0707.4211/ms.tex
1: % For ApJ submission -----------------------------------------
2: %\documentclass[12pt,preprint]{aastex}
3: %\def\baselinestretch{1.3}
4: 
5: \documentclass{emulateapj}
6: \usepackage{apjfonts}
7: \lefthead{HAN} 
8: \righthead{CHANNELS OF PLANETARY LENSING}
9: 
10: %==== CUSTOMIZED LATEX MACROS ========================================
11: 
12: \newcommand{\vect}[1]{\ensuremath{\mbox{\boldmath $#1$}}}
13: 
14: 
15: % Equation align
16: \def\eqalign#1{\null\,\vcenter{\openup\jot
17:         \ialign{\strut\hfil$\displaystyle{##}$&$
18:         \displaystyle{{}##}$\hfil \crcr#1\crcr}}\,}
19: 
20: %=======================================================================
21: 
22: 
23: \begin{document}
24: \title{Expansion of the Planet Detection Channels in Next-Generation 
25: Microlensing Surveys}
26: 
27: 
28: \author{Cheongho Han}
29: \affil{Program of Brain Korea 21, Department of Physics, Institute 
30: for Basic Science Research, Chungbuk National University, Chongju 
31: 361-763, Korea;\\
32: cheongho@astroph.chungbuk.ac.kr}
33: 
34: 
35: 
36: % ==================================================================
37: 
38: %\submitted{Submitted to The Astrophysical Journal}
39: 
40: \begin{abstract}
41: We classify various types of planetary lensing signals and the 
42: channels of detecting them.  We estimate the relative frequencies 
43: of planet detections through the individual channels with special 
44: emphasis on the new channels to be additionally provided by future 
45: lensing experiments that will survey wide fields continuously at 
46: high cadence by using very large-format imaging cameras.  From this 
47: investigation, we find that the fraction of wide-separation planets 
48: that would be discovered through the new channels of detecting 
49: planetary signals as independent and repeating events would be 
50: substantial.  We estimate that the fraction of planets detectable 
51: through the new channels would comprise $\sim 15$ -- 30\% of all 
52: planets depending on the models of the planetary separation 
53: distribution and mass ratios of planets.  Considering that a 
54: significant fraction of planets might exist in the form of free-floating 
55: planets, the frequency of planets to be detected through the new channel 
56: would be even higher.  With the expansion of the channels of detecting 
57: planet, future lensing surveys will greatly expand the range of planets 
58: to be probed.
59: \end{abstract}
60: 
61: \keywords{gravitational lensing -- planets and satellites: general}
62: 
63: 
64: 
65: \section{Introduction}
66: 
67: Microlensing is one of the important techniques that can detect and 
68: characterize extrasolar planets.  The technique is important especially 
69: in the detections of low-mass planets and it is possible to detect 
70: Earth-mass planets from ground-based observations.  The capability of 
71: the microlensing technique has been demonstrated by the recent detections 
72: of four planets \citep{bond04, udalski05, beaulieu06, gould06}.  Among 
73: them, one (OGLE-2005-BLG-390Lb) is the lowest mass planet ever detected
74: among those orbiting normal stars.
75: 
76: 
77: Planetary lensing signal lasts a short period of time; several days 
78: for a Jupiter-mass planet and several hours for an Earth-mass planet.  
79: To achieve the observational frequency required for the detection of 
80: the short-lived planetary signal, current microlensing planet search 
81: experiments are being operated in an observational setup, where survey 
82: observations (e.g., OGLE: \citet{udalski03}, MOA: \citet{bond02a}) issue 
83: alerts of ongoing events and subsequent follow-up observations (e.g., 
84: PLANET: \citet{albrow01}, MicroFUN: \citet{dong06}) intensively monitor 
85: the alerted events.  Under this strategy, however, only planetary signals 
86: occurring during the lensing magnification of the source star can be 
87: effectively monitored.  These signals are produced by planets having 
88: projected separations from the star similar to the Einstein radius of 
89: the primary star.  As a result, only planets located in a narrow region 
90: of separation from the host star can be effectively detected under the 
91: current planetary lensing searches.
92: 
93: 
94: The range of planets detectable with the microlensing technique will 
95: be expanded with next-generation lensing experiments that will 
96: survey wide fields continuously at high cadence by using very 
97: large-format imaging cameras.  Several such surveys in space and on 
98: ground are being seriously considered.  {\it Microlensing Planet 
99: Finder} ({\it MPF}), that succeeds the original concept of {\it 
100: Galactic Exoplanet Survey Telescope} ({\it GEST}) \citep{bennett02}, 
101: is a space mission proposed to NASA's Discovery Program with the main 
102: goal of searching for a large sample of extrasolar planets by using 
103: gravitation lensing technique \citep{bennett04}.  The `Earth-Hunter' 
104: project is a ground-based microlensing survey that plans to achieve 
105: $\sim 10$ minute sampling by using a distributed network of multiple 
106: wide-field ($\sim 2^\circ\times2^\circ$) telescopes (A.\ Gould, private 
107: communication). Recently, the MOA group begin a strategy observing a 
108: fraction of their fields very frequently by using a recently 
109: upgraded 1.8 m telescope with a $2.2\ {\rm deg}^2$ field of view.
110: These surveys dispense with the alert/follow-up mode of searching for 
111: planets and instead simultaneously obtain densely and continuously 
112: sampled light curves of all microlensing events in the field-of-view.  
113: With these surveys, the efficiency of planet detection will greatly 
114: improve thanks to the enhanced monitoring frequency and continuous 
115: sampling.  In addition, the future lensing experiments will be able 
116: to open new channels of planet detections because source stars are 
117: monitored regardless of their magnifications.  In this paper, we 
118: classify various channels of detecting planetary lensing signals 
119: and estimate the relative frequencies of planet detections through the 
120: individual channels with special emphasis on the new channels that 
121: will be provided by future lensing surveys.
122: 
123: 
124: The format of the paper is as follows.  In \S\ 2, we describe basics 
125: of planetary microlensing.  In \S\ 3, we present various types of 
126: planetary lensing perturbations and classify the channels of detecting 
127: them.  In \S\ 4, we estimate the relative frequencies of detecting 
128: planets through the individual channels.   We discuss the importance
129: of the new planet detection channels to be provided by future lensing 
130: surveys.  We summarize the results and conclude in \S\ 5.
131: 
132: 
133: 
134: % ==========================================================================
135: 
136: \section{Basics of Planetary Microlensing}
137: 
138: 
139: 
140: 
141: Due to the existence of a planetary companion, description of the  
142: planetary lensing behavior requires formalism of binary lensing 
143: \citep{witt90, witt95}.  Because of the small mass ratio of the 
144: planet, the light curve of a planetary lensing event is well described 
145: by that of a single lens of the primary star for most of the event 
146: duration.  However, a short-duration perturbation can occur when the 
147: source star passes the region around the caustics. The caustics are 
148: important features of binary lensing and they represent the set of 
149: source positions at which the magnification of a point source becomes 
150: infinite.  The caustics of binary lensing form a single or multiple 
151: sets of closed curves where each of which is composed of concave 
152: curves (fold caustics) that meet at points (cusps).  For a planetary 
153: case, there exist two sets of disconnected caustics: `central' and 
154: `planetary' caustics. 
155: 
156: 
157: 
158: The single central caustic is located close to the host star.  It 
159: has a wedge shape with four cusps, where two are located on the 
160: star-planet axis and the other two are located off the axis.  The 
161: size of the central caustic as measured by the separation between 
162: the on-axis cusps is represented by \citep{chung06}
163: \begin{equation}
164: \Delta\xi_{\rm cc} \sim {4q\over (s-s^{-1})^2},
165: \label{eq1}
166: \end{equation}
167: where $q$ is the planet/star mass ratio and $s$ is the star-planet 
168: separation normalized by the Einstein radius of the planetary lens 
169: system, $\theta_{\rm E}$.  Then the caustic size
170: %, which is directly 
171: %proportional to the cross-section of the planetary perturbation region, 
172: becomes maximum when $s\sim 1.0$.  In the limiting case of a very 
173: wide-separation planet ($s\gg 1.0$) and a close-in planet ($s\ll 1.0$), 
174: the caustic size decreases, respectively, as
175: \begin{equation}
176: \Delta\xi_{\rm cc} \propto
177: \cases{
178: s^{-2}  & for $s \gg 1.0$, \cr
179: s^2     & for $s \ll 1.0$.\cr }
180: \label{eq2}
181: \end{equation}
182: For a given mass ratio, a pair of central caustics with separations 
183: $s$ and $s^{-1}$ are identical to the first order of the approximation
184: where the planet-induced anomalies is treated as a perturbation 
185: \citep{dominik99, griest98,an05}.  The central caustic is always 
186: smaller than the planetary caustic.  Since the central caustic is 
187: located close to the primary lens, the perturbation induced by the 
188: central caustic always occurs near the peak of high-magnification 
189: events.
190: 
191: 
192: The planetary caustic is located away from the host star.  The center 
193: of the planetary caustic is located on the star-planet axis and the 
194: position vector to the center of the planetary caustic from the primary 
195: lens position is related to the lens-source separation vector, ${\bf s}$, 
196: by 
197: \begin{equation}
198: {\bf r}_{\rm pc}={\bf s}\left(1-{1 \over s^2}\right).
199: \label{eq3}
200: \end{equation}
201: Then, the planetary caustic is located on the planet side, i.e.\ 
202: ${\rm sign} ({\bf r}_{\rm pc})= {\rm sign}({\bf s})$, when $s>1.0$, and 
203: on the opposite side, i.e.\ ${\rm sign} ({\bf r}_{\rm pc})=-{\rm sign}
204: ({\bf s})$, when $s<1.0$.  When $s>1.0$, there exists a single planetary 
205: caustic and it has a diamond shape with four cusps.  When $s<1$, there 
206: are two caustics and each has a triangular shape with three cusps.  
207: The size of the planetary caustic is related to the planet parameters by
208: \begin{equation}
209: \Delta\xi_{\rm pc} \propto 
210: \cases{
211: q^{1/2}\left[ s (s^2-1)^{1/2} \right]^{-1}             & for $s > 1$,\cr
212: q^{1/2}(\kappa_0-\kappa_0^{-1}+\kappa_0s^{-2})\cos\theta_0  & for $s < 1$,\cr
213: }
214: \label{eq4}
215: \end{equation}
216: where $\kappa(\theta)=\left\{[\cos 2\theta\pm (s^4-\sin^2 2\theta)^{1/2}]
217: /(s^2-s^{-2})\right\}^{1/2}$, $\theta_0 = [\pi \pm \sin^{-1}(3^{1/2}s^2/2)]
218: /2$, and $\kappa_0=\kappa(\theta_0)$ \citep{han06}.  We note that the 
219: dependence of the planetary caustic size on the mass ratio is $\Delta 
220: \xi_{\rm pc}\propto q^{1/2}$, while the dependence of the central caustic 
221: size is $\Delta \xi_{\rm cc}\propto q$.  Therefore, the decay rate of 
222: the planetary caustic with the decrease of the planet mass is slower 
223: than that of the central caustic.  The planetary caustic is located 
224: within the Einstein ring of the primary star when the planet is located 
225: in the range of separation from the star of $0.6 \lesssim s \lesssim 1.6$.  
226: The size of the caustic is maximized when the planet is located in this 
227: range, and thus this range is called as the `lensing zone' \citep{gould92, 
228: griest98}.  In the limiting cases of planetary separation, the size of 
229: the planetary caustic deceases similar to the central caustic, i.e.
230: \begin{equation}
231: \Delta\xi_{\rm pc} \propto
232: \cases{
233: s^{-2}  & for $s \gg 1.0$, \cr
234: s^2     & for $s \ll 1.0$.\cr }
235: \label{eq5}
236: \end{equation}
237: In the limiting case of $s\ll 1.0$, the two lens components work as if 
238: they are a single lens.  In the limiting case of $s\gg 1.0$, on the other 
239: hand, the star and planet work as if they are two independent lenses. 
240: 
241: 
242: 
243: 
244: % ==========================================================================
245: 
246: 
247: % Figure 1 
248: \begin{figure}[t]
249: \epsscale{1.2}
250: \plotone{f1.eps}
251: \caption{\label{fig:one}
252: Schematic representation of the planetary signals with various types.
253: For the details of the classification scheme, see \S\ 3.
254: }\end{figure}
255: 
256: 
257: \section{Classification of Planet Detection Channels} 
258: 
259: Planetary lensing signal takes various forms depending on the 
260: characteristics of the planetary system, especially on the star-planet 
261: separation, and the source trajectory with respect to the positions 
262: of the star and planet.  In this section, we classify various types 
263: of planetary lensing signals and the channels of detecting them.
264: 
265: 
266: Type I perturbation shows up as a perturbation to the smooth light 
267: curve of the primary-induced lensing event (see the top panel in 
268: Figure~\ref{fig:one}).  This type of perturbation is produced by 
269: planets with projected star-planet separations similar to the Einstein 
270: radius of the primary star, i.e.\ $s\sim 1.0$.  As a result, the 
271: channel of detecting planets through the detection of type I perturbation 
272: is often referred as the `resonant channel'.  The resonant channel is 
273: the prime channel of detecting planets in current planetary lensing 
274: searches, which are based on survey/follow-up mode.  Depending on 
275: whether the perturbation is induced by the planetary or central 
276: caustic, the perturbation takes place on the side or near the peak 
277: of the light curve.
278: 
279: 
280: Type II perturbation is produced by a planet with a projected separation 
281: from the primary star substantially larger than the Einstein radius of 
282: the primary star ($s\gg 1.0$) and it occurs when the source trajectory 
283: passes both the effective magnification regions of the primary star 
284: and planet.  The planetary signal is the planet-induced lensing light 
285: curve itself that is well separated from the light curve of the primary 
286: (see the second panel in Figure~\ref{fig:one}).  Since two successive 
287: events are produced by the star and planet, respectively, this channel 
288: is often referred as the `repeating channel' \citep{distefano99}.  The 
289: planetary signal occurs long after or before the event induced by the 
290: primary star, and thus type II perturbation is difficult to be detected 
291: by the current planetary lensing searches monitoring only during the 
292: time of primary-induced lensing magnification.
293: 
294: 
295: Type III perturbation is produced also by a wide-separation planet,
296: but it occurs when the source trajectory passes only the effective 
297: magnification region of the planet.  Then, the planetary signal is the
298: independent lensing light curve produced by the planet itself (see the 
299: third panel in Figure~\ref{fig:one}).  We refer this channel of planet 
300: detection as the `independent channel'.  
301: %Another population of planets 
302: %that can be detected through the independent channel are free-floating 
303: %planets \citep{bennett02, han04}.  
304: Type III perturbation is also difficult 
305: to be detected by the current planetary lensing searches.
306: 
307: 
308: % Figure 2 
309: \begin{figure*}[ht]
310: \epsscale{0.80}
311: %\begin{figure}[t]
312: %\epsscale{1.25}
313: \plotone{f2.eps}
314: \caption{\label{fig:two}
315: Maps of planetary signal detectability as a function of the source 
316: position for planetary systems with various projected star-planet 
317: separation.  The detectability is quantified as the ratio of the 
318: fractional deviation of the planetary lensing light curve from that 
319: of the single lensing event of the primary star to the photometric 
320: precision.  The maps are centered at the position of the primary lens 
321: star and the planet is located on the left.  The green (green curve) 
322: is drawn at the level of $D=3$.  The solid circle centered at the 
323: primary star in each map represents the Einstein ring.  The closed 
324: figures drawn by red curves are the caustics.  All lengths are in 
325: units of the Einstein radius of the planetary system and $\xi$ and 
326: $\eta$ represent the coordinates that are parallel and normal to the 
327: star-planet axis, respectively.  Each planetary perturbation region 
328: is approximated by an elliptical region and the green ellipse represents 
329: its boundary.  The planetary lens system has a common planet/star mass 
330: ratio of $q=3\times 10^{-3}$.  
331: }\end{figure*}
332: %}\end{figure}
333: 
334: 
335: 
336: 
337: Wide-separation planets can also be detected through another channel.
338: Type IV perturbation is produced by a wide-separation planet and it 
339: occurs when the source trajectory passes very close to the primary 
340: star.  The planetary signal is then a brief perturbation near the peak 
341: of the high-magnification lensing light curve produced by the primary 
342: star (see the fourth panel in Figure~\ref{fig:one}).  We refer this 
343: channel as the `wide central channel'.  High-magnification events 
344: are of highest-priority in the current microlensing follow-up 
345: observations because of their high sensitivity to planets
346: \citep{griest98, albrow00, bond02b, rattenbury02, abe04, yoo04, 
347: dong06} and thus type IV perturbation can be detected from current 
348: follow-up observations.
349: 
350: 
351: Type IV perturbation is produced by a planet with a star-planet 
352: separation substantially smaller than the Einstein radius of the 
353: primary star and it occurs when the source trajectory passes very 
354: close to the primary star.  The planetary signal is very similar to 
355: the type IV perturbation (see the fourth panel in Figure~\ref{fig:one}) 
356: due to the $s\leftrightarrow s^{-1}$ symmetry of the central caustic 
357: \citep{dominik99, an05}.  As a result, distinguishing the two types 
358: of perturbation is often difficult, especially for perturbations 
359: induced by low-mass planets with mass ratios $q\lesssim 10^{-4}$ 
360: \citep{chung06}.
361: 
362: 
363: 
364: % ==========================================================================
365: 
366: 
367: 
368: \section{Frequency of the Individual Channels}
369: 
370: In the previous section, we presented various types of planetary 
371: perturbations and the channels of detecting them.  Then, what would 
372: be the relative frequencies of detecting planets through the individual 
373: channels from future lensing surveys?  In this section, we answer to 
374: this question.
375: 
376: 
377: The rate of planet detection is proportional to the cross-section of 
378: the planetary perturbation region, $\sigma$.  We, therefore, estimate 
379: the relative frequencies by computing the cross-sections of the planetary 
380: perturbation regions of the corresponding types.  We proceed our 
381: computation according to the following procedure.
382: \begin{enumerate}
383: \item
384: First, we make maps of perturbation induced by planets with various 
385: separations and mass ratios.
386: \item
387: Second, we classify the types of perturbation based on the planetary 
388: separation and the location of the perturbation regions.  We then 
389: estimate the average cross-sections of the perturbation regions belonging 
390: to the individual categories based on the constructed perturbation maps.  
391: \item
392: Finally, we compute the relative frequencies of detecting planets through 
393: the individual channels by convolving the cross-sections with the model 
394: distributions of planetary separation.  
395: \end{enumerate}
396: The details of the individual processes are described in the following
397: subsections.
398: 
399: 
400: 
401: \subsection{Maps of Planetary Signal Detectability}
402: 
403: The map of planetary perturbation represents the region of planet-induced 
404: lensing perturbations as a function of source position.  The quantity 
405: that has been often used to represent the perturbation region is the 
406: `fractional deviation' of the planetary lensing light curve from that 
407: of the single lensing event of the primary star, i.e.,
408: \begin{equation}
409: \epsilon = {A-A_0 \over A_0},
410: \label{eq6}
411: \end{equation}
412: where $A$ and $A_0$ represent the lensing magnifications with and 
413: without the planet, respectively.  With this quantity, however, one
414: cannot consider the variation of the photometric precision depending 
415: on the lensing magnification and source brightness.  To consider this,
416: we construct the map of perturbation with the quantity 
417: defined as the ratio of the fractional deviation, $\epsilon$, to the 
418: photometric precision, $\sigma_\nu$, i.e, 
419: \begin{equation}
420: {\cal D}={\left\vert \epsilon\right\vert \over \sigma_\nu};\qquad
421: \sigma_\nu = { ( AF_{\nu,{\rm S}}+F_{\nu,{\rm B}})^{1/2} 
422: \over (A-1)F_{\nu,{\rm S}}},
423: \label{eq7}
424: \end{equation}
425: where $F_{\nu,{\rm S}}$ and $F_{\nu,{\rm B}}$ represent the photon 
426: counts from the source star and blended background stars, respectively.  
427: Under this definition of the planetary perturbation, ${\cal D}=1$ 
428: implies that the planetary signal is equivalent to the photometric 
429: precision.  Hereafter, we refer the quantity ${\cal D}$ as the 
430: `detectability'.
431: 
432: 
433: To construct the map of detectability, we choose a representative 
434: Galactic bulge event.  Following the result of simulations of Galactic 
435: bulge events, e.g.\ \citet{han03, han05}, we choose a representative 
436: event as the one produced by a lens with the primary lens mass of 
437: $m=0.3\ M_\odot$ and the distances to the lens and source of $D_{\rm L}=6$ 
438: kpc and $D_{\rm S}=8$ kpc, respectively.  For the observational condition, 
439: we take the space-based lensing survey by using the {\it MPF} mission as 
440: a reference experiment.  The prime target source stars to be monitored 
441: by the {\it MPF} survey are main-sequence stars and thus we choose a 
442: main-sequence source star with an $I$-band absolute magnitude of $M_I=4.8$, 
443: which corresponds to a K0 star.  With the assumed amount of extinction 
444: toward the Galactic bulge field of $A_I=1.0$, this correspond to the 
445: apparent magnitude of $I=20.3$.  Following the specification of the 
446: {\it MPF} mission, we assume that the photon acquisition rate is 13 
447: photons per second for an $I=22$ star and photometry is done on each 
448: combined image with an exposure time $t_{\rm exp}=10$ minutes.  We 
449: assume that blending is not important due to high resolution from 
450: space-based observation.  Finite size of the source star might affect 
451: the planet detectability \citep{bennett96}.  However, the populations 
452: of planets of our interest are the ones to be detected through the new 
453: channels (i.e., repeating and independent channels) and most of them 
454: would be giant planets because of their larger cross-sections.
455: Considering that the angular size of the source star is a few percent 
456: of the angular Einstein radius of a giant planet, finite size of the 
457: source star has little effect on the planet detectability.  We therefore 
458: do not consider finite-source effect in our analysis.
459: 
460: 
461: Figure~\ref{fig:two} shows constructed maps of detectability for some 
462: example planetary systems.  The maps are centered at the position of 
463: the primary lens star and the planet is located on the left.  The 
464: contour (green curve) is drawn at the level of $D=3$, within which the 
465: planetary signal is detected with a $3\sigma$ confidence level.  The 
466: solid circle centered at the primary star in each map represents the 
467: Einstein ring and the close figure drawn by red curves represent the 
468: caustics.  All lengths are normalized in units of the Einstein radius 
469: corresponding to the mass of the primary star, $\theta_{\rm E}$.  
470: 
471: 
472: % Figure 3 
473: \begin{figure}[t]
474: \epsscale{1.20}
475: \plotone{f3.eps}
476: \caption{\label{fig:three}
477: Cross-section of planetary perturbation region as a function of the
478: normalized star-planet separation.  The segments marked by different 
479: tones of shade under the curve represent the types of the related
480: planetary perturbations.  We note that the cross-section is normalized 
481: by the angular Einstein radius of the planet.  The assumed planet/star 
482: mass ratio is $q=3\times 10^{-3}$.
483: }\end{figure}
484: 
485: 
486: 
487: \subsection{Cross-Section of Perturbation}
488: 
489: With the constructed maps of perturbation, we then estimate the 
490: average cross-sections of the perturbation regions of the individual 
491: types.  For this, we classify the types of perturbation based on the 
492: planetary separation and the location of the perturbation region.
493: 
494: 
495: The followings are the criteria for the classification.  First, we 
496: classify all perturbations induced by planets located in the lensing 
497: zone ($0.6 \leq s \le 1.6$) into type I.  If a perturbation is induced 
498: by a wide-separation ($s>1.6$) planet, the perturbation region is 
499: divided into two parts; one around the planetary caustic and the other 
500: around the central caustic.  If the perturbation is caused by the 
501: central caustic, it is classified into type IV perturbation.  Among 
502: the perturbations caused by the planetary caustics of wide-separation 
503: planets, the fraction of the type II perturbation is geometrically 
504: estimated as $\sin(2/\pi) \sin^{-1}( u_{\rm th}/s)$, where $u_{\rm th}$ 
505: is the radius of the effective lensing magnification region of the 
506: primary star.  We adopt $u_{\rm th}=1.5$.  Then the rest of perturbations 
507: induced by the planetary caustics of wide-separation planets are 
508: classified into type III.  Finally, perturbations induced by planets 
509: with separations $s<0.6$ are classified into type V.
510: 
511: 
512: Once the types of the individual perturbation regions are determined, 
513: we then estimate the average cross-sections of the perturbation regions.  
514: A straightforward approach to estimating the cross-section would be 
515: first drawing many light curves resulting from source trajectories with 
516: various combinations of the distance to the trajectory from the center 
517: of the perturbation region and orientation angles, then checking the 
518: detectability of the planet-induced perturbations for the individual 
519: light curves, and finally estimating the cross-section as an angle-averaged 
520: value.  However, this requires a large amount of computation time.  
521: Fortunately, the perturbation region is confined around caustics and 
522: its boundary is approximated by an ellipse.  We therefore estimate the 
523: cross-section by approximating the perturbation region as an elliptical 
524: region.  With this approximation, the cross-section of the perturbation 
525: region is estimated as the angle-averaged cross-section of the ellipse, 
526: i.e.\ 
527: \begin{equation}
528: \langle\sigma\rangle = {1\over \pi}\int_0^\pi 
529: \left( a_{\rm p}^2\sin^2 \theta + b_{\rm p}^2\cos^2\theta \right)^{1/2} 
530: d\theta = {2\over \pi} a_{\rm p} E(e),
531: \label{eq8}
532: \end{equation}
533: where $a_{\rm p}$ and $b_{\rm p}$ are the semimajor and semiminor axes 
534: of the elliptical boundary of the perturbation region, $e=(1-b_{\rm p}^2
535: /a_{\rm p}^2)^{1/2}$ is the eccentricity of the ellipse, and $E$ 
536: represents the complete elliptical integral of the second kind.  We 
537: determine the semimajor and semiminor axes of the ellipse as the widths 
538: of the perturbation region enclosed by the detectability contour with a 
539: level of $D=3$ along and normal to the star-planet axis, respectively.  
540: If the perturbation region is composed of multiple segments, we approximate 
541: the individual segments with different ellipses.  In Figure 2, we present 
542: the elliptical boundaries of perturbation regions (green curves) on the 
543: top of the detectability map.
544: 
545: 
546: Figure~\ref{fig:three} shows the determined cross-section of the 
547: planetary perturbation region as a function of the normalized star-planet 
548: separation for a planetary lens with a mass ratio $q=3\times 10^{-3}$, 
549: which corresponds to a Jupiter-mass planet around a primary star with a
550: mass $m=0.3\ M_\odot$.  We note that the cross-section is normalized by 
551: the angular Einstein radius corresponding to the mass of the planet, 
552: $\theta_{\rm E,p}=q^{1/2} \theta_{\rm E}$.  The segments marked by 
553: different tones of shade under the curve represent the types of the 
554: related perturbations.  From the figure, one finds that the cross-section 
555: vanishes in the limiting case of $s\rightarrow 0$.  This is because 
556: the star and planet work as if they are a single lens in this limit.  
557: In the limiting case of the other end ($s\rightarrow \infty$), on the 
558: other hand, the planet acts as an independent lens and thus the 
559: cross-section converges into the value corresponding to the cross-section 
560: of the effective magnification region of the planet.  
561: 
562: 
563: 
564: 
565: \begin{deluxetable*}{ccrrrrr}
566: \tablecaption{Relative Frequency\label{table:one}}
567: \tablewidth{0pt}
568: \tablehead{
569: \multicolumn{1}{c}{planetary separation} &
570: \multicolumn{1}{c}{planet/star} &
571: \multicolumn{5}{c}{types of planetary perturbation} \\
572: \multicolumn{1}{c}{distribution model} &
573: \multicolumn{1}{c}{mass ratio} &
574: \multicolumn{1}{c}{type I} &
575: \multicolumn{1}{c}{type II} &
576: \multicolumn{1}{c}{type III} &
577: \multicolumn{1}{c}{type IV} &
578: \multicolumn{1}{c}{type V} 
579: }
580: \startdata
581: $\alpha=1.0$ & $q=5.0\times 10^{-3}$ & 48.3\%  & 3.8\%  & 26.5\%  & 12.8\%  &  8.6\% \\
582:              & $q=1.0\times 10^{-3}$ & 56.8\%  & 4.1\%  & 25.9\%  &  6.9\%  &  6.2\% \\
583:              & $q=5.0\times 10^{-4}$ & 59.4\%  & 4.2\%  & 25.6\%  &  5.4\%  &  5.3\% \\
584: \smallskip
585:              & $q=1.0\times 10^{-4}$ & 64.3\%  & 4.0\%  & 24.3\%  &  3.7\%  &  3.7\% \\
586: 
587: $\alpha=1.5$ & $q=5.0\times 10^{-3}$ & 58.2\%  & 2.5\%  & 12.7\%  & 10.0\%  & 16.6\% \\
588:              & $q=1.0\times 10^{-3}$ & 67.2\%  & 2.9\%  & 12.7\%  &  5.3\%  & 11.9\% \\
589:              & $q=5.0\times 10^{-4}$ & 69.9\%  & 3.0\%  & 12.7\%  &  4.2\%  & 10.0\% \\
590:              & $q=1.0\times 10^{-4}$ & 75.0\%  & 2.8\%  & 12.1\%  &  2.9\%  &  7.2\% 
591: 
592: %$\alpha=2.0$ & $5.0\times 10^{-3}$ & 58.0\%  & 1.4\%  &  5.6\%  &  6.8\%  & 28.2\% \\
593: %             & $1.0\times 10^{-3}$ & 68.0\%  & 1.7\%  &  6.0\%  &  3.6\%  & 20.6\% \\
594: %             & $5.0\times 10^{-4}$ & 71.4\%  & 1.8\%  &  6.1\%  &  2.8\%  & 17.8\% \\
595: %             & $1.0\times 10^{-4}$ & 77.1\%  & 1.8\%  &  6.1\%  &  2.0\%  & 13.0\% 
596: \enddata
597: \tablecomments{
598: Relative frequencies of detecting planets through various channels.
599: The frequencies are estimated under the assumption that the star-planet 
600: separation follows a power-law distribution of $dN/da \propto a^{-\alpha}$,
601: where $\alpha$ represents the power of the distribution.
602: }
603: \end{deluxetable*}
604: 
605: 
606: 
607: \subsection{Relative Frequencies}
608: 
609: Once the average cross-sections of the individual types of perturbation
610: is computed, we then estimate the relative frequencies of detecting planets
611: through the individual channels of planet detections.  This is done 
612: by convolving the cross-section with model distributions of planetary 
613: separation.
614: 
615: 
616: We model the distribution of star-planet separation as a power-law
617: function of the form
618: \begin{equation}
619: {dN\over da} \propto a^{-\alpha},
620: \label{eq9}
621: \end{equation}
622: where $a$ is the semimajor axis of the planet orbit.  There is little 
623: consensus about the power of the distribution.  From the analysis of 
624: observed extrasolar planets detected by radial velocity surveys, 
625: \citet{tabachnik02} claimed $\alpha\sim 1$.  On the other hand, 
626: \citet{hayashi95} claimed that the surface density distribution of the 
627: minimum mass solar nebula is well described with $\alpha\sim 1.5$.  
628: %\citet{kuchner04} argued that multi-planet extrasolar planetary 
629: %systems indicate $\alpha\sim 2.0$.  
630: We, therefore, test two different powers of $\alpha=1.0$ and 1.5.
631: We note that larger absolute value of the power implies that planets 
632: are populated in the inner region.  Then the fraction of planetary 
633: events detectable through the type I perturbation increases while the 
634: fractions through the type II, III, and IV decrease.  We assume that 
635: planets are distributed up to a distance of 100 AU.  
636: %We note that the 
637: %fraction of planets detectable through the independent channel depends 
638: %critically on the choice of this limit.  
639: Once the semimajor axis of 
640: the planet orbit is determined, the projected star-planet separation 
641: is determined under the assumption of a circular orbit and random 
642: orientation of the orbital plane.  The projected separation is related 
643: to the intrinsic separation by
644: \begin{equation}
645: \tilde{a} = a(\sin^2 i \cos^2 \varphi + \cos^2 i)^{1/2},
646: \label{eq10}
647: \end{equation}
648: where $i$ is the inclination angle of the orbital plane and $\varphi$ 
649: is the phase of the planet on the orbital plane.
650: 
651: 
652: In Table~\ref{table:one}, we present the relative frequencies of 
653: detecting planets through the individual channels.  From the table, we 
654: find that the frequency of detecting planets through the new channels 
655: to be provided by future lensing surveys would be substantial.  We 
656: estimate that the fraction of planets detectable through the independent 
657: and repeating channels would comprise $\sim 15$ -- 30\% of all planets 
658: depending on the models of the planetary separation distribution and 
659: mass ratios of planets.  Considering that the total number of planets 
660: expected to be detected from five-year lensing surveys in space would 
661: be several thousands \citep{bennett04}, the number of planets detectable 
662: through the new channels would be of the order of hundred and can reach 
663: up to a thousand.  We note that the estimation in Table~\ref{table:one} 
664: is based only on planets bound to primary stars.  
665: 
666: 
667: The new channels to be provided by future lensing surveys are important
668: for better understanding of planet formation and evolution processes.
669: Planets located $\gtrsim 5$ AU from host stars could not have been  
670: detected by any of the methods currently being used for planet 
671: searches.  Being able to detect planets in this range, therefore,
672: microlensing method would provide complete sample of planets.  Another 
673: population of planets that can be detected through the new channels 
674: are free-floating planets \citep{bennett02, han04}.  It is believed 
675: that a good fraction of planets have been ejected from their planetary 
676: systems during or after the epoch of planet formation \citep{zinnecker01}.
677: Another possible origin of these planets would be the accretion of gas 
678: similar to star formation process \citep{boss01}.  Since these planets 
679: were not included in our analysis, the relative frequency of planet 
680: detection through the new channels would be even larger if these planets 
681: are common.
682: 
683: 
684: 
685: \section{Conclusion}
686: 
687: We classified various types of planetary lensing signals and the 
688: channels of detecting them.  We estimated the relative frequencies 
689: of planet detections through the individual channels with special 
690: emphasis on the new channels that will be additionally provided by 
691: future lensing surveys.  From this investigation, we found that the 
692: fraction of wide-separation planets that would be discovered through 
693: the new channels of detecting planetary signals as independent and 
694: repeating events would be substantial.  We estimated that the fraction 
695: of planets detectable through the new channels would comprise $\sim 15$ 
696: -- 30\% of all planets depending on the models of the planetary separation 
697: distribution and mass ratios of planets.  Considering that a significant 
698: fraction of planets might exist in the form of free-floating planets, 
699: the frequency of planets to be detected through the new channel would 
700: be even higher.  We, therefore, demonstrate that future lensing surveys 
701: will greatly expand the range of planets to be probed.
702: 
703: 
704: \acknowledgments 
705: This work was supported by the grant (C00072) of the Korea Research
706: Foundation.
707: 
708: 
709: 
710: \begin{thebibliography}{99}
711: \frenchspacing
712: 
713: \bibitem[Abe et al.(2004)]{abe04}
714: Abe, F., et al.\ 2004, Science, 305, 1264
715: 
716: \bibitem[Albrow et al.(2000)]{albrow00}
717: Albrow, M.\ D., et al.\ 2000, \apj, 535, 176
718: 
719: \bibitem[Albrow et al.(2001)]{albrow01}
720: Albrow, M.\ D., et al.\ 2001, \apj, 556, L113
721: 
722: \bibitem[An(2005)]{an05}
723: An, J.\ H.\ 2005, \mnras, 356, 1409
724: 
725: \bibitem[Beaulieu et al.(2006)]{beaulieu06}
726: Beaulieu, J.\ P., et al.\ 2006, Nature, 439, 437
727: 
728: \bibitem[Bennett(2004)]{bennett04}
729: Bennett, D.\ P.\ 2004, BAAS, 205, 11.26
730: 
731: \bibitem[Bennett \& Rhie(1996)]{bennett96}
732: Bennett, D. P., \& Rhie, S.\ H.\ 1996, \apj, 472, 660
733: 
734: \bibitem[Bennett \& Rhie(2002)]{bennett02}
735: Bennett, D.\ P., \& Rhie, S.\ H.\ 2002, \apj, 574, 985
736: 
737: \bibitem[Bond et al.(2002)]{bond02a}
738: Bond, I.\ A., et al.\ 2002, \mnras, 331, L19
739: 
740: \bibitem[Bond et al.(2002)]{bond02b}
741: Bond, I.\ A., et al.\ 2002, \mnras, 333, 71
742: 
743: \bibitem[Bond et al.(2004)]{bond04}
744: Bond, I.A., et al.\ 2004, \apj, 606, L155
745: 
746: \bibitem[Boss(2001)]{boss01}
747: Boss, A.\ P.\ 2001, \apj, 551, L167
748: 
749: \bibitem[Chung et al.(2006)]{chung06}
750: Chung, S.-J., et al.\ 2006, \apj, 650, 432
751: 
752: \bibitem[Di Stefano \& Scalzo(1999)]{distefano99}
753: Di Stefano, R., \& Scalzo, R.\ A.\ 1999, \apj, 512, 579
754: 
755: \bibitem[Dominik(1999)]{dominik99}
756: Dominik, M.\ 1999, \aap, 349, 108
757: 
758: \bibitem[Dong et al.(2006)]{dong06}
759: Dong, S., et al.\ 2006, \apj, 642, 842
760: 
761: \bibitem[Gould \& Loeb(1992)]{gould92}
762: Gould, A., \& Loeb, A.\ 1992, \apj, 396, 104
763: 
764: \bibitem[Gould et al.(2006)]{gould06}
765: Gould, A., et al.\ 2006, \apj, 644, L37
766: 
767: \bibitem[Griest \& Safizadeh(1998)]{griest98}
768: Griest, K., \& Safizadeh, N.\ 1998, \apj, 500, 37
769: 
770: \bibitem[Han(2004)]{han04}
771: Han, C.\ 2004, \apj, 604, 372
772: 
773: \bibitem[Han(2006)]{han06}
774: Han, C.\ 2006, \apj, 638, 1080
775: 
776: \bibitem[Han \& Gould(2003)]{han03}
777: Han, C., \& Gould, A.\ 2003, \apj, 592, 172
778: 
779: \bibitem[Han \& Gould(2005)]{han05}
780: Han, C., \& Gould, A.\ 2005, \apj, 447, 53
781: 
782: \bibitem[Hayashi(1995)]{hayashi95}
783: Hayashi, S.\ S.\ 1995, Ap\&SS, 224, 479
784: 
785: %\bibitem[Kuchner(2004)]{kuchner04}
786: %Kuchner, M.\ J.\ 2004, \apj, 612, 1147
787: 
788: \bibitem[Rattenbury et al.(2002)]{rattenbury02}
789: Rattenbury, N.\ J., Bond, I.\ A., Skuljan, J., \& Yock, P.\ C.\ M.\ 
790: 2002, \mnras, 335, 159
791: 
792: \bibitem[Tabachnik \& Tremaine(2002)]{tabachnik02}
793: Tabachnik, S., \& Tremaine, S.\ 2002, \mnras, 335, 151
794: 
795: \bibitem[Udalski(2003)]{udalski03}
796: Udalski, A.\ 2003, Acta Astronomica, 53, 291
797: 
798: \bibitem[Udalski et al.(2005)]{udalski05}
799: Udalski, A., et al.\ 2005, \apj, 628, L109
800: 
801: \bibitem[Witt(1990)]{witt90}
802: Witt, H.\ J.\ 1990, \aap, 236, 311
803: 
804: \bibitem[Witt \& Mao(1995)]{witt95}
805: Witt, H.\ J., \& Mao, S.\ 1995, \apj, 447, L105
806: 
807: \bibitem[Yoo et al.(2004)]{yoo04}
808: Yoo, J., et al.\ 2004, \apj, 616, 1204
809: 
810: \bibitem[Zinnecker(2001)]{zinnecker01}
811: Zinnecker, H.\ 2001, Microlensing 2000: A New Era of Microlensing 
812: Astrophysics, ASP Conference Proceedings, eds. J.\ W.\ Menzies \& 
813: P.\ D.\ Sackett, San Francisco, 239, 223
814: 
815: \end{thebibliography}
816: 
817: \end{document}
818: