1: \documentclass[12pt,preprint]{aastex}
2:
3: \def\sun{\odot}
4: \def\earth{\oplus}
5: \begin{document}
6: \title{Formation of Kuiper Belt Binaries}
7: \shortauthors{Schlichting & Sari}
8: \shorttitle{Formation of Kuiper Belt Binaries }
9: \author{Hilke E. Schlichting and Re'em Sari}
10: \affil{California Institute of Technology, MC 130-33, Pasadena, CA 91125}
11: \email{hes@astro.caltech.edu, sari@tapir.caltech.edu}
12:
13: \begin{abstract} The discovery that a substantial fraction of Kuiper Belt
14: objects (KBOs) exists in binaries with wide separations and roughly equal
15: masses, has motivated a variety of new theories explaining their
16: formation. \citet{GLS02} proposed two formation scenarios: In the first, a
17: transient binary is formed, which becomes bound with the aid of dynamical
18: friction from the sea of small bodies ($L^2s$ mechanism); in the second, a
19: binary is formed by three body gravitational deflection ($L^3$
20: mechanism). Here, we accurately calculate the $L^2s$ and $L^3$ formation rates
21: for sub-Hill velocities. While the $L^2s$ formation rate is close to previous
22: order of magnitude estimates, the $L^3$ formation rate is about a factor of 4
23: smaller. For sub-Hill KBO velocities ($v \ll v_H$) the ratio of the $ L^3 $ to
24: the $ L^2s $ formation rate is $ 0.05 (v/v_H)$ independent of the small
25: bodies' velocity dispersion, their surface density or their mutual
26: collisions. For Super-Hill velocities ($v \gg v_H$) the $L^3$ mechanism
27: dominates over the $L^2s$ mechanism. Binary formation via the $L^3$ mechanism
28: competes with binary destruction by passing bodies. Given sufficient time, a
29: statistical equilibrium abundance of binaries forms.
30:
31: We show that the frequency of long-lived transient binaries drops
32: exponentially with the system's lifetime and that such transient
33: binaries are not important for binary formation via the $L^3$ mechanism,
34: contrary to \citet{A07}. For the $L^2s$ mechanism we find that the typical
35: time, transient binaries must last, to form Kuiper Belt binaries (KBBs) for a
36: given strength of dynamical friction, $D$, increases only logarithmically with
37: $D$. Longevity of transient binaries (with lifetimes $\geqslant 15
38: \Omega^{-1}$ as suggested by \citet{A05}) only becomes important for very weak
39: dynamical friction (i.e. $D \lesssim 0.002$) and is most likely not crucial
40: for KBB formation.
41: \end{abstract}
42:
43: \keywords {Kuiper Belt --- planets and satellites: formation}
44:
45: \section{INTRODUCTION} One of the many intriguing discoveries in the Kuiper
46: Belt is that a substantial fraction of its largest members are binaries. 48
47: such systems are currently known (for a comprehensive review see
48: \citet{N07}). Broadly speaking, we can identify two classes of Kuiper Belt
49: binaries (KBBs). The first class consists of small satellites around the
50: largest Kuiper Belt objects (KBOs) and the second of roughly equal-mass
51: binaries with wide separations. The existence of the first class of binaries
52: is most likely explained by the standard formation scenario involving a
53: collision and tidal evolution, as has been proposed for the formation of the
54: Moon and the Pluto-Charon system \citep{HD75,CW76,MK89}. This formation
55: scenario fails however for the second class of KBBs, since it cannot account
56: for their wide separations. This has motivated a variety of new theories for
57: the formation of comparable mass KBBs
58: \citep[e.g.][]{W02,GLS02,F04,A05,A07}. \citet{W02} proposed a new formation
59: mechanism for KBBs consisting of a collision between two bodies inside the
60: Hill sphere of a third. However, in the Kuiper Belt, gravitational scattering
61: between the two intruders is about 100 times\footnote{For this estimate we
62: used $\alpha\sim 10^{-4}$ and assumed that the velocity dispersion of the KBOs
63: at the time of binary formation is less than their Hill velocity, see \S 2 for
64: details} more common than a collision. Binary formation by three body
65: gravitational deflection ($L^3$ mechanism), as proposed by \citet{GLS02},
66: should therefore dominate over such a collisional formation
67: scenario. \citet{GLS02} proposed a second binary formation scenario: it
68: consists of the formation of a transient binary, which becomes bound with the
69: aid of dynamical friction from the sea of small bodies. This is called the
70: $L^2s$ mechanism. \cite{A05} and \cite{A07} suggest that transient binaries
71: that spend a long time in their mutual Hill sphere, near a periodic orbit,
72: form the binaries in the $L^2s$ and $L^3$ mechanism. We address and
73: investigate the relative importance of these long-lived transient binaries for
74: the $L^2s$ and $L^3$ formation mechanism and find that they are most likely
75: not significant for the overall binary formation in the Kuiper Belt. Finally,
76: \citet{F04} proposed a binary formation mechanism which involves a collision
77: between two large KBOs which creates a small moon. An exchange reaction
78: replaces the moon with a massive body with high eccentricity and large
79: semi-major axis.
80:
81: In this paper, we accurately calculate the $L^2s$ and $L^3$ formation rates
82: for sub-Hill KBO velocities and discuss how these rates are modified for
83: super-Hill velocities. This allows us to determine for which physical
84: parameters and velocity regime each mechanism dominates the binary
85: formation. Further, we calculate the frequency of long-lived transient
86: binaries and assess their importance for the overall KBB formation.
87:
88: Our paper is structured as follows: In \S 2 we outline our assumptions,
89: explain our choice of parameters and define variables that will be used
90: throughout this paper. We calculate the $L^3$ and $L^2s$ formation rates for
91: sub-Hill KBO velocities in \S 3 and \S 4 respectively. We compare the $L^2s$
92: and $L^3$ formation rates in the sub-Hill velocity regime in \S 5. In \S 6 we
93: discuss how these formation rates are modified for super-Hill KBO
94: velocities. The frequency of long-lived transient binaries and their
95: significance for the overall KBB formation is calculated in \S 7. Summary and
96: conclusions follow in \S 8.
97:
98:
99: \section{DEFINITIONS AND ASSUMPTIONS}
100: The Hill radius denotes the distance from a KBO at which the tidal forces due
101: to the Sun and the gravitational force due to the KBO, both acting on a test
102: particle, are in equilibrium. It is given by
103: \begin{equation}\label{e1}
104: R_{H} \equiv a \left( \frac{M}{3 M_{\sun}}\right) ^{1/3}
105: \end{equation}
106: where $a$ is the semi-major axis and $M$ the mass of the KBO. $M_{\sun}$ is
107: the mass of the sun. We use the `two-group approximation' \citep{GLS02,GLS04}
108: which consists of the identification of two groups of objects, small ones,
109: that contain most of the total mass with surface mass density $\sigma$, and
110: large ones, that contain only a small fraction of the total mass with surface
111: mass density $\Sigma \ll \sigma $. We assume $\sigma \sim 0.3 \rm{g~cm^{-2}}$
112: which is the extrapolation of the minimum-mass solar nebular to a heliocentric
113: distance of $40\rm{AU}$. Estimates from current Kuiper Belt surveys
114: \citep{TB03,TJL01} yield $\Sigma \sim 3 \times 10^{-4} \rm{g~cm^{-2}}$ for
115: KBOs with radii of $R \sim 100~\rm{km}$. We use this value of $\Sigma$,
116: assuming that $\Sigma$ during the formation of KBBs was the same as it is
117: now. Our choice for $\Sigma$ and $\sigma$ is also consistent with results from
118: numerical coagulation simulations by \citet{KL99}.
119:
120: Large bodies grow by the accretion of small bodies. Large KBOs viscously stir
121: the small bodies, increasing the small bodies' velocity dispersion $u$. As a
122: result $u$ grows on the same timescale as $R$ provided that mutual collisions
123: among the small bodies are not yet important. In this case, $u$ is given by
124: \begin{equation}\label{e16}
125: \frac{u}{v_H} \sim \left( \frac{\Sigma}{\sigma \alpha} \right)^{1/2} \sim 3
126: \end{equation}
127: where $\alpha = R/R_{H}\sim 10^{-4}$ at $40\rm{AU}$ \citep{GLS02}. $v_H$ is
128: the Hill velocity of the large bodies which is given by $v_H = \Omega R_H$
129: where $\Omega$ is the orbital frequency around the sun. The velocity $v$ of
130: large KBOs increases due to mutual viscous stirring, but is damped by
131: dynamical friction from the sea of small bodies such that $v < u$. Balancing
132: the stirring and damping rates of $v$ and substituting for $u$ from equation
133: \ref{e16}, we find
134: \begin{equation}\label{e17}
135: \frac{v}{v_H} \sim \alpha^{-2} \left(\frac{\Sigma}{\sigma}\right)^{3} \sim
136: 0.1.
137: \end{equation}
138: For our choice of parameters, we have that $v < v_{H}$ during the epoch of
139: formation of bodies with $R\sim 100\rm{km}$. Additionally, we argue that $v$
140: could not have exceeded $v_{H}$ significantly during satellite formation in
141: the Kuiper Belt: If $v_{esc} > v > v_{H}$, where $v_{esc}$ is the escape
142: velocity from the large bodies, then the timescale for mutual collisions is
143: \begin{eqnarray}\label{e22}
144: {\tau_{coll}\sim 0.13 \left(\frac {\Sigma}{3 \times 10^{-4}
145: \rm{g~cm^{-2}}}\right)^{-1} \left(\frac{\rho}{1\rm{g~cm^{-3}}}\right)
146: \left(\frac{R}{100\rm{km}}\right)\left(\frac{\alpha}{1
147: \times 10^{-4}}\right)\left(\frac{v}{v_{H}}\right)^{2} {} } \nonumber\\
148: {} \left(\frac{\Omega}{7.9 \times
149: 10^{-10}\rm{s^{-1}}}\right)^{-1} \rm{Gyr} {}.
150: \end{eqnarray}
151: Equation \ref{e22} shows that the collision timescale among the largest KBOs
152: ($R>100\rm{km}$) would have been excessively long if $v \gg v_{H}$ during
153: satellite formation. The ubiquity of small satellites around KBOs, that have
154: radii as large as $\sim 1000\rm{km}$, \citep{BDB06,BS07} and the Pluto-Charon
155: system \citep{WSM06} suggests that $v < v_{H}$ during their formation, since
156: their origin is best explained by a giant impact
157: \citep[e.g.][]{SWS06,BBRS07}. This is supported further by the recent
158: discovery of a collisional family belonging to $\rm{EL}_{61}$
159: \citep{BBRS07}. We therefore focus our work on the shear-dominated velocity
160: regime ($v<v_{H}$). However, we discuss how our results would be modified if
161: $v>v_H$.
162:
163:
164: \section{$L^3$ FORMATION RATE}
165: A transient binary forms when two large KBOs penetrate each other's Hill
166: sphere. This transient binary must lose energy in order to become
167: gravitationally bound. In the $L^3$ mechanism the excess energy is carried
168: away by an encounter with a third massive body. We calculate the binary
169: formation rate via the $L^3$ mechanism in the shear-dominated velocity
170: regime. Since the growth of inclinations is suppressed in the shear-dominated
171: velocity regime the disk of KBOs is effectively two-dimensional
172: \citep{WS1993,R2003,GLS04}. We therefore restrict this calculation to two
173: dimensions. As initial condition, we assume that all bodies are on circular
174: orbits. We chose to work in the rotating frame with the $x$-axis pointing
175: radially outwards and the $y$-axis in the prograde direction. For a
176: gravitational deflection of three equal-mass bodies, the $L^3$ formation rate
177: per body is
178: \begin{equation}\label{e19}
179: FR_{L^3}=\int_{\gamma=-\infty}^{\infty} \int_{b_2>b_1}^{\infty}
180: \int_{b_1=0}^{\infty}\left(\frac{\Sigma}{\frac{4 \pi}{3}\rho R^{3}}\right)^{2}
181: \frac{3}{2} b_1 \Omega F_{L^3}(b_1,b_2,\gamma) db_1 db_2 d\gamma.
182: \end{equation}
183: $\Sigma/(4\pi \rho R^3/3)$ is the surface number density of the KBOs. $b_1$
184: and $b_2$ are the relative initial separations in the $x$-direction between
185: bodies 1 and 2 and bodies 1 and 3 respectively. $\gamma$ is the offset in the
186: $y$-direction body 3 would have when bodies 1 and 2 would encounter each other
187: had their relative velocity been solely due to the Kepler shear of the disk:
188: $3 b_1 \Omega /2$. Finally, $F_{L^3}(b_1,b_2,\gamma)$ is a function that takes
189: on the value 1 if the encounter resulted in the formation of a binary between
190: any two of the three KBOs involved and 0 otherwise. The choice of limits on
191: the integrals in equation \ref{e19} ensures no double counting of the
192: binaries. Expression (\ref{e19}) can be written as
193: \begin{equation}\label{e20}
194: FR_{L^3}=A_{L^3} \left( \frac {\Sigma}{\rho R} \right)^2
195: \alpha^{-4} \Omega
196: \end{equation}
197: where
198: \begin{equation}\label{e4}
199: A_{L^3}=\left( \frac{27}{32 \pi^2} \right) \int_{\gamma=-\infty}^{\infty}
200: \int_{b_2>b_1}^{\infty} \int_{b_1=0}^{\infty} F_{L^3}(b_1,b_2,\gamma)
201: \left(\frac{b_1}{R_{H}}\right) \left( \frac{db_1}{R_{H}}\right) \left(
202: \frac{db_2}{R_{H}} \right) \left(\frac{d\gamma}{R_H}\right).
203: \end{equation}
204: Expression (\ref{e20}) agrees with the order of magnitude estimate of
205: \citet{GLS02} if we set $A_{L^3}=1$. It is the value of the constant
206: $A_{L^3}$ we determine here. Since we are interested in close encounters
207: among the KBOs, their interaction is well described by Hill's equations
208: \citep{H78,GT80,PH86} that we modify to include three equal mass bodies
209: besides the Sun. The equations of motion, with length scaled by $R_{H}$ and
210: time by $\Omega^{-1}$, for body 1 are given by
211: \begin{equation}\label{e2}
212: \ddot{x}_{1}-2
213: \dot{y}_{1} -3x_{1}=-\frac{3 (x_{1}-x_{2})}
214: {((x_{1}-x_{2})^2+(y_{1}-y_{2})^2)^{3/2}}-\frac{3 (x_{1}-x_{3})}
215: {((x_{1}-x_{3})^2+(y_{1}-y_{3})^2)^{3/2}} \end{equation}
216: \begin{equation}\label{e3}
217: \ddot{y}_{1}+2\dot{x}_{1}=-\frac{3
218: (y_{1}-y_2)} {((x_{1}-x_2)^2+(y_{1}-y_2)^2)^{3/2}}-\frac{3 (y_{1}-y_{3})}
219: {((x_{1}-x_{3})^2+(y_{1}-y_{3})^2)^{3/2}}.
220: \end{equation}
221: The subscripts 1, 2 and 3 label the $x$- and $y$-coordinates of KBO 1, 2, and
222: 3 respectively. Similar equations of motion can be obtained for bodies 2 and
223: 3. $F_{L^3}(b_1,b_2,\gamma)$ is calculated by numerically integrating the
224: equations of motion. A binary formation event is detected in the following
225: way: The equations of motion of the three bodies are integrated until a time
226: that corresponds to a separation of at least $30 R_H$ between all three bodies
227: (after their conjunction), assuming that their relative velocity is solely due
228: to their Keplerian sheer (i.e. ignoring the actual gravitational interaction
229: between the bodies), plus an additional time of $120 \Omega^{-1}$. If after
230: this time the separation between two bodies is still less than $3R_H$ a binary
231: is considered to have formed. We chose a separation of $3 R_H$ instead of
232: $R_H$ to allow for binary orbits that reach slightly outside $R_H$. Numerical
233: integrations are terminated early if the separation between KBOs becomes less
234: than $10^{-4} R_H$ and these events are not counted towards the binaries
235: formed. This serves two purposes: first of all, $10^{-4} R_H$ roughly
236: corresponds to the separation at which physical collisions occur in the Kuiper
237: Belt. Secondly, by introducing a minimum separation, we prevent divergence in
238: the equations of motion. This cut-off limits, strictly speaking, the validity
239: of the value of $A_L^3$ calculated here to binary formation at heliocentric
240: distances of $\sim 40\rm{AU}$ since the separation in units of $R_H$,
241: corresponding to collisions among the KBOs, is inversely proportional to the
242: heliocentric distance. In order to determine $A_{L^3}$ we need to cover the
243: three dimensional parameter space spanned by $b_1$, $b_2$ and $\gamma$. We
244: chose a spacing of $0.1 R_{H}$ for all three parameters. $12.5 R_H$ is chosen
245: as the upper limit for $b_1$ and $b_2$, the upper limit for $|\gamma|$ is $25
246: R_H$. The given limits and resolution require numerical integrations of $\sim
247: 4 \times 10^6$ orbits. We obtain
248: \begin{equation}\label{e5}
249: A_{L^3}=0.28 \pm 0.01
250: \end{equation}
251: where $0.01$ is the estimated Poisson error. We repeated the calculation for
252: $A_{L^3}$ with randomly chosen grid points for $b_1$, $b_2$ and $\gamma$ and
253: the same number of numerical integrations and confirmed that the value of
254: $A_{L^3}$ is insensitive to the grid points chosen. $A_{L^3}$ tends
255: to $0.35$ in the limit that the bodies are treated as point masses
256: (i.e. the limit that the cut-off tends to zero). We will use $A_{L^3}=0.28$
257: since it corresponds to the physically relevant situation in the Kuiper
258: Belt. This yields a binary formation rate of
259: \begin{eqnarray}\label{e6}
260: {FR_{L^3}=(6.3 \pm 0.2) \times 10^{-8} \left(\frac {\Sigma}{3 \times 10^{-4}
261: \rm{g~cm^{-2}}}\right)^2 \left(\frac{\rho}{1\rm{g~cm^{-3}}}\right)^{-2}
262: \left(\frac{R}{100\rm{km}}\right)^{-2} {} } \nonumber\\
263: {} \left(\frac{\alpha}{1 \times 10^{-4}}\right)^{-4}
264: \left(\frac{\Omega}{7.9 \times 10^{-10}\rm{s^{-1}}}\right) \rm{yr^{-1}} {},
265: \end{eqnarray}
266: which is smaller by $1/A_{L^3} \sim 4$ than the order of magnitude estimate of
267: \citet{GLS02}.
268:
269: \section{$L^2s$ FORMATION RATE}
270: So far, we have only considered binary formation due to an encounter with a
271: third body that carries away the excess energy. However, binary formation
272: might also occur due to dynamical friction generated by the sea of small
273: bodies ($L^2s$ mechanism). The random velocity of large KBOs is damped due to
274: gravitational interactions with many small bodies. Since it is not feasible to
275: examine the interactions with each small body individually, their net effect
276: is modeled by an averaged force which acts to damp the large KBOs'
277: non-circular velocity. We parameterize the strength of the damping by a
278: dimensionless quantity $D$ defined as the fractional decrease in non-circular
279: velocity due to dynamical friction over a time $\Omega^{-1}$:
280: \begin{equation}\label{e41}
281: D\sim \frac{ \sigma}{\rho R} \left(\frac{u}{v_H}\right)^{-4} \alpha^{-2} \sim
282: \frac{\Sigma}{\rho R} \alpha^{-2} \left(\frac{v}{v_H}\right)^{-1}.
283: \end{equation}
284: The first expression is simply an estimate of dynamical friction by the sea of
285: small bodies assuming $u>v_H$. The second expression describes the mutual
286: excitation among the large KBOs for $v<v_H$. These two expressions can be
287: equated since the stirring among the large KBOs is balanced by the damping due
288: to dynamical friction. In fact, if $v$ is defined as the product of the median
289: eccentricity and the orbital velocity, we can calculate the exact relationship
290: between $D$ and $(v/v_H)$ since the velocity distribution in the
291: shear-dominated velocity regime has been fully determined (see
292: \citet{CS06,CSS07}). Defining $v$ as the product of the median eccentricity
293: and the orbital velocity, we obtain
294: \begin{equation}\label{e51}
295: D=4.1 \frac{\Sigma}{\rho R} \alpha^{-2} \left(\frac{v}{v_H}\right)^{-1}.
296: \end{equation}
297: For $\rho \sim 1 \rm{g~cm^{-3}}$ and our estimates for $(v/v_H)$, $\Sigma$ and
298: $R$ from \S 2 we find $D \sim 0.12$. We calculate the binary formation rate
299: for equal mass bodies via the $L^2s$ mechanism in shear-dominated velocity
300: regime. As in \S 3, we restrict this calculation to two dimensions with
301: circular motion as initial conditions for the large KBOs and use the same
302: coordinate system as in \S 3. The binary formation rate per body via the
303: $L^2s$ mechanism can be written as
304: \begin{equation}\label{e199}
305: FR_{L^2s}=\int_{b=0}^{\infty}\left(\frac{\Sigma}{\frac{4 \pi}{3}\rho
306: R^{3}}\right) \frac{3}{2} b \Omega F_{L^2s}(D,b) db
307: \end{equation}
308: where $\Sigma/(4\pi \rho R^3/3)$ is the surface number density of the KBOs and
309: $b$ is the relative initial separation in the $x$-direction between the two
310: KBOs. $F_{L^2s}(D,b)$ is a function that takes on the value 1 if the
311: encounter resulted in the formation of a binary for a given $D$ and $b$ and 0
312: otherwise. Equation \ref{e199} can be written as
313: \begin{equation}\label{e9}
314: FR_{L^2s}=A_{L^2s} D \left(\frac{\Sigma}{\rho R}\right) \alpha^{-2} \Omega
315: \end{equation}
316: where
317: \begin{equation}
318: A_{L^2s}=D^{-1}\left(\frac{9}{8 \pi}\right) \int_{b=0}^{\infty} F_{L^2s}(D,b)
319: \left(\frac{b}{R_H}\right) \left(\frac{db}{R_H}\right).
320: \end{equation}
321: \citet{GLS02} showed, using numerical integrations, that $FR_{L^2s}$ is indeed
322: proportional to $D$. Here we want to determine the actual value of
323: $A_{L^2s}$. In Hill coordinates the equations of motion of the two KBOs can be
324: decomposed into their center of mass motion and their relative motion with
325: respect to one another. The relative motion of two equal mass KBOs, including
326: the dynamical friction term, is governed by
327: \begin{equation}\label{e10}
328: \ddot{x}-2 \dot{y}
329: -3x=-\frac{6 x}{(x^2+y^2)^{3/2}}-D \dot{x}
330: \end{equation}
331: \begin{equation}\label{e11}
332: \ddot{y}+2
333: \dot{x}=-\frac{6 y}{(x^2+y^2)^{3/2}} -D (\dot{y}+1.5x).
334: \end{equation}
335: where $x$ and $y$ correspond to the relative separation between the two KBOs
336: in the $x$ and $y$ direction respectively. Again, length has been scaled by
337: $R_{H}$ and time by $\Omega^{-1}$. Equations \ref{e10} and \ref{e11} are
338: integrated for different values of $D$ and impact parameters ranging from $2.2
339: R_H$ to $3.2 R_H$. Impact parameters outside this range result in a distance
340: of closest approach between the two KBOs of more than $R_H$. Figure \ref{fig1}
341: shows that the rate of binary formation is proportional to $D$. The value of
342: $A_{L^2s}$, estimated from the line of best fit, is 1.4. This yields a binary
343: formation rate of
344: \begin{eqnarray}\label{e12}
345: {FR_{L^2s}= 1.3 \times 10^{-5} \left(\frac{D}{0.12}\right) \left(\frac
346: {\Sigma}{3 \times 10^{-4}\rm{g~cm^{-2}}}\right)
347: \left(\frac{\rho}{1\rm{g~cm^{-3}}}\right)^{-1}
348: \left(\frac{R}{100\rm{km}}\right)^{-1} {} } \nonumber\\ {}
349: \left(\frac{\alpha}{1 \times 10^{-4}}\right)^{-2} \left(\frac{\Omega}{7.9
350: \times 10^{-10}\rm{s^{-1}}}\right) \rm{yr^{-1}} {}.
351: \end{eqnarray}
352: Using equation \ref{e41} we can retrieve the scalings of
353: \citet{GLS02}. Although, we know the exact $L^2s$ formation rate for
354: a given $D$ and have an exact expression for $D$ in terms of $v$ (see equation
355: \ref{e51}), the relation between $v$ and the actual physical parameters,
356: i.e. the numerical coefficient in equation \ref{e17}, which is needed for a
357: precise value of $D$ is uncertain to a factor of order unity.
358:
359: Contrary to claims by \citet{A05} and \citet{A07}, the $L^2s$ mechanism does
360: predict a mass-ratio selection. This can be seen from the first part of
361: equation \ref{e41}. For a given $u$, we have that $D \propto R^3$ since
362: $v_H\alpha^{-1/2} \sim v_{esc}\propto R$. Large KBOs experience stronger
363: dynamical friction (larger $D$) than smaller ones. This is not at all
364: surprising and is a general feature of dynamical friction \citep{C43,BT87}. We
365: can write $D=D_0M$ where $D_0\sim \sigma G^2/u^4$. $D_0$ is a constant
366: independent of the KBO mass for a given $\sigma$ and $u$. For two KBOs with
367: masses $M_1$ and $M_2$, the position of body 1 essentially coincides with the
368: center of mass of the two bodies provided that $M_1/M_2 \gg 1$. In the limit
369: that the KBOs random velocity tends to zero and that $D_0M_2 \ll 1$ we can
370: place body 1 at the origin of the Hill coordinate system and treat the center
371: of mass as stationary throughout the interaction. In this limit we find
372: that the relative motion of the two KBOs is governed by
373: \begin{equation}\label{e777}
374: \ddot{x}-2 \dot{y} -3x=-\frac{3x}{(x^2+y^2)^{3/2}}-2D_0M_2 \dot{x}
375: \end{equation}
376: \begin{equation}\label{e778}
377: \ddot{y}+2 \dot{x}=-\frac{3y}{(x^2+y^2)^{3/2}} -2D_0M_2 (\dot{y}+1.5x)
378: \end{equation}
379: where length is scaled by $R_H$ of KBO 1 and time is scaled by $\Omega^{-1}$.
380: For extreme mass-ratio binaries the relevant strength of the dynamical
381: friction that enters equations \ref{e777} and \ref{e778} is twice that acting
382: on the small body (i.e. $2D_0M_2$), and significantly less than that acting on
383: the large body (i.e. $D_0M_1$). The $L^2s$ mechanism therfore favors the
384: formation of comparable mass binaries from the largest available bodies over
385: high mass-ratio ones. It is an open question, whether this preference for
386: comparable mass binaries remains after the Kuiper Belt mass spectrum during
387: their formation and their survival probability are accounted for.
388:
389:
390: \section{COMPARISON OF $L^2s$ AND $L^3$ FORMATION RATES}
391: We are now able to compare the binary formation rates for the $L^2s$ and
392: $L^3$ mechanism for sub-Hill velocities.
393: The ratio of the $L^3$ to $L^2s$ formation rates is
394: \begin{equation}\label{e13}
395: \frac{FR_{L^3}}{FR_{L^2s}}=0.20 D^{-1} \frac{\Sigma}{\rho R} \alpha^{-2}
396: =0.05\frac{v}{v_H}
397: \end{equation}
398: where we substituted for $D$ using the exact relationship from equation
399: \ref{e51}. It is remarkable that this expression depends explicity only
400: on $v/v_H$ and is independent of what sets $D$. It is therefore independent
401: of the velocity dispersion of the small bodies, their surface density or the
402: importance of collisions among the small bodies. We therefore conclude that
403: for $v\ll v_H$, binaries in the Kuiper Belt formed primarily due to dynamical
404: friction rather than three body encounters. Figure \ref{fig1} shows the
405: $L^2s$ and $L^3$ formation rates as a function of $D$. For our estimate of
406: $(v/v_H)\sim 0.1$, we have that $ FR_{L^3}/FR_{L^2s} \sim 0.005 $.
407:
408:
409: \section{Super-Hill Velocity: $v>v_H$}
410: Obviously there is some uncertainty in what the actual values of $\sigma$ and
411: $\Sigma$ were during binary formation. For a few times larger value of $
412: \Sigma $ with $ \sigma $ unchanged, we enter the regime in which $v>v_H$ (this
413: can be seen from equation \ref{e16}). Although it
414: is rather unlikely that $v \gg v_H$ during binary formation (see \S 2) we
415: discuss here briefly how this would affect the $L^2s$ and $L^3$ formation
416: rates.
417:
418: For $v>v_H$ the velocity dispersion of the large bodies is still set by the
419: balance between their mutual stirring and the damping due to dynamical
420: friction generated by the sea of small bodies. Therefore, dynamical friction
421: shrinks the orbit of a KBB with a mutual orbital velocity $v_B$ at a rate
422: \begin{equation}\label{e42}
423: D\Omega\sim \frac{\Sigma}{\rho R} \alpha^{-2} \Omega
424: \left(\frac{v}{v_H}\right)^{-4}
425: \end{equation}
426: where we assume that $v_B<u$. For $v_B \lesssim v$, binaries are broken up by
427: passing KBOs at a rate
428: \begin{equation}\label{e101}
429: R_{break}\sim \frac{\Sigma }{\rho R} \alpha^{-2} \Omega
430: \left(\frac{v}{v_H}\right)^{-2}\left(\frac{v_B}{v_H}\right)^{-2}.
431: \end{equation}
432: The ratio of these two rates yields
433: \begin{equation}\label{e600}
434: \frac{D\Omega}{R_{break}}\sim \left(\frac{v_B}{v}\right)^{2}.
435: \end{equation}
436: Since the ratio in equation \ref{e600} is $<1$ for $v_B<v$, we conclude that
437: KBBs with separations $R_B> R_H (v_H/v)^2$ (i.e. KBBs with $v_B < v$) tend to
438: be broken up by passing KBOs. Binaries with separations of $R_{crit}= R_H
439: (v_H/v)^2$ or less, tend to survive. The cross section for the $L^3$ mechanism
440: is therefore reduced with respect to the sub-Hill case. The probability of
441: having a KBO within $R_{crit}$ of a given KBO is $(\Sigma \Omega)/(\rho R^3
442: v)~ R_{crit}^3$ where $(\Sigma \Omega)/(\rho R^3 v)$ is the volume number
443: density of KBOs. The flux of KBOs into area $R_{crit}^2$ is $(\Sigma
444: \Omega)/(\rho R^3 v)~v R_{crit}^2$. The super-Hill formation rate for tight
445: binaries with separations $\sim R_{crit}$, via the $L^3$ mechanism, is
446: therefore
447: \begin{equation}\label{e167}
448: FR_{L^3}\sim \left( \frac {\Sigma \Omega}{\rho R^3 v} \right)^2 v R_{crit}^5
449: \sim \left( \frac {\Sigma}{\rho R} \right)^2 \alpha^{-4}
450: \left(\frac{v_H}{v}\right)^{11} \Omega
451: \end{equation}
452: (see also \citet{N07}). In addition to tight binaries with separations of
453: $R_{crit}$ and less, there exist a second class of binaries with larger
454: separations. Binaries with separations $R_B>R_{crit}$ are constantly created
455: and destroyed via the $L^3$ mechanism. KBBs can form from two KBOs that
456: approach each other with relative velocity $v_B \lesssim v$ while a third KBO
457: removes energy, through gravitational scattering, enabling the KBOs to get
458: bound. Since we are selecting bodies with relative velocities $\sim v_B$ or
459: less, the number of KBOs that can form binaries with separation $R_B= R_H
460: (v_H/v_B)^2$ is reduced by $\sim (v_B/v)^3$. The formation rate for binaries
461: with separation $R_B=R_H (v_H/v_B)^2$ is
462: \begin{equation}\label{e169}
463: FR_{L^3}(R_B>R_{crit})\sim \left( \frac {\Sigma}{\rho R} \right)^2
464: \alpha^{-4} \left(\frac{v_H}{v}\right)^6 \left(\frac{v_H}{v_B}\right)^{5}
465: \Omega.
466: \end{equation}
467: These wider binaries ($R_B>R_{crit}$) have a higher formation rate compared to
468: the tight ones which have a separation $\sim R_{crit}$. The ratio of the
469: formation rate (equation \ref{e169}) to the destruction rate (equation
470: \ref{e101}) yields an equilibrium abundance of binaries per KBO at any given
471: time that is given by
472: \begin{equation}\label{e170}
473: \frac{N_{KBB}}{N_{KBO}}\sim \frac{\Sigma}{\rho R} \alpha^{-2}
474: \left(\frac{v_H}{v}\right)^4 \left(\frac{v_H}{v_B}\right)^{3}.
475: \end{equation}
476: The number of binaries scales as $(R_B/R_H)^{3/2}$. Binaries with separation
477: $R_B$ are therefore $(R_B/R_{crit})^{3/2}\sim(v/v_B)^3$ times more common than
478: those with separation $R_{crit}$ provided there is sufficient time for the
479: equilibrium to be established. The same statistical equilibrium abundance can
480: be derived using phase space arguments. The phase space number density of KBOs
481: is $(\Sigma \Omega)/(\rho R^3 v^4)$. The phase space volume corresponding to a
482: binary separation $R_B$ and velocity $v_B$ is $R_B^3 v_B^3=R_H^3 v_H^3
483: (v_H/v_B)^3$. Multiplying the KBO phase space number density by the binary
484: phase space volume yields a statistical equilibrium abundance per KBO of
485: \begin{equation}
486: \frac{N_{KBB}}{N_{KBO}}\sim \frac{\Sigma \Omega}{\rho R^3 v^4} R_H^3 v_H^3
487: \left(\frac{v_H}{v_B}\right)^3\sim \frac{\Sigma}{\rho R} \alpha^{-2}
488: \left(\frac{v_H}{v}\right)^4 \left(\frac{v_H}{v_B}\right)^{3},
489: \end{equation}
490: which is in agreement with the binary abundance derived in
491: (\ref{e170}). Whether any of these binaries would survive the dynamical
492: excitation of the Kuiper Belt remains an open question.
493:
494: The $L^2s$ mechanism fails in creating binaries with separations $\sim
495: R_{crit}$ since dynamical friction is not able to dissipate sufficient energy
496: for tight binaries to form. Dynamical friction is only able to assist in
497: the formation of binaries with wide separations ($\sim R_H$) that form from
498: KBOs that happen to approach each other with low relative velocities ($\sim
499: v_H$). This reduces the number density of KBOs that can participate in
500: binary formation by a factor of $\sim (v_H/v)^{3}$. In this case, the $L^2s$
501: formation rate is given by
502: \begin{equation}\label{e201}
503: FR_{L^2s}(R_B\sim R_H)\sim D \left(\frac{\Sigma}{\rho R}\right) \alpha^{-2}
504: \left(\frac{v_H}{v}\right)^4\Omega \sim \left(\frac{\Sigma}{\rho R}\right)^2
505: \alpha^{-4} \left(\frac{v_H}{v}\right)^8\Omega
506: \end{equation}
507: where we have substituted for $D$ from equation \ref{e42} in the last
508: step. These wide binaries face the same fate as the wide ones formed via the
509: $L^3$ mechanism in that they will be broken up quickly due to scattering of
510: other large bodies. However, the $L^2s$ mechanism does not even contribute
511: significantly to the binary equilibrium abundance calculated in equation
512: \ref{e170} since $FR_{L^2s}(R_B\sim R_H)/FR_{L^3}(R_B\sim R_{H})\sim (v_H/v)^2
513: \ll 1$. Therfore, the $L^2s$ mechanism does not play an important role in KBB
514: formation if super-Hill velocities prevail.
515:
516: In summary, the $L^3$ mechanism forms tight binaries, that tend to be saved
517: from break up, at a rate that is reduced by a factor of $(v_H/v)^{11}$
518: compared to the sub-Hill case. In addition, the $L^3$ mechanism forms wider
519: binaries $(R_B>R_{crit})$, at a higher rate that is 'only' reduced by a factor
520: of $(v_H/v)^{6}(v_H/v_B)^5$ relative to the sub-Hill rate. These wide binaries
521: are constantly created and destroyed leading to an equilibrium abundance of
522: binaries that scales as $(R_B/R_H)^{3/2}$. The $L^2s$ mechanism is not
523: important if KBOs have super-Hill velocities.
524:
525:
526: \section{FREQUENCY OF LONG-LIVED TRANSIENT BINARIES AND THEIR SIGNIFICANCE FOR
527: BINARY FORMATION} \citet{A05} propose that transient binaries, that
528: spent a time of $15 \Omega^{-1}$ ($\sim 600 \rm{yr}$ at $40\rm{AU}$) or longer
529: in their mutual Hill sphere, near a periodic orbit, are responsible for binary
530: formation in the $L^2s$ and $L^3$ mechanism. Here, we determine how the
531: frequency of long-lived transient binaries depends on the transient binary
532: lifetime. This allows us to quantify the importance of long-lived transient
533: binaries for the overall binary formation. Finally, we address the
534: significance of long-lived transient binaries for the $L^2s$ and $L^3$
535: formation mechanism.
536:
537: \subsection{Frequency of long-lived Transient Binaries}
538: First, we assess how common long lived transient binaries are. We integrate
539: equations \ref{e10} and \ref{e11} without the dynamical friction term and
540: determine the time $t_{3R_{H}}$ over which the separation between the two KBOs
541: is less than $3 R_{H}$ for all KBOs that approach one another to $R_H$ and
542: less. We chose to calculate the time the two KBOs spent with a separation of
543: less than $3R_{H}$ to allow for orbits that reach slightly outside of $R_{H}$
544: but return back to within $R_{H}$ during the encounter. We integrate $\sim
545: 10^5$ orbits in total with impact parameters ranging from $2.2 R_{H}$ to $3.2
546: R_{H}$. Impact parameters outside this range result in a distance of closest
547: approach between the two KBOs of more than $R_H$. As initial conditions, we
548: assume that the orbits of the bodies are circular. Figure \ref{fig2} shows
549: that the frequency of transient binaries decreases exponentially with the
550: transient binary lifetime, $t_{3R_H}$. The line of best fit yields a
551: differential transient binary frequency, valid for $t_{3R_H}\gtrsim 1
552: \Omega^{-1}$, of
553: \begin{equation}\label{e14}
554: \frac{d (FR_{tb})}{d(t_{3R_H} \Omega)}= 1.0 \times 10^{-(0.25 t_{3R_H}
555: \Omega)}\frac{\Sigma}{\rho R} \alpha^{-2} \Omega.
556: \end{equation}
557: The frequency of transient binaries that spend a time of $ \gtrsim 15
558: \Omega^{-1}$ with a separation of less than $3R_H$ is 3 orders of magnitude
559: smaller than that of short-lived ones with $t_{3R_H} \gtrsim 3 \Omega
560: ^{-1}$. The analysis discussed here was carried out assuming that the KBOs
561: approach each other with relative velocities $v_{rel} < v_H$. Long-lived
562: transient binaries do not exist for bodies that encounter each other at
563: $v_{rel} \gg v_H$. This can be understood by looking at the Jacobi
564: constant. The Jacobi constant in Hill coordinates with length scaled by $R_H$
565: and time by $\Omega^{-1}$ is given by
566: \begin{equation}\label{500}
567: C_J=3x^2-z^2 +\frac{12}{(x^2+y^2+z^2)^{1/2}} -\dot{x}^2 -\dot{y}^2 -\dot{z}^2.
568: \end{equation}
569: KBOs that approach each other with $v_{rel}\gg v_H$ at $R_H$ will experience
570: at most one encounter before they separate. Evaluation of their Jacobi
571: constant at $R_H$ yields that it is large and negative. In order to experience
572: multiple encounters, KBOs must approach each other with $v_{rel}\sim v_H$ at
573: $R_H$ which corresponds to $ C_J$ of order unity. Since the Jacobi constant is
574: a conserved quantity, we can be sure that no long-lived transient binaries
575: exist for KBOs that encounter each other at $v_{rel}\gg v_H$. Long-lived
576: transient binaries therefore offer no solution to the fine tuning problem,
577: contrary to claims by \cite{A07}. For KBOs with a given velocity distribution
578: there always exist a few bodies that have $v_{rel}<v_H$ even if $v\gg
579: v_H$. Such bodies can give rise to long-lived transient binaries in the
580: same way that they can form wide binaries (see \S 6 for details), but the
581: frequency of transient binaries due to such bodies is reduced by a factor of
582: $(v_H/v)^4$.
583:
584: \subsection{Importance of long-lived Transient Binaries in the $L^3$ Formation Mechanism}
585: \cite{A07} claim that the probability of binary formation from transient
586: binaries with $t_{3R_H}\lesssim 2.5 \Omega^{-1}$ is extremely small and they
587: therefore include only transient-binaries with $t_{3R_H}\gtrsim 5 \Omega^{-1}$
588: in the main set of their integrations. However, their conclusion, that the
589: probability of binary formation from transient binaries with $t_{3R_H}\lesssim
590: 2.5 \Omega^{-1}$ is extremely small, is due to a bias in their initial
591: conditions that discriminates against binary formation from transient binaries
592: with $t_{3R_H}\lesssim 5 \Omega^{-1}$. The shortcoming of their analysis is
593: due to the fact that they launch the third body from an initial separation $>
594: 38 R_H$ when the first two KBOs come within a few $R_H$ of each
595: other\footnote{The numerical values stated by \citet{A07} are multiplied by a
596: factor of $2^{1/3}$ to compensate for the different definitions of $R_H$
597: }. Since \citet{A07} select the initial conditions for the third body such
598: that it penetrates within $2.5 R_H$, the largest impact parameter is $\sim 4.5
599: R_H$. The minimum time it takes for the third body to come within a few $R_H$
600: of the transient binary is therefore $ \sim 38 R_H/(1.5 \times 4.5 R_H \Omega)
601: \sim 5.6 \Omega^{-1}$. The third body therefore only reaches the vicinity of
602: the transient binary for $t_{3R_H}\gtrsim 5.6 \Omega^{-1}$, but it is exactly
603: this proximity of the third body that is required for binary formation by
604: strong gravitational scattering. This explains why \citet{A07} find such a
605: small probability for binary formation by transient binaries with
606: $t_{3R_H}\lesssim 5.6 \Omega^{-1}$. The range of impact parameters that lead
607: to binary formation is comparable for short- and long-lived transient
608: binaries. This means that the transient binary lifetime is the only advantage
609: long-lived transient binaries have compared to short-lived ones, in terms of
610: binary formation likelihood. However, as we show in \S 7.1, the frequency of
611: transient binaries drops exponentially as a function of their lifetime. The
612: ratio of the binary formation rate due to short-lived transient binaries
613: ($t_{3R_H}\gtrsim 3 \Omega^{-1}$) compared to that due to long-lived ones
614: ($t_{3R_H}\gtrsim 15 \Omega^{-1}$), is therefore
615: \begin{equation}
616: \frac{FR(t_{3R_H}\gtrsim 3 \Omega^{-1})}{FR(t_{3R_H}\gtrsim 15
617: \Omega^{-1})}\sim \frac{3 \Omega^{-1}}{15 \Omega^{-1}} \frac{10^{-(0.25 \times
618: 3)}}{10^{-(0.25 \times 15)}} \sim 200.
619: \end{equation}
620: Although the binary formation rate scales linearly with transient-binary
621: lifetime, the frequency of transient binaries drops exponentially as a
622: function of its lifetime. Therefore, long-lived transient binaries are not
623: important for binary formation via the $L^3$ mechanism.
624:
625: \subsection{Importance of long-lived Transient Binaries in the $L^2s$ Formation Mechanism}
626: In general, KBOs that spend a longer time in the Hill sphere, lose more energy
627: due to dynamical friction, and are therefore more likely to be
628: captured. However, they might not be responsible for the majority of the
629: binary formation, if the frequency for long-lived transient binaries is
630: sufficiently small. To address this question we determine the typical time
631: $t_{Typ}$, required for a transient binary to become bound with the aid of
632: dynamical friction. We define $t_{Typ}$ as the time it takes for 50\% of all
633: the KBOs, that form a binary, to become bound for a given strength of
634: dynamical friction $D$. $t_{Typ}$ is measured from the point at which the
635: relative separation between the two KBOs is less than $3R_H$. We determine
636: $t_{Typ}$ in the following way. First, we integrate the same equations as in
637: \S 4 (i.e. equations \ref{e10} and \ref{e11}). We switch off the dynamical
638: friction at different times and continue the evolution of the KBOs until $t =
639: 1000 \Omega^{-1}$. This process is repeated until we find the time for which
640: dynamical friction has to act for 50\% of all KBOs, that form a binary, to
641: become bound. A transient binary is considered to have become bound when it
642: remains a binary (i.e. relative separation $ < 3 R_H$) until $t = 1000
643: \Omega^{-1}$. We repeat this for different $ D $ in order to reveal the
644: relationship between $t_{Typ}$ and $D$. Again, impact parameters are chosen to
645: range from $2.2 R_H$ to $3.2 R_H$. Figure \ref{fig3} shows that, for $D
646: \gtrsim 0.002$, the typical time for permanent capture does not depend
647: linearly on the strength of the dynamical friction $ D $, but shows a weaker
648: logarithmic dependence. $t_{Typ} $ only ranges from $ \sim 2 \Omega^{-1} $ for
649: $ D \sim 0.2$ to $\sim 10 \Omega^{-1} $ for $ D \sim 0.002 $. Furthermore,
650: figure \ref{fig3} shows a noticeable break around $ D \sim 0.001$; for $D
651: \lesssim 0.001$, $t_{Typ}$ increase significantly to $ 20 \Omega^{-1} $ and
652: more. From this, we conclude that longevity of the transient binary (as
653: discussed by \citet{A05} with $t_{3R_H} \geqslant 15 \Omega^{-1} $) becomes
654: only important for very weak dynamical friction (i.e. $ D \lesssim 0.002 $)
655: and is most likely not crucial for KBB formation. In \S 4 we estimate $D\sim
656: 0.12$, in which case longevity of transient binaries ($ t_{3R_H}\geqslant 15
657: \Omega^{-1} $) is unlikely to be a major requirement for binary formation.
658:
659: \section{SUMMARY AND CONCLUSIONS}
660: We accurately determine the $ L^2s $ and $ L^3 $ formation rates for
661: $v<v_H$. We find that while the $L^2s$ formation rate is close to previous
662: order of magnitude estimates, the $L^3$ formation rate is about a factor of 4
663: smaller. For $v\ll v_H$, the ratio of the $ L^3 $ to the $ L^2s $ formation
664: rates is $\sim 0.05(v/v_H)$ and is independent of what sets $D$. It is
665: therefore independent of the velocity dispersion of the small bodies, their
666: surface density or the importance of collisions among the small bodies. For
667: sub-Hill KBO velocities, binaries in the Kuiper Belt formed primarily due to
668: dynamical friction rather than three body encounters. For super-Hill KBO
669: velocities ($v \gg v_H$) the $L^2s$ mechanism becomes unimportant. The $L^3$
670: mechanism forms tight binaries that tend to be saved from break up at a rate
671: that is reduced by a factor of $(v_H/v)^{11}$ compared to the sub-Hill
672: case. In addition, the $L^3$ mechanism forms wider binaries $(R_B>R_{crit})$,
673: at a higher rate that is `only' reduced by a factor of
674: $(v_H/v)^{6}(v_H/v_B)^5$ relative to the sub-Hill rate. These wide binaries
675: are constantly created and destroyed leading to an equilibrium abundance of
676: binaries that scales as $(R_B/R_H)^{3/2}$. Whether and how any of these wide
677: binaries would survive the dynamical excitation of the Kuiper Belt remains an
678: open question.
679:
680: In addition, we determine the frequency of long-lived transient binaries. We
681: show that the frequency of long-lived transient binaries drops exponentially
682: with the system's lifetime for $v_{rel}<v_H$. About 1000 transient binaries
683: occur with $ t_{3R_H} \gtrsim 3 \Omega ^{-1} $ for each transient-binary with
684: $ t_{3R_H} \gtrsim 15 \Omega ^{-1} $. The long-lived transient binaries
685: investigated by \citet{A05} and \citet{A07} are therefore very
686: rare. Long-lived transient binaries are not important for binary formation via
687: the $L^3$ mechanism, since the binary formation rate scales only linearly with
688: transient-binary lifetime, but the frequency of transient binaries drops
689: exponentially as a function of its lifetime. Long-lived transient binaries do
690: not exist for $v_{rel}\gg v_H$. We show that the apparent shortage of
691: binaries forming from short-lived transient binaries (i.e. $t_{3R_H}\lesssim
692: 2.5 \Omega^{-1}$) found by \citet{A07} can be explained by a bias in their
693: initial conditions that discriminates against binary formation from transient
694: binaries with $t_{3R_H}\lesssim 5 \Omega^{-1}$. Finally, to assess the
695: importance of long-lived transient binaries in the $L^2s$ mechanism, we
696: determine the typical time $ t_{Typ} $ required for a transient binary to
697: become bound with the aid of dynamical friction. We show that longevity of
698: the transient binary (as discussed by \citet {A05} with $ t_{3R_H} \geqslant
699: 15 \Omega^{-1}$) only becomes important for very weak dynamical friction
700: (i.e. $D \lesssim 0.002$). We estimate $D\sim 0.12$, in which case longevity
701: of transient binaries ($ t_{3R_H}\geqslant 15 \Omega^{-1} $) is unlikely to be
702: a major requirement for binary formation.
703:
704:
705: \acknowledgments We thank Peter Goldreich for stimulating discussions and the
706: anonymous referee for valuable comments that helped to clarify the
707: manuscript. Some of the numerical calculations presented here were performed
708: on Caltech's Division of Geological and Planetary Sciences Dell
709: cluster. R. S. is an Alfred P. Sloan Fellow and a Packard Fellow.
710:
711:
712: \begin{thebibliography}{27}
713: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
714:
715: \bibitem[{{Astakhov} {et~al.}(2005){Astakhov}, {Lee}, \& {Farrelly}}]{A05}
716: {Astakhov}, S.~A., {Lee}, E.~A., \& {Farrelly}, D. 2005, \mnras, 360, 401
717:
718: \bibitem[{{Binney} \& {Tremaine}(1987)}]{BT87}
719: {Binney}, J. \& {Tremaine}, S. 1987, {Galactic dynamics} (Princeton, NJ,
720: Princeton University Press, 1987, 747 p.)
721:
722: \bibitem[{{Brown} {et~al.}(2007){Brown}, {Barkume}, {Ragozzine}, \&
723: {Schaller}}]{BBRS07}
724: {Brown}, M.~E., {Barkume}, K.~M., {Ragozzine}, D., \& {Schaller}, E.~L. 2007,
725: \nat, 446, 294
726:
727: \bibitem[{{Brown} \& {Suer}(2007)}]{BS07}
728: {Brown}, M.~E. \& {Suer}, T.-A. 2007, \iaucirc, 8812, 1
729:
730: \bibitem[{{Brown} {et~al.}(2006){Brown}, {van Dam}, {Bouchez}, {Le Mignant},
731: {Campbell}, {Chin}, {Conrad}, {Hartman}, {Johansson}, {Lafon}, {Rabinowitz},
732: {Stomski}, {Summers}, {Trujillo}, \& {Wizinowich}}]{BDB06}
733: {Brown}, M.~E., {van Dam}, M.~A., {Bouchez}, A.~H., {Le Mignant}, D.,
734: {Campbell}, R.~D., {Chin}, J.~C.~Y., {Conrad}, A., {Hartman}, S.~K.,
735: {Johansson}, E.~M., {Lafon}, R.~E., {Rabinowitz}, D.~L., {Stomski}, Jr.,
736: P.~J., {Summers}, D.~M., {Trujillo}, C.~A., \& {Wizinowich}, P.~L. 2006,
737: \apjl, 639, L43
738:
739: \bibitem[{{Cameron} \& {Ward}(1976)}]{CW76}
740: {Cameron}, A.~G.~W. \& {Ward}, W.~R. 1976, in Lunar and Planetary Institute
741: Conference Abstracts, Vol.~7, Lunar and Planetary Institute Conference
742: Abstracts, 120--+
743:
744: \bibitem[{{Chandrasekhar}(1943)}]{C43}
745: {Chandrasekhar}, S. 1943, \apj, 97, 255
746:
747: \bibitem[{{Collins} \& {Sari}(2006)}]{CS06}
748: {Collins}, B.~F. \& {Sari}, R. 2006, \aj, 132, 1316
749:
750: \bibitem[{{Collins} {et~al.}(2007){Collins}, {Schlichting}, \& {Sari}}]{CSS07}
751: {Collins}, B.~F., {Schlichting}, H.~E., \& {Sari}, R. 2007, \aj, 133, 2389
752:
753: \bibitem[{{Funato} {et~al.}(2004){Funato}, {Makino}, {Hut}, {Kokubo}, \&
754: {Kinoshita}}]{F04}
755: {Funato}, Y., {Makino}, J., {Hut}, P., {Kokubo}, E., \& {Kinoshita}, D. 2004,
756: \nat, 427, 518
757:
758: \bibitem[{{Goldreich} {et~al.}(2002){Goldreich}, {Lithwick}, \& {Sari}}]{GLS02}
759: {Goldreich}, P., {Lithwick}, Y., \& {Sari}, R. 2002, \nat, 420, 643
760:
761: \bibitem[{{Goldreich} {et~al.}(2004){Goldreich}, {Lithwick}, \& {Sari}}]{GLS04}
762: ---. 2004, \araa, 42, 549
763:
764: \bibitem[{{Goldreich} \& {Tremaine}(1980)}]{GT80}
765: {Goldreich}, P. \& {Tremaine}, S. 1980, \apj, 241, 425
766:
767: \bibitem[{{Hartmann} \& {Davis}(1975)}]{HD75}
768: {Hartmann}, W.~K. \& {Davis}, D.~R. 1975, Icarus, 24, 504
769:
770: \bibitem[{{Hill}(1878)}]{H78}
771: {Hill}, G.~W. 1878, Am. J. Math., 1, 5
772:
773: \bibitem[{{Kenyon} \& {Luu}(1999)}]{KL99}
774: {Kenyon}, S.~J. \& {Luu}, J.~X. 1999, \aj, 118, 1101
775:
776: \bibitem[{{Lee} {et~al.}(2007){Lee}, {Astakhov}, \& {Farrelly}}]{A07}
777: {Lee}, E.~A., {Astakhov}, S.~A., \& {Farrelly}, D. 2007, ArXiv e-prints, 705
778:
779: \bibitem[{{McKinnon}(1989)}]{MK89}
780: {McKinnon}, W.~B. 1989, \apjl, 344, L41
781:
782: \bibitem[{{Noll} {et~al.}(2007){Noll}, {Grundy}, {Chiang}, {Margot}, \&
783: {Kern}}]{N07}
784: {Noll}, K.~S., {Grundy}, W.~M., {Chiang}, E.~I., {Margot}, J.-L., \& {Kern},
785: S.~D. 2007, ArXiv Astrophysics e-prints
786:
787: \bibitem[{{Petit} \& {Henon}(1986)}]{PH86}
788: {Petit}, J.-M. \& {Henon}, M. 1986, Icarus, 66, 536
789:
790: \bibitem[{{Rafikov}(2003)}]{R2003}
791: {Rafikov}, R.~R. 2003, \aj, 125, 942
792:
793: \bibitem[{{Stern} {et~al.}(2006){Stern}, {Weaver}, {Steffl}, {Mutchler},
794: {Merline}, {Buie}, {Young}, {Young}, \& {Spencer}}]{SWS06}
795: {Stern}, S.~A., {Weaver}, H.~A., {Steffl}, A.~J., {Mutchler}, M.~J., {Merline},
796: W.~J., {Buie}, M.~W., {Young}, E.~F., {Young}, L.~A., \& {Spencer}, J.~R.
797: 2006, \nat, 439, 946
798:
799: \bibitem[{{Trujillo} \& {Brown}(2003)}]{TB03}
800: {Trujillo}, C.~A. \& {Brown}, M.~E. 2003, Earth Moon and Planets, 92, 99
801:
802: \bibitem[{{Trujillo} {et~al.}(2001){Trujillo}, {Jewitt}, \& {Luu}}]{TJL01}
803: {Trujillo}, C.~A., {Jewitt}, D.~C., \& {Luu}, J.~X. 2001, \aj, 122, 457
804:
805: \bibitem[{{Weaver} {et~al.}(2006){Weaver}, {Stern}, {Mutchler}, {Steffl},
806: {Buie}, {Merline}, {Spencer}, {Young}, \& {Young}}]{WSM06}
807: {Weaver}, H.~A., {Stern}, S.~A., {Mutchler}, M.~J., {Steffl}, A.~J., {Buie},
808: M.~W., {Merline}, W.~J., {Spencer}, J.~R., {Young}, E.~F., \& {Young}, L.~A.
809: 2006, \nat, 439, 943
810:
811: \bibitem[{{Weidenschilling}(2002)}]{W02}
812: {Weidenschilling}, S.~J. 2002, Icarus, 160, 212
813:
814: \bibitem[{{Wetherill} \& {Stewart}(1993)}]{WS1993}
815: {Wetherill}, G.~W. \& {Stewart}, G.~R. 1993, Icarus, 106, 190
816:
817: \end{thebibliography}
818:
819: \clearpage
820:
821: \begin{figure}
822: \plotone{f1.eps}
823: \caption{Binary formation rate as a function of dynamical friction strength $
824: D $. The crosses correspond to the formation rate via the $ L^2s $
825: mechanism and the dashed horizontal line corresponds to the $ L^3 $
826: formation rate for $ (\Sigma/\rho R)\alpha^{-2} = 3 \times 10^{-3}$. The $
827: L^2s $ formation rate is proportional to $D$. In \S 4 we estimate that
828: $D\sim 0.12$ as a result of which $ FR_{L^3}/FR_{L^2s} \sim 0.005 $.
829: }
830: \label{fig1}
831: \end{figure}
832:
833: \begin{figure}
834: \plotone{f2.eps}
835: \caption{Differential transient binary frequency $\Delta (FR_{tb})$ as a
836: function of the transient binary lifetime $t_{3R_H}$ in the shear dominated
837: velocity regime. $t_{3R_H}$ is the time the transient binary separation is
838: less than $3 R_H$. The frequency decreases exponentially with $
839: t_{3R_H}\Omega$.
840: }
841: \label{fig2}
842: \end{figure}
843:
844:
845: \begin{figure}
846: \plotone{f3.eps}
847: \caption{The variation of $t_{Typ}$, the time taken for 50\% of binaries to
848: get bound, plotted against strength of dynamical friction $D$. For about two
849: orders of magnitude change in $D$ ($D \sim 0.2$ to $D \sim 0.002$) $t_{Typ}$
850: only changes from $\sim 2 \Omega^{-1}$ to $\sim 10 \Omega^{-1}$. A rapid rise
851: in $t_{Typ}$ occurs for $D \lesssim 0.001$.
852: }
853: \label{fig3}
854: \end{figure}
855:
856:
857: \end{document}
858: