A hybrid particle code has been used to examine how Titan's interaction with Saturn's magnetosphere is affected by the presence of negative ions in Titan's ionosphere. The simulations self-consistently include a version of Titan's ionosphere represented by 8 generic positive ion species, over 40 ion-neutral chemical reactions, ion-neutral collisions and Hall and Pederson conductivities. A model consisting of 6 generic negative ion species is also included. The presence of negative ions is found to alter the conductivity of Titan's ionosphere, changing the loss rates of the ionospheric species and modifying the topology of Titan's ion tail.
 One of the most unexpected discoveries of the Cassini Plasma Spectrometer (CAPS) was the persistent presence of negative ions in Titan's ionosphere with masses up to ∼15000 amu. [cf. Coates et al., 2007, 2009, 2010; A. Coates, private communication]. These negative ions indicate the existence of heavy hydrocarbon and nitrile molecules. It has been suggested that these ions may be precursors of aerosols in Titan's atmosphere and may precipitate to the surface as tholins [cf. Coates et al., 2010, 2007; Waite et al., 2007]. The observed negative ions are found in mass groups of 10–30, 30–50, 50–80, 80–110, 110–200 and 200+ amu/q. Analysis by Coates et al.  found an altitude and latitude dependence of the mass of the heavy negative ions. The maximum negative ion mass peaked at low altitude and high latitude. A weaker dependence of the maximum ion mass on the solar zenith angle was also found, with the highest masses near the terminators.
 There remain many outstanding chemistry issues with regard to the production and loss of the negative ions [cf. Coates et al., 2007; Waite et al., 2007; Vuitton et al., 2009], however, these issues are not addressed here. The focus of the paper is on the dynamical consequences due to the presence of the negative ions; specifically does the presence of negative ions in Titan's ionosphere have an effect on Titan's plasma interaction? If so what are those effects?
 The remainder of this paper will address this main question. The next section will describe the simulations. The simulation results will be presented in the following section. In subsequent sections a brief discussion and some conclusions will be presented.
2. The Simulations
 Titan's plasma interaction can be broken down into three interacting regions: 1) the magnetospheric flow, 2) Titan's ionosphere and 3) Titan's neutral atmosphere. The simulations are composed of a number of models describing the physical processes in each region. A brief description is given here. A detailed description can be found in Ledvina et al. .
 The dynamics of the magnetospheric and ionospheric plasma are modeled using the hybrid approximation. The plasma is assumed to be charge-neutral. The ions are treated as kinetic particles, while the electrons are treated as a massless fluid. A full set of electromagnetic plasma waves up to and including the low frequency part of the Whistler waves are retained. The quasi-neutrality assumption means that charge separation and electrostatic instabilities are removed from the simulations. Further discussion about the hybrid approximation, solution algorithms and applications can be found inLedvina et al. , and references therein.
 When the neutral densities become significant, ion-neutral and electron-neutral collisions will modify the electric field via the Hall and Pederson conductivities. These processes are important in Titan's ionosphere and indeed the conductivities have been measured [Rosenqvist et al., 2009]. The Hall and Pederson conductivities are included in the simulations. The electron-ion and electron-neutral collision frequencies were taken fromMitchner and Kruger . The ion-neutral collision frequencies were taken fromSchunk and Nagy .
 The neutral atmosphere used in the simulations is assumed to be stationary, spherically symmetric and consist of 11 neutral species, H, H2, CH4, HCN, N2, C2H2, C2H4, C2H6, C3H4, C4H2 and HC3N. Of these species the density profiles of CH4 and N2 were taken from the INMS profiles observed during the TA encounter [Waite et al., 2005]. The neutral profiles for the other species came from the model of Toublanc et al. . There is a large variability in the neutral density profiles measured by INMS over the course of the CASSINI mission, as well as roughly a factor of 3 discrepancy between INMS measurements and those made by other Cassini instruments. The neutral profiles used here lie within the range of the Cassini measurements.
 Titan's ionosphere is controlled by a complex ion-neutral chemical network [cf.Keller et al., 1998; Fox and Yelle, 1997]. It is not currently feasible to include such a detailed chemistry networks into three-dimensional simulations. Instead we use a simplified version of theKeller et al.  chemical network (T. E. Cravens, private communication). The reduced model groups several similar species into 7 generic ion species (see Table 1). Photoionization and secondary electron impact ionization rates of the L+, M+ and H1+ species are computed as a function of altitude and solar zenith angle using a solar minimum flux. Electron impact ionization from Maxwellian electrons with ne = 0.1 cm−3 and Te= 100 eV moving along parabolic magnetic field lines from Saturn is used to create a night side ionosphere. This is consistent with the electron spectra measured during the TA encounter. Other electron spectrum can be used but the choice will not alter the results presented here. All other species are created via ion-neutral chemical reactions of which there are over 40 in the model (seeLedvina et al.  for details). The ionospheric chemistry is solved on a spherical grid that extends radially from 750–2700 km altitude, with Δr = 50 km and Δθ, Δφ = 1.9°.
Table 1. The Ion Loss Rates to Saturn's Magnetosphere for Each Simulationa
Without Negative Ions (#/s)
With Negative Ions (#/s)
% Change in the L.R. With Negative Ions
Note the decrease in the loss rates of the primary ion species (L+, M+ and H1+) and an increase in the loss rates of the other species when negative ions are present.
7.0 × 1020
6.0 × 1020
1.0 × 1024
8.0 × 1023
2.0 × 1024
1.9 × 1024
6.0 × 1023
7.4 × 1023
2.0 × 1023
2.7 × 1023
6.0 × 1022
7.5 × 1022
2.0 × 1019
3.0 × 1019
2.1 × 1021
8.0 × 1021
6.5 × 1021
1.5 × 1021
1.8 × 1020
6.0 × 1017
 The negative ion simulation includes six generic species. The first, N1− has a mass of 20 amu/q and represents contributions from CN-, NH2− and O-. The contributions from NCN-, HNCN- and CH3− are represented by N2− with a mass of 40 amu/q. The combined densities of C5H5−, C6H- and C6H5− are represented by N3−whose mass is 70 amu/q. Polyyines, high order nitriles, PAHs and cyano-aromatics are represented in the simulations by N4−, N5− and N6− with masses of 95, 155 and 200 amu/q respectively. These generic ions are assumed to have no chemical activity; they are loaded at the start of the simulation with spherically symmetric density profiles consistent with the T16 observations [Coates et al., 2007]. New particles are added to maintain these density profiles. Charge neutrality is maintained by reducing the electron density to compensate. The negative ions are free to respond to the fields and create current systems of their own during the simulation.
 This is a first-order approach to representing the negative ions in Titan's ionosphere. As discussed earlier, the actual distribution of negative ions is not symmetric. A more consistent and complete approach awaits a more complete negative ion chemistry model. However, since the goal of this research is to determine if negative ions affect Titan's plasma interaction the assumptions seem reasonable.
 The neutral collision operators are located on a spherical grid with the same dimensions as the chemistry grid. Every particle whose position overlaps this neutral grid is subject to collisions with the neutrals depending on their relative velocities, their density and mass dependent collision cross-sections. These ion-neutral collisions result in a frictional drag applied to each particle.
 The effects that negative ions have on Titan's plasma interaction are investigated by comparing the results of two hybrid simulations. In simulation 1 Titan's ionosphere is created using the interacting chemistry of 7 positive ion species. In simulation 2 the 6 negative ion species are added. Both simulations were run out to 3500 s with the same time step.
 The simulations of Titan's interaction with Saturn's magnetosphere are performed in Titan centered Cartesian coordinates extending ±10 RT in each direction. A cell size of 250 km is used to resolve the electric and magnetic fields. Titan is located at noon Saturn local time. Ions are injected upstream of Titan with a mass of 16 amu and a density of 0.2 cm−3. We take the simplest approach possible and load the upstream ions with a beam distribution and an E × B drift speed of 120 km/s with respect to Titan. The incident magnetic field is 5 nT and is perpendicular to Titan's orbital plane. The convection electric field points from Saturn towards Titan and the sun. The electron temperature is evolved with an energy equation down to an altitude of 1400 km. Below the electron temperature profile is constant and specified by the Langmuir probe results [Ågren et al., 2009].
 The contours of the total positive ion density in the orbital plane for the first simulation are shown in Figure 1a. The results of the second simulation are shown in Figures 1b and 1c. Figure 1b shows the total positive ion density and Figure 1c shows the total negative ion density. Differences in the simulations due to the negative ions show up in this plane since it contains the convection electric field. The presence of the negative ions changes the tail structures. Within 2 RT downstream of Titan the width of the positive ion tails are identical in both simulations. This is not the case further downstream where the presence of negative ions results in a narrower ion tail (24% at −8 RT). The deflection angle of the ion tail is also less when negative ions are present. The negative ion motion is towards Saturn, while the positive ion motion is away from Saturn. Few negative ions are found in Titan's wake region.
Figure 2 shows the number density for the 40 amu (a) and the 70 amu (b) negative ions. For comparison the number densities of the 44 amu (c) and the 70 amu (d) positive ions are also shown. The trajectories of the negative ions are more concentrated and not very well intermixed. They resemble ideal test particle trajectories in uniform electric and magnetic fields. This is in contrast to the intermixed trajectories/densities of the positive ions represented in Figures 1a and 1b. The major pickup location of the negative ions is on the opposite side of the major pickup location of the positive ions. The total negative ion density in Saturn's magnetosphere is small. This low density current does not produce a significant feedback on the fields. The negative ion trajectories are separated by amu, see Figures 2a and 2b. In contrast the positive ions have a much larger density and a stronger feedback on the fields. This aids in the mixing of the positive ion trajectories and the densities shown in Figures 1a, 1b, 2c, and 2d. The results show that the negative ions will be picked up by Saturn's magnetosphere but at densities that are typically an order of magnitude less than the incident magnetospheric plasma. The densities are still large enough to be detected by CAPS (lower limit of 0.02 cm−3, A. Coates, private communication) if it is looking into the negative pickup ion beam.
 Placing the negative ions in the simulation changed the rates that the positive ions are picked up by, or lost, to Saturn's magnetosphere. The loss rates for each ion species and the percent they change when the negative ions are present are listed in Table 1. There is a decrease in the ion loss rates of the first 3 ion species (L+, M+ and H1+). While the loss rates of the other species (H2+, MHC+, HHC+ and HNI+) are increased.
 The L+, M+ and H1+ are created primarily by photo and electron impact ionization. The ionization rates are the same in both simulations. Ideally then the ion loss rates should also be the same in each simulation. However, they are not the same leading to the question of why the difference? One possibility is that conservation of momentum is involved. When negative ions are not present in the simulation the magnetospheric plasma is deflected around Titan. The change in momentum of this plasma is balanced by the momentum of the pickup ions coming off of Titan. When negative ions are present they get picked up by the magnetospheric flow but on the opposite side of Titan. The positive ions are no longer are responsible for conserving all of the momentum of the magnetospheric ions. Fewer of the lighter (L+, M+ and H1+) positive ions are picked up as a result.
 With regard to the increase in the loss rate of H2+, MHC+, HHC+ and HNI+ we speculate that a decrease in the ion loss rates of the light species leads to more light ions available to chemically react in the ionosphere, thus creating more of the heavier positive ion species. Enhancing the amounts of the heavier ion species may allow more of these species to be picked up by Saturn's magnetosphere and hence increasing their loss rates as seen Table 1.
 The ion pickup process is controlled by the electric field. The presence of the negative ions in Titan's ionosphere changes the electric field in Titan's ionosphere/atmosphere. The magnitude of the electric field near Titan is shown in Figure 3. The electric field is much stronger in the nightside ionosphere when negative ions are not present. Negative ions reduce the amount of charge carried by the electrons, thus changing the Hall and Pederson conductivities of the ionosphere. The net result is a change in the ionospheric electric fields.
 The presence of negative ions in Titan's ionosphere effects Titan's plasma interaction. The deflection of the positive ion tail was less and the tail was thinner when the negative ions were present. The simulation containing the negative ions showed a reduction in the loss rates of the L+, M+ and H1+ species and an increase in the loss rates of the heavier species (H2+, MHC+, HHC+ and HNI+). The densities of these species are chemically dependent on the lighter species suggesting that the dynamical consequences of the negative ion can affect the ionospheric chemistry. It was also found that the presence of negative ions changes the electric field within Titan's ionosphere. This result is likely due to changes in conductivities and subsequently the current systems within the ionosphere.
 The simulation results showed that negative ions can be picked up by Saturn's magnetospheric electric fields at levels detectable by CAPS. The loss rates are typically an order of magnitude less than the loss rates of the comparable mass positive ion. For the most part negative ions are picked up from the hemisphere of Titan opposite the positive ions pickup region. The negative ion trajectories do not show the intermixing that the positive ions do. Magnetospheric processes that might strip an electron from the negative ions are not included in the simulations and therefore we cannot address whether the negative ions being pickup will continue to exist in Saturn's magnetosphere.
 These results strongly suggest that negative ions have a global effect on Titan's plasma interaction and a dynamical effect on Titan's ionospheric chemistry. Given this result a chemical model of the creation, loss and distribution of the negative ions is desperately needed. Once this model is included in the simulations a better assessment of the consequences of the negative ions on Titan's plasma interaction, its ionospheric chemistry and its ionospheric dynamics can be obtained.
 The authors of this paper would like to acknowledge the support from NASA grants NNH09CE73C, and NNX08AK95G. In addition, the authors would like to acknowledge the computational support provided by the NASA Advanced Supercomputer, NAS, scientific computing facility at NASA Ames Research Center, Moffett Field, CA. The Editor thanks one anonymous reviewer for his/her assistance in evaluating this paper.