The Influence of the Galactic Bar on the Dynamics of Globular Clusters

We make use of recent estimates for the parameters of the Milky Way's halo globular clusters and study the influence of the galactic bar on the dynamics of these clusters by computing their orbits. We use both an axisymmetric and non-axisymmetric galactic potentials, which include the rotating elongated bar/bulge structure. We account for observational errors both in the positions and in the velocities of the globular clusters and explore the influence of the bar on cluster's evolution. This is contained in the angular momentum-total energy plane, (Lz,E), which is widely exploited as an indicator of the groups of globular clusters that originated from the same accretion event. Particular attention is devoted to the Gaia-Sausage/Enceladus and Pontus structures identified recently as two independent accretion events. Our study shows that it is not possible to identify GSE and Pontus as different merger events.


Introduction
Modern cosmological models predict that the halo of our galaxy was formed by and evolved through the continuous accretion and disruption of star clusters, small satellite galaxies [1,2], and gas accretion events [3]. Our galaxy, the Milky Way, is a unique and excellent laboratory [4] in which to investigate how this process occurred by studying examples of such accretions, such as the ongoing merging process of the Sagittarius dwarf galaxy [5]; echoes of the past accretion events, e.g., Gaia-Sausage/Enceladus (GSE) at z ∼ 2 [4,6]; or even the first fall beginnings of the Magellanic Clouds [7]. As a result of such mergers, the halo of Milky Way galaxy, or at least part of it, was built with the products of such accretions as the stellar streams or the globular clusters of extra-galactic origin.
Our Galaxy consists of a few subsystems-a galactic halo, thick and the thin disks and an elongated bar/bulge [8,9]-that have stars with different spatial distributions, velocity dispersions and chemical compositions. Moreover, even within one subsystem, a few subgroups that have different origins can be often distinguished [4,10,11]. According to Naidu et al. [4], for example, most of the stars in the inner halo of the Milky Way seem to be associated with the GSE accretion event. A comprehensive study of traces of the past mergers is important, therefore, to understand how our Galaxy was formed and how it will evolve in the future.
Based on the last data releases, attempts were made to decipher the nature of globular clusters in the Milky Way using the energy-vertical angular momentum space (L z , E), where E is a total energy and L z is the component of the angular momentum of a cluster in the direction perpendicular to the galactic disk. Such attempts were made under the assumption that these objects have values for the energy and z-component of the angular momentum close to the parent accretion event. This idea, proposed by Helmi and de Zeeuw [23], is based on the assumption that the clusters conserve (L z , E) for billions of years after their accretion onto our galaxy. This picture assumes that the galactic potential was static during billions of years of galactic evolution. In addition, the authors did not take into account such effects as the dynamical friction, which significantly changes the energy and the angular momentum of the satellite during its accretion, and the influence of the Milky Way's elongated bar/bulge potential, which may affect the positions of the clusters on the (L z , E)-plane, especially for those clusters that have small pericentric radii.
The aforementioned (L z , E) classification approach allowed researchers to detect several merger events of the satellite galaxies, such as GSE [6,11], Sequoia [24], Sagittarius [5], Kraken [25] and Pontus [26,27]. However, Pagnini et al. [28], using dissipation-less N-body simulations that reproduce the accretion of one or two satellites with their globular cluster population on a Milky Way-type galaxy, it was found that accreted globular clusters do not show dynamical coherence; that is, they do not concentrate in kinematic spaces, if the mass of the progenitor satellite galaxy is approximately about 10% of the mass of the Milky Way. This seems quite contrary to the identification methods presented above. However, for small accreted satellite galaxies, accretion products (globular clusters) can be confined in a tight range of energies and momentum. In this case, other mechanisms (in particular galactic bar) may alter the identification of clusters in the same accretion event.
Malhan et al. [26] used the extended approach and determined the action angles J= (J R , J φ , J z ) for known globular clusters, together with their total energies E. By analysing the distributions of the clusters in (J φ = L z , E), (J R , E), (J z , E) and (J φ = L z , J R ), the authors concluded that clusters with similar values of these quantities belong to the same accretion event.
The Milky Way potential, however, is neither stationary nor axisymmetric. In the central parts, the Milky Way potential is influenced by the non-axisymmetric and non-stationary potential of the galactic bar, so the dynamics of the clusters that have small enough pericentric distances will be strongly affected by bar potential after a few passages. In addition, the positions and the velocities of the clusters are determined with errors, which also can significantly affect the identification of the clusters in (L z , E)-space. It should be noted here, that we will limit ourselves to studying only (L z , E)-space, because the calculation of J R and J z values in the non-axisymmetric and time-dependent potential with elongated bar/bulge component is not possible. Since the Hamiltonian becomes time-dependent, canonical transformation to action-angle coordinates cannot be obtained [29].
In this paper, we study the combined influence of the Milky Way bar potential and the observational errors to test the identification of GSE and Pontus as two different accretion events Malhan et al. [26]. We recall that the significant influence of the bar on the orbits of objects has been shown in the many previous studies: for globular clusters in Chemel et al. [30]; for stellar streams in Hattori et al. [31] and Hunt and Bovy [32]. Accordingly, the layout of the paper is as follows. Section 2 describes available observational data for the globular clusters of the Milky Way halo; Section 3 discusses the model we adopted to study the dynamics of a globular cluster in Milky Way axisymmetric and non-axisymmetric time-dependent potentials. Section 4 describes the our results of evolution of the globular clusters in (L z , E) space. Section 5 summarises the results of our study.

Observational Data
To demonstrate the influence of the galactic bar on the positions of the globular clusters on the (L z , E) diagram, we selected two groups of globular clusters that were identified in the literature as the members of two progenitor dwarf galaxies-Pontus and GSE [26,27]. The Pontus group consists of seven globular clusters, and the GSE group consists of 13 clusters listed in Table 1. In addition, we include in our considerations the tentative globular cluster NGC 6864/M 75, indicated by Malhan et al. [26] as a possible candidate for both groups of globular clusters. Table 1 provides parameters of the globular clusters, together with their observational errors (inside the parentheses) taken from [17,18]. To convert the heliocentric measurements provided in Table 1 into the values in the Galactocentric reference frame, the astropy package [33][34][35] was used. We adopted for the solar Galactocentric distance the value of R = 8.122 kpc taken from GRAVITY Collaboration et al. [36], and for the velocity components of the Sun, the values of V = (12.9, 245.6, 7.78) km/s taken from Drimmel and Poggio [37], Reid and Brunthaler [38]. The displacement of the Sun from the mid-plane of the Galaxy, the value Z = 20.8 pc was taken from Bennett and Bovy [39]. Figure 1 shows the positions of the globular clusters on the angular momentum-energy plane (L z , E), calculated with the axisymmetric galactic potential McMillan17 using data taken from Table 1. The potential model is described in more detail in Section 3. Figure 1 shows also the influences of the observational errors on the positions of the globular clusters on the (L z , E)-plane calculated using one hundred randomly chosen coordinates, consistent with the errors in Table 1(inside the parentheses). We have assumed a normal distribution with the standard deviations set by the errors.  Table 1. We have assumed a normal distribution with the standard deviations set by the errors.
. As one can readily see in Figure [26] came to the conclusion that these groups of clusters belong to different merger events. Besides the values L z = J φ and E, Malhan et al. [26] considered the action-angles J R and J z as well. In this study, we refrained from this procedure, since the calculation of these values in an non-axisymmetric and time-dependent potential is not possible. We will instead focus our efforts on the influence of the galactic bar/bulge on evolution of the objects on the (L z , E)-plane. We notice that all objects listed in Table 1 have small pericentric radii (r peri < 3 kpc), which makes a strong influence of the bar on these objects. We notice also that the globular clusters from Table 1 have both prograde and retrograde orbits, which is caused, probably, by the internal velocity dispersion of globular clusters that were accreted almost radially on the Milky Way galaxy [26,27].

The Models
We conducted simulations of the dynamics of the globular clusters using the astrophysical package Galpy [40]. The integration of orbits was performed using a Dormand-Prince (dop853) integrator [41], which is an eighth order Runge-Kutta family method with adaptive timestep. This method was implemented in the package Galpy, and for 1000 orbital periods, it conserved energy at the level of ∆E/E = 10 −8 in the case of an axisymmetric time-independent potential.
We integrated the orbits during 4 Gyr forward in time starting from the present epoch, supposing that building of Galaxy was stopped approximately 4 Gyr ago [42]. Therefore, the bar potential remained approximately constant during the last 4 Gyr. The simulations show that bar stops its evolution after the thickening phase and remains almost unchanged as long as the calculations are performed [43,44]. Therefore, 4 Gyr is enough time to demonstrate that even with the fixed bar potential, the quantities used to pinpoint the past accretion events are not conserved. It also follows from the above that it remains not important to integrate orbits forward or backward in time if we want to demonstrate the influence of the bar.

Galactic Potential Models
To study globular cluster dynamics in the Milky Way potential, we used two models. The first model has an axisymmetric and time-independent potential, whereas the second one has a non-axisymmetric time-dependent potential with an elongated bar/bulge.

Axisymmetric Time-Independent Potential
Following Malhan et al. [26], first we modelled the Milky Way using the axisymmetric potential described by McMillan [45](McMillan17). The model included the bulge, the thick and thin stellar disks, two gaseous disks (H I and H 2 ) and the dark matter halo.
The axisymmetric bulge potential was described by the density profile from Bissantz and Gerhard [46] and can be represented as: where r = R 2 + (z/q) 2 ; the parameters α and r 0 were equal to 1.8 and 0.075 kpc, respectively; r cut = 2.1 kpc; and the axis ratio q was equal to 0.5. The central volume density of the bulge: ρ 0,b = 9.84 × 10 10 M kpc −3 , which gives a total bulge mass equal to M b = 9.23 × 10 9 M .
The Milky Way disk consists of two subsystems, the thin and the thick [47] exponential disks both described by density profiles with the form: Here, z d , R d and Σ 0 are the vertical scale height, the radial scale length and central surface density, respectively. The vertical scale heights of the disks are fixed to the values z d,thin = 300 pc and z d,thick = 900 pc, and the scale lengths for the thin and thick disks are assumed to be 2.5 kpc and 3.02 kpc, respectively. With the central surface densities of the thin and of the thick disks taken to be Σ 0,d,thin = 896 M pc −2 and Σ 0,d,thick = 183 M pc −2 , the total masses of the thin and of thick disks can be estimated as M d = 2πΣ 0 R 2 d and equal to 3.518 × 10 10 M and 1.048 × 10 10 M , respectively. A possible central 'hole' in stellar density is not considered.
The McMillan17 potential also includes two gaseous disks: the atomic (H I) one and the molecular (H 2 ) one. The density distributions in both gaseous disks are described by this equation: The potential of the dark matter halo is described by the simple density profile given by this equation: where x = r/r h , and r h = 19.6 kpc is the scale length of the density distribution, and ρ 0,b = 0.00854 × 10 10 M kpc −3 is a characteristic scaling density of the halo. With the parameter γ = 1, Equation (4) [45]. To integrate the equations of motion, it is necessary to calculate the potentials using known density distributions. The NFW potential has analytical representation. The potential of the bulge was calculated using the basis-function expansion of the self-consistent-field method of Hernquist and Ostriker [49]. Potentials for disks were calculated using a modified technique of Hernquist and Ostriker [49], which was presented in Kuijken and Dubinski [50] and implemented in theGalpy package [40].

Non-Axisymmetric Time-Dependent Potential with Elongated Bar/Bulge
In this model, we replaced the axisymmetric bulge with the non-axisymmetric elongated bar/bulge component.
Following Chemel et al. [30] and Yeh et al. [51], we approximate the bar/bulge by a rotating ellipsoid with a density distribution given by a Ferrers model [40]: where m 2 = x 2 a 2 + For the bar semi-axes, we chose the values a = 5 kpc, b = 2 kpc and c = 1 kpc, which gives a good approximation for the elongated bar/bulge observed in the Milky Way (see, for example, Model M85 in Portail et al. [8,9] where and ∆ 2 (u) = (a 2 + u)(b 2 + u)(c 2 + u), where n is integer number, which is equal to two, according to the degree in Equation (5). λ is the unique positive solution of ζ 2 (λ) = 1, outside the bar (ζ ≥ 1). Inside the bar, λ = 0 (for the more details see Pfenniger [56]). The Milky Way bar rotates with constant angular velocity, and the major axis is tilted with respect to the Sun, as shown in Figure 2. A number of studies estimated the bar orientation relative to the Sun ( [8,9,54,57,58]) and its angular velocity ( [57,[59][60][61][62]). Debattista et al. [59] were the first to estimated the angular velocity of the galactic bar: Ω b = 59 ± 5 km s −1 kpc −1 . Recent data that used Gaia proper motion measurements allow one to estimate the pattern speed of the bar more accurately, although the spread in the measurements remains considerable. Sanders et al. [57] using the VVV Infrared Astrometric Catalogue (VIRAC) and Gaia DR2 proper motions obtained the value of Ω b = 41 ± 3 km s −1 kpc −1 . Clarke et al. [60] obtained Ω b = 37.5 km s −1 kpc −1 [61]. Sormani et al. [62], using hydrodynamic gas simulations in barred Milky Way potential, found that the angular velocity of the bar is Ω b = 40 km s −1 kpc −1 . We chose in our nonaxisymmetric model the value of Ω b = 40 km s −1 kpc −1 , which is in a good agreement with Portail et al. [8,9] and Bovy et al. [63].
Recent measurements of the bar orientation θ relative to the Sun lie in the range of 23 • − 33 • [8,9,54,57]. We used in our simulations the value θ = 28 • , which is in a good agreement with the most recent estimates [9,54,58]. Figure 3 shows the Milky Way rotation curves in our barred and non-barred models. The model with the bar has systematically higher values of the rotational velocity compared to the axisymmetric model McMillan17 due to the fact that in the model with the bar, we replaced the axisymmetric bulge in McMillan17 with an elongated bar/bulge whose mass is about two times larger than the mass of the bulge in the McMillan17 model. Rotation curves in Figure 3 were derived (in the plane z=0) from the equation v(r) = (r dΦ/dr) 1/2 , where Φ is the total potential.

Results and Discussion
We discuss in this section the dynamics of the halo globular clusters and the influence of the galactic bar on their positions in the (L z , E)-plane. Figure 4 shows the (L z , E)-positions of the globular clusters calculated using data taken from Table 1 for barred and non-barred galactic potential models. Figure 4 also shows one hundred alternative positions of each cluster considering observational errors. As one can readily see, the clusters have similar values of L z and E with minor differences. Note, however, that while in the axisymmetric and time-independent potential, the values of L z and E are conserved, in time-dependent non-axisymmetric potential, these values are not conserved in time. We will discuss in detail the non-conservation of L z and E below.
First, let us consider the influence of the galactic bar on the positions of the clusters on (L z , E)-plane using mean values of their parameters (asterisks in Figure 4) in time-dependent non-axisymmetric potential. To do this, we integrated the orbits of each cluster during 4 Gyr in the non-axisymmetric potential that includes the rotating bar/bulge. The evolution of each globular cluster on the (L z , E)-diagram is shown in Figure 5. The initial positions of clusters in (L z , E) space are indicated in Figure 5 with asterisks. As can be seen in the Figure, in the non-axisymmetric potential of the bar, rotating with the constant angular velocity Ω b , the clusters oscillate along a straight line that has the same slope for each cluster. This is due to the fact that in the non-axisymmetric time-dependent potential, the z-component of the angular momentum L z and the total energy E of the object are not conserved. Instead, the Jacobi integral described by the equation: is conserved for each cluster (see, e.g., Hattori et al. [31]). Equation (9) shows that in case of constant angular velocity of the bar, the total energy E is a linear function of L z , resulting in the clusters oscillating in Figure 5 along parallel lines.
As can be seen in Figure 5, the positions of GSE and Pontus clusters are significantly affected by the Milky Way bar, which blur their positions on the (L z , E) diagram over time. Energy variation for clusters is up to 15%, and variation of the angular momentum is more than 100%. It should also be noted that some Pontus objects reach energy values of the GSE group.
We then took into account the errors in observational data of coordinates and proper motions, leading to the uncertainties in the initial galactocentric coordinates and the velocities of the the globular clusters. To do this, we randomly generated one hundred sets of the coordinates in the 6D phase space for each globular cluster. They were drawn assuming Gaussian errors in the measured coordinates and velocities, and then we calculated the positions of the clusters on the (L z , E)-plane. Before discussing the evolution of the globular clusters on the (L z , E)-plane, a closer look at the dynamics of one globular cluster in the Pontus group-NGC 288-is necessary. Figure  6 shows time dependence of the angular momentum L z and the total energy E for the mean values of the initial conditions (red line), while also taking into account observational errors (100 grey dotted lines). As one can see, the angular momentum and the energy of the cluster are significantly affected by the time-dependent potential of the rotating bar. The values oscillate with time and correlate with each other according to Equation   Figure 7 shows the three-dimensional orbits of the globular cluster NGC 288 in the axisymmetric potential McMillan17 and in the non-axisymmetric potential with the rotating bar. Figure 8 shows the orbits of the cluster projected onto the three galactocentric planes. The mean orbits are shown as red lines, whereas one hundred alternative orbits consistent with the observational errors of these initial conditions are shown in grey. We can see that the bar significantly affects the dynamics of the globular cluster: the bar significantly randomises the orbits and destroys orbital boxes (diagram R vs. z in Figure 8), confirming the significant influence of the bar stressed by Chemel et al. [30] and Hattori et al. [31]. The cluster NGC 288 has a small pericentric distance similar to the other clusters considered in this paper. For the clusters that have larger pericenters, the influence of the bar will obviously decrease due to the increase in the distance of the symmetry of the equipotential surfaces of the bar.  McMillan17 (left panel) and the non-axisymmetric rotating bar/bulge potential (right panel): the red lines show the orbits for the mean initial conditions. The grey lines show one hundred orbits consistent with the observational errors in these initial conditions. Finally, let us consider the joint influence of the bar and of the observational errors on the evolution of the clusters in the (L z , E)-plane. To do this, we took one hundred representations of the orbits using the errors in Table 1 and represented by dots in Figure 4. This is illustrated in Figure 9, which shows the evolution of the clusters-represented by one hundred dots for each clusters in Figure 4-in the non-stationary potential of the rotating bar, and under the influence of the uncertainties in the initial kinematical data of the clusters. It is difficult to conclude that these clusters belong to two different accretion events, namely, Pontus and GSE, as was suggested by Malhan et al. [26]. The "tentative" cluster NGC 6864 overlaps with both groups. Thus, it cannot be concluded that GSE and Pontus clusters belong to different accretion events and so to different progenitor galaxies. The same can be said about the origin of the tentative object NGC 6864. It is also impossible to make firm conclusion on which group the tentative object NGC 6864 belongs to. In more general terms, we conclude that it is difficult to identify accretion events using objects that have close passages around the galactic centre. For those objects, the bar significantly affects the values of E and L z . One should also take into account that during accretion of a dwarf galaxy, dynamical friction constantly changes its energy and momentum, making the identification of the original accretion events even more difficult.
Finally, we did not take into account the influence of other important physical phenomena that can also affect the positions of the clusters on the L z , E-diagram such as the change of the galactic potential due to galaxy's accretion and secular evolution, the change of the angular velocity of bar during its evolution, the influence of the spiral structure, and the interaction of globular clusters with satellite galaxies on evolution of the globular clusters. We will attempt to include the effect of these phenomena in the future.

Summary
In this study, we concluded that the bar significantly affects the dynamics of the globular clusters with small enough pericentric distances. We found that the bar potential renders the orbits of such clusters quite chaotic, confirming the results obtained by Chemel et al. [30]. If one takes into account the bar potential and the observational errors in the heliocentric coordinates and in the proper motions of the clusters, it is evident that the positions of the clusters on L z , E-diagram are not tight anymore, making it impossible to identify the cluster's association with a particular accretion event or progenitor galaxy. This result is especially true for those clusters that pass close enough to the galactic centre and have the perihelion less than about 3 kpc, i.e., where the influence of the bar is stronger. To stress it yet again, in this work we provided strong evidence that a proper accounting of the influence of the bar, together with observational errors, is mandatory. We have applied this to two particular accretion events-the GSE and the Pontus-and clearly showed that it is impossible to discriminate among different parent merger events.