Icarus 208 (2010) 824–836 Contents lists available at ScienceDirect Icarus journal homepage: www.elsevier.com/locate/icarus Three-body capture of irregular satellites: Application to Jupiter Catherine M. Philpott a,*, Douglas P. Hamilton a, Craig B. Agnor b a Department of Astronomy, University of Maryland, College Park, MD 20742-2421, USA b Astronomy Unit, School of Mathematical Sciences, Queen Mary University of London, London E14NS, UK article info abstract Article history: We investigate a new theory of the origin of the irregular satellites of the giant planets: capture of one Received 19 October 2009 member of a 100-km binary asteroid after tidal disruption. The energy loss from disruption is sufficient Revised 24 March 2010 for capture, but it cannot deliver the bodies directly to the observed orbits of the irregular satellites. Accepted 26 March 2010 Instead, the long-lived capture orbits subsequently evolve inward due to interactions with a tenuous cir- Available online 7 April 2010 cumplanetary gas disk. We focus on the capture by Jupiter, which, due to its large mass, provides a stringent test of our model. Keywords: We investigate the possible fates of disrupted bodies, the differences between prograde and retrograde Irregular satellites captures, and the effects of Callisto on captured objects. We make an impulse approximation and discuss Jupiter, Satellites Planetary dynamics how it allows us to generalize capture results from equal-mass binaries to binaries with arbitrary mass Satellites, Dynamics ratios. Jupiter We find that at Jupiter, binaries offer an increase of a factor of 10 in the capture rate of 100-km objects as compared to single bodies, for objects separated by tens of radii that approach the planet on relatively low-energy trajectories. These bodies are at risk of collision with Callisto, but may be preserved by gas drag if their pericenters are raised quickly enough. We conclude that our mechanism is as capable of pro- ducing large irregular satellites as previous suggestions, and it avoids several problems faced by alterna- tive models. Ó 2010 Elsevier Inc. All rights reserved. 1. Introduction This is problematic, however, because if the gas disk does not rar- efy substantially in 100–1000 years, the orbits of the new satel- 1.1. Previously suggested capture models lites will decay inward, leading to collisions with the planet or its regular satellites. Furthermore, the atmospheres of Uranus With discoveries accelerating in the last decade, we now know and Neptune have only a few Earth-masses of hydrogen and he- of over 150 satellites orbiting the giant planets. About one-third of lium at present, so their gas disks could not have been as extensive these are classified as regular, with nearly circular and planar or- or long-lived as those of Jupiter and Saturn. A likely outcome of this bits. It is thought that these satellites are formed by accretion in model, then, is that satellite capture should have been different at circumplanetary disks. The majority of the satellites, however, Jupiter and Saturn than at Uranus and Neptune; however, current are irregular and follow distant, highly eccentric and inclined observational estimates suggest roughly equal efficiencies (Jewitt paths. It is widely believed that irregular satellites originated in and Sheppard, 2005). With a model similar to that of Pollack heliocentric orbits and were later captured by their planets, but et al. (1979), C´ uk and Burns (2004a) found that Jupiter’s largest the details of how this occurred are still uncertain. At least seven irregular satellite, Himalia, would evolve inward to its current orbit different models have been proposed, involving dissipative forces, in 104–106 years. This tenuous gas, however, may make capture collisions, resonances, and three-body effects. Each model has its difficult. own strengths and weaknesses. In another model, planetesimals are captured when the mass of In one long-standing theory, planetesimals are slowed as they the planet increases (Heppenheimer and Porco, 1977; Vieira Neto punch through the gas disk surrounding a young, growing planet et al., 2004). This mass growth causes the planet’s escape velocity (Pollack et al., 1979). For this mechanism to be efficient, the gas to increase, rendering a previously free planetesimal bound to the must be sufficiently dense to capture the planetesimals in one pass. planet. For this method to be effective, the planet’s mass must in- crease substantially during the time that planetesimals linger near the planet, 100–1000 years. However, in most planet formation * Corresponding author. Fax: +1 301 314 9067. models (e.g. Pollack et al., 1996), giant planet growth is hypothe- E-mail addresses: [email protected] (C.M. Philpott), [email protected] sized to take place on timescales many orders of magnitude longer d.edu (D.P. Hamilton), [email protected] (C.B. Agnor). 0019-1035/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.icarus.2010.03.026 C.M. Philpott et al. / Icarus 208 (2010) 824–836 825 than required by this capture scenario. Furthermore, Uranus and tion preferentially occurred when the two components were Neptune’s gas deficiency implies that their growth was of very aligned with the planet. In addition, they found that permanent short duration. Thus, our current understanding of planetary for- capture occurred most often when the binary was disrupted 90° mation makes this model improbable. after it passed between the Sun and Jupiter in its orbit. The observation that the four giant planets contain approxi- mately the same number of irregular satellites (accounting for 1.2. Capture from 100-km binaries observational biases; Jewitt and Sheppard, 2005) has led to a re- newal of interest in capture theories that do not depend strongly All of the above models have promising aspects coupled with on the planet’s formation process. In one such scenario, a planetes- important limitations. In this work, we seek to combine the best imal collides with a current satellite or another planetesimal in the features of several models into a viable capture scenario. In partic- vicinity of the planet, resulting in its capture (Colombo and Frank- ular, we examine binaries (as in Agnor and Hamilton (2006a), Vok- lin, 1971). Though collisions were certainly more common in the rouhlicky´ et al. (2008), and Gaspar et al. (2010)) as a way to early Solar System than they are today, if they resulted in enough augment capture from low-velocity orbits resulting from three- energy loss to permit capture, they would likely also have cata- body interactions like those studied by Astakhov et al. (2003). strophically disrupted the bodies. Nevertheless, the fragments While Vokrouhlicky´ et al. (2008) studied exchange reactions in might then have become independent satellites. the context of an assumed initial planetesimal population, we fo- Astakhov et al. (2003) examined low-energy orbits that linger cus on assessing the viability of the mechanism itself. Our goal is near Jupiter and Saturn. While these bodies are not permanently to determine how various parameters of the model affect its plau- captured, the authors found that some of them were stable for sibility. We examine its viability at Jupiter, as a number of the thousands of years, long enough to allow a weak dissipative force above models suggest that capturing at the largest gas giant is such as gas drag to complete the capture process. However, the especially difficult. overall percentage of temporary captures that do not escape is As the largest of the existing irregular satellites are 80– small, and many of these bodies are threatened by collision with 110 km, capture of objects in this size range is particularly interest- the planets’ large outer satellites (e.g., Callisto and Titan). ing. Since it is likely that the irregular satellite population contains Agnor and Hamilton (2006a) examined the capture of Triton collisional families (Nesvorny´ et al., 2003; Sheppard and Jewitt, from an exchange reaction between a binary pair and Neptune. 2003), it may be the case that only the largest objects were cap- Their motivation stemmed from the newly-discovered abundance tured, while the smaller satellites formed later, via collisions. For of binaries in small-body populations. Currently, it is estimated this reason, we focus our investigation on capturing the 100- that binaries account for 30% of Kuiper belt objects (KBOs) with km progenitors. inclinations <5°, 5% of the rest of the KBOs (Noll et al., 2008), In order to stabilize and shrink the resulting capture orbits, a and 2% of large main belt asteroids (diameters >20 km; this per- dissipation source is required; we suggest a tenuous version of centage increases for smaller objects; Merline et al., 2007). In Ag- the gas drag originally proposed by Pollack et al. (1979). Two of nor and Hamilton’s capture model, a binary is tidally disrupted Jupiter’s irregular satellites, Pasiphae and Sinope, as well as Sat- and one of its members, Triton, is captured as a satellite. This pro- urn’s satellites, Siarnaq and Narvi, and Uranus’ Stephano are found cess is most effective for large satellites like Triton, with radius in resonances or near-resonances that may require just such a 1350 km. However, the largest of the other irregular satellites are weak dissipative force (Whipple and Shelus, 1993; Saha and Tre- more than 10 times smaller than Triton: Himalia at Jupiter is maine, 1993; C´ uk et al., 2002; Nesvorny´ et al., 2003; C´ uk and 85 km in radius, Saturn’s largest irregular, Phoebe, is 110 km, Burns, 2004b; Beaugé and Nesvorny´ , 2007). Uranus’s Sycorax is 80 km, and Neptune’s Halimede and Neso Furthermore, a tenuous circumplanetary disk is consistent with are only 30 km each.