1: % Paper for ApJL, Aug. 2003
2: % Julian van Eyken
3:
4: \documentclass[12pt,preprint]{aastex} %>>> use for US letter paper
5: %\documentclass{emulateapj}
6:
7: \newcommand {\kms}{{\rm km}\,{\rm s}^{-1}}
8: \newcommand {\pc}{{\rm pc}}
9: \newcommand {\Mpc}{{\rm Mpc}}
10: \newcommand {\Msun}{{\rm M_\sun}}
11: \newcommand {\Lsun}{{\rm L_{\sun}}}
12: \newcommand {\ms}{\rm ms^{-1}}
13: \newcommand {\ET}{{\it ET}}
14:
15: \shorttitle{First Planet Confirmation with \ET }
16: \shortauthors{van Eyken et al.}
17:
18: %\received{2003 September 29}
19: \begin{document}
20:
21: \title{First Planet Confirmation with a Dispersed Fixed-Delay Interferometer}
22:
23: \author{J. C. van Eyken\altaffilmark{1}, J. Ge\altaffilmark{1}, S. Mahadevan\altaffilmark{1}, C. DeWitt\altaffilmark{1}}
24: \affil{Dept. of Astronomy \& Astrophysics, Penn State University, 525
25: Davey Laboratory, University Park, PA 16802
26: %\email{vaneyken@astro.psu.edu, jian@astro.psu.edu,
27: %suvrath@astro.psu.edu, dewitt@astro.psu.edu}
28: }
29: \altaffiltext{1}{Visiting Astronomer, Kitt Peak National Observatory, National Optical Astronomy Observatory, which is operated by the Association of Universities for Research in Astronomy, Inc. (AURA) under cooperative agreement with the National Science Foundation. E-mails: vaneyken@astro.psu.edu, jian@astro.psu.edu, suvrath@astro.psu.edu, dewitt@astro.psu.edu.}
30:
31:
32: \begin{abstract}
33: The {\it Exoplanet Tracker} is a prototype of a new type of
34: fibre-fed instrument for performing high precision relative
35: Doppler measurements to detect extra-solar planets. A combination
36: of Michelson interferometer and medium resolution spectrograph,
37: this low-cost instrument facilitates radial velocity measurements
38: with high throughput over a small bandwidth ($\sim300$\AA), and
39: has the potential to be designed for multi-object operation with
40: moderate bandwidths ($\sim$ 1000\AA). We present the first planet
41: detection with this new type of instrument, a successful
42: confirmation of the well established planetary companion to 51
43: Peg, showing an rms precision of $11.5\ms$ over five days. We also
44: show comparison measurements of the radial velocity stable star,
45: $\eta$ Cas, showing an rms precision of $7.9\ms$ over seven days.
46: These new results are starting to approach the precision levels
47: obtained with traditional radial velocity techniques based on cross-dispersed echelles. We
48: anticipate that this new technique could have an important impact
49: in the search for extra-solar planets.
50: \end{abstract}
51:
52: \keywords{instrumentation: interferometers --- instrumentation:
53: spectrographs --- techniques: radial velocities --- stars: individual
54: (51 Peg, Eta Cas) --- planetary systems }
55:
56:
57: %------------------------------------------------------------------
58:
59: \section{INTRODUCTION}
60: Of the more than one hundred extra-solar planets that have been found to date,
61: the vast majority have been found using the radial velocity (RV)
62: technique (e.g. \citet{Butler1996, Baranne1996}). Current
63: approaches to making RV measurements rely on using very high
64: resolution echelle spectrographs, employing cross-correlation or fits to line profiles in
65: stellar spectra to determine Doppler shifts in the centroids of
66: the lines. While the RV technique has been the most successful
67: technique for locating extra-solar planets, traditional echelles
68: have relatively low light throughput, large instrument volume, and
69: tend to be very expensive. In addition, they cover only a single object
70: in each observation. The low light throughput limits survey
71: sensitivity to relatively bright stars and single object operation
72: leads to slow survey speeds.
73:
74: The {\it Exoplanet Tracker} (\ET) is a prototype of a new type of
75: fibre-fed RV instrument based on a dispersed fixed-delay
76: interferometer, a combination of a Michelson interferometer
77: followed by a low or medium resolution post-disperser. This
78: combination has been suggested for spectroscopic applications as
79: early as the 1890's \citep{EdserButler}, and was proposed for
80: precision Doppler planet searches by D. J. Erskine in 1997
81: \citep{ErskineGe, GeErskine}; a similar approach is discussed in \citet{Mosser2003}. The effective resolution of the
82: instrument is determined primarily by the interferometer, so the
83: post-dispersing spectrograph can be of much lower resolution than
84: in traditional techniques, and consequently can have much higher
85: throughput \citep{GeTheory, Ge2003a, Ge2003b}.
86:
87: The cost of the instrument is comparitively low, and furthermore it operates
88: in a single-order mode: a single spectrum only takes up one strip
89: along the CCD detector. Spectra from multiple stars can be lined up
90: at once on a single detector to increase survey speed \citep{GeTheory}.
91: In combination with a wide field multi-fibre telescope,
92: multi-object surveying should therefore be achievable
93: \citep{SuvrathSPIE}, with the potential to rapidly increase the number of
94: known extra-solar planets.
95:
96: %Our current understanding of extra-solar planets is limited by the
97: %small sample size available. With the ability to search faint
98: %stars with very high survey speed, it should be possible to
99: %rapidly increase the number of known extra-solar planets.
100:
101: The instrument works by producing a long-slit stellar spectrum
102: `channeled' with fringes, also known as Edser-Butler fringes
103: \citep{EdserButler, SummerSchool, GeTheory}. Sinusoidal
104: interference fringes are formed along the slit direction wherever
105: there are spectral lines. Doppler shifts in the underlying
106: spectrum result in directly proportionate phase shifts in these
107: fringes. Hence, we measure shifts of the sinusoids in the slit
108: direction, rather than shifts of the spectrum itself in the
109: dispersion direction as in traditional techniques. By fitting sine
110: functions to the CCD response along each wavelength channel on the
111: detector and combining the results from all channels, we are able
112: to accurately measure any changes in the Doppler shift of an
113: object. The concept is described in more detail in
114: \citet{GeTheory, vaneykenSPIE}.
115:
116: In this paper, we report Doppler RV curves of the known planet-bearing
117: star, 51 Peg \citep{51PegDiscovery}, and a RV stable star, $\eta$ Cas with \ET~at the KPNO
118: 2.1m telescope. This represents the first
119: planet detection using this independent new technique.
120:
121:
122: %------------------------------------------------------------------
123:
124: \section{OBSERVATIONS}
125:
126: The observations were conducted with a prototype of \ET~during an
127: engineering run at the KPNO 2.1m telescope in August 2002, built
128: largely from cheap off-the-shelf components. The observations and details of the
129: instrument setup were reported in \citet{Ge2003a}.
130:
131: The spectrograph
132: operating resolution was measured at $R\approx4540$. Using a KPNO
133: 1k$\times$3k back-illuminated CCD gave a wavelength coverage of
134: $\sim270$\AA\ centred around 5445\AA. The image was spread over
135: $\sim300$ pixels in the slit direction, giving a total of around
136: 12 periods of fringing. An iodine vapour cell was inserted into the
137: beam as a Doppler zero velocity reference, with its temperature
138: stabilised to $60\pm0.1^\circ{\rm C}.$
139:
140: During the run, we were able to obtain regular observations of a
141: number of stars, including known planet bearing stars 51 Peg,
142: $\upsilon$ And and HD209458; RV stable stars $\eta$ Cas, $\tau$
143: Ceti and and 31 Aql; and a bright star, $\alpha$ Boo, over a
144: period of about seven days \citep{Ge2003a}. %The weather was good for
145: %the most part, with typical seeing of around 1.7 arc-sec.
146: In this
147: paper, only results from 51 Peg and $\eta$ Cas are reported.
148:
149:
150: %----------------------------------------------------------------
151: \section{DATA ANALYSIS}
152:
153: Raw spectra were first trimmed and dark subtracted using standard
154: {\it IRAF} routines, with bias being subtracted along with the
155: darks in one step. Pixel-pixel flatfielding was performed using
156: non-fringing quartz-lamp continuum spectra as flatfields, where
157: the fringes were eliminated by rapidly oscillating the interferometer PZT mirror during the exposure.
158:
159: The rest of the data reduction was then performed using custom
160: software written in the {\it IDL} data analysis language, by {\it
161: Research Systems Inc.} Images were `self illumination corrected' using an algorithm to
162: extract the underlying continuum illumination function from each
163: image, which is divided out. This avoids problems with changes in
164: the illumination over time. The spectra were then corrected for
165: slant so that the slit direction was exactly aligned with the CCD
166: pixel axes. They were then low-pass Fourier filtered in order to
167: remove the interferometer comb, the series of parallel fringes
168: that would be present if pure white light were to be observed and
169: which contains no Doppler information itself.
170:
171: After these pre-processing steps, the phase and visibility were determined for each wavelength channel by fitting a sin wave to each
172: column of the CCD image, each pixel being weighted according to
173: the number of counts in the original non-flatfielded data on the
174: assumption of photon noise dominated error. Since fringe spatial
175: frequency varies only slowly as a function of wavelength, we fit a
176: smooth function to the frequencies obtained from the
177: sinusoid fits, and then performed a second pass with the frequencies
178: fixed to match this function, helping to reduce random errors.
179:
180: To a good approximation, the combined iodine/star data frames can
181: be considered a linear summation of the complex visibilities of
182: the individual iodine and stellar spectra (where complex
183: visibility is defined as ${\bf V} = Ve^{i\phi}$, with $V$ the
184: fringe visibility and $\phi$ the phase offset). Pure stellar and
185: pure iodine template spectra were taken at the beginning of each
186: observation, and these were used to mathematically extract the
187: phase shifts of the star and the iodine individually, and hence
188: calculate the intrinsic stellar velocity shift corrected for
189: instrumental drifts.
190:
191: Finally, the RV due to the motion of the Earth was subtracted to
192: leave an intrinsic stellar relative velocity curve. Currently the
193: exposure time is taken to be the centre of the exposure (although
194: this is by no means necessarily ideal).
195:
196: Error bars are based on the standard statistical curve-fitting
197: errors determined during measurement of phase and visibility. The
198: errors are translated to error bars through calculations
199: appropriate to the algorithms used to extract the final intrinsic
200: stellar RV. They are expected to give a reasonable guide to the
201: random scatter expected in the data, although they may not catch
202: all systematic errors.
203:
204: On closer inspection, the data were found to show
205: variation in the fringe phase and visibility along the
206: length of the slit. We therefore cut the spectra into three slices
207: along the dispersion direction and treated each slice separately,
208: in order to obtain sinusoidal fits less affected by this
209: systematic error. A weighted average of the three results was then obtained to give a final RV plot.
210:
211:
212: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
213:
214: \section{RESULTS}
215:
216: Part of the raw fringing spectrum for 51 Peg with the overlayed
217: iodine spectrum is shown in figure \ref{fig:rawspectra}, obtained
218: in 25 minutes at visual magnitude 5.5 with S/N per pixel in the
219: central strip of around 50. Typical exposure times for $\eta$ Cas
220: (mag 3.5) were 30 min at an S/N of 80 (including iodine cell losses).
221:
222: %\subsection{51 Peg B}
223: Figure \ref{fig:51Peg} shows the radial velocity variation
224: measured for 51 Peg after diurnal motion is subtracted.
225: The zero point is chosen arbitrarily. Over-plotted is the expected
226: curve extrapolated from the most recently determined orbital
227: parameters \citep{51peg_orbit}. The same data are listed in table
228: \ref{tab:51peg}. S/N per pixel ratios obtained were in the range
229: $\sim$40--60 for star+iodine spectra. The templates used for the
230: processing are from the night of August 16 (August 17 UT), and S/N for
231: the iodine and star templates were approximately 300 and 70
232: per pixel respectively (for the central strip).
233: Averaging over the three detector strips gives an rms deviation
234: from the predicted curve
235: of $11.5\ms$. The value of the reduced $\chi^2$ is
236: 2.70.\footnote{These results represent a substantial improvement
237: over our previously reported measurements \citep{vaneykenSPIE},
238: due in part to using all three detector strips and also to several improvements in
239: the reduction software.}
240:
241:
242:
243: %\subsection{$\eta$ Cas}
244:
245: Residuals after diurnal correction for the star $\eta$ Cas are
246: shown in figure \ref{fig:EtaCas} and table \ref{tab:EtaCas}, using
247: templates from the night of Aug 15 (Aug 16 UT). $\eta$ Cas is a
248: known RV stable star (W. D. Cochran 2002, private communication) and is
249: therefore expected to show zero shift at our current level of precision. The three image strips are
250: averaged, weighted according to flux. The rms scatter is $7.9\ms$,
251: with a reduced $\chi^2$ of 2.03. Typical S/N per pixel in the central
252: strip is around
253: 70--90 for star+iodine spectra, 270 for the iodine template, and 100
254: for the star template.
255:
256:
257: %\subsection{Throughput}
258:
259: Under 1.5 arc-sec seeing conditions, we obtained a total
260: instrument throughput of $\sim4\%$, from above the atmosphere to
261: the detector, including sky, telescope transmission, fibre loss,
262: instrument and iodine cell transmission, detector quantum
263: efficiency, and using only one interferometer output. Excluding
264: slit loss, the transmission of the instrument itself from fibre to
265: detector was 19\%.
266:
267: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
268:
269: \section{DISCUSSION}
270: \label{sec:discussion} It is possible to make an estimate of
271: the photon limited error for this instrument using the following
272: analysis. Following \citet{GeTheory}, the error $\sigma_i$ in
273: velocity from a single wavelength channel $i$ due to photon noise
274: alone can be calculated as a function of fringe visibility,
275: $\gamma_i$, and total photon flux in the channel, $F_i$. Using a slightly
276: more accurate derivation than given in \citet{GeTheory} gives the
277: relation:
278:
279: \begin{equation}\label{eqn:jian} \sigma_i = \frac{c\lambda}{\pi d\gamma_i\sqrt{2F_i}},
280: \end{equation}
281:
282: where $c$ is the speed of light, $\lambda$ is the operating
283: wavelength, and $d$ is the path difference between the
284: interferometer arms. The combined error over all
285: channels is then given by:
286:
287: \begin{equation}\label{eqn:individual} \sigma = [\sum_{i} 1/\sigma_i^2]^{-\onehalf}.
288: \end{equation}
289:
290: This gives us an estimate of the error due to the photon noise in
291: one complete fringing spectrum. In order to estimate the photon
292: error for the final iodine reference corrected RV measurement, we
293: must combine the errors from the templates, $\sigma_{\rm I_2,
294: template}$ and $\sigma_{\rm star, template}$, and from the
295: combined star+iodine data. We treat the combined data as
296: consisting of two separate components, with errors $\sigma_{\rm
297: I_2, data}$ and $\sigma_{\rm star, data}$. The final relative
298: velocity measured, $V$, is given by $V = (V_{\rm star,
299: data}-V_{\rm star, template}) - (V_{\rm I_2, data} - V_{\rm I_2,
300: template}) $, and so the final error is obtained by quadrature
301: addition:
302:
303: \begin{equation}\label{eqn:final} \sigma_V^2 = \sigma_{\rm star, data}^2 + \sigma_{\rm
304: star, template}^2 + \sigma_{\rm I_2, data}^2 + \sigma_{\rm I_2,
305: template}^2. \end{equation}
306:
307: We find the errors due to the templates using equations \ref{eqn:jian} and \ref{eqn:individual}. To find the errors $\sigma_{\rm I_2, data}$
308: and $\sigma_{\rm star, data}$, we take the errors calculated for
309: the templates, and scale these to find the values that they should
310: have at the S/N of the data, noting from equation \ref{eqn:jian}
311: that the errors scale as the reciprocal of S/N (where
312: S/N$=\sqrt{F}$).
313:
314: The results of these calculations are shown in table
315: \ref{tbl:errors}. We find final photon limiting precisions
316: (averaged over all data points) of $11.0\ms$ for 51 Peg and
317: $8.1\ms$ for $\eta$ Cas. Within the uncertainty in the rms
318: residual values obtained for the data due to the small number of
319: data points ($11.5\pm2.1\ms$ for 51Peg and $7.9\pm1.2\ms$ for
320: $\eta$ Cas), we find a good match with the data and conclude that
321: we have reached the photon limit: the reduction software has
322: successfully extracted the maximum possible information from the data. It is
323: important to note, however, that these values for the photon limit
324: are those expected {\em given} the fringe visibility that was
325: obtained. Various instrument effects (for example defocus) can
326: reduce the visibility from its optimum and hence reduce the
327: precision. It is therefore possible that the intrinsic limit is
328: somewhat lower for an ideally optimised instrument.
329:
330:
331:
332: We note the large contribution to the errors due to the iodine
333: reference. Though the iodine can be measured to very high accuracy
334: for the template since a quartz lamp is used for illumination, the
335: iodine in the combined star/iodine images has much lower S/N, and
336: this becomes an important source of error. In both the 51 Peg and
337: the $\eta$ Cas cases, the error due to the iodine is comparable to
338: that of the star itself.
339:
340: Given this photon limit, the error bars in the data appear to be
341: underestimated (leading to the large values for the reduced $\chi^2$).
342: A possible cause of this is the low pass Fourier filtering
343: that is done to remove the interferometer comb. In addition to
344: removing the comb, filtering has the effect of smoothing the
345: photon noise in the data, reducing the residuals in the sinusoid
346: fits to the fringes and thereby reducing the resulting error
347: estimates for each fringe. This may be an artificial
348: effect, however, which in fact does not improve the precision of
349: the fits. The
350: extent to which this effect occurs is under investigation.
351:
352:
353:
354: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
355:
356: \section{CONCLUSION}
357:
358: We have achieved $11.5\ms$ RV precision over five days of
359: observations of 51 Peg, obtaining results in excellent agreement
360: with previously measured orbital parameters due to its planetary
361: companion. We have also obtained measurements of a RV stable star,
362: $\eta$ Cas, showing that we can reach a precision of $7.9\ms$ over
363: seven days, using a simple and inexpensive prototype. The rms residuals
364: match the expected photon limited errors for the instrument, given
365: the fringe visibilities obtained, and show that the precision we
366: are able to obtain with \ET~is becoming comparable with current
367: traditional echelle techniques. For comparison, rms scatters
368: obtained previously for 51 Peg have been
369: $13\ms$\citep{51PegDiscovery}, $5.2\ms$ \citep{51PegConfirmation},
370: and $11.8\ms$ \citep{51peg_orbit}.
371:
372:
373: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
374:
375: \acknowledgments The authors are grateful to Richard Green, Skip
376: Andree, Daryl Wilmarth and the KPNO staff for their generous
377: support and advice, and to Dominique Naef for very helpful input. The authors are also grateful to Stuart Shaklan, Michael Shao and Chas Beichman
378: for their encouragement and support, and Bill Cochran, Larry Ramsey and Eric Feigelson for many useful discussions. This work is supported
379: by the National Science Foundation with grant AST-0243090, the
380: Penn State Eberly College of Science and JPL.
381: J. V. E. and S. M. acknowledge travel support from KPNO; S. M. acknowledges the JPL Michelson Fellowship funded by NASA.
382:
383: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
384:
385: \begin{thebibliography}{}
386:
387: \bibitem[Baranne et al.(1996)]{Baranne1996} Baranne, A.~et al.\
388: 1996, \aaps, 119, 373
389: \bibitem[Butler et al.(1996)]{Butler1996} Butler, R.~P., Marcy,
390: G.~W., Williams, E., McCarthy, C., Dosanjh, P., \& Vogt, S.~S.\ 1996,
391: \pasp, 108, 500
392: \bibitem[Edser \& Butler(1898)]{EdserButler}Edser, E. \& Butler, C. P. 1898, Phil. Mag., 46, 207
393: \bibitem[Erskine \& Ge(2000)]{ErskineGe} Erskine, D.~J.~\& Ge,
394: J.\ 2000, ASP Conf.~Ser.~195: Imaging the Universe in Three Dimensions, 501
395: \bibitem[Ge(2002)]{GeTheory} Ge, J.\ 2002, \apjl, 571, L165
396: \bibitem[Ge, Erskine, \& Rushford(2002)]{GeErskine} Ge, J.,
397: Erskine, D.~J., \& Rushford, M.\ 2002, \pasp, 114, 1016
398: \bibitem[Ge et al.(2003a)]{Ge2003a} Ge, J., van Eyken, J.~C.,
399: Mahadevan, S., DeWitt, C., Ramsey, L.~W., Shaklan, S.~B., \& Pan, X.\ 2003a,
400: \procspie, 4838, 503
401: \bibitem[Ge et al.(2003b)]{Ge2003b} Ge, J., Mahadevan, S., van Eyken, J.,
402: DeWitt, C., \& Shaklan, S. 2003b, in ASP Conf. Ser. 294, Scientific Frontier in Research in Extrasolar
403: Planets, ed. Deming, D., \& Seager, S. (San Fransico: ASP), 573
404: \bibitem[Lawson(2000)]{SummerSchool}Lawson, P. R. 2000, in Principles of Long Baseline Stellar Interferometry, ed. P. R. Lawson (Pasadena: JPL Publications), 113
405: \bibitem[Mahadevan et al.(2003)]{SuvrathSPIE} Mahadevan, S., Ge, J.,
406: van Eyken, J. C., DeWitt, \& Shacklan, S. 2003, \procspie, in press
407: \bibitem[Marcy et al.(1997)]{51PegConfirmation} Marcy, G.~W., Butler,
408: R.~P., Williams, E., Bildsten, L., Graham, J.~R., Ghez, A.~M., \& Jernigan,
409: J.~G.\ 1997, \apj, 481, 926
410: \bibitem[Mayor \& Queloz(1995)]{51PegDiscovery} Mayor, M.~\& Queloz,
411: D.\ 1995, \nat, 378, 355
412: \bibitem[Mosser, Maillard, \& Bouchy(2003)]{Mosser2003} Mosser,
413: B., Maillard, J., \& Bouchy, F.\ 2003, \pasp, 115, 990
414: \bibitem[Naef et al.(2003)]{51peg_orbit}Naef, D., Mayor, M., Beuzit,
415: J. L., Perrier, C., Queloz, D., Sivan, J. P., Udry, S. 2003, \aap, in press
416: \bibitem[van Eyken et al.(2003)]{vaneykenSPIE} van Eyken, J. C., Ge,
417: J., Mahadevan., S., DeWitt, C., \& Ren, D. 2003, \procspie, in press
418:
419: \end{thebibliography}
420:
421: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
422:
423: %%%%%%%
424: \clearpage
425: \begin{deluxetable}{rcc}
426: \tablecaption{RV measurements for 51 Peg \label{tab:51peg}}
427: \tablehead{\colhead{JD} & \colhead{Velocity} & \colhead{Error} \\
428: \colhead{$-2450000$} & \colhead{$(\ms)$} & \colhead{$(\ms)$} }
429:
430: \startdata
431:
432: $ 2500.8665$ & $ 35.4$ & $ 7.1$ \\
433: $ 2500.8879$ & $ 46.8$ & $ 6.8$ \\
434: $ 2500.9094$ & $ 29.4$ & $ 6.8$ \\
435:
436: $ 2501.9077$ & $ 18.1$ & $ 6.8$ \\
437: $ 2501.9263$ & $ 30.8$ & $ 7.0$ \\
438: $ 2501.9441$ & $ 29.9$ & $ 7.0$ \\
439:
440: $ 2502.9392$ & $ -78.1$ & $ 8.4$ \\
441: $ 2502.9573$ & $ -61.6$ & $ 8.3$ \\
442: $ 2502.9751$ & $ -57.9$ & $ 7.8$ \\
443:
444: $ 2503.9250$ & $ -50.9$ & $ 6.7$ \\
445: $ 2503.9480$ & $ -57.9$ & $ 6.8$ \\
446: $ 2503.9692$ & $ -27.6$ & $ 6.6$ \\
447:
448: $ 2504.8960$ & $ 44.2$ & $ 6.5$ \\
449: $ 2504.9177$ & $ 28.1$ & $ 6.6$ \\
450: $ 2504.9392$ & $ 15.4$ & $ 6.8$
451:
452: \enddata
453: \end{deluxetable}
454: %%%%%%%%
455:
456: %%%%%%%%%%%%%%
457: \clearpage
458: \begin{deluxetable}{rcc}
459: \tablecaption{RV measurements for $\eta$ Cas \label{tab:EtaCas}}
460: \tablehead{\colhead{JD} & \colhead{Velocity} & \colhead{Error} \\
461: \colhead{$-2450000$} & \colhead{$(\ms)$} & \colhead{$(\ms)$} }
462: \startdata
463: $ 2498.8318$ & $ 6.4$ & $ 5.7$ \\
464: $ 2498.8545$ & $ 8.3$ & $ 5.1$ \\
465: $ 2498.8823$ & $ -11.2$ & $ 5.4$ \\
466: $ 2498.8936$ & $ 10.9$ & $ 5.3$ \\
467:
468: $ 2499.7839$ & $ 6.8$ & $ 6.3$ \\
469: $ 2499.7952$ & $ 2.9$ & $ 6.0$ \\
470: $ 2499.8064$ & $ -3.6$ & $ 5.8$ \\
471: $ 2499.8174$ & $ 1.7$ & $ 5.8$ \\
472: $ 2499.8284$ & $ 1.4$ & $ 5.9$ \\
473:
474: $ 2501.8701$ & $ -5.9$ & $ 5.3$ \\
475: $ 2501.8777$ & $ -9.6$ & $ 5.4$ \\
476: $ 2501.8850$ & $ -15.3$ & $ 5.5$ \\
477: $ 2501.8926$ & $ 7.7$ & $ 5.4$ \\
478:
479: $ 2502.8528$ & $ -0.0$ & $ 4.9$ \\
480: $ 2502.8643$ & $ -1.7$ & $ 4.9$ \\
481: $ 2502.8752$ & $ 5.4$ & $ 5.1$ \\
482: $ 2502.8860$ & $ -10.8$ & $ 5.4$ \\
483:
484: $ 2504.8093$ & $ -1.4$ & $ 5.7$ \\
485: $ 2504.8201$ & $ -10.1$ & $ 6.4$ \\
486: $ 2504.8311$ & $ 2.5$ & $ 5.4$ \\
487: $ 2504.8401$ & $ 11.7$ & $ 5.9$
488:
489: \enddata
490: \end{deluxetable}
491:
492: %%%%%%%
493: \clearpage
494: \begin{deluxetable}{lcc}
495: \tablecaption{\label{tbl:errors}Mean photon limited error
496: estimation}
497: \tablehead{\colhead{Component} & \colhead{Star} & \colhead{Iodine} \\
498: \colhead{ } & \colhead{($\ms$)} & \colhead{($\ms$)} }
499: \tablecolumns{3}
500:
501: \startdata \sidehead{51 Peg}
502: Templates & 4.7 & 1.2 \\
503: Data & 6.7 & 7.2 \\
504: Combined & \multicolumn{2}{c}{11.0} \\
505:
506: \sidehead{$\eta$ Cas}
507: Templates & 4.2 & 1.2 \\
508: Data & 5.3 & 4.3 \\
509: Combined & \multicolumn{2}{c}{8.1}
510:
511: \enddata
512: \end{deluxetable}
513: %%%%%%%%
514:
515: \clearpage
516:
517: %%%%%%%
518: \begin{figure}
519: \plotone{figure1.eps} %, raw_51peg.eps, raw_hd209458.eps}
520: \caption{\label{fig:rawspectra} Raw fringing spectrum of 51 Peg
521: with iodine, obtained at KPNO on the night of 2002 Aug 14. (mag
522: 5.5, S/N$\sim50$ per pixel)}
523: \end{figure}
524: %%%%%%%%
525:
526: %%%%%%%
527: \begin{figure}
528: \plotone{figure2.eps}
529: \caption{\label{fig:51Peg} Radial velocity measurements for 51
530: Peg, with the predicted curve over-plotted. RMS residuals
531: are $11.5\ms$.}
532: \end{figure}
533: %%%%%%%%
534:
535:
536: %%%%%%%
537: \begin{figure}
538: %%\epsscale{1.1}
539: \plotone{figure3.eps}
540: \caption{\label{fig:EtaCas} RV measurements for $\eta$ Cas, an RV
541: stable star, expected to show zero shift. RMS residuals are
542: $7.9\ms$.}
543: \end{figure}
544: %%%%%%%%
545:
546:
547: \end{document}
548: