The water vapor plume on the geological-active south-polar region of the moon Enceladus is recognized as the main source of Saturn's neutral torus centered on the Enceladus orbit. The composition of the torus is dominated by water group species. Recent in situ Cassini plasma spectrometer measurements indicate the existence of freshly produced, slow and non-thermalized water group ions throughout the Enceladus torus including regions far from the moon. We report the results of modeling spacecraft-plasma interactions in the environment relevant for the Enceladus torus to show that new-born non-thermalized ions will inevitably be captured by the electric fields arising around the charged spacecraft. The associated plasma configuration can directly impact the plasma measurements and thus is important for reliable interpretation of data obtained by Cassini instruments in the Enceladus torus. The simulation results appear to be partially supported by Cassini observations and can provide new insights into intricate process of Enceladus-plasma interactions.
 The plume of gas, water vapor and icy grains emanating from the south polar region of the icy moon Enceladus imaged by Cassini in 2005 [Porco et al., 2006] represents one of the major discoveries made during the Cassini mission. Models predict [Johnson et al., 2005] that the Enceladus plume produces a radially narrow ∼ 1RS (Saturn radius RS = 60268 km) and dense torus of water-group neutral atoms and molecules centered on Enceladus' orbit. Charge exchange collisions replace a fraction of the corotating ions with the new and slower ion population. This remarkable process of production of fresh ions was first detected by the Cassini plasma spectrometer (CAPS) during equatorial Cassini flybys in 2005 [Tokar et al., 2008]. It was found that the newly-created ions move near the local Kepler speed, and are “picked-up” by Saturn's magnetic field. The fresh ions identified in the CAPS measurements via the characteristic ring-like signatures in the ion velocity distributions have been observed throughout the Enceladus torus including regions far from the icy moon [Tokar et al., 2008]. The purpose of this study is to explore spacecraft (SC) - plasma interactions in the environment relevant for the Enceladus torus. Using formalism, developed in our earlier paper [Yaroshenko et al., 2011], we apply a three-dimensional particle-in-cell (PIC) self-consistent code to study how the Saturn's magnetosphere interacts with an orbiter. If previously, the SC charging in Saturn's magnetosphere has been studied in the assumption that all plasma species corotate with the planetary magnetic field [Olson et al., 2010; Yaroshenko et al., 2011], the simulations now involve two types of the water group ions: those that corotate with the planetary magnetic field and the new-born non-thermalized ions moving with nearly Kepler speed. The resulting novel type of plasma distribution around SC is investigated for the parameters consistent with estimates deduced from CAPS measurements atr/RS ∼ 4 (r being an orbital distance from Saturn) [Tokar et al., 2008; Sittler et al., 2006, 2008].
 CAPS measurements in the inner magnetosphere performed in the last quarter of 2005 (revolutions 16–19) are of special interest since Cassini traversed the dense Enceladus torus almost azimuthally around the moon orbit (within the vertical distances away from the equatorial plane of Z ≤ 0.03RS) and the instrument reliably registered the new-born slow ions throughout the Enceladus torus [Tokar et al., 2008]. As an example we give in Figure 1 the equatorial projection of the revolution 19 in Saturn Equatorial coordinates (http://www-pw.physics.uiowa.edu/jbg/cas.html for Cassini ephemeris database, University of Iowa). Cassini passed along the Enceladus torus at spacecraft event time SCET358:19:00–24:00h on 24, December 2005 traveling at distances r/RS ∼ (4.8 − 4.6). The CAPS data provide the following estimations of the core plasma parameters: the total plasma density of ne,0 ∈ (40, 70) cm−3 and the plasma temperatures: Te ∈ (1, 2) eV and Ti ∈ (20, 30) eV [Tokar et al., 2008, 2009].
Table 1contains the input parameters adopted in the numerical analysis. For the freshly-produced ions we take the average massmi = 18 amu, the ion density nb0 ≃5 cm−3 and the thermalization time of ∼ 103 s [Tokar et al., 2008]. At the time scales of a few ion plasma periods (sufficient to acquire an equilibrium SC charge and plasma configuration around an object in our PIC simulations) the new-born ions thus remain cold, having the temperatures close to the temperature of the parent neutral gas. In the Enceladus rifts, the gas temperatureTg is predicted to be in the range Tg ∈ (150, 170) K [Halevy and Steward, 2008] and we assign Tb ≃ 200 K for the freshly produced ions.
Table 1. Model Parameters
 The SC is exposed to a corotating flow and to a flow of the new-born ions with respective relative velocitiesV0c,b = Vc,b − VSC. Here VSC stands for the SC velocity vector while the vectors Vc and Vb denote the azimuthal plasma flow velocities, given by the corotating speed Vc = Ωpr (with Ωp being the planetary rotation frequency) and by the Kepler speed Vb = (GMS/r)1/2 (with MSbeing Saturn's mass), respectively. We aim at describing the SC - plasma interactions in a model which has an advantage of bringing out the qualitative features and thus assume that the orbiter trajectory is azimuthal (e.g., similarly to the part shown inFigure 1 at SCET358:19:00–24:00 h). For estimates V0c,b we take the typical azimuthal speed of Cassini during the equatorial flybys VSC = 14 km/s as given by Tokar et al. . The Kepler speed at the Enceladus orbit is Vb ≃ 12.6 km/s, whereas the thermal plasma corotates with Saturn at Vc ≃ 39 km/s. Thus at distances r/RS ∼ 4 in the orbiter frame of reference, the corotating plasma overtakes SC at the speed V0,c ≃ 25 km/s while the new-born ions lag behind with the velocity ofV0,b ≃ −2 km/s. Note that these numbers for V0c,b slightly differ from exact values at SCET358:19:00–24:00 h.
 To illustrate the simulation results we introduce the local coordinate system given in the insertion of Figure 1. The relative corotating flow velocity, V0,c, is in the positive x(azimuthal) direction, while the new-born ions have oppositely directed velocityV0,b; α denotes the angle between V0,b and the direction towards the Sun. Note that for both ion species the respective flow velocities V0c,b are supersonic, while for the electrons the flow velocity V0,c remains always subsonic. In such a case, the plasma electrons are mainly responsible for the plasma screening [Fortov et al., 2004], and thus the electron Debye length λD is used for normalization in Table 1 and Figure 1 presenting the results.
 We employ a three dimensional particle-in-cell DiP3D code, which calculates the self-consistent potential on an object and wake formation in the electrostatic approximation in intricate, collisionless plasma environments [Miloch, 2010; Miloch et al., 2010]. The concept of collisionless plasma needs to be elaborated for the Enceladus torus, where the density of the neutrals peaks. Taking the neutral gas densities nn ∼ 108 cm−3 for the gas plume [Cravens, 2008] and nn ∈ (103, 104) cm−3 for the extended torus [Cravens et al., 2011] and typical ion-neutral and electron-neutral cross sectionsσin ∈ (10−14, 10−15)cm2 and σen ∈ (10−15, 10−16)cm2, yields the mean free paths which obey ordering . Consequently there appears a huge gap between the smallest ion mean free path of ∼ 10 km and the typical screening length of λD ∼ 1 m, and thus the plasma in the Enceladus torus can still be treated as collisionless. We therefore use the same numerical procedure and approximations as developed in Yaroshenko et al. for the modeling of Cassini charging during the Saturn orbit insertion (SOI) flyby, but advance the code to include the counterstreaming cold water-group ions. The orbiter is described as a spherical conducting object with the characteristic sizea ≃ 6.6 m which is charged by the core plasma fluxes, by photoemission due to direct photons and by a flow of the new-born cold ions. Such a simplified model captures the main physical features and thus, as a first order approximation, can provide insight into intricate process of SC-plasma interactions in the Enceladus torus.
 To avoid any confusion with the downstream and upstream directions for the two counterstreaming ion flows, these terms are related to the motion of the more numerous corotating ions. Unless otherwise stated, the area x > a/2 corresponds to the region downstream of the orbiter and x < −a/2 denotes the region upstream of the spacecraft. In this paper we will refer to the wake as a region where the plasma species density is reduced by more than 50% of its ambient value. The characteristic length scales of perturbations are defined as lϕ,e,i,b, where ϕ stands for potential, and e, i, bfor electrons, core ions, and new-born ion species, respectively.
3. Results and Discussion
 In Figure 2 we show the potential and plasma density distributions around the orbiter for the parameters given in Table 1. The SC potential achieves a significant negative value of |ϕSC| ∼ 6Te[eV], which is consistent with the tendency revealed by the electrostatic potential in supersonically flowing plasmas in numerical simulations [Fortov et al., 2004; Hutchinson, 2005]. As can be inferred from Figure 2a, which shows the two-dimensional contour plot of the potential distribution, a high |ϕSC| results in a small asymmetry of the potential configuration outside the orbiter. Most of the potential drop occurs at length scales of lϕ ∼ 2λD (downstream) and lϕ ∼ 3λD (upstream) in the white area in Figure 2a. Thermal electrons, having low energy Te < e|ϕ|, cannot overcome the negative potential around the obiter, and the ne-depletion region is characterized by the same length scalesle ∼ 3λD (upstream) le ∼ 2λD (downstream) as illustrated Figure 2b. There is no evidence of the increased electron density on the Sun-exposed side of the orbiter due to the photoelectron production, it is noticeable only a small asymmetry in thene-contour lines (above the noise amplitude).
 The most inhomogeneous structures are found in densities of both ion species. The core ion distribution does not reveal significant ni-density perturbations upstream, but is characterized by a pronounced elongated wake ofli ∼ 5λD downstream of the orbiter (Figure 2c). There is no sign of the focusing effect for the core ions: their kinetic energy Ek ∼ 100 eV significantly exceeds the thermal energy Eth ∼ 20 eV as well as the electrostatic energy e|ϕSC| ≤ 10 eV, and the ions display a typical geometrical wake similar to those found for the water group ions at SOI trajectory [Yaroshenko et al., 2011]. The resulting plasma structure around SC is not quasineutral, and a probe mounted at the orbiter most likely will measure ni ≠ ne. The Langmuir probe (LP) is attached to Cassini by ∼1.5 m deployable boom [Gurnett et al., 2004] and in Figures 2b and 2c we schematically indicate its position. Such an instrument would register ne/n0,e ∼ 0.3–0.4 and ni/n0,i ∼ 0.5–0.7, respectively, providing a significant (ne/ni ∼ 0.7) misfit of the electron and ion densities. Interestingly enough, the Cassini probe measurements often show a similar tendency close to the Enceladus orbit [Wahlund et al., 2009] displaying misfits ne/ni up to 60–50% during some traversals through the E ring nearby r/RS ∼ 4 [e.g., Yaroshenko et al., 2009, Figures 1 and 2]. Whether this effect is solely a manifestation of the inhomogeneous plasma structure at distances ∼ λD from the SC surface, as discussed above, or perhaps, the electron depletion reflects the presence of negatively charged dust near the moon's orbit [Yaroshenko et al., 2009] requires further detailed studies of the LP data, their dependence on LP position, correlations with Cassini dust impact data, etc.
 The new-born non-thermalized ions (Eth ∼ 0.02 eV), have a flow energy Ek ≤ 1 eV less than e|ϕSC| ∼ 6 eV. The electric field associated with a so high SC potential, effectively captures the incoming cold ions providing the maximum density enhancement near the orbiter surface with a ratio of the captured - to the primary ion densitiesnb/n0,b ≥ 3 (Figure 2d). The corotating wake fields also confine the cold ions at x > a/2 but not so effectively, leading to nb/n0,b ≤ 1.5. However, these wake fields significantly widen an area of the enhanced cold ion population to spatial scales larger than the SC size in the direction transverse to the ion flow, providing lb ∼ 8–10λD as shown in Figure 2d.
 Note that the process of ion capturing has nothing in common with the standard ion focusing, where the region of an enhanced ion density appears due to the bending of ion trajectories by the wake electric field caused by the same ion flow [see, e.g., Svenes and Trøim, 1994; Miloch et al., 2008, 2009; Miloch, 2010]. Our modeling reveals a new effect, when the cold new-born ions are captured by the fields perturbations produced mainly due to the SC charge and the corotating wake. The region of the enhanced densitynbappears upstream for the new-born ions (x > a/2). In their downstream region (x < −a/2), there is a clear Mach cone related to a significant nbdepletion (i.e., the wake of the new-born ion species), extending up tolb ∼ 5λDfrom the orbiter surface. Therefore for azimuthal flybys, the plasma-SC interaction results in a complicated plasma structure with a self-consistent charge separation: the core plasma constituents have significantly reduced densities in the close downstream vicinity of the SC (x > a/2), while the ensuing potential perturbations e|ϕ| ≥ Tb[eV] effectively captured the new-born ions in these “uninhabited” regions. On the other hand, the non-thermalized ions display significant reduction and highly variable density perturbations in their downstream region (x < −a/2) and thus are hardly reliable measurable in the area behind the orbiter, in their own wake. Apparently, the observed density gradients can impact the plasma density measurements onboard Cassini. In particular, our results indicate that the instrument mounted at the position close to the angular position of CAPS (Figure 2d) could measure the fresh ion density enhanced by a factor of ∼2. However, this estimate might change if a realistic SC geometry (including SC booms and aerials) is taken into account.
 For non-azimuthal Cassini trajectories, the main feature of the SC-plasma interactions will remain a formation of the corotating wake downstream of SC. The relative velocitiesV0,c and V0,bwill not be anti-parallel any more as it is inFigure 1 but will form a certain angle. The incoming cold ions will be captured by a negatively charged SC upstream of their motion within the distances of ∼ λD from the orbiter surface. The corotating wake will also be able to provide the cold ion capturing, but a strength of this effect, spatial extension and asymmetry in the nb-distribution will crucially depend on the specific values ofV0,c and V0,b and an angle between the two velocity vectors. Furthermore the full structure around the orbiter will not be quasineutral at scales of ∼ λD, and LP mounted at the boom of such a length can be exposed to very different plasmas: deeply in the corotating wake ni ∼ ne ≪ n0,e; n0i or even ni < ne ≪ n0,e; n0,i, while outside of the wake ne < ni ∼ n0,i.
 Although the presented results can be considered as a first order approximation for modeling Cassini-plasma interactions in the Enceladus torus, they can explain localization of the new-born ions in the corotating shadow of Enceladus as has been observed during close flybys of the moon [Tokar et al., 2009]. Physically we may reconstruct our arguments as follows: The Enceladus plume is naturally the strongest source of the fresh cold ions. These ions being deflected by the planetary magnetic field downstream of the plume [Kriegel et al., 2009] can reach the region of the Enceladus wake fields. Here they form a spatially confined structure similar to that shown in Figure 2d thus providing a local enhancement of the nb-density above a certain threshold, sufficient to be measured by the Cassini instrument. The orbiter does not create its own plasma structure in the moon wake (since the plasma is strongly depleted) but registers the fresh ions that are mainly rammed into the CAPS field-of-view by the spacecraft motion as reported byTokar et al. . Interestingly, our model naturally explains the large scales of nb-enhancement in the direction perpendicular to the ion flow observed by CAPS (similar to that shown inFigure 2d).
 We conclude that our numerical simulations of the Cassini-plasma interactions in the Enceladus torus, reveal a formation of a new kind of the plasma configuration with self-consistent charge separation of the plasma species in the electric field of SC. Such a structure can significantly affect the plasmas measurements of the Cassini instruments, especially when the instrument view directions cover either upstream or downstream directions for some plasma species. The results obtained for the azimuthal flyby yield that the CAPS and LP instruments could register an increased density of the new-born water group ions and a rather large difference in the electron and ion densities.
 The authors thank anonymous referees for constructive comments.
 The Editor thanks an anonymous reviewer for assisting in the evaluation of this paper.