1: \documentclass[apjl]{emulateapj}
2: %\documentstyle[emulateapj5]{article}
3: %\documentclass[12pt,emulateapj]{aastex}
4:
5: %\documentclass[12pt,preprint]{aastex}
6: %\documentstyle[12pt,aasms4]{article}
7:
8: %\documentstyle[aas2pp4]{article}
9: %\usepackage{apjfonts}
10:
11: %\documentclass[12pt,preprint]{aastex}
12: %\usepackage{emulateapj5}
13: %\documentstyle[12pt,aasms4]{article}
14:
15: \shorttitle{The Nature of LMC-5}
16: \shortauthors{Drake et al.}
17:
18: \begin{document}
19: \title{Resolving the Nature of the LMC Microlensing Event LMC-5}
20:
21: \author{A.~J. Drake\altaffilmark{1,2,3}, K.~H. Cook\altaffilmark{3} {\sc and} S.C. Keller\altaffilmark{4}}
22:
23: \altaffiltext{1}{Dept. of Astrophysical Sciences, Princeton University, Princeton, NJ 08544}
24: \altaffiltext{2}{Depto. de Astronomia, P. Universidad Catolica, Casilla 104, Santiago 22, Chile}
25: \altaffiltext{3}{Lawrence Livermore National Laboratory, Livermore, CA 94550}
26: \altaffiltext{4}{Research School of Astronomy and Astrophysics, ANU, Canberra, ACT 2611, Australia}
27:
28:
29: \begin{abstract}
30:
31: We present the results from an analysis of Hubble Space Telescope High
32: Resolution Camera data for the Large Magellanic Cloud microlensing event
33: MACHO-LMC-5. By determining the parallax and proper motion of this object
34: we find that the lens is an M dwarf star at a distance of
35: $578^{+65}_{-53}$pc with a proper motion of $21.39 \pm 0.04 $ mas/yr.
36: Based on the kinematics and location of this star it more likely to be
37: part of the Galactic thick disk than thin disk population. We confirm that
38: the microlensing event LMC-5 is a jerk-parallax event.
39:
40: \end{abstract}
41: \keywords{ stars: low-mass -- Galaxy : halo -- dark matter}
42:
43: \section{\sc Introduction}
44:
45: For over a decade astronomers have been observing the Magellanic Clouds in
46: order to determine the fraction of the dark matter in our Galaxy that may be
47: in the form of Massive Compact Halo Objects (MACHOs). The discovery of a
48: significant number of microlensing events in the first two years of the
49: MACHO project lead to an uncertain initial estimate that approximately half
50: of the halo was composed of MACHOs (Alcock et al.~1997). With 3.7 years of
51: additional data this estimate decreased to $\sim 20\%$ of the halo (Alcock
52: et al.~2000). While it appears these objects make up a significant fraction
53: of the mass in the Galactic halo, little is known about their nature other
54: than that their most probable masses lie in the range of 0.15 and 0.9$M_\sun$
55:
56: In order to obtain the most accurate information which can be gained from a
57: microlensing event, it is useful to accurately determine the flux of the
58: star that was lensed. This is made difficult for sources in crowded fields
59: such as LMC because they are usually blended with neighboring stars. In
60: many cases each ``object'' identified in ground-based observations consists
61: of the blend of a number of stars (Alcock et al.~2000). The exact location
62: of the source star is also poorly known because of blending. To determine
63: the locations of the sources Alcock et al.~(2001a, 2001b) analyzed the MACHO
64: project images using Difference Image Analysis (Alcock et al.~1999, 2001a).
65: With these positions they were able to subsequently identify and photometer
66: the microlensing source stars in observations taken with {\it Hubble Space
67: Telescope} (hereafter HST) Wide Field Planetary Camera 2 (WFPC2).
68:
69: Among the events discovered by the MACHO project toward the Magellanic
70: clouds was event LMC-5. This event had a high magnification ($\sim$50) and
71: was detected in the light curve of Macho object 6.5845.1091 which is located
72: at $\alpha= 05^{\rm h}\!16^{\rm m}\!41\fs 1$, $\delta =-70\arcdeg\!
73: 29\arcmin\! 18\arcsec$ (J2000).
74: %\fd\fh\fm\fs
75: Gould, Bachall \& Flynn (1997) suggested that the baseline color of this
76: event was not consistent with an LMC source star and they proposed that the
77: anomalous color could be attributed to the source being blended with a M
78: dwarf in the Galactic disk. As the likelihood of finding and M dwarf within
79: the seeing disk is small they proposed an M dwarf could be the lens in the
80: foreground. Alcock et al.~(1997) found that the color of the source was in
81: agreement with the colors and magnitudes of LMC stars, but the event was
82: indeed blended with a red object. Most of the LMC microlensing events found
83: by Alcock et al.~(1997) were blended to some extent. When the source star
84: in this microlensing event was identified in HST observations it was
85: discovered that there was a faint red star nearby. The probability of
86: finding an unrelated foreground M dwarf near the microlensing source star is
87: $\sim 1$ in 10000 (Alcock et al.~2001b). With this in mind, it was thought
88: very likely that this object was the lens.
89:
90: The LMC-5 event shows the clear sign of microlensing parallax. In these
91: events, the motion of the Earth during the event changes the shape of the
92: microlensing light curve from the classical Paczy\'nski form (Gould 1992,
93: Alcock et al.~1995). The presence of parallax enables limits to be placed on
94: the mass and location of the microlens. The parallax fit for this event
95: yielded a lens motion direction that was consistent with the red star having
96: been the lens. However, the solution also suggested that the lens was
97: likely to be a sub-stellar object of 0.036M$_\odot$ ($\leq$ 0.097M$_\odot$
98: at 3-$\sigma$ significance) at a distance of $\sim 200$pc. Alcock et
99: al.~(2001b) derived a separate distance estimate for the lens using the
100: objects color and spectral type. First a spectrum of the lens-source
101: combination was obtained and it was found that the prospective lens was an
102: M4V or M5V type star. The V-I color of the object was determined from HST
103: WFPC2 photometry. This color was converted into an absolute magnitude using
104: the $M_{V}$ vs $V-I$ relation of Reid (1991) for M dwarfs. The distance was
105: then obtained from the observed and absolute magnitudes while the errors in
106: distance were estimated from the dispersion in M dwarf magnitudes
107: $V-I\sim3$. The result was that the object was at $650 \pm 190$pc, in stark
108: contrast to the microlensing parallax solution.
109:
110: If the red object is indeed the lens, a measurement of the parallax should
111: confirm the distance inferred from the color of the object in the HST
112: images. In addition, a measurement of the magnitude and direction of the
113: proper motion should agree with the initial estimate which assumed
114: the red object was the lens, and that the relative separations of
115: the two objects represented its proper motion.
116:
117: To resolve the nature of the candidate lens we undertook a program of
118: observations with HST's Advanced Camera System High Resolution Camera
119: (hereafter ACS and HRC, respectively). In the meantime, a new solution to
120: the LMC-5 puzzle was proposed by Gould (2004) based on the recent work of
121: Smith, Mao and Paczy\'nski (2003). By exploring the phase space of ``vector
122: microlens parallax'' in a geocentric reference frame, Gould (2004)
123: discovered a second solution to the microlensing parallax which varied from
124: the original solution of Alcock et al.~(2001b) by less than 0.1 in fit
125: $\chi^2$.
126:
127: The microlens vector parallax of this second solution differs from that of
128: the first solution by the so called "jerk parallax", a vector whose
129: direction lies perpendicular to the direction of the Earth's acceleration
130: and whose magnitude (for LMC events) is about $(4/3) ({\rm yr/2\pi t_{\rm e}}) \sim 2.4$.
131: Events exhibiting these so-called "jerk-parallax degeneracies" are expected
132: to be rare for microlensing toward the LMC, unless the lens resides in the
133: Galactic disk. In the case of the LMC-5 microlensing event, Gould (2004)'s
134: jerk-parallax solution is in agreement with the lens distance and direction
135: estimated from the HST photometry.
136:
137: The solution of Gould (2004) does not rule out the possibility that the
138: initial solution to the microlensing fit was the correct one, since both are
139: equally good fits to the lensing light curve. However, when this solution
140: is considered in combination with the other evidence from the HST data it is
141: much more likely that the lens is a sub-stellar object not detected in the
142: HST data. In this paper we will show with certainty that Gould (2004)'s
143: solution is the correct one.
144:
145:
146: \section{\sc Observations}
147:
148: We obtained images with the HST's HRC in July 2002 and January 2003. The
149: observations were taken approximately six months apart to maximize the
150: parallactic offset of the lens relative to proper motion vector. Each set
151: of observations consists of 6 images of the source - lens field, allowing us
152: to perform robust cosmic ray rejection and to determine very accurate
153: centroids for each object. The observations taken in 2002 used the F606W
154: and F814W filters, while the observations in 2003 were taken in the F606W
155: filter alone. The duration of each of the exposures was 400 seconds.
156:
157:
158: \section{\sc Analysis}
159:
160: The HRC images contain significant distortion in the form of a skew
161: due to the off-axis location of the ACS. To determine the distortion
162: corrected location of the stars in the HRC images, we followed the analysis
163: of Anderson and King (2004). In this process, the standard flat fielded
164: (flt) HRC images simply were fed into Anderson and King's ``img2xym'' task.
165: This program finds stars within the images and fits each with an effective
166: Point Spread Function (ePSF) which is based on an instrumental PSF modified
167: by the sub-pixel offset of each star's center.
168: The ePSF varies between observational filters so the correct starting PSF
169: must be chosen. The centroid location and flux of each star is determined
170: in the fitting process. However, because of the large amount of image
171: distortion the instrumental magnitudes and locations require correction to
172: an undistorted system where the offset and the changing effective pixel area
173: are corrected. The ``img2xym'' task also performs these steps to provide
174: corrected stellar locations and instrumental magnitudes. The RMS scatter
175: after the corrections of Anderson and King (2004) is $< 0.01$ pixels, or
176: about 0.25mas, for the brightest stars in each image.
177:
178: For each image, we determined the offset between the source star and the
179: assumed lens. We combined the results for each photometric band separately,
180: and estimated the uncertainties in these positions based on the scatter in
181: their locations. In addition, we combined these locations with the lower
182: resolution results obtained by Alcock et al.~(2001b) from analysis of HST
183: WFPC2 observations taken in June 1999. %PC1
184: It was not possible to estimate the parallax with the prior HST data since
185: there was only one known location (the WFPC2 point) and one assumed position
186: at the source star during the microlensing event.
187:
188: %\placetable{tab1}
189:
190: In Table 1 we present a summary of the observations used in this
191: analysis. We fitted the proper motion and parallax of the object using the
192: times and locations of the measurements. This fit places the M dwarf at an
193: offset of ($\Delta X$,$\Delta Y$) $=$ ($2.2\pm9.9$, $-1.6\pm7.8$) mas at the
194: time of the microlensing event. The reduced $\chi^2$ value of this fit is $<
195: 0.1$. This suggests that the errors in the locations are over estimated, and
196: the the real uncertainty in this offset if much smaller. The main
197: contributor to the uncertainty in the location is the error in the HST WFPC2
198: location. However, this result makes it quite certain that the red object
199: is indeed the lens. The source star itself is not stationary but moving
200: with the proper motion of the LMC which is
201: ($\mu_{\alpha {\rm LMC}}$, $\mu_{\delta {\rm LMC}}$) = (1.68, 0.34) mas/yr
202: (van der Marel et al.~2002). However, in the case of microlensing events we
203: are only interested in the motion of the lens relative to the source.
204:
205: With the lens identified, an additional constraint for determining the
206: proper motion and trigonometric parallax of the lens was derived from the
207: fact that we know that the source star and the lens must be collinear in our
208: line-of-sight at the time of the microlensing event peak amplification. We
209: fitted the locations again to determine the proper motion and parallax of
210: the lens with the inclusion of this additional point. We find the proper
211: motion of the lens relative to the source to be
212: ($\mu_{\alpha {\rm SL}}$, $\mu_{\delta {\rm SL}}$) = ($17.56 \pm 0.04$, $-12.22\pm0.02$ ) mas/yr.
213: The position angle of the proper motion vector is $\theta = 124.8\arcdeg$.
214: This in exellent agreement with the direction of Gould (2004)'s solution
215: of $123.9\arcdeg$. We note that the direction of proper motion in ecliptic
216: coordinates was incorrectly given by Alcock et al.~(2001b) as $\theta_{sky} = -91.6\arcdeg$,
217: rather than $\theta_{sky} =-105.7\arcdeg$. It appears that $\Delta\lambda$ was used
218: to determine the direction of motion instead of $\rm \Delta\lambda cos(\beta)$.
219: %It appears the factor of $cos(\beta)$ was neglected in the coordinate tranformation.
220: %$138.4\arcdeg$ (Gould 2004) rather than $124\arcdeg$ from the the HST WFPC data.
221:
222:
223: We find the parallax of the lens to be $\pi_{\rm L} = 1.73\pm0.18$ mas.
224: Therefore, the lens lies at a distance of $578^{+65}_{-53}$ pc. This result
225: is in agreement with the previous photometric estimate of Alcock et
226: al.~(2001b) ($650$ pc). The fact that we have been able to measure a $\sim
227: 2$ mas parallax to $\sim$10\% uncertainty is a good demonstration of the
228: astrometric accuracy that can be achieved with the HST HRC instrument.
229:
230: %\placefigure{lmc5}
231:
232: \begin{figure}[ht]
233: \epsscale{1.1}
234: \plotone{f1c.eps}
235: \figcaption{\label{lmc5}
236: Motion of the microlens relative to the position MACHO source star
237: 6.5845.1091 at the time of the microlensing peak magnification. The insert
238: presents an expanded view of dotted region where locations were determined
239: with the HRC instrument.
240: The dot at (0,0) shows the location of the source star and the large errors
241: bars near (0,0) shows the position of the object at the time of the
242: microlensing event determined from the initial parallax and proper motion
243: fit. See the text for further details.}
244: \end{figure}
245:
246: In Figure \ref{lmc5} we present the fit to the motion of the LMC-5 lens
247: corrected for the source star motion. The solid line shows the fit to the
248: HST data including the source star location as a point, while the dashed
249: line shows the proper motion and parallax that is expected with the lens
250: distance of 200pc as determined from the original microlensing parallax fit.
251: The fit shown in this figure is slightly more constrained than it appears
252: since the times of the measurements are an important part of the fitting
253: process. The F606W and F814W HRC points taken in July 2002 should lie very
254: close to each other. However, as the figure shows they are significanty
255: offset. We have checked for systematic errors in the transformations of
256: Anderson and King (2004) by matching a large number of stars between the
257: F606W and F814W frames. The coordinates of these stars matched within the
258: centroid uncertainties while the lens appeared to be slightly offset between
259: bands. However, the current level of uncertainty is too large to tell
260: whether the offset is real or due to an unquantified localized distortion.
261:
262: If we assume the lens undergoes average Galactic foreground reddening for
263: the LMC of E(B-V) = 0.06 (Oestreicher, Gochermann, Schmidt-Kaler 1995), we
264: find $M_{V}= 13.68$, consistent with an M5 dwarf star.
265: Our results are in agreement with the spectra and $M_{V}$ presented by
266: Alcock et al.~(2001b). It is very difficult to observationally
267: rule out the possibility that a 0.036M$_\odot$ object at 200pc as the lens
268: since the object could be fainter than 30th magnitude in V band (Baraffe et al.~1998).
269: This mass lies below the limits where models are accurate. However, K band
270: observations may be more promising.
271: The fact that we know that foreground M dwarfs are extremely
272: rare in fields toward the LMC (Alcock et al.~2001b) suggests this object is
273: very likely the lens. From our initial parallax and proper motion fit we
274: known that the object was within a few milli-arcseconds of the source at the
275: time of the lensing event. Our results agree with the new jerk-parallax
276: solution discovered by Gould (2004). Therefore, there is almost no doubt
277: that the M dwarf we have observed is the indeed the lens.
278:
279: For the lens we find the space velocity components (U,V,W) to be
280: (43.2,-55.7,29.0) $\rm km.s^{-1}$ corrected to the Local Standard of Rest.
281: These values are much higher than those presented by Alcock et al.~(2001b)
282: as they assumed a distance of 200pc and are quite high for thin disk M
283: dwarfs. However, the displacement from the Galactic plane ($\sim$300pc) and
284: the velocity are consistent with both thin and thick disk stars. We have
285: simulated the stellar population of disk and halo stars toward the LMC field
286: following the method of Vallenari, Bertelli, and Schmidtobreick (2000).
287: Using a common range of disk parameters (scale height, scale length, etc.).
288: We find that the lens is slightly more likely to be a thick disk star
289: ($\sim 50\pm30\%$) than a thin disk one. Clearly the likelihood is strongly
290: dependent on adopted parameters for the Galactic components. The kinematics
291: are also in good agreement with the thick disk kinematics derived by Chiba
292: and Beers (2000).
293:
294:
295: \section{Conclusions}
296:
297: We have analyzed HRC data for LMC microlensing event LMC-5 and we find that
298: the lens in this microlensing event is an M dwarf star. Based on our
299: analysis we can confirm that the jerk-parallax solution to the microlensing
300: light curve discovered by Gould (2004) is correct. The kinematics of this
301: star suggest that it is most likely a part of the Galactic thick disk
302: population rather than part of the dark halo. This is the first time that
303: any microlens has been identified with such certainty. However, this
304: discovery does not affect the current estimates of the mass fraction of the
305: Galactic dark halo in the form of MACHOs, since some microlensing events due
306: to foreground disk stars are expected in all LMC microlensing models.
307:
308: We would like to thank an anonymous referee for his many helpful suggestions.
309: We would also like to thank Jay Anderson who generously made his
310: results and analysis programs available to us prior to their release.
311: Support for this publication was provided by NASA through proposal numbers
312: GO-9394 and from the Space Telescope Science Institute, which is operated by
313: the Association of Universities for Research in Astronomy, under NASA
314: contract NAS5-26555.
315: This work was performed under the auspices of the U.S.~Department of Energy
316: National Nuclear Security Administration by the University of California,
317: Lawrence Livermore National Laboratory under contract W-7405-Eng-48. The
318: work done by A.~J.~D. is supported by Chilean FONDECYT grant 1030955.
319:
320: %\newpage
321: % TABLE1.TEX
322:
323:
324: \begin{thebibliography}{}
325:
326: \bibitem[\protect\citeauthoryear{{Alcock}, et~at.}{{Alcock}, et~al.}{1995}]{Alcock95}
327: {Alcock}, C., et~al. 1995, 454, L125
328: \bibitem[\protect\citeauthoryear{{Alcock}, et~at.}{{Alcock}, et~al.}{1997}]{Alcock97}
329: {Alcock}, C., et~al. 1997, ApJ, 491, 436 %Microlensing
330: \bibitem[\protect\citeauthoryear{{Alcock} et~at.}{{Alcock} et~al.}{1999}]{Alcock99}
331: {Alcock}, C., et~al. 1999, ApJ, 521, 602
332: \bibitem[\protect\citeauthoryear{{Alcock}, et~at.}{{Alcock}, et~al.}{2000}]{Alcock00}
333: {Alcock}, C., et~al. 2000, ApJ, 542, 281 %Microlensing
334: \bibitem[\protect\citeauthoryear{{Alcock} et~at.}{{Alcock} et~al.}{2001a}]{Alcock01a}
335: {Alcock}, C., et~al. 2001a, ApJ, 552, 582 %location of lenses
336: \bibitem[\protect\citeauthoryear{{Alcock} et~at.}{{Alcock} et~al.}{2001b}]{Alcock01b}
337: {Alcock}, C., et al. 2001b, Nat, 414, 617 %lmc5
338: \bibitem[\protect\citeauthoryear{{Anderson} \& {King}}{{Anderson} \& {King}}{2004}]{Anderson04}
339: {Anderson}, J., \& {King}, I. 2004, in 2002 HST Calibration Workshop, ed. S. Arribas, A. Koekemoer,
340: \& B., Whitmore (Baltimore: STScI), in prep.
341: \bibitem[\protect\citeauthoryear{{Baraffe}, {Chabrier}, {Allard} \& {Hauschildt}}{{Baraffe}, et~al.}{1998}]{Bar98}
342: {Baraffe}, I., {Chabrier}, G., {Allard}, F., {Hauschildt}, P.H. 1998, A\&A, 337, 403
343: \bibitem[\protect\citeauthoryear{{Chiba} \& {Beers}}{{Chiba} \& {Beers}}{2000}]{Chiba}
344: {Chiba}, M., Beers, T.C. 2000, AJ, 119, 2843
345: \bibitem[\protect\citeauthoryear{{Gould}}{{Gould}}{1992}]{Gould92}
346: {Gould}, A. 1992, ApJ, 392, 442
347: \bibitem[\protect\citeauthoryear{{Gould},{Bachall}, \& {Flynn}}{{Gould} et~al.}{1997}]{Gould97}
348: {Gould}, A., {Bachall}, J.N., \& {Flynn}, C. 1997, 392, 442
349: \bibitem[\protect\citeauthoryear{{Gould}}{{Gould}}{2004}]{Gould04}
350: {Gould}, A. 2004, ApJ, in press, astro-ph/0311548
351: \bibitem[\protect\citeauthoryear{{Oestreicher}, {Gochermann}, \& {Schmidt-Kaler}}{{Oestreicher}, et~al.}{1995}]{Oes95}
352: {Oestreicher}, M., {Gochermann}, J., \& {Schmidt-Kaler}, T. 1995, A\&AS,112, 4950
353: \bibitem[\protect\citeauthoryear{{van der Marel}, {Alves},{Hardy}, \& {Suntzeff}}{{van der Marel}. et~al.}{2002}]{Van02}
354: {van der Marel}, R., {Alves}, D.R., {Hardy}, E., \& {Suntzeff}, N. 2002 AJ, 124, 2639
355: \bibitem[\protect\citeauthoryear{{Reid}}{{Reid}}{1991}]{Reid91}
356: {Reid}, N. 1991, AJ, 102, 1428
357: \bibitem[\protect\citeauthoryear{{Vallenari}, {Bertelli}, {Schmidtobreick}}{{Vallenari} et al.}{2000}]{Vallenari00}
358: {Vallenari}, A., {Bertelli}, G., {Schmidtobreick}, L. 2000, AJ, 102, 1428
359: \bibitem[\protect\citeauthoryear{{Smith}, {Mao}, \& {Paczy\'nski}}{{Smith}, et~al.}{2003}]{Smith03}
360: {Smith}, M., {Mao}, S., \& {Paczy\'nski}, B. 2003, MNRAS, 339, 925
361:
362: \end{thebibliography}
363:
364: \begin{deluxetable}{llll}
365: %\vspace*{6cm}
366: %\hspace*{4cm}
367: %\vskip{5cm}
368: \tablecaption{Astrometric data for LMC-5.\label{tab1}}
369: %\footnotesize
370: %\small
371: \tablewidth{0pt}
372: %\tablewidth{275pt}
373: \tablehead{\colhead{Observations} & \colhead{$\Delta \alpha$} & \colhead{$\Delta \delta$} & \colhead{Time}\\
374: \colhead{} & \colhead{$\arcsec$} & \colhead{$\arcsec$} & \colhead{days}}
375: \startdata
376: WFPC2 (F555W \& F814W) & $0.1110\pm0.0038$ & $-0.0748\pm0.0026$ & 2288.16\nl
377: ACS HRC (F606W) & $0.16772\pm0.00021$ & $-0.11263\pm0.00024$ & 3442.55\nl
378: ACS HRC (F814W) & $0.16618\pm0.00055$ & $-0.11192\pm0.00023$ & 3442.64\nl
379: ACS HRC (F606W) & $0.17484\pm0.00037$ & $-0.12183\pm0.00028$ & 3621.52
380: \enddata
381: \tablecomments{
382: Col. (1), instrument and filters used in observations.
383: Cols. (2) \& (3), relative offsets between source and lens in
384: right ascension (great-circle) and declination, respectively.
385: Col. (4), observation time relative to peak time of lensing event
386: (JD=2449023.9).
387: }
388: \end{deluxetable}
389:
390: \end{document}
391:
392:
393:
394:
395:
396:
397:
398: