1: %\documentstyle[emulateapj]{article}
2: \documentclass[12pt,preprint]{aastex}
3: %\documentclass[emulateapj,twocolumn,article]{aastex}
4: \usepackage{graphics}
5:
6: %\def\gtorder{\stackrel{\textstyle>}{\sim}}
7: %\def\ltorder{\stackrel{\textstyle<}{\sim}}
8: \def\gtorder{\mathrel{\raise.3ex\hbox{$>$}\mkern-14mu
9: \lower0.6ex\hbox{$\sim$}}}
10: \def\ltorder{\mathrel{\raise.3ex\hbox{$<$}\mkern-14mu
11: \lower0.6ex\hbox{$\sim$}}}
12:
13: \def\Msun{\>{\rm M_{\odot}}}
14:
15: \shorttitle{Search for Comets Around HD209458}
16: \shortauthors{Mendelowitz, Ge, Mandell \& Li}
17:
18:
19: \begin{document}
20: \title{A Search for Sodium Absorption from Comets Around HD209458}
21: \author{Caylin Mendelowitz\altaffilmark{1}, Jian Ge\altaffilmark{1},
22: Avi M. Mandell\altaffilmark{1}, and Aigen Li\altaffilmark{2}}
23: \altaffiltext{1}{Pennsylvania State University, Department of
24: Astronomy \& Astrophysics, University Park, PA 16802, USA Email:
25: caylinm@astro.psu.edu, jian@astro.psu.edu, mandell@astro.psu.edu}
26: \altaffiltext{2}{Theoretical Astrophysics Program, Steward
27: Observatory, and Lunar \& Planetary Laboratory, University of Arizona,
28: Tucson, AZ 85721, USA Email: agli@lpl.arizona.edu}
29:
30: \begin{abstract}
31: We monitored the planet-bearing solar-type star HD209458 for sodium
32: absorption in the region of the stellar NaI D1 line that would be
33: indicative of cometary activity in the system. We observed the star
34: using the HET HRS with high S/N and high spectral resolution for 6
35: nights over the course of two years, from July 2001 to July 2003. From
36: modelling we determine a 20\% likelihood of a detection, based on a
37: predicted number of comets similar to that of the solar system. We
38: find that our analytical method is able to recover a signal and our
39: S/N is sufficient to detect this feature in the spectral regions on
40: either side of the core of the D1 line, where it is most likely to
41: appear. No significant absorption was detected for any of the nights
42: based on a 3$\sigma$ detection limit. We derive upper limits on the
43: column density of sodium of $\lesssim 6 \times 10^{9}$ cm$^{-2}$ for a
44: signal in the region around the line core and $\lesssim 2 \times
45: 10^{10}$ cm$^{-2}$ for a signal in the core of the photospheric D1
46: line. These numbers are consistent with the sodium released in a
47: single periodic comet in our own system, though higher S/N may be
48: necessary to uncover a signal in the core of the D1 line. Implications
49: for cometary activity in the HD209458 system are discussed.
50:
51: \end{abstract}
52:
53: \keywords{planetary systems:formation and detection- stars:individual(HD209458) - techniques:spectroscopic - comets:general}
54:
55: \section{Introduction}
56: The Falling Evaporating Bodies (FEB) scenario \citep{Lagrange92} was
57: introduced as an explanation for time variable circumstellar
58: absorption features observed in the spectral lines of the A-type star
59: $\beta$ Pictoris \citep{Lagrange86,Beust00}. Similar features have
60: also been observed in other early-type main sequence stars and some
61: Herbig Ae/Be stars, which are considered $\beta$ Pictoris precursors
62: \citep{Welsh98,Grady97,Grinin96III}. According to the FEB hypothesis,
63: solid bodies with sizes typical of solar system comets, are evaporated
64: within the vicinity of the star producing absorption lines when
65: observed along the line of sight \citep{Ferlet87,Lagrange88}. The high
66: frequency of infalling comets is a phenomenon invoked in models of the
67: early solar system and corresponds to the clearing out stage in a
68: system's formation \citep{Welsh98,Beust98}. The mechanism behind the
69: infall of these bodies is thought to be from the gravitational
70: perturbation by at least one planet \citep{Beust00}. Monitoring known
71: planet-bearing solar analogs for comet activity would then give
72: insight into the evolution of an exosolar system and such a wide study
73: would investigate how comet populations evolve in time and with
74: stellar mass. Also, searches for comet disks and clouds orbiting other
75: stars offers a new method for inferring the presence of planetary
76: systems \citep{Stern93}.
77:
78: Sungrazing comets and periodic comets are quite frequent in our own
79: system. Since the first light of the SOlar and Heliospheric
80: Observatory (SOHO) Large Angle and Spectrometric COronograph (LASCO)
81: on 1995 December 30 over 600 new sungrazing comets have been
82: discovered ($\sim$60/year) \citep{SOHOweb}. Also, the number of
83: catalogued periodic comets totals $\sim$900 with half having perihelia
84: $\le$ 1 AU \citep{Biswas00}. Comets evaporate a significant amount of
85: material as they approach perihelion in the form of a coma and
86: tails. For example, Halley's comet loses $\sim$0.001 of its mass every
87: perihelion passage ($\sim10^{11}$kg) \citep{Zeilik98}. Hale-Bopp is a
88: richer dust factory: in its 1997 apparition, it lost a total dust mass
89: of $\sim3 \times 10^{13}$ kg \citep{Jewitt99}. The absorption features
90: expected for many species known to be present in comets are quite weak
91: due to the small column densities. \citet{Cremoneseal97} discovered
92: the sodium tail in comet C/1995 O1 Hale-Bopp, which is a third type of
93: cometary tail consisting of neutral atoms. Sodium has also been
94: observed in the tail of some other bright comets and in the comae of
95: comets that were located at around 1 AU from the Sun (e.g., comet
96: 1P/Halley) \citep{Wanatabe03}. Sodium is a good candidate for
97: extra-solar detection because it is detectable even when the column
98: density is low \citep{Cremoneseal97}, and sodium has already been
99: detected in absorption from gas in the $\beta$ Pictoris disk
100: \citep{Hobbs85}, in variable absorption in some Herbig Ae/Be stars
101: \citep{Sorelli96}, and in the atmosphere of the exosolar planet HD
102: 209458b \citep{Charbonneau02}.
103:
104: HD209458 is a solar-type star with a close-in Jovian-type planet. It
105: is oriented edge-on to our line of sight making it an ideal candidate
106: for absorption spectroscopy and the detection of orbiting bodies. This
107: has initiated several new exoplanet discoveries. It is the first
108: exosolar system for which repeated transits across the stellar disk
109: have been observed \citep{Charbonneau00}. This allowed
110: \citet{Charbonneau02} to make the first observation of an exosolar
111: planet atmosphere through the detection of neutral sodium
112: absorption. More recently, \citet{VidalMadjar03} detected atomic
113: hydrogen which is associated with the upper atmosphere of the
114: planet. Here we observe HD209458 for neutral sodium absorption from a
115: comet crossing the line of sight. We present spectra of HD209458 in
116: the vicinity of the Na D1 line (5895.9243 \AA). In $\S$2 we present
117: the data and describe the data reduction technique. We analyze the
118: data and remove the photospheric contribution of sodium. In $\S$3 we
119: use our simulation of a cometary sodium absorption feature to examine
120: its position in relation to the photospheric sodium line and its shape
121: based on expected values of the column density, in order to define a
122: search region. A synthetic spectrum is analyzed to check the
123: reliability of our method. In $\S$4 we examine the residuals, discuss
124: the results and derive upper limits for the column density of sodium
125: on each night. We use modelling to determine the likelihood of
126: detecting a comet and conclude with possible explanations for the
127: non-detection and a discussion of comet activity in the system.
128:
129: \section{Observations \& Analysis}
130: Observations of HD209458 were made with the High Resolution
131: Spectrograph (HRS) at the 9.2 meter Hobby-Eberly Telescope (HET)
132: \citep{Tull98}. The HRS is a fiber-coupled echelle spectrograph, using
133: an R-4 echelle mosaic with cross-dispersing gratings. The camera
134: images onto a mosaic of two 2K x 4K CCDs with 15 $\mu$m pixels and a
135: gap between them that spans $\sim$72 pixels. The spectrograph was used
136: in high-dispersion mode, with an average resolving power of $R =
137: \lambda/\Delta\lambda \sim 115,000$ measured from thorium lines,
138: corresponding to a resolution element of $\sim$0.049 \AA\ in the
139: region of interest. The spectral coverage for the cross disperser
140: setting used is between 4076 \AA\ and 7838 \AA. Observation dates and
141: parameters are listed in Table \ref{observations}. The star HD12235
142: was also observed to serve as a reference system. We observed for two
143: nights on 2002 August 28 and 31, using the same instrumental settings
144: as HD209458. Stellar parameters are listed in Table \ref{params}.
145:
146: Images typically had an integration time of 5 to 25 minutes depending
147: on brightness, and a total integration time between 1 and 2 hours. A
148: Thorium-Argon lamp was observed before and after each sequence of
149: exposures. Due to sodium emission lines in the flatfield illumination
150: lamp, it was necessary to use extremely high S/N observations of
151: rapidly rotating B and A stars as flat fields. These calibration stars
152: were observed as near in time to the program stars as possible. The
153: data were reduced using the publically available IRAF software
154: package. Continuum normalization was performed by selecting continuum
155: regions in each image without obvious absorption lines or bad pixels
156: and fitting a higher order polynomial to these sections. Images were
157: then combined with appropriate rejection algorithms. Wavelength
158: calibration of the spectra was obtained using the Thorium-Argon lamp
159: comparison spectra for each data set. The 2002 December 9 data was
160: missing a calibration star so the lamp was used for flatfielding, with
161: the sodium emission lines removed using a polynomial fit to obtain a
162: featureless spectrum. The S/N per pixel in the continuum for each
163: image is listed in Table \ref{upplim}. Figure \ref{plotone} shows the
164: reduced spectra for HD209458 for each of the 6 nights. The spectra
165: have been shifted to the rest frame of HD209458.
166:
167: The D1 line was used for analysis since the D2 line is heavily
168: contaminated by telluric lines \citep{Ge98}. The broad photospheric
169: sodium line was removed using a polynomial fit (see Figure
170: \ref{plotone}), including multiple rejection iterations and the
171: normalized spectrum was searched for possible narrow absorption
172: features up to 0.7 \AA\ from the photospheric line center. Several
173: absorption features above 3$\sigma$ were attributed to telluric
174: features that were removed imperfectly due to changes in observing
175: conditions between exposures of HD209458 and the calibration
176: star. This is the case for the 2002 October 10 spectrum where there is
177: a significant absorption feature around 5896.5 \AA; the feature is
178: also present in the calibration star used for flatfielding. For 2001
179: July 7 observing conditions were deteriorated by the fourth exposure,
180: which is the cause of the absorption feature near 5895.86 \AA. No
181: absorption features greater than 3$\sigma$ due to cometary activity
182: were identified.
183:
184: \section{Modelling}
185: Simulations were run to produce estimates of the position and shape of
186: a sodium absorption feature due to solar-like comets in the HD209458
187: system. Constraints on spatial and chemical characteristics of
188: cometary bodies were taken from observations of short and long period
189: comets in our own system, mainly the comets Halley and Hale-Bopp
190: \citep{Swamy97,Zeilik98,Cremoneseal97}. Although sungrazing comets
191: also display tail and coma features, they are much smaller than
192: periodic comets, only a few meters to a few tens of meters
193: \citep{Iseli02} (whereas periodic comets are a few kilometers to
194: hundreds of kilometers), and therefore will not release a significant
195: amount of dust. The position of the feature will depend on the
196: velocity of the comet when detected; as a comet moves within $\sim$3
197: AU the cometary nucleus begins to display coma formation followed by
198: the extension of gas and ion tails. Full display of the comet tail
199: occurs between 1.0 and 1.5 AU, \citep{Biswas00} with sodium extending
200: beyond 10$^{5}$ km from the nucleus \citep{Oppenheimer80}. In
201: Hale-Bopp the length of the sodium tail was measured to be 3$\times
202: 10^{7}$ km \citep{Cremonese97}. In order to obtain the maximum column
203: density along the line of sight, a theoretical comet velocity was
204: calculated specifically for an orientation aligned with the sodium
205: tail, which points roughly anti-sunward. Since the radius and
206: luminosity of HD209458 are larger than for the Sun, the orbital
207: distance for detecting the sodium tail was set to within 3 AU of the
208: star.
209:
210: The wavelength shift due to the orbital motion of the comet with
211: respect to the parent star is described by the radial velocity of the
212: comet:
213: \begin{equation}
214: v_{r}=(2a/P)(e\sin\theta)(1-e^{2})^{-1/2}
215: \end{equation}
216: where $e$ is the eccentricity, $a$ is the semi-major axis, $P$ is the
217: period and $\theta$ is measured so that $\theta$=0 is at
218: perihelion. The distance from the star is given by:
219: \begin{equation}
220: r = \frac{a(1-e^{2})} {e\cos\theta+1}
221: \end{equation}
222: The range of velocity shifts for a spectral feature have been
223: calculated using the parameters of the short-period (SP) comet Halley
224: ($e$=0.967, a=18 AU, P=76 years, perihelion distance q=0.587 AU) and
225: the long-period (LP) comet Hale-Bopp ($e$=0.9951, a=187 AU, P=2550
226: years, q=0.914 AU). Figure \ref{plottwo} shows the wavelength and
227: corresponding doppler shift for both types of comet. The velocity and
228: thus the wavelength shift for Halley peaks at 1.17 AU, with a maximum
229: shift of 0.53 \AA\ from the photospheric line center. For Hale-Bopp
230: the maximum velocity is reached at 1.83 AU, which results in a maximum
231: wavelength shift of 0.43 \AA. The majority of the orbit, from
232: $\sim$1-3 AU, for both types of comet will produce a wavelength shift
233: of $\sim$0.4 \AA. Since the D1 line is $\sim$0.6 \AA\ wide, a feature
234: will most likely be in the wing or continuum region. We also take into
235: consideration the periodic comets with perihelia less than those of
236: Halley and Hale-Bopp ($\sim$220 comets \citep{Biswas00}), which would
237: produce a greater wavelength shift. Our search for Na absorption will
238: be $\pm$0.7 \AA\ around the stellar Na D1 line.
239:
240: The depth and shape of the feature depends on the column density and
241: temperature of the absorbing material. A model of the signal was
242: created using a Voigt absorption profile, which was then convolved
243: with the instrumental spectral resolution. Figure \ref{plotthree}
244: shows model absorption lines based on the core and continuum upper
245: limits derived for our best S/N data, 2002 December 9, in
246: $\S$4.1. Estimates of the column density of sodium from evaporating
247: comets in other systems are on the order of $10^{10}$ cm$^{-2}$ from
248: HR10 (FEB) \citep{Welsh98} and from $\beta$ Pictoris (disk)
249: \citep{Hobbs85}. The thermal broadening can be calculated by assuming
250: a gas temperature for the comet coma and tail. The estimated
251: temperature for the sodium tail of a comet, from solar system
252: observations, is 150K \citep{Swamy97}, and atomic parameters for the
253: sodium D1 line were taken from \citet{Morton91}. The inherent Doppler
254: width at 150K corresponds to only 0.03 \AA, and therefore a single
255: radial velocity absorption component would be spectrally unresolved,
256: with the instrumental FWHM of 0.049 \AA\ of the HRS. The equivalent
257: width of a sodium absorption line with a column density of 10$^{10}$
258: cm$^{-2}$ is $\sim$1 m\AA.
259:
260: \subsection{Synthetic Detection: Check on the Analysis}
261: As a check of our analytical method we input our synthetic absorption
262: features into the observed spectrum and recover the absorption through
263: fitting the broad stellar sodium D1 line profile. We use the data from
264: 2002 November 23 since telluric features have been successfully
265: removed and the S/N is the highest (They were not removed in 2002
266: December 9). Since the detection limit is lower in the continuum than
267: in the core of the line, we place the feature in the continuum
268: (S/N=237) with a column density of 6.4$\times 10^{9}$ cm$^{-2}$ and in
269: the core (S/N=88) with a column density of 1.7$\times 10^{10}$
270: cm$^{-2}$ derived in $\S$4.2 as the 3$\sigma$ upper limits for this
271: date. We also place a feature in the wing of the line with a column
272: density of 9.3 $\times 10^{9}$ cm$^{-2}$ based on a calculated S/N of
273: 162. The feature was added at the stage just before the removal of the
274: photospheric D1 line. Our method of polynomial fitting was able to
275: recover the features at the core, wing and also continuum region. The
276: residuals and recovered absorption features are shown in Figure
277: \ref{plotfour}. The equivalent width of the recovered features are
278: 1.6$\pm$0.6, 1.0$\pm$0.3, and 0.62$\pm$0.2 m\AA, for the core, wing,
279: and continuum respectively. These are all consistent with the input
280: values of 1.7, 0.9, and 0.6 m\AA, for the core, wing, and continuum
281: respectively. This demonstrates proof that the stellar D1 line profile
282: fitting technique can recover narrow sodium features from comet
283: events.
284:
285: \section{Results and Discussion}
286: \subsection{Upper Limits on Sodium Column Density by Comets}
287: The residual after the photospheric sodium line is removed are shown
288: in Figure \ref{plotfive}. No signal was detected in any of the spectra,
289: but upper limits were determined based on a 3$\sigma$ detection
290: limit. The 1$\sigma$ upper limit to the equivalent width for the
291: undetected feature was determined by:
292: \begin{equation}
293: W_{\lambda} = \int \frac{F_{c} - F_{\lambda}}{F_{c}} d\lambda
294: \approx \Delta\lambda \times \sigma \\
295: \end{equation}
296: where $F_{c}$ is the normalized flux in the continuum, $F_{\lambda}$
297: is the normalized absorption line, $\Delta\lambda$ is the spectral
298: resolution element 0.049 \AA, and $\sigma$ is:
299: \begin{equation}
300: \sigma(\lambda) = \frac{1} {F_{\lambda}I_{c}(\lambda)}
301: \{F_{\lambda}I_{c}(\lambda) + p\varepsilon^{2}\}^{1/2} \approx \frac{1}{S/N}
302: \end{equation}
303: $I_{c}(\lambda)$ = $N_{c}$ $\times$ gain, where $N_{c}$ is the number
304: of counts in the continuum and gain=0.6320 \emph{e}$^{-}$ ADU$^{-1}$,
305: $\varepsilon$ is the readnoise 2.70 \emph{e}$^{-}$, and $p$ is the
306: pixel number in the cross-slit direction used for sampling a
307: spectrum. It is appropriate to assume that the absorption line is in
308: the linear portion of the curve of growth, where the column density is
309: proportional to the equivalent width. The integrated number of sodium
310: atoms per square centimeter in the line of sight is then given by:
311: \begin{equation}
312: N = \frac{mc^{2}}{\pi e^{2} f} \int_{0}^{\infty} \frac{F_{c}(v) -
313: F(v)}{F_{c}(v)} dv = 1.130 \times 10^{20} \frac{W_{\lambda}}
314: {f\lambda^2} cm^{-2},\\
315: \end{equation}
316: where $f$ is the oscillator strength ($f$=0.318 for 5895.9243
317: \citep{Morton91}, and $W_{\lambda}$ and $\lambda$ are in units of
318: Angstroms. The upper limits for each date are given in Table
319: \ref{upplim}. In the continuum region, the column density for each
320: night ranges from $\sim6\times 10^{9}-2\times 10^{10}$ cm$^{-2}$,
321: whereas in the region of the D1 line core upper limits range from
322: $\sim2\times 10^{10}-7\times 10^{10}$ cm$^{-2}$. This appears to be
323: consistent with the solar comet sodium column density (e.g., sodium
324: absorption from an occultation of the solar system comet
325: Giacobini-Zinner resulted in upper limits of N(Na{\footnotesize I}) $<
326: 2 \times 10^{11}$ cm$^{-2}$, based on solar abundance ratios for
327: K{\footnotesize I} \citep{Schempp89}). For comparison, we also show
328: the residual for the reference star HD12235 in Figure
329: \ref{plotsix}. This star was reduced in exactly the same manner as
330: HD209458 and the stellar sodium D1 line was fit and removed using the
331: same methods as for HD209458. As expected, no features were present in
332: the spectra on either night. The 3$\sigma$ upper limits are 1.1
333: $\times 10^{10}$ cm$^{-2}$ (2002 August 28), and 9.4 $\times 10^{9}$
334: cm$^{-2}$ (2002 August 31), which are in agreement with our results
335: for HD209458.
336:
337:
338: \subsection{Detection Probability}
339: Since no signal was detected we examine the likelihood of observing a
340: comet crossing based on the amount of time the comet spends within an
341: observable window. We only consider the orbit within 3 AU since comet
342: features will not be significant beyond this distance from the
343: star. The amount of time the comet spends in front of the star is
344: determined by the diameter of the star, D$_{sun}$, and the tangential
345: velocity, v$_{tan}$. This time will vary depending on the orientation
346: of the orbit to our line of sight. The radius of HD209458 is
347: 1.18R$\odot$ $\approx 8.21 \times 10^{5}$ km, which is comparable to
348: the width of the sodium tail, 8$\times 10^{5}$ km
349: \citep{NASAHaleBopp}. For the total time in front of the star within a
350: year for any comet in the system we multiply by the average number of
351: periodic comets per year $\sim$25, $\sim$10 of which are SP comets
352: (Halley type) and $\sim$15 of which are LP comets (Hale-Bopp type)
353: \citep{Swamy97}. We double this since our observations are over the
354: course of two years. The time that any SP comet spends in front of the
355: star over two years ranges between 7 days and 35 days for v$_{tan}$
356: varying from 11 km s$^{-1}$ to 54 km s$^{-1}$, with an average time of
357: in front of the star of 10 days. For LP comets the time ranges between
358: 13 days and 42 days for v$_{tan}$ varying from 14 km s$^{-1}$ to 44 km
359: s$^{-1}$, with an average time of 18 days in front of the star. Since
360: our observing time is small compared to the time comets spend in front
361: of the star and observing a comet can be considered a random process,
362: the probability can be efficiently described by a Poisson distribution
363: $P(t,\tau)=1-e^{-\mu}$ where $e^{-\mu}$ is the probability of
364: observing zero events over the total combined observing window, and
365: $\mu=(\tau N)$ is the average number of events for the combined
366: observations. $\tau$ is defined as the average number of comets
367: observable at any time, which can be calculated by multiplying the
368: total amount of time that a single comet would spend in the line of
369: sight by the estimated number of comets per year and then dividing by
370: the the total time (one year). N is the total number of chances to see
371: a comet, defined as the total number of observing sessions. An
372: observing session is defined as the period of time needed to reach a
373: S/N necessary to detect sodium absorption. This calculation assumes
374: that the observable windows for the different comets do not overlap, a
375: reasonable assumption considering the estimated number of comets and
376: amount of time each one spends in front of the star. Since the
377: estimated number of comets per year can be considered to be constant
378: over our monitoring period, the chances of observing a comet are
379: insensitive to the elapsed time between observations, and depend only
380: on the average number of comets per year in the system and the total
381: number of observing sessions.
382:
383: The probability as a function of the number of observing sessions for
384: our long and short cometary models are plotted in Figure
385: \ref{plotseven}. The probability of detecting either type of periodic
386: comet from 6 observations (in two years, 2001-2003), ranges from 16\%
387: to 55\%, with an average probability of 20\%. According to these
388: probabilities then, we would need to make $\sim$20 observations to
389: achieve a 50\% likelihood of observing a comet. As a more conservative
390: estimate we only consider the scenario when we are looking directly
391: into the sodium tail, which would maximize the column density. This
392: then decreases D$_{sun}$ to essentially the stellar radius and gives a
393: probability of about half the values listed above and we would need to
394: make double the amount of observations to be certain of a 50\% chance.
395: The probability could also vary depending on the number of comets in
396: the system. In Figure \ref{ploteight} we show how the number of comets
397: affects the probability using the average, minimum and maximum time a
398: comet spends in front of the star and our current number of
399: observations. There would need to be $\sim$150 total comets per year
400: for us to have a 50\% chance with our current number of observations.
401:
402: This implies that it would be necessary to either continue with more
403: observations of this system or expand our target list to cover many
404: more systems in order to have a better chance of detecting a comet
405: event.
406:
407:
408: \subsection{Theoretical Column Density}
409: For comparison, we have also conducted theoretical calculations on the
410: column density we would expect from typical comet events. Using
411: measurements of comet Hale-Bopp we calculate an approximate column
412: density of sodium in the tail. Using the production rate and lifetime
413: of neutral sodium atoms listed in \citet{Cremoneseal97} of
414: Q(Na{\footnotesize I})$ \simeq 5 \times 10^{25}$ atoms s$^{-1}$ and
415: $\tau=1.69 \times 10^{5}$ s at 1 AU (which for Hale-Bopp is very close
416: to perihelion, q=0.914), we estimate the column density of the tail to
417: be 4 $\times 10^{8}$ cm$^{-2}$. Soon after the discovery by
418: \citet{Cremoneseal97} of the sodium tail, \citet{Wilson98} discovered
419: that there was a second sodium tail (termed the diffuse sodium tail)
420: superimposed on the dust tail with a production rate of sodium very
421: similar to that determined for the pure sodium tail. If we consider
422: the sodium in both sodium tails and the contribution of sodium in
423: the coma, (which was estimated to be Q(Na{\footnotesize I})$ \simeq 1
424: \times 10^{25}$ atoms s$^{-1}$ by \citet{Combi97}), the total
425: production rate increases to 1 $\times 10^{26}$ atoms s$^{-1}$ and the
426: column density to $\sim 8\times 10^{8}$ cm$^{-2}$. \citet{Combi97}
427: argue that, based on solar abundances and the production rate of
428: oxygen (10$^{30}$ cm$^{-2}$), the total production rate of sodium
429: should be a few times 10$^{27}$ cm$^{-2}$, which would lead to a
430: column density of $\sim$8$\times 10^{9}$ cm$^{-2}$. Based on our upper
431: limits we would be able to detect a feature of this column density if
432: it were offset from the rest wavelength of the star. This is likely
433: since from Figure \ref{plottwo} it is shown that the majority of the
434: orbit will produce at least a 0.4 \AA\ shift. The upper limits
435: determined in the previous section are on the order of the column
436: density of sodium from a single periodic comet, which verifies that we
437: would be able to detect the passage of a single comet. With the
438: probability, this suggests an upper limit of $\sim$300 total comets
439: per year that come within 3 AU of HD209458 (based on a 95\%
440: probability(2$\sigma$)). Considering that the $\beta$ Pictoris system
441: has $\sim$300-1000 FEB's per year \citep{Lagrange92,Ferlet93}, this
442: upper limit is consistent with a more evolved system.
443:
444: \section{Conclusion}
445: We monitored the planet-bearing solar-type star HD209458 for comet
446: activity using the HET over the course of 2 years with 6 observations
447: with S/N ratios between 70 and 258. Our observations of the sodium D1
448: line were examined for variable absorption components, as seen in
449: $\beta$ Pictoris and other stars, which are considered to be
450: signatures of this type of activity. From modelling, we determined
451: that a feature with a detectable column density of $\sim 10^{10}$
452: cm$^{-2}$ would have a FWHM of $\sim$0.05 \AA\ and a $W_{\lambda}$ of
453: $\sim$1 m\AA\ and could be shifted up to $\sim$0.5 \AA\ from the
454: photospheric line center. Simulations based on typical solar system
455: periodic comets showed that the probability of a detection from 6
456: observations was $\sim$20\% and was insensitive to the total elapsed
457: time between observations. The probabilities derived suggest that less
458: than $\sim$ 300 total comets per year come within 3 AU of HD209458 or
459: one would have likely been detected. No absorption features greater
460: than 3$\sigma$ due to cometary activity were identified. Our upper
461: limits for the sodium column density for each observation are
462: consistent with the column density expected from the sodium tail of
463: solar-type comets. Some possible reasons for the non-detection are
464: \flushleft
465: \begin{enumerate}
466: \item The small number of observations. This system is expected to have
467: comet activity on a similar scale as the solar system since it is a
468: $\sim$5 Gyr old solar-type star \citep{Cody02}. With only 6
469: observations in 2 years, our chances of witnessing a comet crossing
470: were quite small. We plan to continue to monitor this star with many
471: more observations.
472: \item A comet was passing but it was not detectable to us. The
473: inclination of the comets orbit might have been such that it was not
474: within our line of sight. Though, in the solar system, 87\% of short
475: period comets have inclinations within 30$^{\circ}$ and 81\% have
476: perihelia latitudes less than 10$^{\circ}$ from the ecliptic (all of
477: them are within 30$^{\circ}$) \citep{Biswas00}. Long period comet
478: inclinations are more scattered about the ecliptic, with only 47\%
479: having inclinations less than 90$^{\circ}$ and 63\% having perihelia
480: latitudes less than 30$^{\circ}$ \citep{Biswas00}. So, inclination may
481: factor into the probability when considering long period comets.
482: \item The sodium feature was at the rest wavelength of the star. This
483: would occur if the comet was detected at perihelion, which would be a
484: less likely circumstance for periodic comets since they spend only a
485: small fraction of their time there. If it were the case though, based
486: on estimates of the column density of the sodium tail from a single
487: comet, our S/N in the core of the D1 line might not be sufficient to
488: uncover a signal. For our future observations we plan to obtain higher
489: S/N data.
490: \end{enumerate}
491: Continued monitoring of HD209458 will occur over the next
492: year, and observations are planned for several other planet-bearing
493: stars. Through continued observations we can improve our upper limits
494: for the number of comets in the planetary system around HD209458 and
495: increase our chances of detecting a cometary event.
496:
497:
498: \acknowledgements
499:
500: We would like to thank John Debes for helping to initiate the
501: observations and for the use of his absorption profile code. We also
502: thank Robin Ciardullo for his helpful discussions concerning
503: probability. This work was supported by NASA grants NGA5-12115,
504: NAG5-11427, and NAG5-13320, NSF grants AST-0138235 and AST-0243090 and
505: the Penn State Eberly College of Science.
506:
507:
508: \begin{thebibliography}
509:
510: \bibitem[Beust {\sl et al.}(2000)]{Beust00} {Beust, H. and Morbidelli,
511: A. 2000, Icarus, 143, 170-188}
512: \bibitem[Beust {\sl et al.}(1998)]{Beust98} {Beust, H., Lagrange,
513: A.M., Crawford, I.A., Goudard, C., Spyromilio, J.,\& Vidal-Madjar,
514: A. 1998, A\&A, 338, 1015-1030}
515: \bibitem[Biswas(2000)]{Biswas00} {Biswas, S.N. 2000,
516: Bull. Astr. Soc. India, 28, 539-578}
517: \bibitem[Charbonneau {\sl et al.}(2002)]{Charbonneau02} {Charbonneau,
518: D., Brown, T.M., Noyes, R.W., \& Gilliland, R.L. 2002, ApJ, 568,
519: 377-384}
520: \bibitem[Charbonneau {\sl et al.}(2000)]{Charbonneau00} {Charbonneau,
521: D., Brown, T.M., Latham, D.W., \& Mayor, M. 2000, ApJ, 529, L45-L48}
522: \bibitem[Combi {\sl et al.}(1997)]{Combi97} {Combi, M.R., DiSanti, M.A., \&
523: Fink, U. 1997, Icarus, 130, 336-354}
524: \bibitem[Cody \& Sasselov(2002)]{Cody02} {Cody, A.M. \& Sasselov, D.D.
525: 2002, ApJ, 569, 451-458}
526: \bibitem[Cremonese(1997)]{Cremonese97} {Cremonese, G. 1997,
527: Space. Sci. Reviews, 90, 83-89}
528: \bibitem[Cremonese {\sl et al.}(1997)]{Cremoneseal97} {Cremonese, G.,
529: Boehnhardt, H., Crovisier, J., Rauer, H., Fitzsimmons, A., Fulle, M.,
530: Licandro, J., Pollacco, D., Tozzi, G.P., \& West, R.M. 1997, ApJ, 490,
531: L199-L202}
532: \bibitem[Ferlet {\sl et al.}(1993)]{Ferlet93} {Ferlet, R.,
533: Lagrange-Henri, A.M., Beust, H., Vitry, R., Zimmermann, V.P., Martin,
534: M., Char, S., Belmahdi, M., Clavier, J.P., Coupiac, P., Foing, B.H.,
535: Sevre, F., \& Vidal-Madjar, A. 1993, A\&A, 267, 137-144}
536: \bibitem[Ferlet {\sl et al.}(1987)]{Ferlet87} {Ferlet, R., Hobbs,
537: L.M., \& A. Vidal-Madjar 1987, A\&A, 185, 267-270}
538: \bibitem[Ge(1998)]{Ge98} {Ge, J. 1998, Ph.D Dissertation, The
539: University of Arizona}
540: \bibitem[Grady {\sl et al.}(1997)]{Grady97} {Grady, C.A., Sitko, M.L.,
541: Bjorkman, K.S., Perez, M.R., Lynch, D.K., Russell, R.W., \& Hanner,
542: M.S. 1997, ApJ, 483, 449}
543: \bibitem[Grinin {\sl et al.}(1996)]{Grinin96III} {Grinin, V.P., Kozlova, O.V.,
544: The, P.S., \& Rostopchina, A.N. 1996, A\&A, 309, 474-480}
545: \bibitem[Hobbs {\sl et al.}(1985)]{Hobbs85} {Hobbs, L.M., Albert, C.E., \&
546: Gry, C. 1985, ApJ, 293, L29-L33}
547: \bibitem[Ibukiyama \& Arimoto(2002)]{Ibukiyama02} {Ibukiyama, A. \&
548: Arimoto, N. 2002, A\&A, 394, 927-941}
549: \bibitem[Iseli {\sl et al.}(2002)]{Iseli02}{Iseli, M., Kuppers, M.,
550: Benz, W., \& Bochsler, P. 2002, Icarus, 155, 350-364}
551: \bibitem[Jewitt \& Matthews(1999)]{Jewitt99} {Jewitt, D. \&
552: Matthews, H. 1999, ApJ, 117, 1056-1062}
553: \bibitem[Krishna Swamy(1997)]{Swamy97} {Krishna Swamy, K.S. 1997,
554: Physics of Comets 2nd Ed.}
555: \bibitem[Lagrange {\sl et al.}(1992)]{Lagrange92} {Lagrange-Henri,
556: A.M., Gosset, E., Beust, H, Ferlet, R., \& Vidal-Madjar, A. 1992,
557: A\&A, 264, 637-653}
558: \bibitem[Lagrange {\sl et al.}(1988)]{Lagrange88} {Lagrange-Henri, A.M.,
559: Vidal-Madjar, A., \& Ferlet, R. 1988, A\&A, 190, 275-282}
560: \bibitem[Lagrange {\sl et al.}(1986)]{Lagrange86} {Lagrange, A.M.,
561: Ferlet, R., \& Vidal-Madjar, A. 1986, ESA Proc. Eight years of UV
562: Astro. with IUE, 569-571}
563: \bibitem[Morton(1991)]{Morton91} {Morton, D.C. 1991, ApJ
564: Suppl. Series, 77, 119-202}
565: \bibitem[NASA (Hale-Bopp website)]{NASAHaleBopp} {NASA Goddard Space
566: Flight Center, Hale-Bopp website,
567: http://www.gsfc.nasa.gov/ftp/pub/pao/releases/1997/97-40.htm}
568: \bibitem[Oppenheimer(1980)]{Oppenheimer80} {Oppenheimer, M. 1980, ApJ,
569: 240, 923-928}
570: \bibitem[Schempp {\sl et al.}(1989)]{Schempp89} {Schempp, W.V. \&
571: Hayden Smith, W.M. 1989, PASP, 101, 431-433}
572: \bibitem[SOHO (Sungrazer website)]{SOHOweb} {SOHO sungrazer website,
573: http://sungrazer.nascom.nasa.gov/}
574: \bibitem[Sorelli {\sl et al.}(1996)]{Sorelli96} {Sorelli, C., Grinin,
575: V.P., \& Natta, A. 1996, A\&A, 309, 155-162}
576: \bibitem[Stern(1993)]{Stern93} {Stern, Alan S. 1993, Southwest
577: Research Institution, Instrumentation and Space Res. Div., Semiannual
578: Report No. 2}
579: \bibitem[Tull(1998)]{Tull98} {Tull, R.G. 1998, SPIE, 3355, 387-398}
580: \bibitem[Vidal-Madjar {\sl et al.}(2003)]{VidalMadjar03}
581: {Vidal-Madjar, A., Lecavelier des Etangs, A., Desert, J.M., Ballester,
582: G.E., Ferlet, R., Hebrard, G., \& Mayor, M. 2003, Letters to Nature,
583: 422, 143-146}
584: \bibitem[Wanatabe {\sl et al.}(2003)]{Wanatabe03} {Watanabe, J., Kawakita, H.,
585: Furusho, R., \& Fujii, M. 2003, ApJ, 585, L159-L162}
586: \bibitem[Welsh {\sl et al.}(1998)]{Welsh98} {Welsh, B.Y., Craig, N.,
587: Crawford, I.A., \& Price, R.J. 1998, A\&A, 338, 674-682}
588: \bibitem[Wilson {\sl et al.}(1998)]{Wilson98} {Wilson, J.K., Mendillo, M., \&
589: Baumgardner, J. 1998, Geophys. Res. Letter, 25, 225}
590: \bibitem[Zeilik \& Gregory(1998)]{Zeilik98} {Zeilik, M. \&
591: Gregory, S.A. 1998, Intro. Astro. \& Astroph. 4th ed.}
592: \end{thebibliography}
593:
594:
595: \begin{deluxetable} {ccccc}
596: \tabletypesize{\scriptsize}
597: \tablewidth{0pt}
598: \tablecaption{Observations \label{observations}}
599: \tablehead{
600: \colhead{Star} &
601: \colhead{Observation Date} &
602: \colhead{Heliocentric Velocity} &
603: \colhead{Number of Exposures} &
604: \colhead{Time per Exposure} \\
605: \colhead{} &
606: \colhead{} &
607: \colhead{km/s} &
608: \colhead{} &
609: \colhead{seconds}
610: }
611: \startdata
612: HD209458 & 2001 Jul07 & 21.29 & 4 & 500 \\
613: & 2001 Jul09 & 20.78 & 2 & 900 \\
614: & 2002 Oct10 &-16.4 & 2 & 1500 \\
615: & 2002 Nov23 &-26.36 & 2 & 1500 \\
616: & 2002 Dec09 &-26.46 & 2 & 1500 \\
617: & 2003 Jul14 &19.60 & 1 & 3600 \\
618: %\hline
619: %HD22484 & 27Dec01&&1&380 \\
620: HD12235 & 28Aug02 & 23.60 & 2 & 1200 \\
621: & 31Aug02 & 22.70 & 3& 1000 \\
622: \enddata
623: \end{deluxetable}
624:
625: \begin{deluxetable} {cccccccc}
626: \tabletypesize{\scriptsize}
627: \tablewidth{0pt}
628: \tablecaption{Stellar Parameters \label{params}}
629: \tablehead{
630: \colhead{Object} &
631: \colhead{Age} &
632: \colhead{Spectral Type\tablenotemark{a}} &
633: \colhead{Radius\tablenotemark{b}} &
634: \colhead{Teff\tablenotemark{b}} &
635: \colhead{L\tablenotemark{b}} &
636: \colhead{Vmag\tablenotemark{a}} &
637: \colhead{Radial Velocity\tablenotemark{a}} \\
638: \colhead{} &
639: \colhead{Gyr} &
640: \colhead{} &
641: \colhead{R$_{\odot}$} &
642: \colhead{K} &
643: \colhead{L$_{\odot}$} &
644: \colhead{} &
645: \colhead{km/s}
646: }
647: \startdata
648: HD209458 & 5.2\tablenotemark{b}& GOV & 1.18 & 6000 & 1.61 & 7.7 & -14.8 \\
649: %HD22484 && F9IV-V &&&& 4.28 & 27.6 \\
650: %\hline
651: HD12235 & 5.39\tablenotemark{c}& G2IV &&&& 5.9 & -17.4 \\
652: \enddata
653: \tablenotetext{a}{Reference-SIMBAD data retrieval system (of the Astronomical Data center in Strasbourg, France)}
654: \tablenotetext{b}{Reference-\citet{Cody02}}
655: \tablenotetext{c}{Reference-\citet{Ibukiyama02}}
656: \end{deluxetable}
657:
658: \begin{deluxetable} {ccccc}
659: \tabletypesize{\scriptsize}
660: \tablewidth{0pt}
661: \tablecaption{Sodium Column Density Upper Limits of
662: HD209458 System.\label{upplim}}
663: \tablehead{
664: \colhead{Observation Date} &
665: \colhead{S/N in continuum} &
666: \colhead{N(NaI) Upper Limit} &
667: \colhead{S/N in line core} &
668: \colhead{N(NaI) Upper Limit} \\
669: \colhead{} &
670: \colhead{} &
671: \colhead{cm$^{-2}$} &
672: \colhead{} &
673: \colhead{cm$^{-2}$}
674: }
675: \startdata
676: 7July01 & 70 & 2.2$\times10^{10}$ & 23 & 6.5$\times10^{10}$ \\
677: 9July01 & 110 & 1.4$\times10^{10}$ & 36 & 4.3$\times10^{10}$ \\
678: 10Oct02 & 198 & 7.6$\times10^{9}$ & 67 & 2.3$\times10^{10}$ \\
679: 23Nov02 & 237 & 6.4$\times10^{9}$ & 88 & 1.7$\times10^{10}$ \\
680: 9Dec02 & 258 & 5.8$\times10^{9}$ & 88 & 1.7$\times10^{10}$ \\
681: 14July03 & 204 & 7.4$\times10^{9}$ & 67 & 2.2$\times10^{10}$ \\
682: \enddata
683: \end{deluxetable}
684:
685:
686: \begin{figure}
687: \begin{center}
688: \includegraphics[scale=0.8]{f1.ps}
689: \caption{\label{plotone} HD209458 on all 6 nights and HD12235. The fit to
690: the D1 line is overplotted with a dashed line. The spectra have been
691: shifted to the rest frame of the system.}
692: \end{center}
693: \end{figure}
694:
695: \begin{figure}
696: \includegraphics[scale=0.5]{f2.ps} \includegraphics[scale=0.5]{f3.ps}
697: \caption{\label{plottwo} Wavelength and corresponding doppler shift due
698: to the tail velocity for (Left) Short period comets and (Right) Long
699: period comets. The above plots show that the greatest wavelength shift
700: expected would be 0.53 \AA\ from the photospheric line center.}
701: \end{figure}
702:
703: \begin{figure}
704: \includegraphics[scale=0.5]{f4.ps}
705: \includegraphics[scale=0.5]{f5.ps}\\
706: \caption{\label{plotthree} Theoretical Sodium Absorption Line. (Left)
707: Column density of 5.8 $\times 10^{9}$ cm$^{-2}$ based on upper limit
708: of best S/N in the continuum ($\sim$260 for 2002 Dec 09). Dashed lines
709: represent 3$\sigma$. (Right) Column density of 1.7 $\times 10^{10}$
710: cm$^{-2}$ based on upper limit of best S/N in the core of the D1 line
711: ($\sim$90 for 2002 Dec 09). Dashed lines represent 3$\sigma$.}
712: \end{figure}
713:
714: \begin{figure}
715: \begin{center}
716: \includegraphics[scale=0.9]{f6.ps}
717: \caption{\label{plotfour} Residual from polynomial fit to the 2002
718: November 23 stellar sodium D1 profile. The absorption feature is a
719: sodium synthetic feature in the continuum region with N(NaI)=6.4
720: $\times 10^{9}$ cm$^{-2}$ (Top) in the wing with N(NaI)=9.3 $\times
721: 10^{9}$ cm$^{-2}$ (Middle) and in the core with N(NaI)=1.7 $\times
722: 10^{10}$ cm$^{-2}$ (Bottom). Column densities from 3$\sigma$ upper
723: limits listed in Table 3. Dashed lines represent 3$\sigma$.}
724: \end{center}
725: \end{figure}
726:
727: \begin{figure}
728: \includegraphics[scale=0.5]{f7.ps}
729: \includegraphics[scale=0.5]{f8.ps}
730: \includegraphics[scale=0.5]{f9.ps}
731: \includegraphics[scale=0.5]{f10.ps}
732: \includegraphics[scale=0.5]{f11.ps}
733: \includegraphics[scale=0.5]{f12.ps}
734: %\includegraphics[scale=0.5]{hd22484.7A.resid.ps}
735: \caption{\label{plotfive} Residual from polynomial fit to the
736: photospheric D1 profile. Dashed line represents 3$\sigma$. In 2002
737: October 10 the absorption feature at 5896.5 \AA\ is from a telluric
738: line that was poorly removed.}
739: \end{figure}
740:
741: \begin{figure}
742: \includegraphics[scale=0.5]{f13.ps}
743: \includegraphics[scale=0.5]{f14.ps}
744: \caption{\label{plotsix}Reference star HD12235. Residual from
745: polynomial fit to the photospheric D1 line. Dashed line represents
746: 3$\sigma$.}
747: \end{figure}
748:
749: \begin{figure}
750: \includegraphics[scale=0.5]{f15.ps}
751: \includegraphics[scale=0.5]{f16.ps}
752: \caption{\label{plotseven} Probability of detection based on number of
753: observations and total number of comets per year=10 SP and 15
754: LP. Dashed line=maximum probability, solid line=average and
755: dotted/dashed line=minimum. (Left) Short period comets (Right) Long
756: period comets.}
757: \end{figure}
758:
759: \begin{figure}
760: \includegraphics[scale=0.5]{f17.ps}
761: \includegraphics[scale=0.5]{f18.ps}
762: \caption{\label{ploteight} Probability of detection based on number of
763: comets per year and our total number of observations=6. Dashed
764: line=maximum probability, solid line=average and dotted/dashed
765: line=minimum. (Left) Short period comets. (Right) Long period comets.}
766: \end{figure}
767:
768: %\clearpage
769: %\begin{figure}
770: %\includegraphics[scale=0.5]{hd209.7July01.noall.eps}
771: %\includegraphics[scale=0.5]{hd209.9July01.noall.eps}
772: %\includegraphics[scale=0.5]{hd209.10Oct02.noall.eps}
773: %\includegraphics[scale=0.5]{hd209.23Nov02.noall.eps}
774: %\caption{\label{flatcomp} HD209458 before telluric line removal. (Top
775: %Left) solid line = 7July01 HD209458 before flatfielding and doppler
776: %correction, dashed line = calibration star used for flatfielding
777: %(HR7557). (Top Right) solid line=9July01 HD209458 before flatfielding
778: %and doppler correction, dashed line= calibration star used for
779: %flatfielding (HR7557). (Bottom Left) solid line = 10Oct02 HD209458
780: %before flatfielding and doppler correction, dashed line = calibration
781: %star used for flatfielding ($\alpha$ Aql). At 5896.5 the telluric
782: line %in HD209458 is larger than in the $\alpha$ Aql resulting in the
783: %residual feature of Figure \ref{plotfive}. (Bottom Right) solid line
784: = %23Nov02 HD209458 before flatfielding and doppler correction, dashed
785: %line = calibration star used for flatfielding (HR1544).}
786: %\end{figure}
787:
788: \end{document}
789: