1:
2: %% manuscript produatces a one-column, double-spaced document:
3:
4: \documentclass[preprint]{aa}
5: %\documentclass[preprint]{aa}
6:
7: %% preprint2 produces a double-column, single-spaced document:
8: \usepackage{graphicx}
9: \usepackage{natbib}
10: \newcommand{\be}{\begin{eqnarray}}
11: \newcommand{\ee}{\end{eqnarray}}
12: \newcommand{\vdag}{(v)^\dagger}
13: \newcommand{\myemail}{pfalzner@ph1.uni-koeln.de}
14: \newcommand {\nbody}{\textsc{\mbox{nbody6}}}
15: \newcommand {\nbodypp}{\textsc{\mbox{nbody6\raise.4ex\hbox{\tiny++}}}}
16: \newcommand {\COri} {\mbox{$\theta^1{\rm{C}}\:{\rm{Ori}}$}}
17: \newcommand {\rclose} {$r_{\rm{close}}$ }
18: \newcommand {\vclose} {$v_{\rm{close}}$ }
19: \newcommand {\eclose} {$e_{\rm{close}}$ }
20: \newcommand {\Msun} {\mbox{M$_{\odot}$}}
21: \newcommand {\Lsun} {\mbox{L$_{\odot}$} }
22: \newcommand {\Macc} {$\stackrel{.}{M}_{acc}$ }
23:
24:
25: \begin{document}
26:
27: \title{Quasi-binarity of massive stars in young dense clusters - the case of the ONC}
28: \author{S.~Pfalzner \and C.~Olczak}
29: \institute{I. Physikalisches Institut, University of Cologne, Z\"ulpicher Str. 77, 50937 Cologne, Germany}
30: \authorrunning{Pfalzner \& Olczak}
31: \titlerunning{Quasi-binarity of massive stars}
32:
33:
34: \abstract
35: %
36: {Observations indicate that in young stellar clusters the binary fraction for
37: massive stars is higher than for solar mass stars. For the Orion Nebula Cluster
38: (ONC) there is a binary frequency of $\sim$ 50\% for solar-mass stars
39: compared to 70-100\% for the massive O- and B-stars.}
40: %
41: {We explore the reasons for this discrepancy and come up with two possible answers:
42: a) a primordially higher binarity of massive stars could be inherent to the star
43: formation process or b) the primordial binary rate might be the same for solar-mass
44: and massive stars, but the higher capture cross section of the massive stars possibly
45: leads to the formation of additional massive binaries in the early cluster development. Here we
46: investigate the likelihood of the latter scenario in detail using the ONC as an example.}
47: %
48: {N-body simulations are performed to track the capture events in an ONC-like cluster.}
49: %
50: {We find that whereas low-mass stars rarely form bound systems through capture, the dynamics
51: of the massive stars - especially in the first 0.5 Myrs - is dominated by a rapid
52: succession of ``transient binary or multiple
53: systems''.
54: In observations the transient nature of these systems would not be apparent, so
55: that they would be rated as binaries. At 1-2 Myrs, the supposed age of the ONC, the
56: ``transient'' massive systems become increasingly stable,
57: lasting on average several 10$^6$ yrs. Despite the ONC being so young, the observed binary
58: frequency for massive stars --- unlike that of solar-mass stars --- is not identical to the
59: primordial binary frequency but is increased by at least 10-15\% through dynamical interaction
60: processes. This value might be increased to at least 20-25\% by taking disc effects into account.}
61: {The primordial binary frequency
62: could well be the same for massive and solar mass stars because the observed difference
63: can be explained by capture processes alone.}
64:
65:
66: \keywords{clusters - binaries - ONC}
67:
68: \maketitle
69:
70:
71: \section{Introduction}
72:
73: Observations predominantly in the 1990's found the fraction of field stars in multiple systems to
74: be high --- for solar-mass stars $\sim$ 55$\%$, \cite{duquennoy:91} and for M-dwarf low-mass stars
75: $\sim$ 35-42$\%$ \cite{fischer:92,reid:97}.
76: %
77: To explain the high binary rate of field stars, gravitational capture between two single stars
78: was initially discussed as
79: a possibility. However, it was acknowledged that while some binaries might actually
80: form this way, capture cannot be the primary binary formation process due to the low likelihood
81: of such an event.
82: %The reason for this low likelihood is that three objects are actually required for
83: %such an event to happen, as conservation of energy rules out a single gravitating body capturing
84: %another.
85:
86: Observations of pre-main sequence stars support the theory that binaries
87: are already formed during star formation. The multiplicity of young pre-main-sequence (PMS) stars
88: in many regions (especially in loose T-associations like Taurus) seems to be systematically higher than that
89: of their main-sequence counterparts \citep[e.g.][and references therein]{mat:00}.
90: %Observational results \cite{leinert:93,ghez:97,duchene:99} give strong
91: %indications that in some young clusters the low-mass stars might have a
92: %companion star frequency close to 100$\%$.
93: Currently fragmentation of the molecular cloud during the formation of protostars is the accept
94: explanation for the formation of a binary or multiple star systems \cite{boss:92,tohline}
95:
96: However, unsolved problems remain. Denser young stellar clusters like the central (very dense) Trapezium
97: cluster of the Orion Nebula Cluster, IC 348 and NGC 2024 show much lower binary frequencies
98: than above mentioned loose T-associations
99: \cite{prosser:94,padgett:97,bouvier:97,patience:98,petr:98,simon:99,scally:99,duchene:99b,mccaughrean:01,beck:03,liu:03,luhman:05,kraus:07}.
100: There is emerging evidence that high-density regions might have lower binary frequencies
101: or preferentially smaller binary separations \cite{koehler:06}. Essentially two theoretical
102: concepts exist to explain this observational finding: (a) the dynamical
103: disruption of wide binaries (with a separation distance $r_s > $ 100 AU) through close
104: stellar encounters decreases the primordial binary fraction in dense clusters
105: \citep{kroupa:95,kroupa:99} or (b) the influence of the temperature of the molecular
106: cores on the fragmentation mechanism results in higher primordial binary fractions
107: \cite{durisen:94,sterzik:03} in less dense clusters.
108:
109: %These two theoretical concepts lead to different predictions concerning the binary
110: %fraction as a function of the distance to the cluster centre. Whereas the encounter
111: %model would predict a lower binary rate close to the cluster centre as more
112: %binaries would be destroyed due to the higher density there, the
113: %model of different core conditions, would anticipate no dependence on the
114: %distance to the cluster centre.
115:
116: Possibly connected with this problem is the question that is addressed here: why does
117: the binary frequency increase from $\sim$ 50\% for solar-type
118: field stars \cite{duquennoy:91,fischer:92} to $\sim$ 70\% for massive O- and B-stars
119: \cite{abt:90,mason:98})?
120: % more
121: %precisely OB stars have a higher binary (and multiplicity) frequency than solar mass
122: %stars in clusters like the ONC and solar mass field stars have a higher binary frequency
123: %compared to M-dwarfs.
124: %
125: %Further differences concerning the binarity of PMS and field stars are: the separation distribution of
126: %PMS stars shows an over-abundance of binaries with separations in the range of $\sim 100 - 1000$~AU
127: %\citep{mathieu:94,patience:02,duchene:04} by a factor of $\sim 2$.
128: %So that leads to the conclusion that the binary fraction and separation distributions
129: %evolve significantly between young stellar populations and the field.
130: %Equally, the mass ratio distribution of massive stars ($M^* >$ 10 \Msun) appears to depend
131: %on the age of the cluster; those in young clusters being consistent with random sampling from
132: %the IMF and those in dynamically evolved populations favouring equal-mass companions
133: %\citep[see][and references therein]{goodwin:06}.
134: %
135: In this paper we consider the possible reasons for the higher binary frequency of massive stars in
136: comparison to intermediate mass stars. In principle there are only two possibilities:
137: either massive stars are more likely to be binaries to start with, or their binarity increases
138: within the first Myr of their existence in a significant way.
139: In other words, the first possibility would mean that the star formation process is mass-dependent.
140: The question whether this is so is very difficult to answer, since there are still problems explaining
141: how even a single star more massive than $\sim$ 10 \Msun\ can form \citep[for a review see][]{yorke:07}.
142: Here we restrict ourselves to the second possibility, i.e. we start with the assumption that the
143: primordial number of binaries is the same for solar-mass and massive stars and ask whether dynamical
144: processes can lead to a sufficient amount of additional binaries to explain the difference in
145: observed binary frequency between solar-mass and massive stars.
146:
147: To do this we revisit capture processes as a potential candidate for part of the binary formation.
148: In the early 90s, stellar dynamics simulations \citep{clarke:91} of the ONC showed that dynamical
149: interactions in a cluster cannot form a significant enough number of binaries from an initially
150: single-star population to explain a 50\% or more binary population.
151: Indeed, we also find that for solar-type stars, capture processes rarely lead to bound systems.
152: However, it will be demonstrated here that the situation is different for {\em massive} stars.
153: The combination of the higher capture cross section of the massive stars with the fact that they
154: are predominantly located in the high density central regions of the cluster leads to the formation
155: of massive binaries in addition to the primordial existing ones by capture processes in the early
156: cluster development.
157:
158: As a model cluster we chose the Orion Nebula Cluster, because it is one of the densest close-by clusters
159: in the Galaxy, so if capture processes play any role one should find indications for it here.
160: In addition, it is observationally one of the best investigated clusters, so the detailed knowledge
161: of its properties helps to limit the simulation parameters. What is known about the binary statistics in the
162: ONC?
163: %
164: In Orion the binary rate for solar-type stars is $\sim$ 50\% whereas the binary rate
165: for massive stars is $\sim$ 75 \% \cite{preibisch:99,koehler:06}. The latter value has a high
166: uncertainty upwards, as observations of the binary frequency of high-mass stars are usually obtained
167: from small-sized samples and inhomogeneity of the sample and/or selection effects in general cannot
168: be completely excluded.
169: %Spectroscopic methods can only find close binaries, whereas visual methods are limited to wide binaries.
170: Correction for completeness, as done for lower mass stars, is difficult as the necessary assumptions of
171: the binary mass ratio and orbital period distribution cannot be made. This means that the observations
172: could be equally well be interpreted as a 100\% binary rate for massive stars in the ONC \cite{preibisch:99,koehler:06}.
173: %
174: %Investigations by \cite{preibisch:99} suggest that there are on average at least 1.5 companions per
175: %OB star in the Orion Nebula cluster compared to $\sim$ 0.5 companions per low-mass star. This
176: %is confirmed by a recent study by \cite{koehler:06}, who find the binarity of stars with masses
177: %$<$ 2\Msun to be significantly larger than that of low-mass stars.
178: Possibly there exists a trend of
179: decreasing multiplicity from the centre of the ONC outwards \cite{preibisch:99,koehler:06}.
180: % the binary frequency of low-mass stars in the periphery of the ONC
181: %is slightly higher than in the core (but only with low statistical significance) and that
182: One concludes that differences in the formation mechanisms of high-mass and low-mass stars
183: multiple systems must exist.
184: %Whether this is valid, is still an open question.
185:
186: Although capture usually is no longer considered, Moeckel \& Bally (2007a) studying single isolated
187: encounter events of disc-surrounded stars only, suggested that
188: encounter-induced capture could be a common event for massive stars. They estimated
189: that the density in the ONC would suffice for capture processes alone to lead to 50\% binaries.
190: Although we do not include the effect of discs in the simulations presented here, we perform
191: the first systematic study of the stellar dynamics in an ONC-like system with the emphasis on the
192: massive stars. It reveals that {\em massive stars frequently form ``transient bound systems''} (TBS)
193: through gravitational interaction. Single such systems are indistinguishable from long-lasting classical
194: binaries in observations, but in Section 4 we will see that some binary statistics results could be
195: interpreted as effect of a population of capture-formed massive binaries.
196:
197:
198: %Such massive stars can have a major impact on
199: %their surroundings through their high photoionizing uv-flux, feedback on the interstellar
200: %medium through strong mass loss and dynamical interactions with neighbouring (lower mass) stars.
201: %It is the latter that we are investigating here for the ONC as an example cluster.
202:
203: What kind of binaries do form? Binaries can be sub-divided into three dynamical groups according to
204: the ratio between $E_b$, the binding energy of the binary, and $E_{kin}$, the kinetic energy
205: of the encounter with another star of the system: (i)
206: the wide, or soft, binaries with $E_b/E_{kin} <$ 1, (ii) the dynamically active binaries
207: $E_b/E_{kin} \sim$ 1, and
208: (iii) the tight or hard binaries $E_b/E_{kin} >$ 1.
209: %
210: Soft and hard binaries are relatively well understood \cite[see][for a review]{heggie:03} and
211: \cite{heggie:75} and \cite{hill:75} previously summarized the dynamics of these systems with the
212: observation that ``soft binaries soften and hard binaries harden''.
213: %
214: Of these groups the active binaries with intermediate binding energies are least well understood.
215: Such binaries couple efficiently to the cluster, and exchange energy with it. We will see that
216: it is this type of binary that the massive stars in the ONC primarily form in the
217: initial phases of the cluster development.
218: %They do {\em not} form long-lived binaries but belong to the group of active binary (or multiple)
219: %systems.
220: These active binaries or TBS are quite proficient in exchanging partners.
221: %
222: %This unifies the seemingly contradicting findings that dynamical interactions
223: %in a cluster cannot form a significant number of binaries by \cite{clarke} with
224: %the result of \cite{moeckel:06} that dynamical should lead to a binary
225: %fraction of $\sim$ 50\% for massive stars.
226: %showing that the number of long-term binary pairs is relatively low but the rate of
227: %quasi-binaries is very high.
228:
229: Section 2 describes the numerical method applied. In Section 3 we determine the number of TBS formed by
230: capture in the ONC followed by a detailed analysis of their properties.
231: In Section 4 it is discussed how these findings might change the interpretation of the origin of a
232: higher binary rate for massive stars.
233:
234:
235:
236: \section{Method}
237:
238:
239: \begin{figure}
240: %\resizebox{!}{\vsize}{\includegraphics[angle=0]{binarity.eps}}
241: \resizebox{\hsize}{!}{\includegraphics[angle=-90]{density.ps}}
242: \caption{The development of the particle density as a function of time
243: for the central core, the Trapezium region and the entire ONC.}
244: \label{fig:density}
245: \end{figure}
246:
247:
248: For simplicity we start with a system initially consisting only of single stars - the influence
249: of both primordial binaries and discs around the stars are excluded from this first study. Primordial
250: binaries would make it more difficult to track the effect of the capture processes
251: themselves, we will discuss in Section 4 in how the assumption of no primordial binaries
252: effects our results. The influence of discs can only be included if detailed knowledge of
253: star-disc-star-disc interactions leading to binaries exists for the entire parameter space required.
254: To our knowledge only single cases of such interactions have been modelled so far.
255:
256:
257: \begin{figure}
258: %\resizebox{!}{\vsize}{\includegraphics[angle=0]{encount.eps}}
259: \resizebox{\hsize}{!}{\includegraphics[angle=0]{f1.eps}}
260: \caption{Number of encounters with $\epsilon <$ 1 as function of a) radial distance to cluster centre,
261: b) time and c) primary mass M$_1^*$. In b) the number of encounters in the Trapezium and the entire ONC are shown.
262: In c) all encounters with $\epsilon <$ 1 independent of the duration of the bound state are shown, but although
263: the ones that keep at least bound for one (dashed line) and ten (dotted line) orbits.}
264: \label{fig:enc_mass}
265: \end{figure}
266:
267:
268:
269: In the cluster simulations we followed the dynamical development of $\sim$ 4000 stars in an
270: virial equilibrium situation, i.e. $Q_{vir}$=0.5, with a spherical density distribution
271: $\rho(r) \sim r^{-2}$ using \nbodypp \cite{aarseth:book03,spurzem:mnras02}.
272: Gas components and the potential of the background molecular cloud OMC~1 were neglected in these
273: simulations. The most massive
274: star was assigned a mass $M^*$= 50 \Msun\ and all other stars' masses according to the mass
275: distribution given by \cite{kroupa:01},
276: \be
277: \xi(M^*)= \left\{
278: \begin{array}{lrll}
279: 0.035 M^*\,^{-1.3} & \mbox{ if } & 0.08&\le M^*\,<0.5, \\
280: 0.019 M^*\,^{-2.2} & \mbox{ if } & 0.5&\le M^*\,<1.0, \\
281: 0.019 M^*\,^{-2.7} & \mbox{ if } & 1.0&\le M^*\,<\infty.
282: \end{array} \right.
283: \label{eq:imf}
284: \ee
285: %
286: The simulations were performed as described in \cite{olczak:06}.
287: There the quality of the dynamical
288: models was judged by comparing them to observational data at 1-2\,Myr, marking the range of the mean
289: ONC age. The quantities of interest were: number of stars, half-mass radius, number densities,
290: velocity dispersion and projected density profile.
291: In the initial cluster the central number density within 0.05 pc was 2.85 $\times$ 10$^5$ pc$^{-3}$ and
292: the velocity dispersion 2.4 km/s. There is a fast development of the number density (see Fig.~\ref{fig:density})
293: which peaks in the core at $\sim$ 0.03Myr.
294:
295: However, we found it necessary to improve the tracking of the cluster centre for comparison
296: to observational data. Previously we used the centres-of-mass to define the cluster centre. However, as
297: the simulations contain only a limited number of stars,
298: %the centre-of-mass does not serve as a good reference to the cluster centre, because
299: ``escapers'' kicked out of the cluster due to strong interactions, can temporarily shift the
300: centre-of-mass before they are removed at large distances.
301: %
302: Using the (mass or number) density centre is a more reliable alternative, especially
303: since observers usually use either the brightness concentration or the maximum projected number
304: density of objects as a reference point.
305: However,
306: %the density centre as a function of time and cluster properties is a very rough quantity.
307: %using the density centre has as well some disadvantages. First, since star clusters are
308: %dynamically very active,
309: since the local particle density changes strongly with time, so does the estimated
310: density centre, and is additionally influenced by the choice of the number of
311: nearest neighbours to some degree.
312: %
313:
314: Therefore we use the density centre but introduce the following smoothing algorithm:
315: Instead of using all simulation particles to define the centre of mass, particles at large
316: distances from the cluster centre are excluded. On the other hand this sub-sample has to be large enough
317: that proper statistics are possible and that the most massive stars are not too dominant. This is achieved
318: in practice by requiring that the sub-sample contain at least 10\% of the total stellar population
319: and that the most massive star must represent at most 10\% of the mass of this sub-sample.
320: The centre-of-mass of the sub-sample serves as the new cluster centre.
321: This algorithm combines the advantages of keeping the cluster centre close to
322: the density centre and taking a larger stellar sample into account to smooth the strong temporal
323: fluctuations.
324:
325: Having defined the cluster centre in such a way, simulations have been selected as valid representations
326: of the ONC if they fulfilled the following criteria:
327: \begin{itemize}
328: \item{The projected density distribution at 1\,Myr has to match the data from
329: Hillenbrand \& Hartmann (1998) and McCaughrean et al. (2002) within the statistical errors.}
330: \item{The most massive star, \COri, must be located inside the Trapezium Cluster ($R_{\small\textrm{TC}}=0.3$\,pc) for at least 1\,Myr, the estimate of the mean age of the ONC.}
331: \end{itemize}
332:
333: Altogether we have selected 40 simulations according to the above criteria.
334: During the course of these simulations we track all encounters with eccentricity $\epsilon <$ 1,
335: noting the stellar mass ratio, periastron, eccentricity and duration of the bound state. It is not checked
336: for bound systems with more than two partners, as we want to concentrate on the binary rather than
337: the general multiplicity rate in this investigation.
338:
339: Despite the large number of simulations and the resulting several hundred TBS formed per simulation run,
340: one still has to deal with small-number statistics here. There are two reasons for this: the parameter space
341: in mass ratio, periastron, etc. is very large and it is often necessary to restrict the sample to a certain
342: time interval or for systems to stay stable for a certain duration. This leads to the difficulty that
343: the results in the following section are sometimes presented for different time-intervals.
344: We only present results where we have at least 20 events per bin to guarantee statistical significance. One
345: should keep in mind that observations usually have much fewer objects to work with when determining binary properties.
346:
347:
348:
349: \section{Results}
350:
351: \begin{figure}
352: %\resizebox{!}{\vsize}{\includegraphics[angle=0]{binarity.eps}}
353: \resizebox{\hsize}{!}{\includegraphics[angle=0]{f7.eps}}
354: \caption{a) shows the fraction of time a TBS would appear as a binary for different
355: primary stars. The distribution of the stellar masses is not fixed but varies. This done to give a
356: better representation of the IMF on average. Therefore no fixed stellar masses can be given here,
357: but the higher star number is equivalent to lower stellar mass. In b) the percentages of the
358: 5 (dashed line) and 13 (full line) most massive stars are shown that
359: are part of an encounter-induced binary in a time interval of 0.1 Myrs as a function of cluster age.
360: The thin dashed line shows the case where these massive TBS are considered twice.}
361: \label{fig:binarity}
362: \end{figure}
363:
364: All encounter events with an eccentricity $\epsilon <$ 1 --- hitherto called ``capturing encounters'' ---
365: are recorded. It turns out that these capturing encounters mostly happen early on in the cluster
366: development, close to the cluster centre and mainly involve one of the most massive stars
367: (see Fig.~\ref{fig:enc_mass}a-c).
368: This is exactly what one expects: close to the cluster centre the stellar density
369: is highest, even more so early on when the cluster initially contracts (see Fig.~\ref{fig:density})
370: and the most massive stars
371: dominate this area gravitationally. Specifically, during the first Myr there are on
372: average more than 350 capturing encounters in the Trapezium region, in contrast to
373: $\sim$ 150 in the rest of the ONC, which covers a thousand times larger spatial volume (see
374: Fig.~\ref{fig:enc_mass}a). During the first 0.3 Myr of the cluster development more
375: than 300 capturing encounters occur; afterwards less than 15 capturing events happen
376: in any 0.1 Myr time interval (see Fig.~\ref{fig:enc_mass}b).
377: %The single most massive star with the mass $M_1^*$=50\Msun\ has nearly three times as many capturing
378: %encounters as any other mass group, despite the relatively large number of low mass stars (see
379: %Fig.~\ref{fig:enc_mass}c).
380: Whereas the most massive star has on average of the order of 200 capturing
381: encounters in the first 5 Myr, a star with $M_1^*= 1$\Msun\ has less than one. In addition, a higher
382: percentage of the capturing encounters involving the massive stars lead to relatively stable
383: configurations. If the primary is of mass $M_1^*$=50\Msun, about half of the forming systems
384: stay stable for at least one orbit and a quarter for at least 10 orbits. By contrast, for a primary of
385: mass $M_1^*$= 1\Msun\ about a third stay stable for one orbit and less than 10\% for more
386: than 10 orbits. Here and in the following the word ``primary'' is used for the more massive rather
387: than the more luminous star, in contrast to convention in observations.
388:
389: We are now in a position to ask which binarity would be observed for clusters like our ONC model
390: cluster. Fig.~\ref{fig:binarity}a) shows the probability of finding the most massive stars in a
391: bound state as a result of a capturing process. It can be seen that the most massive star would
392: in the majority of cases be found as a partner in a
393: bound state. The second most massive star would be likely to be bound in $\sim$ 30$\%$ of all
394: cases. For stars with lower mass the likelihood to be found in such a capture-induced bound state
395: decreases rapidly with decreasing mass (see star 5 and 50 in Fig.~\ref{fig:binarity}a, where higher numbers
396: are equivalent to lower masses).
397:
398: Fig.~\ref{fig:binarity}b shows the likelihood of the 5 and 13 (number of OB stars in the ONC)
399: most massive stars of the cluster to have at least one companion (here the percentage is shown)
400: as a function of time.
401: At 2 Myrs 10-15$\%$ of the all OB stars and 25-30$\%$ of the five most massive stars would form
402: TBSs due to the interaction dynamics in the cluster and would appearing as
403: binaries. Here a TBS is taken into account only once if both stars were massive.
404: The thin dashed line in Fig.~\ref{fig:binarity}b shows the case where these massive TBS are
405: considered twice. This is equivalent to the likelihood of a specific star to be in a TBS.
406: For the five most massive stars this likelihood to be in a TBS is $\sim$ 40 $\%$ at 2 Myr.
407: Fig.~\ref{fig:binarity}b agrees with the result by Moeckel \& Bally (2007b) who found
408: an expected binary fraction of 12\% for a star with $M_*=20\Msun$ at 0.5 Myr.
409: So for the most massive stars of the cluster there is a high likelihood
410: of it being a TBS and appearing as a binary. The number of TBS is actually sufficient to explain the
411: difference in binary rates between massive and solar-mass stars. But what are the properties of these
412: capture-formed TBSs?
413:
414: We start with the periastra of the transient bound systems.
415: Fig.\ref{fig:peri_mass}a) shows the average periastron as a function of the primary mass $M_1^*$.
416: The values are averaged over the first 5 Myr of the cluster development.
417: It can be seen that the average periastron declines with increasing primary mass from $\sim$
418: 700 AU for stars with $M_1^* < $ 0.1\Msun\ to $\sim$ 60 AU for $M_1^* $=50\Msun, reflecting the fact
419: that massive stars form tighter bound systems more easily.
420: The decline in periastron
421: can be approximated by a linear dependence on log10($M_1^*$).
422:
423:
424: \begin{figure}
425: %\resizebox{!}{\vsize}{\includegraphics[angle=0]{peri_mass.eps}}
426: \resizebox{\hsize}{!}{\includegraphics[angle=0]{f2.eps}}
427: \caption{Average periastron as a function of a) the primary mass $M_1^*$, b) the mass ratio $q = M_2^*/M_1^*$
428: and c) time. Here only encounters where considered were the binary remained bound for at least one entire orbit.
429: These are averaged over all times up to 2Myr. In b) the periastra of the most massive star for the time interval
430: 0.3-3.0 Myr are shown as an exemplary case. c) shows the average periastron for different primaries separately:
431: $M^*_1$=50 \Msun, 30 \Msun $ < M^*_1 <$ 40 \Msun, 2 \Msun $< M^*_1 <$ 30 \Msun, and $M^*_1 <$ 2 \Msun.}
432: \label{fig:peri_mass}
433: \end{figure}
434:
435:
436: \begin{figure}
437: %\resizebox{!}{\vsize}{\includegraphics[angle=0]{frac_enc_m2.eps}}
438: \resizebox{\hsize}{!}{\includegraphics[angle=0]{f6.eps}}
439: \caption{a)The average relative companion mass $M_2^*/M_1^*$ as a function of time for
440: capturing encounters for $M_1^*$ = 50 \Msun. The line is no proper fit, but just meant as a guide line. b) Distribution of companion mass up to
441: 0.1 Myr and 2 - 5 Myr for the most massive star and c) for stars with $M_1^* <$ 10 \Msun.}
442: \label{fig:m2_dist}
443: \end{figure}
444:
445:
446:
447: Fig~\ref{fig:peri_mass}b) shows the average periastron as a function of the mass ratio $q = M_2^*/M_1^*$
448: for the most massive star for the time interval 0.3-3.0 Myr. Due to the random selection process for the
449: stellar masses in the initial
450: cluster set-up, the second most massive star usually has a mass in the range of 30-40 \Msun, so the
451: statistics become increasingly worse for q $>$ 0.7. To a good approximation the average periastron
452: increases linearly with higher mass ratios from $\sim$ 50-65 AU for $q$ = 0.1 to $\sim$ 160 AU for
453: $q$ = 0.7. This reflects the fact that lower mass stars can only replace a more massive star if they orbit on a
454: smaller semi-major axis. In other words, the more equal the masses of the stars bound to each other,
455: the larger the periastron of a relatively
456: stable configuration can be.
457:
458: The average periastron as a function of cluster age is shown in
459: Fig.~\ref{fig:peri_mass}c). Generally there is a wide spread in periastra data obtained in single simulation runs.
460: TBS with large periastra are usually of short duration, so it seems more appropriate to average over
461: the periastron weighted by the duration of the bound state. This reflects the
462: observations which are more likely to detect long-lasting bound systems than short
463: ones. The average periastron of the 50 \Msun-star strongly depends on the cluster age, decreasing from $\sim$
464: 500-600 AU at 0.1 Myr to $\sim$ 20-30 AU at 5 Myr. Exchange processes lead to increasingly stronger bound TBS
465: over time. For the stars in the mass group $M_1^*$ $\sim$ 30-40 \Msun\ the trend is less
466: pronounced, decreasing from $\sim$ 700 AU at 0.1 Myr to $\sim$ 350-450 AU at 5 Myr. For lower mass stars
467: the periastron does not significantly alter over the considered timespan and is the larger for lower
468: primary masses. At 1-2 Myrs, the most likely age of the ONC, the average periastron of the most
469: massive TBS is between 50-200 AU.
470:
471: Next we consider the two properties that observations usually come provide: the
472: average mass ratio $q=M_2^*/M_1^*$ and the binary frequency.
473:
474: Fig.~\ref{fig:m2_dist}a shows the average relative mass ratio $M_2^*/M_1^*$ of the TBS for the most massive stars.
475: It can be seen that in the encounter-averaged case the
476: relative mass of the companion increases from very low values ($< 0.1 M_1^*$) in the early cluster development
477: to $q$-values $\sim$ 0.4 at 3 Myr. If we weigh the q-values by the encounter duration, the general trend
478: remains the same but shifts to somewhat higher values.
479: Our simulations show in addition the difference in this development for primaries with low and high mass.
480: %This relatively high value results from the fact that
481: %relative mass range from which to choose the companion is much smaller.
482: Looking at the distribution of companion masses one sees in Fig.~\ref{fig:m2_dist}b that
483: for the most massive star at early times (t$<$0.2Myr) mass ratios $q <$0.1 dominate,
484: whereas capturing encounters with more massive stars become increasingly important later on in the
485: cluster development.
486: The mass ratio in these TBS develops from an initial preference of low-mass
487: companions to companions with high mass. At an cluster age of 2-5Myr the maximum of $q$ lies in the
488: range of 0.6-0.8, which is even more pronounced, if we consider only TBS which last at least 1000 yrs.
489: The value of $q$ is not the expected equal mass binary value of 1, because the mass of the second most massive
490: star varies in the different simulations between 30-40\Msun . In addition, the most massive star usually
491: captures just one of the 10 most massive stars as companion and not necessarily the second most massive star.
492:
493: This result of high q-values later on in the cluster development correlates very well with the observations
494: which show that for the massive stars the mass ratio $q$ seems to be much smaller in young cluster than for
495: field stars \cite{goodwin:06}. How quickly this development happens will depend crucially on the cluster properties
496: such as its density.
497:
498: %The initial eccentricity seems not to be a crucial parameter in capturing encounters.
499: %Fig~\ref{fig:ecc_m2m1} shows the average eccentricity as a function of
500: %Independently of $M_2^*/M_1^*$ and the cluster age, the initial eccentricity is on average $\sim$ 0.7-0.75.
501: %In most bound systems the eccentricity does not change more than 1-2\% before the system
502: %splits up again. However, there are some systems where the eccentricity changes considerably over time,
503: %but the statistics are too low here to give reliable results on general trends of these processes.
504: %Nevertheless,
505: What about the eccentricity in these bound systems?
506: Fig.~\ref{fig:ecc_peri} shows a very clear connection between eccentricity and periastron:
507: the larger the periastron the smaller the average eccentricity. This is simply what one would expect
508: as a consequence of two-body relaxation. Obviously, whereas a binary with a
509: 1 AU periastron can tolerate a 0.9 eccentricity, a binary with a 2000 AU periastron has on average
510: only a 0.55 eccentricity in the here investigated system. If we consider only systems remaining bound
511: for at least 10 orbits, the decline is even faster, but levels of at about 200 AU at an eccentricity of
512: $\sim$0.5. This faster decline reflects that large periastron TBS only stay stable if their eccentricity is low,
513: otherwise they are likely to undergo strong interactions with the other cluster stars. We saw before that the
514: TBS at 1-2Myrs have on average a periastron of 50-200AU. Using Fig.~\ref{fig:ecc_peri} this is coupled to
515: an average eccentricity $\epsilon \sim$ 0.5-0.6.
516:
517:
518: \begin{figure}
519: %\resizebox{\hsize}{!}{\includegraphics[angle=-90]{ecc_peri.ps}}
520: \resizebox{\hsize}{!}{\includegraphics[angle=-90]{f4.ps}}
521: \caption{Eccentricity as a function of periastron. Here only encounters where considered were the
522: binary remained bound for at least one entire orbit. These are averaged over all times up to 2Myr. Only
523: $M_2^*/M_1^*<1$ are considered. The line is no proper fit, but just meant as a guide line.}
524: \label{fig:ecc_peri}
525: \end{figure}
526:
527:
528: The trend towards more stable configurations, especially for the most massive stars, is most
529: dramatically reflected in the average duration of TBS.
530: %For the most massive star
531: %the average duration of a bound state, $t_{bound}$, increases from $\leq$ 1000 yrs in the early
532: %stages of cluster development ($t_{cluster} <$ 0.3Myr) to several 10$^5$yrs for cluster ages above 1-2 Myrs.
533: %
534: %Even at later stages of the cluster development
535: %the relatively stable phases are sometimes interrupted by periods of shortly bound states
536: %followed by the re-formation of more stable systems.
537: %
538: The dependence of the average duration of a bound state on the stellar mass $M_1^*$ of the primary can be
539: seen in Fig.~\ref{fig:binary_time}. It shows that
540: generally the bound states last on average longer the higher the primary mass. This is most pronounced for stars with
541: $M_1^* > $ 10 - 20 \Msun\ where the bound state lasts on average several times $10^6$yrs in contrast to
542: bound state durations of on average $< 10^5$ yrs for lower mass stars. So apart from the very early
543: stages of the cluster developement ($<$ 0.1Myr) these systems stay stable for many orbital periods.
544:
545: %Fig~\ref{fig:binary_time}b) shows the average duration of TBS as a function of $M_2^*/M_1^*$ for
546: %the most massive star for a cluster age $t_{cluster}$=0.3-3Myr.
547: The duration of the binary phase $t_{bound}$ increases for larger
548: $M_2^*/M_1^*$, these are the more stable configurations that are formed later in the cluster development.
549: For the same reason, we see in our simulations an increase of $t_{bound}$
550: %IF we look at the dependence of the binary duration on the periastron ( see Fig~\ref{fig:dur_peri}) we see
551: %as expected a decline of the binary duration
552: with larger periastra and smaller eccentricities (a strong increase for $\epsilon < 0.2$).
553: This formation of more stable configurations is reflected
554: %in Fig.~\ref{fig:binary_time}c)
555: by a dramatic increase of the average duration of a bound state with cluster age. Whereas bound states last on average
556: only a few 10-100 yrs at $t_{cluster}$ = 0.1 Myr, it rises to a few $10^6$ yrs at $t_{cluster}$ = 2 Myr.
557: So in very young clusters these massive stars would {\em appear} as binaries but are actually just running
558: through a succession of TBS.
559:
560: As the duration of the bound state is a strong function of time so is the number of orbits.
561: %Initially such capturing encounters rarely lead to the completion
562: %of a whole orbit. Even when taking the time-weighted average which would reflect more the
563: %observational values, the average number of orbits increases from $\sim$ 100 at $t_{cluster}$ = 0.1 Myr to several
564: %ten thousand by $t_{cluster}$=2.0 Myr(see Fig.~\ref{fig:orbit_time}).
565: It is initially higher but increases less rapidly for lower mass primary stars. So that after 1 Myr the binaries
566: involving the most massive stars stay bound for a larger
567: number of orbits ($\sim$ several 1000).
568:
569:
570:
571: \begin{figure}
572: %\resizebox{!}{\vsize}{\includegraphics[angle=0]{duration.eps}}
573: \resizebox{\hsize}{!}{\includegraphics[angle=-90]{duration_m1.ps}}
574: \caption{Average duration of a bound state, $t_{bound}$, as function of the primary mass $M_1^*$.
575: The line is no proper fit, but just meant as a guide line.}
576: \label{fig:binary_time}
577: \end{figure}
578:
579:
580:
581: \section{Discussion and Conclusions}
582:
583: The results of our simulations of capturing processes in an ONC-like system can be summarized by the following:
584: %
585: In cluster environments similar to the ONC, massive stars have a much higher probability of
586: involvement in a capturing encounter than solar-mass stars. Assuming a cluster age of 1-2 Myr,
587: at least 10-15\% of the OB stars in the ONC are in a bound state caused by capturing processes,
588: %The properties of the so-formed bound systems involving massive stars
589: %strongly depend on the cluster age with a development towards smaller periastra,
590: %higher mass ratios and longer duration of the bound states.
591: %Assuming a cluster age of 1-2 Myr for the ONC
592: with the average periastron of the so-formed TBS being $r_p \sim$ 50-200 AU,
593: the mass ratio $q \sim$ 0.4-0.5 and the eccentricity $\epsilon \sim$ 0.6.
594: In observations these bound systems would be regarded as normal binaries, even if the bound state is
595: of short duration($\sim 10^{3}$ - 10$^6$ yr).
596:
597: So we have to discuss two questions: i) is the number of TBS formed sufficiently high to explain the difference
598: in binary between solar-type and massive stars and ii) are the observed properties of these TBS in accordance
599: with observations?
600:
601: For simplicity the simulations presented here started with all stars initially being single. In reality,
602: there exists a primordial binary fraction. We saw that for massive stars, capture processes quickly lead
603: to additional bound systems, so that there are two binary
604: populations present - the primordial binaries and the TBS formed by capture processes.
605: As our simulations show that at least 10-15\% of the OB stars in the ONC are in a bound state caused by capturing
606: processes, the observed higher binary frequency for massive stars compared to solar-mass stars
607: must reflect therefore, at least partly, the presence of these transient capture-formed binaries.
608:
609: The value of 10-15\% can be regarded as a lower limit, because most, if not all, stars in the cluster
610: are initially surrounded by discs.
611: The large radii of the discs compared to the stellar size results in a much larger cross-section
612: for encounters, an effect that has not been included here. The larger capture cross-section will
613: influence the overall dynamics, leading to more bound systems.
614: Moeckel \& Bally (2007a,b) studied the disk-assisted capture formation of binaries for stars of $M_1^*$ = 20 \Msun\
615: and estimated that in the Trapezium region, capture could account for $\sim 50 \%$ of massive stars
616: having a companion after 1 Myr.
617: Since the presence of discs around the stars increases the capture cross-section, it is even possible
618: that the fraction of primordial binaries is the same for solar-mass
619: stars as for massive stars and that the observed higher binarity could be {\em exclusively} due to
620: the encounter-induced capture processes. To clarify this, future modelling of the cluster dynamics should
621: include the effect of discs by some means.
622:
623: However, it is most likely that discs only influence capture processes significantly in the very early
624: cluster development, because observational examples for discs around massive stars are usually only found for young,
625: still embedded, protostars \cite{cesaroni:99,tafoya:04,patel:05,kraus:07}. Massive stars in a more
626: evolved stage, like in the Trapezium, are usually not surrounded by discs.
627: Processes like photoevaporation \cite{clarke:07} and gravitational interaction \cite{pfalzner:aa06}
628: are responsible for the disc destruction. Observational and theoretical evidence point to
629: lifetimes of discs around massive stars of approximately $\sim$ 10$^5$ - 10$^6$ years.
630:
631: The second question whether the observed properties of these TBS are in accordance
632: with observations is difficult to answer. The reason is that the ONC contains only 15 OB-stars
633: and as we have shown, the observed binaries are a combination of primordial binaries and
634: those formed in capture processes. We would expect here 2-4 of the massive stars
635: to be part of a TBS - clearly a number where statistics is impossible.
636: Another problem is that observing the Trapezium the surface density of stars is such
637: that the visual binary population at periastra of a few 100 AU is difficult to distinguish
638: from unassociated stars with current spectroscopic methods. However, future interferometric
639: observations of proper motions - with, for example the VLTI, - should be able to provide the
640: neccessary data.
641:
642: Interesting in this context is that recent observations \cite{kraus_s:07} indicate that \COri\ itself is a binary
643: with a massive companion, possibly with \COri\ having a mass of 34 \Msun\ and the companion mass being 15.5 \Msun.
644: Naturally one cannot decide whether a specific binary has formed primordially or as result of a
645: capture process. In this specific case the high mass ratio would speak for a capture process, but the low
646: periastron of $\sim$1.5 AU could occur in both cases - capture-formed or primordial binaries.
647: %As the periastron
648: %is only $\sim$1.5 AU this situation is still comparable to our simulations were we assumed a
649: %50 \Msun\ massive star instead.
650:
651: So although the ONC has the advantage that many observational results are available, it has the
652: disadvantage that it does not contain a high number of massive stars.
653: However, statistical properties in other cluster --- especially in very young ones --- could
654: provide signatures of capture processes. Properties where one would expect observable differences between
655: primordial and capture-formed binaries are the periastron and the mass
656: ratio distribution. Our simulations show that the periastra of the encounter-induced bound
657: systems are usually in the range of a few 10 - 2000 AU for the first Myrs of evolution, so they clearly belong
658: to the so-called wide binaries.
659: The observed over-abundance of wide binaries among PMS stars compared to field stars \cite{patience:02} could
660: possibly be a tracer for these capture-formed binaries. As we have shown here the average periastra in TBS with a massive
661: primary become smaller ($<$ 100 AU) for older clusters, so one should be able to trace
662: this effect by observing the binary frequency of massive stars in clusters of different age but similar structure.
663: The over-abundance of wide binaries should shift to smaller periastron values and become less pronounced with
664: increasing age.
665: The latter is due to the shorter lifetime of the massive stars, so these binaries would naturally disappear
666: if the cluster is beyond a certain age.
667:
668: %Another important signature for capture-formed binaries with massive primary is the mass ratio $q$.
669: Our simulations indicate that capture processes lead to an increase in the mass ratio with cluster age.
670: Investigations of the mass ratio in binaries with massive primary to date give a complex picture: while small $q$
671: (consistent with random pairing) are observed in the ONC \cite{preibisch:99},
672: in rich clusters most O stars are believed to exist as short-period binaries with $q \simeq 1$
673: \cite{garcia:02}. Mass-ratio distributions not consistent with random pairing were also reported
674: by Kouwenhoven (2005) for A and late-type B binaries in Scorpius OB2.
675:
676:
677: So ideally in future this kind of
678: investigation should be performed for a cluster containing a high number of O stars.
679: The question is: what can one expect in systems other than the ONC? In lower density and
680: open clusters the number of bound systems formed by capturing will be too insignificant to contribute
681: to the total binary fraction. However, in high-density clusters like the Arches cluster near
682: the Galactic centre the central mass density ($\sim$ 3 $\times$ 10$^5$ \Msun pc$^{-3}$
683: (Figer et al. 2002)) is about ten times higher than in the
684: Trapezium region, so capturing encounters could play an important role. However, this is
685: difficult to predict because a number of other cluster properties influence the capture rate. Apart from
686: the cluster density, the velocity dispersion and the degree of mass segregation will be significant
687: factors. In the Arches cluster the velocity dispersion is not very well known,
688: only an upper limit of 22 km s$^{-1}$ exists. If the real velocity dispersion is close to this
689: upper limit, it might mean a large number of high velocity encounters, which would decrease
690: the capture rate. The relative importance of these different cluster properties for the capture
691: rate need further investigation.
692:
693: Since the primordial binary fraction seems to depend strongly on the cluster
694: environment, it is important to compare relative values. One way to find indications for the
695: capture-formed massive binaries would be to compare the relative binary frequency of solar-type and massive
696: stars in clusters of different stellar density. If capturing is an important process, the higher
697: frequency of capturing encounters in dense systems would lead to a higher fraction of massive binaries
698: to solar-type binaries there. However, one has to be careful with very dense systems containing many OB-stars,
699: since higher velocity dispersion could counteract this process.
700:
701: The study performed here can only be regarded as a first step towards understanding the
702: significance of capture processes in young dense clusters. In future some points need
703: further investigation. First, we concentrated here on the closest bound partner, whereas observations as well
704: as our simulations indicate that multiple systems are a common feature for massive stars.
705: Second, simulations should start with a finite primordial binary fraction and see how it develops.
706: The massive stars would still capture partners, but it is also known that
707: the dynamical evolution of dense clusters has a destructive effect on primordial binaries
708: \cite{kroupa:95,kroupa:99,kroupa:01}. It would be important to see which of these competing effects dominates.
709: Third, as pointed out above in order to improve the statistics, simulation of systems containing
710: a higher number of OB stars are required.
711: Finally, disc effects have to be included as despite their short lifetimes around massive stars
712: they could not only increase the capture-induced binary rate but also influence
713: the periastron and eccentricity of the formed systems. From our results it is apparent that it is not uncommon for
714: the periastron in capture-formed bound systems to be just a few 10 AU. This means that the secondary
715: star would pass through the disc of the primary. Repeated disc passages can reduce the periastron and
716: eccentricity through gravitational drag. For our
717: simulations this would mean an even faster development in the initial stages.
718:
719:
720: %Recent observations indicate that the binarity of stars might play a key
721: %role in the development of protoplanetary discs. \cite{bouwman:06,prisinzano:06} show
722: %that binary systems have a lower disc frequency than single stars and
723: %concluded that discs dissipation time is $\sim$ 5 Myr for binary systems
724: %and $\sim$ 9 Myr for single star systems.
725:
726: \section*{Acknowledgments}
727:
728: We are grateful to R.Spurzem for providing the Nbody6++ code for the cluster
729: simulations and want to thank H.Zinnecker for his very helpful comments.
730: Simulations were partly performed at the John von
731: Neumann Institute for Computing, Research Centre J\"ulich, Project HKU14.
732:
733:
734: \bibliographystyle{apj}
735: %\begin{thebibliography}{alpha}
736: \begin{thebibliography}{}
737: \bibitem[Aarseth 2003]{aarseth:book03}
738: Aarseth, S. {\em Gravitational N-Body Simulations} 2003, CUP, Cambridge.
739:
740: \bibitem[Abt et al. 1990]{abt:90}
741: Abt, H.A., Gomez, A.E., Levy, S.G. 1990, ApJS, 74, 551.
742:
743: \bibitem[Beck et al. 2003]{beck:03}
744: Beck, T.L., Simon,M., Close, L.M. 2003, \apj, 583, 358.
745:
746: \bibitem[Boss 1992]{boss:92}
747: Boss, A.P. 1992 in "The Realm of Interacting Binary Stars", eds. J. Sahade, G.E. McCluskey, Yoji Kondo, 355.
748:
749: \bibitem[Bouvier 1997]{bouvier:97}
750: Bouvier, J., Rigaut, F., Nadeau, D. 1997, \aap, 323, 139.
751:
752: %\bibitem[Bouwman et al. 2006]{bouwman:06}
753: %Bouwman, J., Lawson, W. A., Dominik, C., Feigelson, E. D., Henning, Th.,
754: %Tielens, A. G. G. M., Waters, L. B. F. M. 2006, astro-ph/0610853.
755:
756: \bibitem[Cesaroni et al. 1999]{cesaroni:99}
757: Cesaroni, R., Felli, M., Jenness, T., Neri, R. Olmi, L., Robberto, M., Testi, L., Walmsley, C.~M.
758: 1999, \aap, 345, 949.
759:
760: \bibitem[Clarke 2007]{clarke:07}
761: Clarke, C.~J. 2007, \mnras, 376, 1350-1356.
762:
763: \bibitem[Clarke \& Pringle 1991]{clarke:91}
764: Clarke, C.~J., Pringle, J.~E. 1991, \mnras, 249, 584.
765:
766: %\bibitem[Duch{\^e}ne 1999]{duchene:99}
767: %Duch{\^e}ne, G. 1999, \aap, 341, 547.
768:
769: \bibitem[Duch{\^e}ne et al. 1999]{duchene:99b}
770: Duch{\^e}ne, G., Bouvier, J., Simon, T. 1999, \aap, 343, 831.
771:
772: \bibitem[Duch{\^e}ne et al. 2004]{duchene:04}
773: Duch{\^e}ne, G., Bouvier, J., Bontemps, S., Andre, P., Motte, F. 2004
774: \aap, 427, 651.
775:
776: \bibitem[Duchquennoy \& Mayor 1991]{duquennoy:91}
777: Duquennoy, A. \& Mayor, M. 1991, \aap, 248, 485.
778:
779: \bibitem[Durisen \& Sterzik 1994]{durisen:94}
780: Durisen, R.H. \& Sterzik, M.F. 1994, \aap, 286, 84.
781:
782: \bibitem[Figer et al. 2002]{figer:02}
783: Figer D. F. et al. 2002, \apj, 581, 258.
784:
785: \bibitem[Fischer \& Marcy 1992]{fischer:92}
786: Fischer, D.A., Marcy, G.W. 1992, \apj, 396, 178.
787:
788: %\bibitem[Ghez et al. 1997]{ghez:97}
789: %Ghez, A. M., McCarthy, D.W., Patience, J., Beck, T. 1997, {\aj}, 481, 378.
790:
791: \bibitem[Garcia \& Mermilliod 2001]{garcia:02}
792: Garcia, B.\& Mermilliod, J. C. 2002, \aap, 386, 122.
793:
794: \bibitem[Goodwin et al. 2006]{goodwin:06}
795: Goodwin, S. P., Kroupa, P., Goodman, A., Burkert, A. 2007,
796: in 'Protostars and Planets V' ed. by Reipurth, D. Jewitt, K. Keil, Univ. Arizona Press, Tucson, p.133.
797:
798: \bibitem[Heggie 1975]{heggie:75}
799: Heggie, D. C. 1975, \mnras, 173,729.
800:
801: \bibitem[Heggie \& Hut 2003]{heggie:03}
802: Heggie, D. C. \& Hut, P., 2003 "The Gravitational Million-Body Problem: A Multidisciplinary Approach to Star
803: Cluster Dynamics", CUP, Cambridge.
804:
805: \bibitem[Hills 1975]{hill:75}
806: Hills, J. G. 1975, \aj, 80, 809..
807:
808: %\bibitem[Hillenbrand 1997]{hillenbrand:aj97}
809: %Hillenbrand, L.~A. 1997, {\aj}, 113, 1733.
810:
811: \bibitem[Hillenbrand \& Hartmann 1998]{hillenbrand:apj98}
812: Hillenbrand, L.~A., Hartmann, L.~W. 1998, {\apj}, 492, 540.
813:
814: \bibitem[K\"ohler et al. 2006]{koehler:06}
815: K{\"o}hler, R., Petr-Gotzens, M.~G., McCaughrean, M.~J., Bouvier, J.,
816: Duch{\^e}ne, G., Quirrenbach, A., Zinnecker, H. 2006, \aap, 458, 461.
817:
818: \bibitem[Kouwenhoven et al. 2005]{kouwenhoven:05}
819: Kouwenhoven, M.B.N., Brown, A.~G.A., Zinnecker, H., Kaper, L.,
820: Portegies Zwart S.F. 2005, \aap, 430, 137.
821:
822: \bibitem[Kraus \& Hillenbrand 2007]{kraus:07}
823: Kraus, A.~M., Hillenbrand, L.~A. 2007, \apj, 662, 413.
824:
825: \bibitem[Kraus et al. 2007]{kraus_s:07}
826: Kraus S., Balega, Y.~Y., Berger, J.-P., Hofmann, K.-H., Millan-Gabet, R.,
827: Monnier, J.~D., Ohnaka, K., Pedretti, E., Preibisch, T., Schertl, D., Schloerb, F.~P.,
828: Traub, W.~A., Weigelt, G. 2007 \aap, 466, 649.
829:
830: \bibitem[Kroupa, Tout \& Gilmore 1993]{kroupa:mnras93}
831: Kroupa, P., Tout, C.~A., Gilmore, G. 1993, \mnras, 262, 545.
832:
833: \bibitem[Kroupa 1995]{kroupa:95}
834: Kroupa, P. 1995, {\mnras}, 277, 1491.
835:
836: \bibitem[Kroupa et al. 1999]{kroupa:99}
837: Kroupa, P., Petr, M., McCaughrean, M. 1999, New Astron., 4, 495.
838:
839: \bibitem[Kroupa 2001]{kroupa:01}
840: Kroupa, P. 2001, \mnras, 322, 231.
841:
842: %\bibitem[Leinert et al. 1993]{leinert:93}
843: %Leinert, C., Zinnecker, H., Weitzel, N., Christou, J., Ridgway, S.~T.,
844: %Jameson, R., Haas, M., Lenzen, R. 1993, New Astron., 4, 495.
845:
846: \bibitem[Liu et al 2003]{liu:03}
847: Liu, W.M., Meyer, M.R., Cotera, A.S., Young, E.T. 2003, \aj, 126, 1665.
848:
849: %\bibitem[Luhman et al. 2000]{luhman:apj00}
850: % {{Luhman}, K.~L., {Rieke}, G.~H., {Young}, E.~T., {Cotera}, A.~S., {Chen}, H., {Rieke}, M.~J., {Schneider}, G., {Thompson}, R.~I.} 2000, {\apj}, 540, 1016.
851:
852: \bibitem[Luhman et al. 2005]{luhman:05}
853: Luhman, K.L., McLeod, K.K., Goldenson, N. 2005, \apj, 623, 1141.
854:
855: \bibitem[Mason 1998]{mason:98}
856: Mason, B.D., Gies, D.R., Hartkopf, W.I. et al. 1998, AJ 115, 821.
857:
858: \bibitem[Mathieu 1994]{mathieu:94}
859: Mathieu, R. D. 1994, Ann. Rev. Astron. Astroph., 32, 465
860:
861: \bibitem[Mathieu et al. 2000]{mat:00}
862: Mathieu, R. D., Ghez, A. M., Jensen, E. L. N., Simon, M. 2000,
863: in Protostars and Planets IV, eds Mannings, V., Boss, A.P., Russell, S. S.
864: Univ.Arizona Press.
865:
866: \bibitem[McCaughren 2001]{mccaughrean:01}
867: McCaughrean, M.J. 2001 in The Formation of Binary Stars, ed H. Zinnecker
868: \& R.D. Mathieu, ASP Conf. Ser. Proc. IAU Symp., 211, 311.
869:
870: \bibitem[McCaughren et al. 2002]{mccaughrean:msngr02}
871: {{McCaughrean}, M., {Zinnecker}, H., {Andersen}, M.,
872: {Meeus}, G., {Lodieu}, N.} 2002, {\em The Messenger}, 109, 28.
873:
874: \bibitem[Moeckel \& Bally 2007a]{moeckel:07a}
875: Moeckel, N. \& Bally, J. 2007, \apj 656, 275.
876:
877: \bibitem[Moeckel \& Bally 2007b]{moeckel:07b}
878: Moeckel, N. \& Bally, J. 2007, \apj 661, L183.
879:
880: %\bibitem[Mouschovias 1977]{mouschovias:apj77}
881: %Mouschovias, T. Ch. 1977, {\em ApJ}, 211, 147.
882:
883: \bibitem[Olczak et al. 2006]{olczak:06}
884: Olczak, C., Pfalzner, S., Spurzem, R. 2006, \apj, 642, 1140.
885:
886: \bibitem[Padgett et al. 1997]{padgett:97}
887: Padgett, D.L., Strom, S.E., Ghez, A. 1997, \apj, 477, 705.
888:
889: \bibitem[Patel et al. 2005]{patel:05}
890: Patel, N.~A., {Curiel}, S., Sridharan, T.~K., Zhang, Q., Hunter, T.~R., Ho, P.~T.~P.,
891: Torrelles, J.~M., Moran, J.~M., Gomez, J.~F., Anglada, G. 2005, Nature, 437, 109.
892:
893: \bibitem[Patience et al. 1998]{patience:98}
894: Patience, J., Ghez, A.M., Reid, I.N., Weinberger, A.J., Matthews, K. 1998, \aj, 115, 1972.
895:
896: \bibitem[Patience et al. 2002]{patience:02}
897: Patience, J., Ghez, A. M., Reid, I. N., Matthews, K. 2002, \aj, 123, 1570.
898:
899: \bibitem[Petr 1998]{petr:98}
900: Petr, M.G., Coude de Foresto, V., Beckwith, S.V.W., Richichi, A., McCaughren, M.J. 1998, \apj, 500, 825.
901:
902: \bibitem[Pfalzner et al. 2006]{pfalzner:aa06}
903: Pfalzner, S., Olczak, C., Eckart, A. 2006, \aap, 454, 811.
904:
905: %\bibitem[Pfalzner 2006]{pfalzner:apj06}
906: %Pfalzner, S. 2006, \apj, 452, L129.
907:
908: \bibitem[Preibisch et al. 1999]{preibisch:99}
909: Preibisch, T., Balega, Y., Hofmann, K.-H. Weigelt, G., Zinnecker, H. 1999, New Astronomy, 7, 531.
910:
911: %\bibitem[Prisinzano et al. 2006]{prisinzano:06}
912: %Prisinzano, L., Damiani, F., Micela, G., Pillitteri, I. 2006, astro-ph/0610901.
913:
914: \bibitem[Prosser et al. 1994]{prosser:94}
915: Prosser, C.F., Stauffer, J.R., Hartmann, L. et al. 1994, \apj, 421, 517.
916:
917: \bibitem[Reid \& Gizis 1997]{reid:97}
918: Reid, I.N., Gizis, J.E. 1997 \aj, 113, 2246.
919:
920: \bibitem[Scally et al. 1999]{scally:99}
921: Scally, A., Clarke, C., McCaughren, M.J. 1999, \mnras, 306, 253.
922:
923: %\bibitem[Shu et al 1987]{shu:aara87}
924: %Shu, F.H., Adams, F.C. Lizano, S. 1987, {\em ARA \& A}, 25, 23.
925:
926: \bibitem[Simon et al. 1999]{simon:99}
927: Simon, M., Close, L.M., Beck, T.L. 1999, \aj, 117, 1375.
928:
929: \bibitem[Spurzem 1999]{spurzem:mnras02}
930: Spurzem, R. 1999, Comp. Astroph., 109, 407.
931:
932: \bibitem[Sterzik et al. 2003]{sterzik:03}
933: Sterzik, M.F., Durisen, R.H., Zinnecker, H. 2003, \aap, 411, 91.
934:
935: \bibitem[Tafoya et al. 2004]{tafoya:04}
936: Tafoya, D., Gomez, Y., Rodriguez, L.~F., 2004, \apj, 610, 827.
937:
938: \bibitem[Tohline et al. 1999]{tohline}
939: {Tohline}, J.~E., {Cazes}, J.~E., {Cohl}, H.~S. 1999
940: in Numerical Astrophysics, Astrophysics and Space Science Library 240, ed. {Miyama}, S.~M.,
941: {Tomisaka}, K., {Hanawa}, T., p.155.
942:
943: \bibitem[Yorke \& Zinnecker 2007]{yorke:07}
944: Yorke, H. W., Zinnecker, H. Z., to appear in Ann. Rev. Astron. Astrophys.
945:
946: \end{thebibliography}
947:
948:
949:
950:
951:
952: %\begin{figure}
953: %\resizebox{\hsize}{!}{\includegraphics[angle=-90]{peri_m1.ps}}
954: %\caption{Average periastron as a function of $M_1^*$. Here only encounters where considered were the
955: %binary remained bound for at least one entire orbit. These are averaged over all times up to 13 Myr.}
956: %\label{fig:peri_m1}
957: %\end{figure}
958: %
959: %\begin{figure}
960: %\resizebox{\hsize}{!}{\includegraphics[angle=-90]{peri_m2m1.ps}}
961: %\caption{Average periastron as a function of the mass ratio $q = M_2^*/M_1^*$. Here only encounters where
962: %considered were the binary remained bound for at least one entire orbit.
963: %These are averaged over all times up to 2Myr. Only $M_2^*/M_1^*<1$ are considered.}
964: %\label{fig:peri_m2m1}
965: %\end{figure}
966: %
967: %\begin{figure}
968: %\resizebox{\hsize}{!}{\includegraphics[angle=-90]{peri_time.ps}}
969: %\caption{a) Average periastron as a function of time for for different primaries: a) $M^*_1$=50 \Msun,
970: %b) 10 \Msun $ < M^*_1 <$ 50 \Msun, c) 2 \Msun $< M^*_1 <$ 10 \Msun, and $M^*_1 <$ 2 \Msun.}
971: %\label{fig:peri_time}
972: %\end{figure}
973:
974:
975: %\begin{figure}
976: %\resizebox{\hsize}{!}{\includegraphics[angle=-90]{orbit_time2.ps}}
977: %\resizebox{\hsize}{!}{\includegraphics[angle=-90]{f5.ps}}
978: %\caption{Average number of orbits of TBS for two mass regimes, $M^*_1$=50 \Msun\ and $M^*_1$<2 \Msun, as a function of time.}
979: %\label{fig:orbit_time}
980: %\end{figure}
981:
982:
983: %\begin{figure}
984: %\resizebox{\hsize}{!}{\includegraphics[angle=-90]{eccent_time.ps}}
985: %\caption{Average eccentricity as a function of time for different primaries: a) $M^*_1$=50 \Msun,
986: %b) 10 \Msun $ < M^*_1 <$ 50 \Msun, c) 2 \Msun $< M^*_1 <$ 10 \Msun, and $M^*_1 <$ 2 \Msun.}
987: %\label{fig:ecc_time}
988: %\end{figure}
989:
990: %\begin{figure}
991: %\resizebox{\hsize}{!}{\includegraphics[angle=-90]{binary_star.ps}}
992: %\caption{Average duration of bound state in an time interval of 0.1Myr as a function of time for different primaries: a) $M^*_1$=50 \Msun,
993: %b) 10 \Msun $ < M^*_1 <$ 50 \Msun, c) 2 \Msun $ < M^*_1 <$ 10 \Msun, and $M^*_1 <$ 2 \Msun.}
994: %\label{fig:binary_star}
995: %\end{figure}
996:
997: %\begin{figure}
998: %\resizebox{\hsize}{!}{\includegraphics[angle=-90]{diff_ecc.ps}}
999: %\caption{Relative change in eccentricity during bound stages}
1000: %\label{fig:diff_ecc}
1001: %\end{figure}
1002:
1003: %Fig.~\ref{fig:diff_ecc} shows the change in eccentricity once the (transient)
1004: %binaries have formed. At seems that high mass binaries are hardened by the cluster environment,
1005: %whereas for lower mass binaries the cluster environment has hardening and weakening effects.
1006:
1007: %\begin{figure}
1008: %\resizebox{\hsize}{!}{\includegraphics[angle=-90]{duration_m1.ps}}
1009: %\caption{Average duration of a binary phase as a function of $M_1$. Here only encounters where considered were the
1010: %binary remained bound for at least one entire orbit. These are averaged over all times up to 2Myr.}
1011: %\label{fig:duration_m1}
1012: %\end{figure}
1013:
1014: %\begin{figure}
1015: %\resizebox{\hsize}{!}{\includegraphics[angle=-90]{ecc_m2m1.ps}}
1016: %\caption{Average eccentricity as a function of $M_2/M_1$. Here only encounters where considered were the
1017: %binary remained bound for at least one entire orbit. These are averaged over all times up to 2Myr. Only
1018: % $M_2/M_1<0$ are considered.}
1019: %\label{fig:ecc_m2m1}
1020: %\end{figure}
1021:
1022: %\begin{figure}
1023: %\resizebox{\hsize}{!}{\includegraphics[angle=-90]{dur_m2m1.ps}}
1024: %\caption{Average duration of binary phase as a function of $M_2/M_1$. Here only encounters where considered were the
1025: %binary remained bound for at least one entire orbit. These are averaged over all times up to 2Myr. Only
1026: % $M_2/M_1<0$ are considered.}
1027: %\label{fig:dur_m2m1}
1028: %\end{figure}
1029:
1030:
1031: %\begin{figure}
1032: %\resizebox{\hsize}{!}{\includegraphics[angle=-90]{dur_peri.ps}}
1033: %\caption{Binary duration as a function of periastron. }
1034: %\label{fig:dur_peri}
1035: %\end{figure}
1036:
1037: %\begin{figure}
1038: %\resizebox{\hsize}{!}{\includegraphics[angle=-90]{dur_ecc.ps}}
1039: %\caption{Duration as function of eccentricity. Here only encounters where considered were the
1040: %binary remained bound for at least one entire orbit. These are averaged over all times up to 2Myr. Only
1041: % $M_2/M_1<0$ are considered.}
1042: %\label{fig:dur_ecc}
1043: %\end{figure}
1044:
1045: \end{document}