astro-ph0507287/ms.tex
1: %% \documentclass[12pt,preprint,myplots,psfig]{aastex}
2: %\documentclass[12pt,preprint]{aastex}
3: %%\documentclass[apjl]{emulateapj}
4: %% \documentclass[manuscript]{aastex}
5: \documentclass[preprint2,myplots,psfig]{aastex}
6: %% \documentclass[preprint2,longabstract]{aastex}
7: %% using \newcommand. Your macros should appear before
8: 
9: \newcommand{\vdag}{(v)^\dagger}
10: 
11: \usepackage{graphics}
12: 
13: \slugcomment{ApJL, submitted}
14: 
15: \shorttitle{M31 Microlensing Confirmation}
16: \shortauthors{Cseresnjes et al.}
17: 
18: \def\ga{\lower 2pt\hbox{$\, \buildrel{\scriptstyle>}\over{\scriptstyle\sim}\,$}}
19: \def\la{\lower 2pt\hbox{$\, \buildrel{\scriptstyle<}\over{\scriptstyle\sim}\,$}}
20: 
21: \begin{document}
22: 
23: \title{
24: {\it HST} Imaging of MEGA Microlensing Candidates in M31\footnote{
25: Based on observations made 1) with the NASA/ESA $HST$, obtained at STScI, which
26: is operated by AURA, Inc., under NASA contract NAS 5-26555. These observations
27: are associated with program \#10273,
28: 2) with the INT operated on La Palma by the ING in the Spanish ORM of the IAC,
29: 3) at KPNO, NOAO, which is operated by AURA, Inc., under cooperative agreement
30: with the NSF.
31: }}
32: 
33: \author{Patrick Cseresnjes\footnote{Columbia Astrophysics Laboratory, Columbia
34: University, 550 W.~120th St., New York, NY~~10027},
35: Arlin P.S.~Crotts,$^2$
36: Jelte T.A.~de Jong\footnote{Kapteyn Astronomical Institute, University of
37: Groningen, PO Box 800, 9700 AV, Groningen, The Netherlands},~
38: Alex Bergier,$^2$
39: Edward A.\ Baltz\footnote{Kavli Inst.~for Particle Astrophysics \&
40: Cosmology, Stanford U., PO Box 20450, MS 29, Stanford, CA~~94309},
41: Geza Gyuk\footnote{Department of Astronomy and Astrophysics, University of
42: Chicago, 5640 S.\ Ellis Ave., Chicago, IL~~60637},
43: Konrad Kuijken,\footnote{Sterrewacht Leiden, University of Leiden, PO Box 9513,2300 RA, Leiden, The Netherlands}$^{~,3}$
44: Lawrence M.~Widrow\footnote{Department of Physics, Queens University, Kingston,
45: ON~~K7L 3N6, Canada}}
46: 
47: \begin{abstract}
48: We investigate $HST$/ACS and WFPC2 images at the positions of
49: five candidate microlensing events from a large survey of variability
50: in M31 (MEGA).  Three closely match unresolved sources, and two produce only
51: flux upper limits.  All are confined to regions of the
52: color-magnitude diagram where stellar variability is unlikely to
53: be easily confused with microlensing.
54: Red variable stars cannot explain these events (although background supernova
55: are possible for two).
56: If these lenses arise in M31's halo, they are due to
57: masses $0.15 < m / M_\odot < 0.49$ (95\% certainty, for a $\delta$-function
58: mass distribution), brown dwarfs for disk lenses, and stellar
59: masses for bulge lenses.
60: \end{abstract}
61: 
62: \keywords{gravitational lensing --- galaxies: individual (M31) ---
63: galaxies: halos --- dark matter}
64: 
65: \section{Introduction}
66: 
67: Galaxian dark matter has been recognized 
68: for over 70 years (Zwicky 1933), and tied in part to the
69: halo for over 30 (Rubin \& Ford 1970).
70: Halo dark matter's nature is still a mystery.
71: Gravitational microlensing can reveal individual
72: dark matter objects of roughly stellar mass (Paczy\'nski 1986).
73: To test this, MACHO observed the Magellanic Clouds for 5.7 years, (Alcock et
74: al.~2000) and EROS (Afonso et al.~2003) did so for 5.
75: The former report microlensing events more common than the
76: known, purely stellar expectation, with lensing fraction $f \approx
77: 20$\% of the dark matter halo mass (8-50\%, with 95\% confidence) of
78: $\sim$0.4 $M_\odot$ masses.
79: EROS found $f$ consistent with zero (but marginally consistent with $f \approx
80: 20$\%).
81: 
82: M31 microlensing 
83: could potentially settle this quandary definitively (Crotts 1992).
84: Since we can explore microlensing across the face of M31, we can use
85: this distribution to distinguish where in the galaxy the lenses arise.
86: Several surveys of M31 microlensing 
87: (Riffeser et al.~2003, Joshi et al.~2004, Calchi-Novati et
88: al.~2005, including MEGA: de Jong et al.~2004 and its predecessor
89: VATT-Columbia: Uglesich et al.~2004),
90: together report $\sim$20 probable microlensing events,
91: and have a tendency to confirm the MACHO result.
92: 
93: With its crowded target stars, M31 microlensing relies
94: on image subtraction to reveal event lightcurves, which removes the baseline
95: flux.
96: Using $HST$ to recover the source flux (e.g. Ansari et al.\ 1999, 
97: Auri\`ere et al.\ 2001), one can compute event amplification,
98: hence Einstein parameters, constraining physical parameters e.g., lens mass.
99: MEGA and VATT-Columbia also use
100: source star color to distinguish microlensing from variable stars, since
101: very red variables (miras and
102: semiregulars) produce outbursts that, with their baselines
103: subtracted, mimic point-source, point-lens (``Paczy\'nski'') light curves
104: (Uglesich et al.~2004).
105: Residual flux from these events, however, is redder than
106: almost all potential microlensing source stars.
107: MEGA will soon publish its microlensing sample, and now is an excellent
108: opportunity to check these event selection criteria.
109: 
110: \begin{figure*}[]
111: \includegraphics[width=16.7cm,height=14.75cm]{f1c.eps}
112: \caption{Left to right: 1) $(1\arcsec .5)^2$ $HST$ 
113: image (f814w band) around microlensing candidates. The circles correspond
114: to $1\sigma$ spread of the individual position estimates. The grid 
115: represents the INT 
116: pixel sampling; for ML-16, the cross point corresponds to an independent
117: estimate as described in the text. 2) full combined light curve 
118: (Filled squares: KP-R, asterisks: KP-I, open circles: INT-r', 
119: open triangles: INT-i');
120: 3) Zoom on the event peak; 4) Lens-mass probability distribution 
121: for a lens in an isothermal halo (full line), in the disk (dashed line), 
122: and in the bulge (dotted line).}
123: \label{fig1}
124: \end{figure*}
125: 
126: \section {Observations and analysis}
127: 
128: \begin{figure}[]
129: \vskip -0.15in
130: \includegraphics[width=8cm,height=8cm]{f2c.eps}
131: \vskip -0.3in
132: \caption{Color-magnitude plot of 5 candidate sources, with
133: upper flux limits as thick lines. The dashed lines enclose the area where
134: LPVs and semi-regulars are expected \cite{brown}.}
135: \label{fig2}
136: \end{figure}
137: 
138: To study candidate events we appeal to superior $HST$ angular resolution:
139: 160 ACS and WFPC2 images, taken in F555W and F814W filters in 16 orbits,
140: cover 0.17 deg$^2$, $\sim$30\% of the MEGA field.
141: Here we study the largest sample of candidate microlensing events, the INT/WFC
142: subsample of MEGA (de Jong et al.~2005), in order to understand and improve
143: ground-based selection criteria.
144: 
145: The analysis used is detailed by Cseresnjes et al.\ (in preparation).
146: We carefully align the $HST$ and ground-based images by matching catalogs of
147: ground-based versus Gaussian-convolved $HST$ sources
148: for each filter combination ($HST$, F555W/F814W versus
149: INT, $r^{\prime}/i^{\prime}$ and KPNO 4m, R/I), providing up to 8
150: different position estimates.
151: For a given ground-based position, the two independent $HST$ positions 
152: (via F555W and F814W) always agree to $\le 0\arcsec .03$ 
153: (typically $0\arcsec.01$), so
154: positional accuracy depends mainly on ground-based data.  
155: The adopted position is a weighted average
156: of individual estimates (Fig. \ref{fig1}). The spread of different estimates
157: for each event are $0\arcsec .02$ to $0\arcsec .08$. Of five microlensing
158: candidates analyzed, we identify three sources and
159: find flux upper limits for two.
160: 
161: $HST$ data were photometered with DAOPHOT (Stetson 1987), as prescribed in 
162: Sirianni et al.\ (2005).
163: The locations of the candidate microlensing sources on a color-magnitude
164: diagram are shown in Fig. \ref{fig2}.
165: For each candidate event, we normalized the differential light curves
166: to R-fluxes, using color-magnitude diagrams and $HST$ baseline fluxes
167: (for the two undetected events, using the baseline flux upper limit),
168: then performed a Paczynski fit in ($u_{0},t_{0},t_{E}$)
169: to the combined light curve (Fig. \ref{fig1}). 
170: For the two undetected events,
171: the resulting $t_{_E}$ corresponds to a lower limit.
172: 
173: Only the Einstein time-scale ($t_{_E}$) constrains
174: lens characteristics, particularly its mass $m$. 
175: For a given time-scale
176: $t_{E}$, the lens-mass probability distribution is 
177: $P(M,t_{E})=({\rm d}\Gamma/{\rm d}t_{E})/\Gamma$, 
178: where $\Gamma$ is the event rate \cite{griest}.
179: We consider alternatively a lens located in the halo of M31, 
180: in the disk or in the bulge. 
181: For halo lensing, 
182: we consider the simple case where the lens is part of a
183: spherical isothermal halo composed of single mass objects, with a density
184: distribution defined as 
185: \begin{equation}
186: \rho_{h} \propto \frac{1}{R^{2}+r_{c}^{2}}
187: \end{equation}
188: where $R$ is the radial distance to the center of M31, and $r_{c}$
189: is a core radius of $5$ kpc. The one-dimensional velocity dispersion
190: of the lenses is set to $170$ km~s$^{-1}$, consistent with a rotation curve
191: of $240$ km~s$^{-1}$. The disk is modelled by a double exponential with 
192: a radial scale length of
193: $6$ kpc and a vertical scale length of $400$ pc. The bulge corresponds to
194: the ``small'' bulge model of Kent (1989), with a velocity dispersion
195: of $150$ km~s$^{-1}$. More details about the model can be found in
196: Baltz, Gyuk, \& Crotts (2003).
197: 
198: \section{Individual events}
199: 
200: \begin{deluxetable}{lccccc}
201: \tabletypesize{\scriptsize}
202: \tablecaption{Event source photometry and Microlensing parameters}
203: \tablewidth{0pt}
204: \tablehead{
205: \colhead{id}         & 
206: \colhead{ML-8}       & 
207: \colhead{ML-10}      & 
208: \colhead{ML-11 (S4)} &
209: \colhead{ML-16 (N1)} &
210: \colhead{ML-18}
211: }
212: \startdata
213: %
214: R.A. (J2000)         &
215: 00:43:24.53          &
216: 00:43:54.87          &
217: 00:42:29.90          &
218: 00:42:51.22          &
219: 00:43:17.27         \\
220: %
221: Declination (J2000)  &
222: 41:37:50.4           &
223: 41:10:33.3           &
224: 40:53:45.6           &
225: 41:23:55.3           &
226: 41:02:13.7          \\
227: %
228: R$_{hst}$            & 
229: $24.94\pm 0.14$      & 
230: $23.36\pm 0.09$      & 
231: $24.86\pm 0.30$      & 
232: $>23.86$             & 
233: $>25.09$            \\
234: %
235: I$_{hst}$            & 
236: $24.34\pm 0.08$      & 
237: $22.31\pm 0.07$      & 
238: $24.71\pm 0.26$      & 
239: $>$23.32             & 
240: $>$24.59            \\
241: %
242: (R-I)$|_{hst}$      & 
243: 0.60                & 
244: 1.05                & 
245: 0.15                & 
246: ~~--~~              & 
247: ~~--~~              \\
248: %
249: (R-I)$|_{lc}$       & 
250: 0.59                & 
251: 1.05                & 
252: 0.21                & 
253: ~~--~~              & 
254: 0.51                \\
255: %
256: $A_{\rm max}$       & 
257: 8.49                & 
258: 4.00                & 
259: 41.93               & 
260: $>$16.01            &  
261: $>$11.42           \\
262: %
263: $t_{_E}$/day        & 
264: $60.6\pm 4.2$       & 
265: $64.7\pm 1.9$       & 
266: $26.1\pm 1.1$       & 
267: $>$6.9              & 
268: $>$86.6            \\
269: %
270: $\chi^{2}/N$        & 
271: 0.89                & 
272: 1.26                & 
273: 1.01                & 
274: 1.29                & 
275: 1.04               \\
276: %
277: $m_{\rm halo}/M_\odot$\tablenotemark{a} &
278: $0.31^{+0.48}_{-0.21}$ &
279: $0.33^{+1.04}_{-0.23}$ &
280: $0.05^{+0.16}_{-0.03}$ & 
281: $>0.00$                & 
282: $>0.62$               \\
283: %
284: $m_{\rm disk}/M_\odot$\tablenotemark{a}  &
285: $0.05^{+1.65}_{-0.03}$ &
286: $0.05^{+0.09}_{-0.04}$ &
287: $0.01^{+0.02}_{-0.01}$ & 
288: $>0.03$                & 
289: $>0.09$               \\
290: %
291: $m_{\rm bulge}/M_\odot$\tablenotemark{a} &
292: $2.85^{+4.03}_{-2.31}$  &
293: $2.00^{+1.66}_{-1.55}$  &
294: $0.36^{+0.31}_{-0.28}$  & 
295: $>0.05$                 & 
296: $>3.73$               \\
297: %
298: comment                & 
299: red clump or SN        & 
300: giant branch           & 
301: very blue              & 
302: undetected(?)          & 
303: in cluster or galaxy  \\
304: 
305: \enddata
306: \tablenotetext{a}{most likely lens mass (with 1$\sigma$ confidence intervals)}
307: \label{ml_table}
308: \end{deluxetable}
309: 
310: {\it ML-8:}
311: this event's position lands within the FWHM of a red clump star's image, 
312: with $R-I$ in 
313: excellent agreement with the peak flux's color in
314: differential light curves ($0.60 \pm 0.16$ vs.\ $0.59$ mag).
315: With the baseline set
316: to this star's flux, a Paczynski fit yields amplification $A = 8.49$,
317: an Einstein time-scale $t_{_E} = 60.6 \pm 4.2$ days. The corresponding
318: lens masses are $m = 0.31^{+0.48}_{-0.21} M_\odot$ for a halo lens, 
319: $m = 0.05^{+1.65}_{-0.03} M_\odot$ for a disk lens, and 
320: $m = 2.85^{+4.03}_{-2.31} M_\odot$ for a bulge lens.\\
321: \indent However, this event lands $\sim0^{\prime\prime}.9$ from the center 
322: of a background
323: galaxy (subtending $\sim1^{\prime\prime}.5 \times 0^{\prime\prime}.3$).
324: Its color, flux and decline rate are consistent with a Type Ia supernova at
325: $z \approx 0.5$, with $\la 1$ mag extinction (see Johnson \& Crotts 2005).
326: One must balance the number of SNe ($\sim 100$ y$^{-1}$ deg$^{-2}$ e.g., Woods
327: \& Loeb 1998) landing within the FWHM disk of a source star of consistent 
328: color we would detect
329: (1.2 arcsec$^{-2}$) versus the number of microlensing events (evidently
330: $\sim$5) landing so close to an $R<23$ galaxy (100 arcmin$^{-2}$ --- Huang et
331: al.\ 2001).
332: The expected number of both kind of events are of the order of a few tenths, 
333: with perhaps microlensing being slightly more likely.
334: 
335: {\it ML-10:}
336: this event lands within the FWHM disk of a giant branch
337: star of color ${\rm R}-{\rm I}=1.05$, in perfect agreement with the
338: microlensing data.
339: This source has 
340: $A = 4.00$ and 
341: $t_{_E} = 64.7 \pm 1.9$ days, corresponding
342: to a halo lens mass $m=0.33^{+1.04}_{-0.23} M_\odot$, a disk lens mass
343: $m=0.05^{+0.09}_{-0.04} M_\odot$, or a bulge lens mass
344: $m=2.00^{+1.66}_{-1.55} M_\odot$.
345: It lands suspiciously close to
346: a region of the CMD common to variables. Still, 
347: the achromaticity of the variation, the well-fit and
348: well-sampled peak ($\chi^{2}/N=1.26$), and the stability of the
349: baseline over $7$ seasons strongly indicate a real microlensing event.
350:  
351: {\it ML-11:}
352: this event lands on a faint blue star 
353: (${\rm R-I}=0.15\pm 0.40$) severely blended with a red clump star.
354: The light curves fit yields a similar ${\rm R-I}=0.21$.
355: Its baseline flux implies $A=41.93$ and $t_{_E}=26.1\pm 1.1$ days. 
356: This event, from Paulin-Henriksson et al.\ (2002), is
357: near M32, suggesting that the lens resides there.
358: If not, the most likely lens mass is $m=0.05^{+0.16}_{-0.03} M_\odot$ in
359: the halo, $m=0.01^{+0.02}_{-0.01} M_\odot$ in the disk, and
360: $m=0.36^{+0.31}_{-0.28} M_\odot$ in the bulge.
361: 
362: {\it ML-16:}
363: this event lands 
364: in a WFPC2 field and was also seen by POINT-AGAPE (Auri\`ere et al. 2001).
365: They publish a color for the event peak based on INT g$^\prime$ and 
366: r$^\prime$ (no i$^\prime$ data are available), corresponding to
367: V-I~$\sim 2.1$.
368: We find no detected source at this position; the nearest detected star 
369: landing $\sim 0\arcsec .1$ away (one WFPC2 pixel) with ${\rm V-I} \simeq 1.1$.
370: Using the flux of this star as an upper limit, we find $A>16.01$, 
371: $t_{_E}>6.9$ days,
372: and $m > 0.003 M_\odot$ being poorly constrained for a halo lens.\\
373: \indent Auri\`ere et al.\ (2001) seem to have isolated a different source
374: star. Their celestial coordinates disagree with ours by $3\arcsec$, but
375: cannot be checked from published data since no image of the source field
376: is provided. In order to check our astrometry, we repeated the procedure
377: we applied to r$^\prime$ data to g$^\prime$ data retrieved from the INT 
378: archive. We also repeated the same procedure (for both r$^\prime$ and
379: g$^\prime$ data) using the WFPC2 images taken by Auri\`ere et al 
380: from the archive. Finally, 
381: one of us (A.C.) made an independent check by choosing $8$ bright 
382: unsaturated, isolated stars as coordinate inputs to IRAF {\it geomap}
383: to construct the coordinate transform. These various estimates agree to
384: better than $0.5$ pixel in the INT data, whereas the nearest star (two
385: of them actually) with consistent colors and magnitudes to that claimed
386: by Auri\`ere et al. are $0\arcsec.6$ (or almost two INT pixels) away. Our
387: position is marginally consistent with the faint source cited above and in 
388: Table 1, but inconsistent with that of Auri\`ere et al. 
389: 
390: {\it ML-18:}
391: this event lands in a bright region, perhaps a
392: cluster or background galaxy.
393: We isolate no source here, 
394: so provide only an upper limit baseline flux, estimated by taking the 
395: brightest pixel within $0^{\arcsec}.05$ and considering that it contains
396: at most $15\%$ of the source flux, as constrained by the point-spread
397: function for ACS.
398: With this flux limit, $A>11.42$, $t_{_E}>86.6$ days, and 
399: $m\ga 0.62\ M_{\odot}$ for a halo lens, $m\ga 0.09\ M_{\odot}$ for 
400: a disk lens, and
401: $m\ga 3.73\ M_{\odot}$ for a bulge lens.
402: 
403: \section {Conclusions}
404: 
405: Of five events in our fields, we find three likely matches, and baseline flux
406: upper limits on the other two.
407: Colors of the three identified sources agree with those obtained
408: from their differential light curves alone.
409: The two upper limits displace these events from the asymtotic giant branch,
410: where confusing mira and semiregular variables can occur.
411: No candidate is a bright red variable.
412: One might interpret ML-8 as a supernova, but a microlensing
413: event is just as probable.
414: We also cannot rule out a supernova as the source for ML-18, which might also
415: coincide with a background galaxy.
416: In a future paper, the complete MEGA data set will fill out ML-18's light
417: curve; unfortunately, we have no additional data on ML-8.
418: 
419: Taking the product of the individual mass probability distributions obtained
420: for each event, these lenses in 
421: a halo model (Baltz, Gyuk \& Crotts 2003)
422: of a single component mass are constrained to $0.15 < m/M_{\odot} <
423: 0.49$ at the 95\% level.
424: M31 microlensing rates may be consistent with pure
425: self-lensing (de Jong et al.\ 2005), so we consider bulge lenses 
426: ($0.64 < m/M_{\odot} < 2.02$), or disk lenses which correspond to 
427: probably unrealistic brown dwarf masses ($0.02 < m/M_{\odot} < 0.06$).
428: 
429: We acknowledge support from STScI (GO 10273) and NSF (grants
430: 0406970 and 0070882).
431: 
432: \begin{thebibliography}{}
433: \bibitem[]{321}Afonso, C.~et al.~2003, A\&A, 400, 951
434: \bibitem[]{322}Alcock, C.~et al.~2000, ApJ, 542, 281
435: \bibitem[Ansari et al. 1999]{ansari}Ansari, R., et al., 1999, A\&A, 344, L49
436: \bibitem[Auri\`ere et al. 2001]{auriere}Auri\`ere, M., et al., 2001, ApJL, 553, L137
437: \bibitem[Baltz et al. 2003]{baltz}Baltz, E.A., Gyuk, G.\ \& Crotts, A.\ 2003, ApJ, 582, 30
438: \bibitem[Brown et al. 2004]{brown} Brown, T., et al., 2004, AJ, 127, 2738
439: \bibitem[]{325}Calchi Novati, S.~et al.~2005, preprint
440: \bibitem[]{326}Crotts, A.~1992, ApJ, 399, L43
441: \bibitem[Griest 1991]{griest} Griest, K., 1991, ApJ, 366, 412
442: \bibitem[]{328}de Jong, J.T.A.~et al.~2004, A\&A, 417, 461
443: \bibitem[]{329}de Jong, J.T.A.~et al.~2005, A\&A, in press
444: \bibitem[Huang et al. 2001]{huang}Huang, J.-S., et al., 2001, A\&A, 368, 787
445: \bibitem[]{330}Johnson, B.~\& Crotts, A.~2005, AJ, in press
446: \bibitem[]{331}Joshi, Y.C., Pandey, A.K., Narasimha \& Sagar, R.~2004, astro-ph/0412550
447: \bibitem[]{332}Paczy\'nski, B.~1986, 302, 1
448: \bibitem[Paulin-Henriksson et al. 2002]{paulin} Paulin-Henriksson, S., et al., 2002, ApJL, 576, 121
449: \bibitem[]{333}Riffeser, A., Fliri, J., Bender, R., Seitz, S.~\& G\"ossl, C.A.~2003, ApJ, 599, L17
450: \bibitem[]{334}Rubin, V.C.~\& Ford, W.K.~1970, ApJ, 159, 379
451: \bibitem[]{335}Sirianni, M., et al., 2005, PASP, in press
452: \bibitem[]{336}Stetson, P., 1987, PASP, 99, 191
453: \bibitem[]{337}Uglesich, R., Crotts, A.P.S., Baltz, E., de Jong, J., Boyle, R.~\& Corbally, C.~2004, ApJ, 612, 877
454: \bibitem[Woods \& Loeb 1998]{woods}Woods, E. \& Loeb, A., 1998, ApJ, 508, 760
455: \bibitem[Zwicky 1933]{zwicky}Zwicky, F., 1933, {\it Helvetica Physica Acta}, 6, 110
456: \end{thebibliography}{}
457: 
458: \end{document}
459: