1: %\documentstyle[12pt,apjpt4]{article}
2: \documentclass[12pt,preprint]{aastex}
3: %\documentstyle[12pt,aaspp4]{article}
4: \def\Rsol{R$_{\odot}$}
5: \def\Msol{M$_{\odot}$}
6: \def\etal{{\it et al. }}
7: \def\kms{km s$^{-1}$}
8: \def\degsq{deg$^2$}
9: \def\perdegsq{deg$^{-2}$}
10: \def\gtsim{\ {\raise-0.5ex\hbox{$\buildrel>\over\sim$}}\ }
11: \def\ltsim{\ {\raise-0.5ex\hbox{$\buildrel<\over\sim$}}\ }
12:
13: \begin{document}
14: %\noindent {\large CONFIDENTIAL FINAL DRAFT}\hfill Version: \today
15: %\noindent {To be submitted to The Astrophysical Journal}\hfill Version: \today
16:
17: \title{
18: Millimagnitude Photometry for Transiting Extrasolar Planetary Candidates IV:
19: The Puzzle of the Extremely Red OGLE-TR-82 Primary Solved\altaffilmark{11}}
20:
21: \author{Sergio Hoyer\altaffilmark{2,1},
22: Sebasti\'an Ram\'{\i}rez Alegr\'{\i}a\altaffilmark{1},
23: Valentin D. Ivanov\altaffilmark{3},
24: Dante Minniti \altaffilmark{1},
25: Grzegorz Pietrzynski\altaffilmark{4,5},
26: Mar\'ia Teresa Ru\'iz\altaffilmark{2},
27: Wolfgang Gieren\altaffilmark{4},
28: Andrzej Udalski\altaffilmark{5},
29: Manuela Zoccali\altaffilmark{1},
30: Rodrigo Carrasco\altaffilmark{6},
31: Rodrigo F. D\'{\i}az\altaffilmark{7},
32: Jos\'e Miguel Fern\'andez\altaffilmark{1,8},
33: Jos\'e Gallardo\altaffilmark{9},
34: Marina Rejkuba\altaffilmark{10},
35: Felipe P\'erez\altaffilmark{1}
36: }
37:
38: \altaffiltext{1}{Department of Astronomy, Pontificia Universidad Cat\'olica,
39: Casilla 306, Santiago 22, Chile; sramirez@astro.puc.cl, dante@astro.puc.cl, mzoccali@astro.puc.cl, fperez@astro.puc.cl}
40: \altaffiltext{2}{Department of Astronomy, Universidad de Chile, Casilla 36-D, Santiago, Chile; shoyer@das.uchile.cl, mtruiz@das.uchile.cl}
41: \altaffiltext{3}{European Southern Observatory, , Alonso de Cordova 3107, Vitacura,
42: Casilla 19001, Santiago 19, Chile; vivanov@eso.org}
43: \altaffiltext{4}{Department of Physics, Universidad de Concepci\'on, Casilla 160-C, Concepci\'on, Chile; pietrzyn@hubble.cfm.udec.cl, wgieren@astro-udec.cl}
44: \altaffiltext{5}{Warsaw University Observatory, Al. Ujazdowskie 4, 00-478 Waszawa, Poland; udalski@astrouw.edu.pl}
45: \altaffiltext{6}{Gemini Observatory, Southern Operations Center, AURA, Casilla 603, La Serena, Chile, rcarrasco@gemini.edu}
46: \altaffiltext{7}{Instituto de Astronom\'{\i}a y F\'{\i}sica del Espacio, CONICET- UBA, CC 67 - Suc 28, Buenos Aires C1428ZAA, Argentina; rodrigo@iafe.uba.ar}
47: \altaffiltext{8}{Harvard-Smithsonian Center for Astrophysics, 60 Garden St. Cambridge, MA 02138, USA; jfernand@gmail.com}
48: \altaffiltext{9}{Centre de Recherche Astronomique de Lyon, Ecole normale
49: superieure de Lyon, 46 allee de Italie, 69364 Lyon, France; jose.gallardo@ens-lyon.fr}
50: \altaffiltext{10}{European Southern Observatory, Karl-Schwarzschild-Str.2, 85748 Garching bei Muenchen, Germany; vivanov@eso.org, mrejkuba@eso.org}
51: \altaffiltext{11}{Based on observations collected with the
52: GEMINI-S Telescope (queue observing, program GS-2005B-Q-9), with the
53: Very Large Telescope at Paranal Observatory
54: (JMF and DM visiting observers), with the
55: ESO New Technology Telescope at La Silla Observatory (SR, FP, and DM visiting
56: observers) for the ESO Programmes 075.C-0427, 075.B-0414, and 076.C-0122.}
57:
58:
59: \begin{abstract}
60:
61: We present precise new $V$, $I$, and $K$-band photometry for the
62: planetary transit candidate star OGLE-TR-82. Good seeing $V$-band
63: images acquired with VIMOS instrument at ESO Very Large Telescope (VLT) allowed us to measure $V=20.6\pm 0.03$~mag star in spite of the presence of a brighter neighbour about 1" away. This faint
64: magnitude answers the question why it has not been possible to
65: measure radial velocities for this object.
66:
67: One transit of this star has been observed with GMOS-S instrument of
68: GEMINI-South telescope in $i$ and $g$-bands. The measurement of the
69: transit allows us to verify that this is not a false positive, to
70: confirm the transit amplitude measured by OGLE, and to improve the
71: ephemeris. The transit is well defined in $i$-band light curve, with a
72: depth of $A_i=0.034$~mag. It is however, less well defined, but deeper
73: ($A_g = 0.1$~mag) in the $g$-band, in which the star is significantly
74: fainter.
75:
76: The near-infrared photometry obtained with SofI array at the ESO New Technology Telescope (NTT)
77: yields $K=12.2\pm 0.1$, and $V-K=8.4\pm 0.1$, so red that it is unlike any other transit candidate
78: studied before. Due to the extreme nature of this object, we have not yet been able to measure velocities for this star, but based on
79: the new data we consider two different possible configurations:
80: (1) a nearby M7V star, or (2) a blend with a very reddened distant red giant.
81: The nearby M7V dwarf hypothesis would give a radius for the companion of $R_p=0.3\pm 0.1 ~R_J$, i.e. the size of Neptune.
82:
83: Quantitative analysis of near-IR spectroscopy finally shows that OGLE-TR-82 is a distant, reddened metal poor early K giant. This result is confirmed by direct comparison with stellar templates
84: that gives the best match for a K3III star.
85:
86:
87: Therefore, we discard the planetary nature of the companion. Based on all the new data, we conclude that this system is a main-sequence binary blended with a background
88: red giant. As a case study, this kind of system that can mimic a planetary transit is a lesson to learn for future transit surveys.
89:
90: \end{abstract}
91:
92: \keywords{Stars: individual (OGLE-TR-82) -- Extrasolar planets: formation}
93:
94: \section{Introduction}
95:
96: Because low mass stars are the most numerous stars in our Galaxy, in
97: spite of their dimness they
98: provide an exciting possibility to detect transits of planets much smaller
99: than Jupiter. This has been pointed out by many authors
100: (see for example Pepper \& Gaudi 2005).
101: The known planets in the Solar neighborhood show a very steep function of mass:
102: low mass planets are much more common around Solar type stars
103: than high mass planets (Butler \etal 2006).
104: Even though low mass planets are more difficult to detect than Jupiters,
105: there have been a few recent discoveries of ``Neptunes" or ``Super Earths" around
106: low mass stars (McArthur \etal 2004, Butler \etal 2004,
107: Santos \etal 2004, Bonfils \etal 2006, Lovis \etal 2006, Beaulieu \etal 2006, Gould \etal 2006).
108:
109: %[GOULD--> Microlens OGLE-2005-BLG-169 Implies That Cool Neptune-like Planets Are Common]
110:
111: More recently,
112: Sahu \etal (2006) found a number of planetary candidates around low-mass
113: stars in the bulge fields. The search for planets is being extended into
114: a new realm: M-dwarfs are the most numerous stars in the Galaxy,
115: and the transiting planets around such stars open a wealth
116: of new possibilities for ground and space-based planet searches.
117: The transit candidate OGLE-TR-82 (Udalski \etal 2002)
118: appears to be an extreme case, that is important to study in the context
119: of the future transit discoveries made by space missions like KEPLER and COROT.
120: We note also that there has been progress on the models triggered by these observational
121: discoveries as well. For example, the formation of Neptunes and super-Earths
122: has been recently studied by % cambio
123: Baraffe \etal (2005, 2006), Kennedy \etal (2006) and Alibert \etal (2006).
124:
125: The OGLE search has provided the largest list of transiting candidates
126: (Udalski \etal 2002a, 2002b, 2002c, 2003). In particular,
127: Udalski \etal (2002c) discovered very low amplitude transits in the
128: $I=16.30$ magnitude star OGLE-TR-82, located in the Carina region
129: of the Milky Way disk, at $RA(2000)=10:58:03.07$, $DEC(2000)=-61:34:25.8$.
130: They monitored 22 individual transits on this star, measuring an amplitude
131: $A_I=0.034$ mag, and a period $P=0.76416$ days.
132: They listed $R_p = 1.09~R_J$, making this a prime Hot Jupiter candidate.
133: Recently, Silve \& Cruz (2006) obtained $M_s=0.65 ~M_{\odot}$, and
134: $R_p = 1.10~R_J$, also singling OGLE-TR-82-b as a possible planetary companion.
135: We followed-up OGLE candidates with infrared and optical photometry
136: with the aim of identifying and characterizing planetary companions
137: to low-mass stars (Gallardo \etal 2005, Fern\'andez \etal 2006, D\'{\i}az \etal 2006).
138: M-dwarfs with hot Jupiters should stand out in a period $vs$ color diagram
139: as objects with short periods and very red colors.
140: Early in the program it was recognized that OGLE-TR-82 was a very interesting
141: candidate. For example, Figure \ref{uno} shows the OGLE periods $vs$ the $I-K$
142: color for several candidates observed in the Carina fields.
143: OGLE-TR-82 is clearly unique in this sample: it is a very red and faint object,
144: both according to 2MASS and to our own photometry.
145: The colors were consistent with a very late M-dwarf,
146: for which the low amplitude transits would imply a very small companion.
147: We then intensively followed up this candidate.
148: However, acquiring the complementary data to characterize OGLE-TR-82
149: and its companion has proven difficult due to the extreme nature of this object.
150:
151: Pont \etal (2004) observed this candidate with UVES/FLAMES at the VLT,
152: arguing that this was possibly a K7-M0V star.
153: They obtained precise velocity measurements for several OGLE targets,
154: discovering three planets in their sample,
155: but they could find no CCF signal for OGLE-TR-82, concluding that this
156: was an unsolved case.
157: Why is there no signal for this object? In this paper we answer the question
158: using optical and infrared photometric data acquired at GEMINI, VLT and NTT.
159:
160: The future space missions like COROT and KEPLER would discover many faint transit
161: candidates, which would need to be followed up and confirmed. The most fundamental
162: follow-up observation is echelle spectroscopy in order to measure radial velocities
163: and derive the companion masses. However, in some special cases the optical echelle
164: spectroscopy would not work. OGLE-TR-82 is one example, and becomes an interesting
165: case study.
166:
167: After a thorough up study, in this paper we discard the planetary
168: nature of OGLE-TR-82, concluding that it is a blend with a reddened distant giant.
169: Section 2 describes the observations and reductions for all the data:
170: optical and infrared photometry and infrared spectroscopy.
171: Section 3 presents the results and discussion. The conclusions are listed in
172: Section 4.
173:
174: \section{The Data}
175:
176: This paper uses data obtained at various facilities, as listed in Table 1, which we
177: discuss in turn.
178:
179: \subsection{Infrared Observations: Photometry and Spectroscopy}
180:
181:
182: OGLE-TR-82 was observed during the nights of May 4 and 5, 2006, using
183: SofI (Son of ISAAC; Moorwood \etal 1998) infrared camera and spectrograph
184: at the ESO NTT. SofI is equipped with a Hawaii HgCdTe detector
185: of 1024$\times$1024 pixels, characterized by a $5.4$ $e/ADU$ gain, a readout noise
186: of 2.1 ADU and a dark current of less than $0.1$ $e/sec$. We used it
187: in the large field camera mode, obtaining $4.9\times 4.9$ $arcmin^2$ field.
188: All measurements were made through the $K_s$-band filter (center = 2.162 $\mu$m
189: and width = 0.275 $\mu$m).
190:
191: The reductions were made using IRAF tasks\footnote{IRAF is distributed by the National Optical Astronomy Observatories, operated by Universities for Research in Astronomy, Inc.}.
192: First of all, the crosstalk
193: correction was applied, taking into account the detector sensitivity difference
194: between the upper and lower half. Then the sky subtraction was applied by subtracting an
195: appropriately scaled mean sky, which was obtained by combining
196: all the offset images (two per target using ``dither-5'' mode) but
197: rejecting outlying pixels.
198: %Then the sky subtraction was applied.
199: %The whole dataset was acquired using ``dither-5" around two offset
200: %positions that included the target. These contiguous sky images were used to
201: %generate local skies close in time for each of the offset images. The
202: %images were averaged combined using $imcombine$
203: %but rejecting outlying pixels. Then, the appropriately scaled sky was subtracted from the images.
204: Finally, we applied flat-field corrections to all images and aligned them. For the flat-fields, we
205: used the correction images provided by the NTT SciOps team\footnote{www.ls.eso.org/lasilla/sciops/ntt/sofi/reduction/flat\_fielding.html},
206: and the alignment was done with $lintran$ and $imshift$.
207:
208: OGLE-TR-82 is listed as 2MASS source 10580297-6134263, with
209: $J=13.479\pm 0.059$,
210: $H=12.077\pm 0.035$, and
211: $Ks=11.538\pm 0.030$.
212: The 2MASS yields a very red color for this star: $J-K=1.94$.
213:
214: OGLE-TR-82 is located in Carina, in the Galactic plane at $l, ~b= (289.8638^o, ~-1.6131^o)$.
215: This region is relatively crowded, hence we find that for some OGLE candidates
216: the 2MASS photometry differs from ours.
217: The 2MASS photometry may be contaminated by the nearby neighbours
218: due to the pixel scale of 2MASS. We obtained the $K$-band magnitude by
219: using deeper and better sampled SofI images.
220: The calibration of our infrared photometry was made using 2MASS.
221: The zero point of the $K$-band photometry should be good to 0.1 mag, which
222: is accurate enough for our purposes.
223:
224: Using aperture photometry with an aperture small enough to
225: exclude nearby neighbours, we find $K=12.20\pm 0.10$ for this star.
226: Even though this is more than half a magnitude fainter than the 2MASS value,
227: in what follows we will adopt this as the final $K$ magnitude for OGLE-TR-82.
228: This allows us to plot OGLE-TR-82 in Figure \ref{uno}, which shows the
229: periods $vs$ the $I-K$ color for several OGLE transits
230: candidates observed in the Carina fields. OGLE-TR-82 was clearly unique in
231: this sample, as a very red object with a short period companion, that warranted
232: further study.
233:
234: Therefore, we also acquired near-IR spectrum with SofI, covering the
235: region from 1 to 2.5 microns. The low-resolution blue and red grisms
236: delivered spectral resolution R$\sim$700, with slit 1 arcsec. This
237: is adequate for spectral classification. The total integration was
238: 480 sec in the blue and 720 sec in the red, split into individual
239: integrations of 240 sec and 180 sec, respectively. The usual
240: observing technique of nodding along two slit positions was used.
241: The data reduction included the following steps: cross talk removal,
242: flat fielding, subtraction of the sky emission, extraction of
243: 1-dimensional spectrum from each individual image, wavelength
244: calibration and combination into a final spectrum for each mode
245: separately. Finally, we removed the telluric absorption observing a
246: solar analog Hip 54715 (see for details Maiolino \etal 1996).
247: The two SofI spectra were joined together by scaling of the overlapping
248: regions from 1.51 to 1.645 microns and the spectra were not flux calibrated.
249:
250:
251:
252:
253:
254: \subsection{VLT Optical Observations and Photometry}
255:
256: The $V$-band VLT observations and photometry are described by Fern\'andez et
257: al. (2006). Briefly, the photometric observations were taken with
258: VIMOS at the Unit Telescope 4 (UT4) of the European Southern
259: Observatory Very Large Telescope (ESO VLT) at Paranal Observatory
260: during the nights of April 9 to 12, 2005. The VIMOS field of view
261: consists of four CCDs, each covering 7$\times$8 arcmin, with a
262: separation gap of 2 arcmin, and a pixel scale of $0.205 ~arcsec/pixel$.
263: We used the Bessell $V$ filter of VIMOS, with $\lambda_0=5460{\rm \AA}$, $FWHM=890{\rm \AA}$.
264:
265: A number of OGLE transit candidates were monitored simultaneously.
266: OGLE-TR-82 was located in one of the four monitored fields, and it
267: was scheduled to have a transit during the first night of our
268: run. However, a later revised ephemerides provided by OGLE
269: (A. Udalski private communication) showed that the transit ocurred
270: after the night was over. Nonetheless, some of the VIMOS images
271: had exquisite seeing, and we can use them for precise astrometry
272: and to measure the magnitude
273: of OGLE-TR-82 in the presence of a brighter optical neighbour located $1$ arcsec away.
274: It was necessary to do PSF photometry using DAOPHOT in IRAF in
275: order to account for this brighter nearby neighbour.
276:
277: The best seeing images ($FWHM=0.5~ \mathrm{arcsec}$) taken near the zenith were
278: selected, and a master optical image was made.
279: The images of OGLE-TR-82 analyzed here are 400$\times$400 pix,
280: or 80 arcsec on a side. This small image contains about 500
281: stars with $15<V<24$.
282: Figure \ref{4juntas}-(a) shows a $20\times 20$ arcsec$^2$
283: portion of this $V$-band master image, showing OGLE-TR-82 with a
284: larger circle, and its neighbour located 1" North with a smaller circle. We measure
285: $V=20.61\pm 0.03$ for OGLE-TR-82, and $V=17.388\pm 0.01$ for its neighbour.
286: The OGLE $I$-band image is shown in Figure \ref{4juntas}-(b) for comparison.
287: both stars have similar magnitudes in this image. Taking $I=16.30\pm 0.10$
288: from Udalski \etal (2002), we obtain a very red color $V-I=4.1\pm 0.1$
289: for OGLE-TR-82. This color indicates that the target is a very late
290: M-dwarf or an early brown dwarf.
291: The $Ks$-band image of Figure \ref{K} shows that the relative brightness of both stars
292: is reversed: OGLE-TR-82 is much brighter than its neighbour in the near-IR.
293:
294: \subsection{GEMINI-South Optical Observations and Photometry}
295:
296: The object is so faint in the optical passbands, that it takes an 8-m telescope
297: to measure the transit accurately. Therefore,
298: a full transit of this target was observed with the GMOS at the GEMINI-S
299: telescope. The observations were acquired in queue mode on 27 January 2006.
300: The GMOS field of view is $5.5'\times 5.5'$, with a scale of $0.0727$ arcsec/pix.
301: This scale is finer than that of VIMOS, but the seeing during the observations
302: was worse than the VIMOS run, which complicates the transit
303: photometry in the presence of the brighter neighbour $1$ arcsec away.
304: The beginning of the transit was noisier because of weather problems,
305: but the night stabilized after the ingress, resulting in a relatively good light curve.
306:
307: We used the $g$ and $i$-band filters (GEMINI filters g-G0325 and i-G0327,
308: respectively), alternating 3 consecutive exposures of each filter in turn.
309: This observation sequence was repeated without interruptions for a period
310: of 5.3 hours. There were 117 images for each of the two filters.
311: The seeing and image quality were not optimal, but they
312: were good enough for the difference imaging photometry.
313: Figures \ref{4juntas}-(c) and \ref{4juntas}-(d) show the best GEMINI-South
314: single images of the OGLE-TR-82 field for comparison with the previous
315: Figures \ref{4juntas}-(a) and \ref{4juntas}-(b). The stellar images are elongated, but obviously of higher
316: resolution than the VLT and OGLE images. It can be appreciated that the source
317: has a faint unresolved companion 1-2 pixels away to the East.
318:
319: The transit light curves were measured following the procedure described for
320: the VLT transits (see Fern\'andez \etal 2006).
321: The $g$-band transit was very difficult to measure, because the star
322: becomes too faint, and it is overwhelmed by the nearby neighbour.
323: There is a factor of 10 difference in counts in each individual image
324: for this star between the $g$ and $i$-bands.
325: %cambio
326: Figure \ref{lc} shows the full $g$ and $i$-band light curve for the GEMINI-South
327: observations, when the OGLE-TR-82 transit was monitored.
328: There are about $30$ points in each of the light curves
329: during in our single transit shown in Figure \ref{lc},
330: %cambio
331: % Figure 5,
332: and the minimum is well sampled, allowing us to measure accurate amplitudes.
333: The resulting light curves (which were not corrected for a linear trend), yield
334: $rms_g=0.005$ and
335: $rms_i=0.002$ mag in the flat portion at the end of the night.
336: Figure \ref{lc} also shows the phased light curve of the OGLE
337: $I$-band photometry (in a similar scale) for comparison. The transit
338: is well sampled in our $i$-band observations, and the scatter appears smaller.
339: We confirm
340: the amplitude of the transit measured by OGLE in the $I$-band, $A_i=0.034 ~mag$,
341: but find that the amplitude in the $g$-band is larger, $A_g=0.10 ~mag$.
342: Already the GEMINI-S data cast doubts on the planetary transit scenario, because
343: of the amplitude difference between the $g$ and $i$-band transits.
344: Also, the $g$-band light curve shows a more triangular eclipse shape, which
345: is more characteristic of a grazing binary.
346: However, the $g$-band light curve was hard to measure, and we cannot
347: discard residual contamination from the neighbours or fluxing problems
348: with the difference image photometry.
349:
350:
351: \subsection{Previous Spectroscopy}
352:
353: \hspace{20pt} As noted in the Introduction, Pont \etal (2004) observed OGLE-TR-82 with
354: UVES/FLAMES at the VLT, but they could find no CCF signal, concluding that this
355: was an unsolved case. They point out that the object could be too red,
356: and suggested further observations. They suggest a K7/M0 dwarf, less
357: extreme than the spectral type found in the next Section.
358:
359: We conclude that there was no signal for this object in the spectroscopic
360: observations because they were made in the $V$-band portion of the spectrum,
361: where the target is simply too faint with $V=20.6$. Using the setup of Pont
362: \etal (2004), this target cannot give useful signal to measure velocities with
363: an accuracy of $\sim 30 ~m/s$ in any reasonable amount
364: of integration time according to the
365: %cambio
366: UVES/FLAMES exposure time calculator.
367: Having answered the mystery about the absence of spectroscopic signal
368: for OGLE-TR-82, we open a more important question: is this star a giant or a dwarf?
369:
370: \section{A planetary transit?}
371:
372: With the present optical and infrared photometry one can estimate
373: some of the stellar parameters: the spectral type, luminosity, mass,
374: radius and distance.
375:
376:
377: \subsection{The Period of OGLE-TR-82-b:}
378:
379: The VLT observations did not show a transit, which was scheduled for the end
380: of the night according to the original OGLE ephemeris.
381: Udalski \etal (2005, private communication) revise the ephemeris for
382: this object, giving mean transit times of:
383: $$JD=2452323.84747 + 0.764244 t.$$
384: This puts the transit beyond the end of the
385: %cambio
386: VIMOS observing night, explaining its absence.
387: At the same time, this is a minor revision that
388: allowed us to recover the transit for the GEMINI observations.
389: With the transit observed at GEMINI-S, the ephemeris is improved. We
390: obtain a period similar to OGLE: $P=0.7643813\pm 0.0000010$ days.
391:
392: \subsection{The Spectral Type of OGLE-TR-82-b:}
393:
394: In a previous work, we have used optical and infrared photometry to
395: characterize OGLE extrasolar planetary companions (Gallardo \etal 2005).
396: We can estimate the spectral type of OGLE-TR-82 first using the multicolor photometry.
397: The stellar parameters from OGLE and the present photometry are listed in Table 2.
398:
399: The optical-infrared color-magnitude diagrams are shown in Figure \ref{cm}
400: for all stars in a Carina field of
401: about $1'\times 1'$. The disk main sequence is
402: very well defined. The target star OGLE-TR-82 is located away from this main sequence,
403: indicating that either this is a nearby late $M$-dwarf, or a very reddened distant giant.
404: Because both alternatives are very
405: different from the spectral type of K7-M0V adopted by Pont \etal (2004),
406: we consider these two possibilities that arise from the photometry.
407:
408: Figure \ref{cc}
409: shows the loci of giants and dwarfs of different spectral types
410: in the $V-I$ $vs$ $I-K$ color-color
411: diagram. The position of OGLE-TR-82 in this diagram is consistent with an
412: unreddened spectral type of M7V,
413: with an error of about one subtype (between M6V and M8V).
414: For this spectral type the mass and radius are
415: $M_s=0.10\pm 0.02 ~M_{\odot}$, and $R_s=0.15\pm 0.02 ~R_{\odot}$.
416: This star lies at the
417: boundary between M-dwarfs and brown dwarfs. In fact, adopting the 2MASS photometry
418: ($V-K=9.1$) instead of our values ($V-K=8.4$) results in an L-type brown dwarf.
419: The spectral type of M7V would yield an absolute magnitude of $M_V=18.60$
420: (Bessell 1991). The resulting distance modulus of $m-M=2.01$ gives a
421: distance of 25 pc, and would make this the nearest OGLE transiting planetary candidate.
422: Most other OGLE transiting candidates are located at distances between a few
423: hundred parsecs and a few kiloparsecs (Gallardo \etal 2005).
424:
425: However, Figure \ref{cc} also shows that if the reddening is severe, $E(B-V)\approx 2 ~mag$, the
426: location of OGLE-TR-82 can also be consistent with a distant red giant.
427: \subsection{The Radius of OGLE-TR-82-b:}
428:
429: If we consider the case of an M7V dwarf, the OGLE $I$-band
430: transit light curve amplitude yields $R_p/R_s = 0.18$.
431: This is a small planetary radius, similar to the radius of Neptune,
432: $R_p=0.3 ~R_J$.
433:
434: This is very different from the two previously published values
435: for the radius of the OGLE-TR-82 companion,
436: both based on the OGLE photometry. Udalski \etal (2002) obtain
437: $R_p=1.1 ~R_J$, and
438: Silva \& Cruz (2006) assume a star with $M= 0.65 ~M_{\odot}$, and obtain
439: a radius $R_p= 1.10 ~R_J$.
440: Both values are much larger than the radius
441: implied for a M7V star, indicating the dramatic change due to the
442: new spectral type obtained for OGLE-TR-82 based on the $V-K$ color of this source
443: if it is indeed a main sequence star.
444:
445: The present photometry shows that the previous
446: attempts to measure the radial velocities of OGLE-TR-82 failed because
447: this object is much too faint in the optical. It is now clear
448: that with $V=20.61$ it would have been impossible to measure.
449: However, velocities in
450: the near-infrared are within reach of the largest telescopes, because
451: the target is very red $K=12.2$. If this were a low mass
452: star with a planet in a tight orbit, these velocity measurements do not need
453: to approach the few m/s accuracy as usually needed for the RV planet searches.
454: In this case, a Jupiter-mass planet would yield a
455: radial velocity semiamplitude of about $1.1 ~km/s$,
456: well within reach of near-infrared spectrographs at large 8m class telescopes.
457: A Saturn-mass planet would give a smaller radial velocity semiamplitude of $300 ~m/s$.
458:
459: The smallest extrasolar planet currently known is
460: HD149206-b, with $R_p=0.7 ~R_J$ (Sato \etal 2005),
461: a Saturn-mass planet with a large dense core.
462: The smallest companions so far have been detected in the bulge
463: fields by Sahu \etal (2006) around stars of $0.5 ~M_{\odot}$.
464: If OGLE-TR-82 were an K7-M0V star (Pont \etal 2004), it
465: would be an extreme case because of its low mass, small size,
466: and short period companion. The planetary option would be still open for this target
467: if we could confirm the M-dwarf of later type.
468:
469: In the absence of spectroscopic velocities, we cannot estimate the
470: mass of the companion. However, in order to give an idea of the
471: possibilities, we note that a companion with $M=1 ~M_J$ would give
472: an orbital semimajor axis of $a = 10.8$ and $R_s = 0.0076 ~AU$.
473: %For a $0.10 ~M_{\odot}$ star, this is
474: %well outside its Roche radius ($R_R \sim 0.002 ~AU$ in this case).
475:
476: There is, however, a problem with this interpretation: the total transit time
477: observed is $t_T\approx 0.04 ~d$, while the stellar and orbital parameters would predict
478: a shorter transit of $t_T=0.027 ~d$.
479:
480: Though interesting as a good planetary candidate at this stage, a deeper study
481: shows that the interpretation is not so simple, as discussed below.
482:
483: \subsection{A triple system?}
484: We have explored different possibilities for the nature of
485: the OGLE-TR-82 companion:
486: a planetary companion to an M-dwarf,
487: a white dwarf transiting in front of a M-dwarf,
488: a brown dwarf transiting in front of a M-dwarf,
489: a grazing M-dwarf tight binary of similar masses, variability like spots,
490: a main sequence star transiting in front of a red giant,
491: and a blend with a reddened background giant star.
492: The last possibility is the only one that agrees with all the available data,
493: as discussed below.
494:
495: \subsection{Blend with a background star hypothesis:}
496:
497:
498: An eclipsing binary blended with a background star yields low amplitude transits
499: because the light of the contaminating star dilutes the eclipse.
500: Alternatively, the eclipses of a distant binary would be diluted by a blend
501: with a foreground star.
502:
503: It is now well known that transit searches are plagued by blends, and
504: that in order to confirm the planetary nature of a transit candidate
505: the radial velocity orbit must be obtained. In the present case the
506: blend option appears the most likely.
507:
508: The best seeing optical images of GEMINI-South (Figures \ref{4juntas}-(c) and \ref{4juntas}-(d) shows an elongated shape even for the faint OGLE-TR-82 star. This is obvious in the $g$-band (Figure \ref{4juntas}-(c)), but less evident in the $i$-band (Figure \ref{4juntas}-(d)).
509: The elongation is in the E-W direction, but the pair is faint and unresolved, we estimate the separation to be approximately 1-2 pixels in the GEMINI-South images.
510: One of the sources appears to be the very red source (to the N), while the
511: other (to the S) shows normal color of typical main sequence field stars.
512: Note that spectroscopic fiber sizes (e.g. HARPS, FLAMES) would not be able
513: to resolve the light form both stars, the spectra would necessarily be the
514: composite spectrum of the blend.
515:
516: We conclude that
517: the red source is a reddened background giant, while the other source is a
518: normal main-sequence binary. The blend reduces the amplitudes of the
519: eclipses of the binary as seen in the $I$-band, mimicking the light curve of
520: a planetary transit.
521:
522: It is usually not possible to distinguish between the cases of a
523: foreground binary blended with a background star $vs.$ a background
524: binary blended with a foreground star.
525: %binary.
526: In the present case, while a
527: late main-sequence star transiting in front of a red giant
528: can produce an eclipse of the measured depth $A=0.034 ~mag$,
529: a blend of a foreground main sequence star with a background binary
530: red giant does not work, because the measured transit time of 1.5 hr
531: is too short compared with typical red-giant sizes, even taking into account
532: the short period of OGLE-TR-82. It would
533: have to be a grazing eclipse with a marked triangular shape.
534:
535: The final solution to the dilemma comes from near-IR spectroscopy obtained
536: with SofI at the NTT.
537:
538: The first question that our near-IR spectra allow us to address is
539: if the star is a giant or a dwarf. The 2-dimensional spectral
540: classification is discussed in detail in Ivanov \etal (2004, see
541: Section 5.2). First, we used Eqs. 2 and 4 to determine the T$_{eff}$
542: and then placed the object on the plots shown in Figure 12 (left) in
543: Ivanov \etal (2004) (Figure \ref{lumclass}). We measured Mg{\sc i} 1.50\,$\mu$m, Si{\sc i}
544: 1.58\,$\mu$m, CO 1.62\,$\mu$m and Mg{\sc i} 1.71\,$\mu$m indices as
545: defined in that paper on the extinction corrected spectrum to obtain
546: 0.11$\pm$0.04 mag, 0.03$\pm$0.01 mag, 0.04$\pm$0.02 mag and 0.06$\pm$0.02 mag,
547: respectively. Note that here and further in this section the
548: uncertainties should be treated with caution because they include
549: only the Poisson errors from the spectra and not any systematic errors
550: that might rise from the sky subtraction or the telluric corrections,
551: for example. Also, we did not correct for any differences due to the
552: lower resolution of our spectra with respect with the library of
553: Ivanov et el. (2004). The derived stellar effective temperatures were
554: 4077 and 5101\,K. We averaged them to obtain T$_{eff}\sim$ 4600 K,
555: suggesting the star has K0 spectral type. The {\it CO\,1.62 -
556: (Mg{\sc i}\,1.50 + Mg{\sc i}\,1.71)/2 vs log\,T$_{eff}$} plot (Fig.
557: \ref{lumclass}, top) firmly places the target
558: among the giants while the {\it CO\,2.29 - (Na{\sc i}\,2.20 +
559: Ca{\sc i}\,2.26)/2 vs log\,T$_{eff}$} plot (Fig. \ref{lumclass}, bottom) is inconclusive.
560:
561: Finally, we applied the metallicity calibration technique of Frogel
562: \etal 2001 (see their Eq. 3) based on the strength
563: of Na{\sc i} 2.20\,$\mu$m, Ca{\sc i} 2.26\,$\mu$m and the 2.3\,$\mu$m
564: CO band. We preferred the spectroscopic method rather than the
565: combination of spectroscopic and photometric indices because our
566: target suffers strong extinction and even a small uncertainty in
567: the extinction would cause significant error in the metallicity
568: estimate. Following the prescriptions of Frogel \etal (2001) and
569: Ram\'{i}rez \etal (1997) we measured the following equivalent widths:
570: 1.5$\pm$0.1, 1.6$\pm$0.1 and 10.2$\pm$2.0\AA, respectively for the
571: Na, Ca and CO features. For the metallicity estimates we tentatively
572: doubled the errors. This yields [Fe/H]$_{ZW}$=$-$1.0$\pm$0.1 in the
573: Zinn \& West (1984) metallicity scale. The transformation of Frogel
574: \etal (2001; Eq. 6) give [Fe/H]$_{CG}$=$-$0.86 in the Carretta \&
575: Gratton (1997) scale. Frogel \etal (2001) also give a metallicity
576: calibration based on the CO strength alone. It yields
577: [Fe/H]$_{ZW}$=$-$1.06$\pm$0.13, in agreement with the first estimate.
578:
579: Of course, the classification criteria of Ivanov \etal (2004) and
580: the calibrations of Frogel \etal (2001) were derived for disks or
581: globular cluster stars with different age, chemical enrichment
582: history, abundance ratios, etc. than the Milky Way halo stars.
583: Therefore, our results should be treated with caution and we can
584: only conclude that the OGLE-TR-82 primary is consistent with being
585: a metal poor red giant.
586:
587:
588: In addition, we used the library of Pickles (1985) to carry out empirical spectral classification by comparing dwarfs and giants templates spectra with the near-IR spectrum of our target. From this way, this spectrum first confirms the extremely red nature
589: of this object, as deduced from the IR photometry. Second, and most
590: importantly, the spectral type obtained is consistent with the K3III, with an estimated
591: absorption $A_V=7 ~mag$.
592:
593: The strong CO band at 2.3 micron makes it impossible to
594: match the host star with the spectrum of any red dwarf from the spectral
595: library of Pickles (1985),
596: except for the latest spectral types M6-M7. However, the specific shape
597: of the continuum for these stars does not match the shape
598: of the continuum of our star (Figure \ref{esp1}).
599: Earlier dwarfs than M5 are also inconsistent with the observed colors,
600: unless one assumes unrealistically larger absorption of
601: $A_V>12-15$ mag, which appears impossible for a nearby star.
602:
603: This leaves us only with the possibility of a
604: distant giant star, supported by the shape of the continuum.
605: The comparison with the giant template spectra shown in Figure %\ref{esp2}%
606: \ref{esp2b} % revision1%
607: %for differents reddenings excludes a match with G5-K1III type stars because of the
608: %observed strenght of the CO band.
609: shows that the best fits are found with
610: K3III giant and $A_V=7 ~mag$, as expected for a distant giant
611: star located in the Galactic plane.
612: %The corresponding absolute magnitude for this spectral type is $M_V=+0.6$, which for
613: %a total absorption of $A_V=6.0$ yields a distance of $D=6.3 ~kpc$.
614: Stars of later types present
615: poor matches because of the strong water vapor features.
616: We caution that some contamination from the nearby blue star is
617: expected in the blue part of the spectra at 1 micron.
618:
619: %Finally, note that the Na and Ca features in the spectrum
620: %are weak, suggesting somewhat lower abundance than that of
621: %the spectral library templates. If this is indeed the case, the
622: %spectral type of the host star may be one-two subtypes later.
623:
624: This spectral type finally rules out a planetary size companion. Even though the mass
625: of the secondary is unconstrained due to the lack of radial velocities,
626: we conclude that this companion is most probably a late-type main sequence star.
627: Thus, the OGLE-TR-82 system is an eclipsing binary,
628: blended with a background reddened K giant.
629:
630:
631:
632: \section{Conclusions}
633:
634: Udalski \etal (2002) discovered low amplitude transits in the main
635: sequence star OGLE-TR-82, which was considered as a prime planetary candidate
636: orbiting a M-type dwarf (Pont \etal 2004, Silva \& Cruz 2006).
637: We find that this object has an extremely red color $V-K=8.4$,
638: making it unique among the OGLE transiting candidates in Carina.
639: Future space based missions like COROT and KEPLER may discover low
640: amplitude transits in a few faint red objects, which may turn
641: out to be difficult to observe with echelle spectrographs. OGLE-TR-82 is
642: one example, a difficult but interesting case that we have followed up here.
643:
644: We acquired good seeing images in the $V$-band with VIMOS at the ESO VLT, and
645: in the $K$-band with SofI at the ESO NTT.
646: We also observed a single transit of this star in the $g$ and $i$-bands with the
647: GEMINI-S telescope.
648: Our data is complemented by data of Udalski \etal (2003)
649: in the $I$-band, who observed 22 transits, but with few points per transit.
650:
651: We conclude that it was not possible to measure velocities for this
652: star in the past, and thus to estimate the mass of the companion,
653: because the star is too faint in the optical ($V=20.6$).
654:
655: The transit amplitudes are well measured, with 30 points in
656: transit for each of the two optical bands. We confirm
657: the amplitude of the transit measured by OGLE in the $I$-band, $A_i=0.034 ~$,
658: but find that the amplitude in the $g$-band is larger, $A_g=0.10 ~mag$.
659:
660: We explore different possibilities for this system.
661: First, the hypothesis of an M7V primary leads to a possible
662: planetary size for the transiting candidate.
663: %We measure a transit amplitude of $A_V=0.XXX\pm 0.0XX$ mag.
664: In this case, based on the new photometric data, and assuming a M7V primary,
665: the radius for the
666: companion would be $R_p=0.3\pm 0.1 ~R_J$, i.e. the size of Neptune.
667: This scenario has the problem that the measured transit time is too long,
668: and that the amplitudes seem to be different in the $g$ and $i$-bands.
669:
670: Second, the
671: alternative explanation consists of a triple system, composed of an eclipsing binary blended
672: with a background red giant.
673: In this case the red giant has to be very reddened
674: and very distant.
675:
676: Clearly, spectroscopic measurements were still needed for this interesting target.
677: These are more efficiently carried out in the near-infrared, given the
678: extreme red color of this object. Quantitative analysis of the near-IR spectrum obtained at the ESO NTT finally proved that OGLE-TR-82 is a distant reddened metal poor early giant. This result is confirmed by direct comparison with stellar templates that gives the best fit with a K3III star with an absorption of $A_V=7 ~mag$.
679: % and distance $D=6.3 ~kpc$.
680: This rules out a planetary size companion.
681:
682: We conclude that this system is a main-sequence binary blended with a background
683: red giant.
684:
685: Based on the $I$-band photometry, the OGLE-TR-82 system perfectly
686: mimics an M dwarf with a Neptune-size companion. After long effort
687: follow-up, we have found that the system is composed of
688: an eclipsing binary blended with a background K giant. There is a lesson to be learned for
689: future transit surveys searching for hot Neptunes and super Earths around
690: late type stars.
691:
692: \acknowledgments
693: SR, JMF, DM, GP, MZ, MTR, WG are supported by Fondap
694: Center for Astrophysics No. 15010003.
695: We thank the GEMINI staff at Cerro Pachon, the ESO staff at Paranal
696: Observatory, and at La Silla Observatory and Ricardo Salinas for his help.
697:
698: \begin{references}
699:
700:
701:
702: %\reference{}{Alard, C. 2000, A\&ASuppl., 144, 363 }
703: %\reference{}{Alard, C., \& Lupton, J. 1998, ApJ, 503,325 }
704: %\reference{}{Alonso, R., \etal 2004, ApJ, 613, L153}
705: \reference{}{Ali, B., Carr, J. S., Depoy, D. L., Frogel, J. A., \& Sellgren, K. 1995, AJ, 110, 2415}
706: \reference{}{Alibert, Y., \etal 2006, 455 L25}
707: \reference{}{Baraffe, I., Chabrier, G., Barman, T., Selsis, F., Allard, F., \& Hauschildt, P. H. 2005, A\&A, 436, L47}
708: \reference{}{Baraffe, I., \etal 2006, A\&A, 450, 1221B }
709: \reference{}{Beaulieu, J.P. \etal 2006, Nature, 439, 437}
710: \reference{}{Bessell, M.S. 1991, A\&A, 89, 35 }
711: \reference{}{Bonfils, X., 2005, A\&A, 443, L15}
712: \reference{}{Bodenheimer, P., Laughlin, G., \& Lin, D. N. C. 2003, ApJ, 592, 555 }
713: \reference{}{Butler, R. P., \etal 2006, ApJ, 646, 505 }
714: \reference{}{Carretta, E., \& Gratton, R. G. 1997, A\&AS, 121, 95}
715: %\reference{}{Bouchy, F., \etal 2004, A\&A, 421, L13 }
716: %\reference{}{Burrows, A., Sudarsky, D., \& Hubbard, W. B. 2003, ApJ, 594, 545 }
717: %\reference{}{Charbonneau, D., Brown, T., Latham, D., \& Mayor, M. 2000, ApJ, 529, L45 }
718: %\reference{}{Claret, A., \& Hauschildt, P. 2003, A\&A, 412, 241 }
719: \reference{}{D{\'i}az, R., \etal 2006, ApJ, submitted}
720: %\reference{}{Farmer \& Agol 2004, ApJ, XXX, XXX}
721: %\reference{}{Farihi, \etal 2005, ApJ, XXX, XXX}
722: \reference{}{Fern\'andez, J. M., \etal 2006, ApJ, 647, 587}
723: \reference{}{Frogel J. A., Stephens, A., Ram\'{i}rez, S., \& DePoy, L. 2001, AJ,
724: 122, 1896}
725: %\reference{}{Fischer, D. A., \etal 2005, ApJ, 620, 481 }
726: \reference{}{Gallardo, J., Minniti, D., Valls-Gabaud, D., \& Rejkuba, M. 2005, A\&A, 431, 707}
727: \reference{}{Gould, A., \etal 2006, ApJ, 644, L37}
728: \reference{}{Ivanov V.D., Rieke, M.J. Engelbracht, C.W. \etal 2004, ApJS, 151, 387}
729:
730: \reference{}{Kennedy \etal 2006, ApJ, in press}
731: \reference{}{Kleinmann, S. G., \& Hall, D. N. B. 1986, ApJS, 62, 501}
732: \reference{}{Legget, S., 1992, ApJS, 82, 351}
733: \reference{}{Lovis, C., \etal 2006, Nature, 441, 305}
734: \reference{}{Maiolino, R., Rieke, G.H. \& Rieke, M.J., AJ, 111, 537}
735: \reference{}{McArthur \etal 2004, ApJ, 614, L81}
736: \reference{}{Moorwood, A., Cuby, J.G., Lidman, C. 1998, The Messenger 91, 9}
737: \reference{}{Origlia, L., Moorwood, A. F. M., \& Oliva, E. 1993, A\&A, 280, 536}
738: \reference{}{Pepper, J., \& Gaudi, S. A. 2005, ApJ, in press (astro-ph/0507548) }
739: \reference{}{Pickles, A.J., 1985, ApJS, 59, 33 }
740: \reference{}{Pickles, A.J., 1998, PASP, 110, 863}
741: \reference{}{Pont, F., Bouchy, F., Queloz, D., Santos, N. C., Melo, C., Mayor, M., \& Udry, S. 2004, A\&A, 426, L15 }
742: \reference{}{Pont, F. 2006, in Tenth Anniversary of 51 Peg-b: Status of and prospects for
743: hot Jupiter studies, ed. L.~{Arnold}, F.~{Bouchy}, \& C.~{Moutou}, 153--164}
744: \reference{}{Ram\'{i}rez, S. V., DePoy, D. L., Frogel, J. A., Sellgren, K., \&
745: Blum, R. 1997, AJ, 113, 1411}
746: \reference{}{Sato, B., \etal 2005, ApJ, 633, 435 }
747: \reference{}{Sahu, K., \etal 2006, Nature, 443, 534}
748: \reference{}{Santos, N., \etal 2004, A\&A, 426, L19}
749: \reference{}{Silva, A. V. R., \& Cruz, P. C. 2005, ApJ, 642, 488 }
750: \reference{}{Udalski, A., Paczy\'nski, B., Zebrun, K., Szymanski, M., Kubiak, M., Soszynski, I., Szewczyk, O., Wyrzykowski, L. \& Pietrzynski, G. 2002a, Acta Astronomica, 52, 1 }
751: \reference{}{Udalski, A., Zebrun, K., Szymanski, M., Kubiak, M., Soszynski, I., Szewczyk, O., Wyrzykowski, L. \& Pietrzynski, G. 2002b, Acta Astronomica, 52, 115 }
752: \reference{}{Udalski, A., Szewczyk, O., Zebrun, K., Pietrzynski, G., Szymanski, M., Kubiak, M., Soszynski, I., Wyrzykowski, L. 2002, Acta Astronomica, 52, 317}
753: \reference{}{Udalski, A., Pietrzynski, G., Szymanski, M., Kubiak, M., Zebrun, K., Soszynski, I., Wyrzykowski, L. 2003, Acta Astronomica, 53, 133}
754: \reference{}{Zinn, R., \& West, M. J. 1984, ApJS, 55, 45}
755: \end{references}
756:
757: \clearpage
758:
759: \begin{deluxetable}{lcl}
760: \tablewidth{0pt}
761: \tablecaption{Log of observations of OGLE-TR-82}
762: \tablehead{\colhead{Instrument} & \colhead{Observation Date} & \colhead{}}
763: \startdata
764: VLT+VIMOS &Apr 2005&$V$-band photometry\\
765: NTT+SofI &May 2005&$Ks$-band photometry\\
766: & & NIR spectroscopy\\
767: GEMINI-S+GMOS &Jan 2006&$g$ \& $i$-band photometry\\
768: \enddata
769: \end{deluxetable}
770:
771:
772: \begin{deluxetable}{lcl}
773: \tablecaption{Parameters of OGLE-TR-82 and its companion}
774: \tablehead{\colhead{Parameter} & \colhead{} &\colhead{References}}
775: \tablewidth{0pt}
776: \startdata
777: $Period$&$0.764244 ~d$ &revised by Udalski \etal 2005\\
778: JDo&$2452323.84747 ~d$ &revised by Udalski \etal 2005\\
779: $A_I$ &$0.034$ &Udalski \etal 2002\\
780: $A_i$ &$0.034$ &This work \\
781: $A_g$ &$ 0.1$ &\\
782: $V$ &$20.61\pm 0.10$&\\
783: $I$ &$16.30\pm 0.10$&Udalski \etal 2002\\
784: $K$ &$12.20\pm 0.10$&\\
785: $V-I$ &$4.31\pm 0.10$ &\\
786: $I-K$ &$4.10\pm 0.10$ &\\
787: $V-K$ &$8.41\pm 0.10$ &\\
788: $t_T$ &$0.063 ~d$ &\\
789: \enddata
790: \end{deluxetable}
791:
792: \clearpage
793:
794: \begin{figure}
795: \plotone{f1.ps}
796: \caption{
797: Period $vs$ $I-K$ color for OGLE transit candidates in the Carina region.
798: The reddest object is OGLE-TR-82, marked with the cross. The left cross on
799: top of the circle is our $I-K$, and the right cross is from 2MASS.
800: }\label{uno}
801: \end{figure}
802:
803:
804:
805: \begin{figure}
806: \begin{center}
807: \leavevmode
808: \plottwo{f2a-s.ps}{f2b-s.ps}
809: \plottwo{f2c-s.eps}{f2d-s.eps}
810: \end{center}
811: \caption{
812: Portions of images in differents bands including OGLE-TR-82 which is the star at the center of the big circle.
813: These images covers $20\times 20$ arcsec.
814: (a): 0.5 arcsec seeing VIMOS $V$-band image. OGLE-TR-82 has $V=20.6$ and the faintest stars seen have $V\sim 24$. The size of the large circle is approximately the size of the fibers
815: used in previous attempts that failed to measure radial velocities for this candidate.
816: Under normal $1$ arcsec seeing conditions, this target is swamped by the
817: nearby neighbour.
818: (b): 0.85 arcsec seeing OGLE $I$-band image. OGLE-TR-82 has $I=16.30$ and the faintest stars seen have $I\sim 22$.
819: (c): 0.5 arcsec seeing GEMINI-South $g$-band image. The faintest stars seen have $g\sim 24$.
820: (d): 0.5 arcsec seeing GEMINI-South $i$-band image. Compare this high resolution image with the OGLE-finding chart shown in (b).
821: }
822: \label{4juntas}
823: \end{figure}
824:
825:
826: \begin{figure}
827: %\plotone{ztr82kdeep.ps}
828: \plotone{f3.ps}
829: \caption{
830: Portion of a 0.7 arcsec seeing SofI $K$-band
831: image including OGLE-TR-82 ($K=12.2$),
832: which is the bright star at the center of the large circle. This image covers
833: $20\times 20$ arcsec as in Figure \ref{4juntas} (scale $0.287 ~arcsec/pix$),
834: and the faintest stars seen have $Ks\sim 18$.
835: }\label{K}
836: \end{figure}
837:
838: \begin{figure}
839: %\plotone{lcgeminiogle.ps}
840: \plotone{f4.ps}
841: \caption{
842: Single transit of OGLE-TR-82 observed with GMOS at the GEMINI-South telescope in the
843: $g$-band (top) and in the $i$-band (middle), compared with the OGLE $I$-band data phased
844: using $P=0.76416 ~d$ shown to the same scale (bottom).
845: }\label{lc}
846: \end{figure}
847:
848: \begin{figure}
849: \plotone{f5.ps}
850: \caption{
851: Optical-infrared color-magnitude diagrams of all stellar objects in a $1'\times 1'$ field in Carina.
852: In both panels, the big square on the right is OGLE-TR-82, while the one on
853: the left is OGLE-TR-113 shown for comparison.
854: The reddening vector corresponding to $A_V=6 ~mag$ is shown for OGLE-TR-82.
855: }\label{cm}
856: \end{figure}
857:
858: \begin{figure}
859: \plotone{f6.ps}
860: \caption{
861: Color-color diagram indicating the fiducial loci of giants (open circles)
862: and dwarfs of different spectral types (full triangles). The observed position
863: of OGLE-TR-82 is marked with the crosses. The top one comes from the
864: 2MASS colors, and the bottom one comes from our photometry. This position
865: is consistent with a late main-sequence star, next to the location of
866: a typical M7V star. The direction of the reddening vector is indicated
867: with the straight line, the length gives the vector corresponding to a total
868: $A_V=6.0 ~mag$. Considering this reddening, the star would lie closer to the
869: location of K-type red giants.}\label{cc}
870: \end{figure}
871:
872:
873: \begin{figure}
874: \plotone{f7.eps}
875: \caption{
876: Two-dimensional spectral classification with H-band features (top) and K-band features (bottom). All indices are in magnitudes. Star symbols indicate supergiants, open circles are giants, and solid dots represent dwarfs and subgiants. The typical ±1 measurement error is shown in the top right corner. The CO bands are defined by Origlia \etal (1993), the Ca and Na indices are from Ali \etal (1995), and the Mg definitions are from Ivanov \etal (2004).
877: }\label{lumclass}
878: \end{figure}
879:
880:
881: \begin{figure}
882: \plotone{f8.eps}
883: \caption{
884: Near-IR spectrum of OGLE-TR-82 corrected for different reddening amounts (red thick curves)
885: compared with templates of dwarf stars from the spectral library of Pickles (1985) (black thin curves). The curves are shifted by a constant for a better comparison (y-axis is in arbitrary units).
886: }\label{esp1}
887: \end{figure}
888:
889: %\begin{figure}
890: %\plotone{f8.eps}
891: %\caption{
892: %Near-IR spectrum of OGLE-TR-82 (red thick curves) corrected for different
893: %reddening amounts (A$_{v}$ used are showed in red numbers)
894: %compared with giant and dwarf star templates (black thin curves) from
895: %the spectral library of Pickles (1985) (left and right panels, respectively).
896: %The curves are shifted by a constant for a better comparison.
897: %}\label{esp2}
898: %\end{figure}
899:
900:
901: \begin{figure}
902: \plotone{f9.eps}
903: \caption{ K-band spectrum of the OGLE-TR-82 primary (thick
904: line) compared with giant templates from the spectral library of Pickles (1998) (thin lines). %The 1$\sigma$ uncertain ***** are also shown (thin lines).%
905: Some more prominent features are marked. Given the signal-to-noise
906: and the resolution of the spectrum, only the stronger features are
907: detected. The curves are shifted by a constant for a better comparison (y-axis is in arbitrary units).
908: }\label{esp2b}
909: \end{figure}
910:
911:
912: \end{document}
913:
914:
915: