in {\it Advanced Stages of Stellar
464: Evolution}, ed.\ Bouvier \& A. Maeder (Geneva Obs., Geneva), 151
465: \bibitem[Rood(1973)]{roo73}   
466: Rood, R.T. 1973, ApJ, 184, 815
467: \bibitem[Pietrinferni et al.(2006)]{pie06}
468: Pietrinferni, A. Cassisi, S., Salaris, M., Castelli, F. 2006,
469: \apj, 642, 797
470: \bibitem[Valenti et al.(2004)]{val04}
471: Valenti, E., Ferraro, F.R., Origlia, L. 2004, \mnras, 351, 1204
472: \bibitem[van Loon(2000)]{van00}
473: van Loon, J. T. 2000, \aap, 354, 125
474: \bibitem[Willson(2000)]{wil00}
475: Willson, L. A. 2000, ARA\&A,  38, 573
476: \end{thebibliography}
477: \end{document}
478: 
479: