1. Introduction
Research Contribution
Received: 10 March 2024; Accepted: 16 October 2024.
Key words and phrases. four-body problem \(\cdot \) Hill's approximation \(\cdot \) celestial mechanics \(\cdot \) contact geometry
MSC 2020 Mathematics Subject Classification. 70G45 \(\cdot \) 70F10 \(\cdot \) 53D35
In order to describe conveniently the dynamics of small bodies attracted by the gravitational field of three bodies in such a triangular central configuration, a restricted four-body problem (R4BP) becomes necessary. There are plenty of results on various models of the R4BP, such as [3] , [8] , [9] , [13] , [11] , [12] , [19] , [22] , [28] , [31] , [35] , [42] , [45] . Relevant for this work is the spatial equilateral circular one, in which three primaries circle around their common center of mass and forming an equilateral triangular configuration. In view of astronomical data associated to such configurations in the solar system, the mass of one of the primaries (the Trojan) is much smaller than the other two primaries. If one equates the mass of the Trojan to zero, the system represents the circular restricted three-body problem (CR3BP). Therefore, to study the dynamics in the vicinity of the Trojan, a practical and intelligent concept is to perform a Hill's approximation in the equilateral circular R4BP. One of Hill's main contributions was the discovery of one periodic solution with period equal to the synodic month of the Moon. Hill's lunar theory was, as Wintner said [47], p. 1 , “considered by Poincaré as representing a turning point in the history of celestial mechanics”. Poincaré sought to make periodic solutions central in the study of the global dynamics, a focus that has persisted since his pioneering work [39] . Extending Hill's concept to the equilateral circular R4BP was performed by Burgos-García and Gidea [12] , which is the central system in this paper. This problem studies the dynamics near the Trojan and pushes the two remaining primaries (e.g., Sun and Jupiter) to infinity, and depends on two parameters, the energy of the system and the mass ratio \(\mu \in [0,\frac {1}{2}]\) of the two primaries at infinity (system is symmetric with respect to \(\mu = \frac {1}{2}\)). The case \(\mu = 0\) corresponds to the classical Hill 3BP, therefore this Hill four-body model generalizes the classical Hill's approach. It is worth noting that this system is different as the one introduced by Scheeres [42] , in which the motion of a spacecraft in the Sun perturbed Earth–Moon system is studied. Moreover, this Hill four-body system was extended in [11] as a problem with oblate bodies modeling the Sun–Jupiter–Hektor–Skamandrios system. The Hill four-body system we consider has four Lagrange points, where \(L_1\) is symmetric to \(L_2\) (lying on the \(x\)-axis), and \(L_3\) is symmetric to \(L_4\) (lying on the \(y\)-axis). If the energy value \(c\) is below the first critical value \(H(L_{1/2})\), then the energy level set has one bounded component (where the origin is contained), which we denote by \(\Sigma _c^b\). This component is non-compact because of a singularity at the origin corresponding to collision. After performing Moser regularization, we obtain a compact 5-dimensional manifold, which we denote by \(\widetilde {\Sigma }_c^b\). The spatial system is invariant under a symplectic involution \(\sigma \) which is induced by the reflection at the ecliptic. The restriction of the spatial problem to the fixed point set Fix(\(\sigma \)) corresponds to the planar problem. In fact, we can restrict the whole procedure to Fix(\(\sigma \)) and obtain a compact 3-dimensional manifold, which we denote by \(\widetilde {\Sigma }_c^b|_{\text {Fix}(\sigma )}\). Our main result in this paper is the following theorem. \begin{align*} \widetilde {\Sigma }_c^b &\cong (S^*S^3,\xi _{st}),\quad \textit {if } c < H(L_{1/2}), \\ \widetilde {\Sigma }_c^b|_{\text {Fix}(\sigma )} &\cong (S^*S^2,\xi _{st}),\quad \textit {if } c < H(L_{1/2}). \end{align*} We consider three point masses (primaries), \(B_1\), \(B_2\) and \(B_3\), moving in circular periodic orbits in the same plane with constant angular velocity around their common center of gravity fixed at the origin, while forming an equilateral triangle configuration (see Figure 1 ). A fourth body \(B_4\) is significantly smaller than the other three and thus a negligible effect on their motion. We set \(B_1\) on the negative \(x\)-axis at the origin of time and assume that the corresponding three masses are \(m_1 \geq m_2 \geq m_3\). It is convenient to choose the units of mass, distance and time such that the gravitational constant is 1, and the period of the circular orbits is \(2\pi \). In these units the side length of the equilateral triangle configuration is normalized to be one, and \(m_1 + m_2 + m_3 = 1\). Moreover, it is convenient to use a rotating frame of reference that rotates with an angular velocity of the orbital angular rate of the primaries. Then, the dynamics of the infinitesimal body \(B_4\) is described by the Hamiltonian \[ H(x,y,z,p_x,p_y,p_z) = \frac {1}{2} \left ( p_x^2 + p_y^2 + p_z^2 \right ) - \frac {m_1}{r_1} - \frac {m_2}{r_2} - \frac {m_3}{r_3} + p_x y - p_y x, \] which is a first integral of the system. An equivalent first integral is the Jacobi integral \(C\) defined by \(C = -2H\). Notice that \(r_i\) indicates the corresponding distance from \(B_4\) to \(i\)-th primary, for \(i=1,2,3\). The general expressions of the position coordinates \((x_i,y_i,0)\) can be seen in [9] . If \(m_3=0\) and \(m_2 = \mu \), then one recovers the constellation of the CR3BP associated to \(B_1\) and \(B_2\), where \(B_3\) is located at the equilateral Lagrange point \(L_4\). Moreover, the phase space is the trivial cotangent bundle \(T^* \left ( \mathbb {R}^3 \setminus \{ B_1,B_2,B_3 \} \right ) = \left ( \mathbb {R}^3 \setminus \{ B_1,B_2,B_3 \} \right ) \times \mathbb {R}^3\), endowed with the standard symplectic form \(\omega = \sum d p_k \wedge d k\) (\(k=x,y,z\)). The flow of the Hamiltonian vector field \(X_H\), defined by \(dH(\cdot ) = \omega ( \cdot , X_H )\), is equivalent to the equations of motion, \(\left \{ \dot {k} = \frac {\partial H}{\partial p_k}, \dot {p}_k = - \frac {\partial H}{\partial k} \right \}\) (\(k=x,y,z\)). We now briefly recall the fundamental steps of Hill's approximation, as performed in [12] where the details can be seen. Let \(B_3\) be the primary (the Trojan), whose mass is much smaller than the other two primaries. The first step is to set the Trojan to the origin. The second step rescales symplectically the coordinates depending on \(m_3^{1/3}\). The third step makes use of a Taylor expansion of the gravitational potential of the Hamiltonian in powers of \(m_3^{1/3}\). Finally, the limiting case for \(m_3 \to 0\) yields the Hamiltonian \[ H (x,y,z,p_x,p_y,p_z) = \frac {1}{2} \left ( p_x^2 + p_y^2 + p_z^2 \right ) + p_x y - p_y x - \frac {1}{r} + \frac {1}{8}x^2 - \frac {3\sqrt {3}}{4}(1-2\mu )xy -\frac {5}{8}y^2 + \frac {1}{2}z^2, \] where \(r = \left ( x^2 + y^2 + z^2 \right )^{\frac {1}{2}}\), \(m_1 = 1 - \mu \) and \(m_2 = \mu \). Notice that if one expands the Hamiltonian of the CR3BP centered at the equilateral Lagrange point \(L_4\), then the quadratic part corresponds to \(H + 1/r\). \begin{align} \label {hamiltonian*1} H (x,y,z,p_x,p_y,p_z) = \frac {1}{2} \left ( p_x^2 + p_y^2 + p_z^2 \right ) + p_x y - p_y x - \frac {1}{r} + a x^2 + b y^2 + \frac {1}{2}z^2, \end{align} \[ a = \frac {1}{2}(1 - \lambda _2),\quad b = \frac {1}{2} (1-\lambda _1),\quad \lambda _1 = \frac {3}{2}(1-d),\quad \lambda _2 = \frac {3}{2}(1+d),\quad d = \sqrt {1 - 3\mu + 3\mu ^2}. \] Since \(d(1-\mu )=d(\mu )\), we can assume that \(\mu \in [0,\frac {1}{2}]\). Notice that \(\lambda _1\) and \(\lambda _2\) are the eigenvalues corresponding to the rotation transformation in the \(xy\)-plane. The quantities \(a,b,\lambda _1,\lambda _2\) and \(d\) are plotted in Figure 2 . The Hamiltonian ( 1 ) consists of the rotating Kepler problem (formed by the Trojan and the infinitesimal body) with a velocity independent gravitational perturbation produced by the two remaining massive primaries (the degree 2 term \(a x^2 + b y^2 + \frac {1}{2}z^2\)) which are sent at infinite distance. By introducing the effective potential \begin{align} \label {effective*potential} U \colon \mathbb {R}^3 \setminus \{ 0 \} \to \mathbb {R},\quad (x,y,z) \mapsto - \frac {1}{r} - \frac {1}{2} \left ( \lambda _2 x^2 + \lambda _1 y^2 - z^2 \right ), \end{align} \begin{align} \label {hamiltonian*2} H (x,y,z,p_x,p_y,p_z) = \frac {1}{2} \left ( (p_x + y)^2 + (p_y - x)^2 + p_z^2 \right ) + U(x,y,z), \end{align} \begin{align} \label {hamiltonian*equation} \ddot {x} - 2 \dot {y} &= - \frac {\partial U}{\partial x} = \left (\lambda _2 - \frac {1}{r^3} \right )x \nonumber \\ \ddot {y} + 2 \dot {x} &= - \frac {\partial U}{\partial y} = \left (\lambda _1 - \frac {1}{r^3} \right )y\\ \ddot {z} &= - \frac {\partial U}{\partial z} = - \left (1 + \frac {1}{r^3} \right )z. \nonumber \end{align} A “symmetry” \(\sigma \) is, by definition, a symplectic or anti-symplectic involution of the phase space which leaves the Hamiltonian invariant, i.e., \begin{align} \label {symmetry*equations} H \circ \sigma = H,\quad \sigma ^2 = \text {id}, \quad \sigma ^* \omega = \pm \omega . \end{align} The reflection at the ecliptic \(\{z = 0\}\) gives rise to a linear symplectic symmetry of ( 1 ), denoted by \begin{align} \label {sigma} \sigma (x,y,z,p_x,p_y,p_z) = (x,y,-z,p_x,p_y,-p_z), \end{align} \[ \pi \colon \Sigma _s \to \Sigma _p,\quad \rho \mapsto \rho | _{\text {Fix}(\sigma )}. \] If \(\rho \in \Sigma _s\), then \(\rho | _{\text {Fix}(\sigma )} \in \Sigma _p\) with the corresponding (anti-)symplectic property. While \(\pi \) is not injective (since \(\pi (\rho _1) = \pi (\rho _2)\)), it is surjective. If \(\rho \in \Sigma _p\) is symplectic (or anti-symplectic), then a symplectic (or anti-symplectic) extension is given by \(z \mapsto z \) and \(p_z \mapsto p_z\) (or \(z\mapsto -z\) and \(p_z \mapsto p_z\)). From the third equation in ( 4 ) it is obvious that all Lagrange points are located at the ecliptic \(\{z=0\}\). Using the projection onto the configuration space given by \begin{align} \label {projection} \pi \colon \mathbb {R}^3 \setminus \{ 0 \} \times \mathbb {R}^3 \to \mathbb {R}^3 \setminus \{ 0 \},\quad (x,y,z,p_x,p_y,p_z) \mapsto (x,y,z), \end{align} \[ L_1 = \left ( \frac {1}{\sqrt [3]{\lambda _2}},0,0 \right ),\quad L_2 = \left ( - \frac {1}{\sqrt [3]{\lambda _2}},0,0 \right ),\quad L_3 = \left ( 0, \frac {1}{\sqrt [3]{\lambda _1}},0 \right ),\quad L_4 = \left ( 0, - \frac {1}{\sqrt [3]{\lambda _1}},0 \right ). \] Note that \(L_{1/2}\) are related to each other by \(\rho _y\) (reflection at the \(y\)-axis), and \(L_{3/4}\) are related to each other by \(\rho _x\) (reflection at the \(x\)-axis). The classical Hill 3BP (\(\mu =0\)) only has \(L_{1/2}\), and especially, if \(\mu \to 0\) then \(\lambda _1 \to 0\), which means that \(L_3\) and \(L_4\) are sent to infinity. Therefore, the presence of a second primary at infinity for \(\mu \in (0,\frac {1}{2}]\) produces the two additional Lagrange points \(L_{3/4}\). Since \(\lambda _2 > \lambda _1\), we have for the critical values \[ H(L_{1/2}) = - \frac {3}{2} \sqrt [3]{\lambda _2} < - \frac {3}{2} \sqrt [3]{\lambda _1} = H(L_{3/4}),\quad \text {for all } \mu \in (0,\frac {1}{2}]. \] We now consider the energy level set \(\Sigma _c := H^{-1}(c)\), for \(c \in \mathbb {R}\). In view of the footpoint projection ( 7 ), the “Hill's region” of \(\Sigma _c\) is defined as \[ \mathscr {K}_c := \pi (\Sigma _c) \subset \mathbb {R}^3 \setminus \{ 0 \}, \] which means that the Hill's region of the energy level set is its shadow under the footpoint projection. Since the first three terms in ( 3 ) are quadratic and hence non-negative, we can obtain the Hill's region by \[ \mathscr {K}_c = \left \{ (x,y,z) \in \mathbb {R}^3 \setminus \{0\} \mid U(x,y,z) \leq c \right \}. \] The topology of the Hill's region depends on the energy level. If \(c < H(L_{1/2})\), then the Hill's region has two connected components, one bounded and one unbounded (see Figure 3 ). We denote the bounded component by \(\mathscr {K}_c^b\) and abbreviate by \begin{align} \label {bounded*component} \Sigma _c^b := \pi ^{-1} (\mathscr {K}_c^b) \cap \Sigma _c \end{align} We now recall some basic definitions and notations from contact geometry, and refer for details to [23] . \[X = \sum _{i=1}^n p_i \frac {\partial }{\partial p_i},\] that is, the radial vector field in fiber direction. Hypersurfaces of contact type. Let \(X\) be a Liouville vector field on a \(2n+2\) dimensional symplectic manifold \((M,\omega )\). Then \(\alpha := \iota _X \omega |_{\Sigma } \) is a contact form on any hypersurface \(\Sigma \subset M\) transverse to \(X\) (i.e., with \(X\) nowhere tangent to \(\Sigma \)). Such hypersurfaces are said to be of “contact type”. To see this, let \(x \in \Sigma \) and let \(\{v_1,...,v_{2n+1} \}\) be a basis of \(T_x\Sigma \). By using the Liouville condition we have, \begin{align} \label {contact} \alpha \wedge (d \alpha )^{\wedge n} (v_1,...,v_{2n+1}) = \iota _X \omega \wedge \omega ^{\wedge n} (v_1,...,v_{2n+1}) = \frac {1}{n} \omega ^{\wedge (n+1)} (X,v_1,...,v_{2n+1}). \end{align} Any hypersurface \(\Sigma \subset M\) has a characteristic foliation \(L\) which is a rank 1 foliation with \(L_x = \text {ker} \left ( \omega | _{T_x \Sigma } \right )\), for \(x \in \Sigma \). If \(\Sigma \) is a energy level set of a Hamiltonian \(H \colon M \to \mathbb {R}\), then for \(x \in \Sigma \) we have that \(X_H(x) \in L_x\). If \(\Sigma \) is of contact type, then \(R_{\alpha }(x) \in L_x\), i.e., the Reeb flow of \(\alpha \) is a reparametrization of the Hamiltonian flow. In the case of \(M=T^* Q\), if the contact form on \(\Sigma \subset T^*Q\) is induced by the transversality of the natural Liouville vector field \(X\) on \(T^*Q\), then the contact structure is called the “standard contact structure” determined by \[\xi _{st} = \text {ker} \alpha _{can},\quad \alpha _{can} := \iota _X \omega _{can} | _{\Sigma } = \lambda _{can} | _{\Sigma }.\] Moreover, in this case the energy hypersurface \(\Sigma \subset T^* Q\) is “fiberwise starshaped”, i.e., for each point \(q \in Q\) the intersection \(\Sigma \cap T_q^*Q\) bounds a starshaped domain in the linear space \(T_q^*Q\), which means that the natural Liouville vector field is transverse to each \(\Sigma \cap T_q^*Q\). We now consider the Liouville vector field on \(T^* \mathbb {R}^3\) given by \begin{align} \label {liouville*vector*field} X = x \frac {\partial }{\partial x} + y \frac {\partial }{\partial y} + z \frac {\partial }{\partial z}. \end{align} \begin{align*} x &= \rho \cos \theta \sin \varphi \\ y &= \rho \sin \theta \sin \varphi \\ z &= \rho \cos \varphi \end{align*} where \(0 \leq \theta \leq 2\pi \), \(0 \leq \varphi \leq \pi \). Since we consider energy level sets below the first critical value, the radius \(\rho \) is always smaller than the distance from \(L_{1/2}\) to the origin, which is \(1 / \sqrt [3]{\lambda _2}\) and always less than 1. Therefore, we assume that the radius \(\rho \) is smaller than 1. Now the effective potential ( 2 ) reads \[ U (\rho , \theta , \varphi ) = - \frac {1}{\rho } - \frac {1}{2} \rho ^2 ( \lambda _2 \cos ^2 \theta \sin ^2 \varphi + \lambda _1 \sin ^2 \theta \sin ^2 \varphi - \cos ^2 \varphi ), \] which is \(\pi \)-periodic in the variables \(\theta \) and \(\varphi \). \[ dU_{\rho } (\theta ,\varphi ) = \rho ^2 (\lambda _2 - \lambda _1) \cos \theta \sin \theta \sin ^2 \varphi d \theta + \rho ^2 \sin \varphi \cos \varphi (\lambda _2 \cos ^2 \theta + \lambda _1 \sin ^2 \theta + 1) d \varphi . \] Since \(\lambda _2 > \lambda _1\), and the term \(\lambda _2 \cos ^2 \theta + \lambda _1 \sin ^2 \theta + 1\) is strictly positive, we find four critical points at \((0,0)\), \((0,\frac {\pi }{2})\), \((\frac {\pi }{2})\) and \((\frac {\pi }{2},\frac {\pi }{2})\). The corresponding Hessians are given by \begin{align*} H_{U_{\rho }} (0,0) &= \begin{pmatrix} 0 & 0 \\ 0 & -\rho ^2(\lambda _2 + 1) \end {pmatrix},\quad H_{U_{\rho }} (0,\frac {\pi }{2}) = \begin{pmatrix} \rho ^2 (\lambda _2 - \lambda _1) & 0 \\ 0 & \rho ^2(\lambda _2 + 1) \end {pmatrix} \\ H_{U_{\rho }} (\frac {\pi }{2},0) &= \begin{pmatrix} 0 & 0 \\ 0 & -\rho ^2(\lambda _1 + 1) \end {pmatrix},\quad H_{U_{\rho }} (\frac {\pi }{2},\frac {\pi }{2}) = \begin{pmatrix} - \rho ^2 (\lambda _2 - \lambda _1) & 0\\ 0 & \rho ^2(\lambda _1 + 1) \end {pmatrix}. \end{align*} Therefore, the function \(U_{\rho }\) attains its minimum at \((\theta ,\varphi ) = (0,\frac {\pi }{2})\). \[ B_r(0) := \{ (x,y,z) \in \mathbb {R}^3 \colon x^2 + y^2 + z^2 \leq r^2 \} \] the ball of radius \(r\) centered at the origin. \begin{align} \label {U*c} U(r,\theta ,\varphi ) \geq U(r,0,\frac {\pi }{2}) = - \frac {1}{r} - \frac {1}{2}r^2 \lambda _2 = - \frac {3}{2}\sqrt [3]{\lambda _2} = H(L_{1/2}) > c. \end{align} \begin{align} \label {equivalence} (\lambda _1 - \lambda _2)\sin ^2 \theta \leq 0 \quad \Leftrightarrow \quad \lambda _2(\cos ^2 \theta - 1) + \lambda _1 \sin ^2 \theta \leq 0 \quad \Leftrightarrow \quad \lambda _2 \cos ^2 \theta + \lambda _1 \sin ^2 \theta \leq \lambda _2. \end{align} \begin{align} \label {partial*U*rho} \frac {\partial U}{\partial \rho } = \frac {1}{\rho ^2} - \rho \left ( \lambda _2 \cos ^2 \theta \sin ^2 \varphi + \lambda _1 \sin ^2 \theta \sin ^2 \varphi - \cos ^2 \varphi \right ) \geq \frac {1}{\rho ^2} - \lambda _2 \rho > 0. \end{align} \[ \frac {\partial ^2 U}{\partial \rho ^2} = - \frac {2}{\rho ^3} + \cos ^2 \varphi - \sin ^2 \varphi \left ( \lambda _2 \cos ^2 \theta + \lambda _1 \sin ^2 \theta \right ) \leq - \frac {2}{r^3} + 1 = -2 \lambda _2 + 1 \leq -3 \leq - \sin ^2 \varphi . \] \begin{align} \label {inequality*to*show*1} d H (X) | _{\Sigma _c^b} > 0. \end{align} \begin{align} \label {dH} dH = & p_x dp_x + p_y dp_y + p_z dp_z + p_x dy + y dp_x - p_y dx - x dp_y \\ & + 2ax dx + 2by dy + z dz + \frac {x}{r^3} dx + \frac {y}{r^3} dy + \frac {z}{r^3} dz. \nonumber \end{align} \begin{align} \label {dHX} dH(X) = p_x y - p_y x + 2ax^2 + 2by^2 + z^2 + \frac {1}{r}. \end{align} \[X = \rho \frac {\partial }{\partial \rho },\] and ( 16 ) reads \begin{align} \label {dHX*2} dH(X) = & p_x \rho \sin \theta \sin \varphi - p_y \rho \cos \theta \sin \varphi + (1 + \lambda _2) \rho ^2 \cos ^2 \theta \sin ^2 \varphi \\ & + (1 - \lambda _1) \rho ^2 \sin ^2 \theta \sin ^2 \varphi + \rho ^2 \cos ^2 \varphi + \frac {1}{\rho }. \nonumber \end{align} \begin{align*} dH(X) = \rho \sin \theta \sin \varphi (p_x + \rho \sin \theta \sin \varphi ) - \rho \cos \theta \sin \varphi (p_y - \rho \cos \theta \sin \varphi ) + \rho \frac {\partial U}{\partial \rho }, \end{align*} which we estimate by using the Cauchy–Schwarz inequality, \begin{align*} dH(X) &\geq \rho \frac {\partial U}{\partial \rho } - \rho \sin \varphi \sqrt { (p_x + \rho \sin \theta \sin \varphi )^2 + (p_y - \rho \cos \theta \sin \varphi )^2 } \\ & = \rho \frac {\partial U}{\partial \rho } - \rho \sin \varphi \sqrt { 2 (H - U) - p_z^2 } \\ & \geq \rho \frac {\partial U}{\partial \rho } - \rho \sin \varphi \sqrt {2 (H - U)}. \end{align*} Therefore, we have \[ dH(X) | _{\Sigma _c^b} \geq \rho \left ( \frac {\partial U}{\partial \rho } - \sin \varphi \sqrt {2(c-U)} \right ). \] Since the right hand side is independent of the momentum coordinates, to prove ( 14 ) it is suffices to show that \begin{align} \label {inequality*to*show*2} \left ( \frac {\partial U}{\partial \rho } - \sin \varphi \sqrt {2(c-U)} \right ) \bigg | _{\mathscr {K}_c^b} > 0. \end{align} \[ U(\rho + \tau , \theta , \varphi ) = c. \] By using Lemma 3.7 and Lemma 3.8 we obtain \begin{align*} \left ( \frac {\partial U}{\partial \rho } (\rho ,\theta ,\varphi ) \right )^2 &= \left ( \frac {\partial U}{\partial \rho } (\rho + \tau ,\theta ,\varphi ) \right )^2 - \int _{0}^{\tau } \frac {d}{dt} \left ( \frac {\partial U}{\partial \rho } (\rho + t,\theta ,\varphi ) \right )^2 dt \\ &> -2 \int _{0}^{\tau } \frac {\partial U}{\partial \rho } (\rho + t,\theta ,\varphi ) \frac {\partial ^2 U}{\partial \rho ^2} (\rho + t,\theta ,\varphi ) dt \\ &\geq 2 \sin ^2 \varphi \int _{0}^{\tau } \frac {\partial U}{\partial \rho } (\rho + t,\theta ,\varphi ) dt \\ &= 2 \sin ^2 \varphi \left ( U(\rho + \tau ,\theta ,\varphi ) - U(\rho ,\theta ,\varphi ) \right ) \\ &= 2 \sin ^2 \varphi \left ( c - U(\rho ,\theta ,\varphi ) \right ). \end{align*} Therefore, by using Lemma 3.7 once more, we imply \[ \frac {\partial U}{\partial \rho } (\rho ,\theta ,\varphi ) > \sin \varphi \sqrt { 2 \left ( c - U(\rho ,\theta ,\varphi ) \right )}, \] which shows ( 18 ) and thereby the proposition. The Hamiltonian ( 1 ) has a singularity at the origin corresponding to collisions, thus the bounded component \(\Sigma _c^b\) of the energy level set is non-compact. Moser [34] observed that the regularized Kepler problem coincides with the geodesic flow on the sphere endowed with its standard metric by interchanging the roles of position and momenta. To remove the singularity in our problem, we use the same concept as introduced by Moser. We abbreviate by \(\mathbf {X} = (x,y,z)\) and \(\mathbf {P}=(p_x,p_y,p_z)\) the corresponding position and momentum coordinates. We use a new time parameter \(s\) and define for an energy value \(c < H(L_{1/2}) = - \frac {3}{2}\sqrt [3]{\lambda _2}\) a new Hamiltonian by \[ s = \int \frac {dt}{|\mathbf {X}|},\quad K_c(\mathbf {X},\mathbf {P}) := |\mathbf {X}| \left ( H(\mathbf {X},\mathbf {P}) - c \right ), \] Notice that the flow of \(H\) at energy level \(c\) corresponds to the flow of \(K_c\) at energy level \(0\). Now we interchange the roles of position and momenta by the symplectic transformation mapping \((\mathbf {X},\mathbf {P})\) to \((-\mathbf {P},\mathbf {X})\). For simplicity of notation, we replace the new coordinates \(\mathbf {X'}=-\mathbf {P}\) and \(\mathbf {P'}=\mathbf {X}\) by \(\mathbf {X}\) and \(\mathbf {P}\). Then, the new transformed Hamiltonian \(\widetilde {K}_c(\mathbf {X},\mathbf {P}) = K_c(-\mathbf {P},\mathbf {X})\) is explicitly given by \begin{align} \label {hamiltonian*Kc} \widetilde {K}_c (\mathbf {X},\mathbf {P}) &= \frac {1}{2}|\mathbf {X}|^2|\mathbf {P}| + |\mathbf {P}|(p_x y - p_y x) - 1 + |\mathbf {P}|(a p_x^2 + b p_y^2 + \frac {1}{2}p_z^2) - |\mathbf {P}|c \\ &= \frac {1}{2} \left ( |\mathbf {X}|^2 + 1 \right )|\mathbf {P}| + (p_x y - p_y x)|\mathbf {P}| -1 + (a p_x^2 + b p_y^2 + \frac {1}{2}p_z^2)|\mathbf {P}| - (c+\frac {1}{2})|\mathbf {P}|. \nonumber \end{align} \begin{gather} \label {stereographic*symplectic} x = \frac {\xi _1}{1-\xi _0},\quad y = \frac {\xi _2}{1-\xi _0},\quad z = \frac {\xi _3}{1-\xi _0}, \\ p_x = \eta _1(1-\xi _0) + \xi _1 \eta _0,\quad p_y = \eta _2(1-\xi _0) + \xi _2 \eta _0,\quad p_z = \eta _3(1-\xi _0) + \xi _3 \eta _0.\nonumber \end{gather} \begin{gather*} \xi _0 = \frac {|\mathbf {X}|^2 - 1}{|\mathbf {X}|^2 + 1},\quad \xi _1 = \frac {2x}{|\mathbf {X}|^2 + 1},\quad \xi _2 = \frac {2y}{|\mathbf {X}|^2 + 1},\quad \xi _3 = \frac {2z}{|\mathbf {X}|^2 + 1}, \\ \eta _0 = \langle \mathbf {X},\mathbf {P} \rangle ,\quad \eta _1 = \frac {|\mathbf {X}|^2+1}{2}p_x - \langle \mathbf {X},\mathbf {P} \rangle x,\quad \eta _2 = \frac {|\mathbf {X}|^2+1}{2}p_y - \langle \mathbf {X},\mathbf {P} \rangle y,\quad \eta _3 = \frac {|\mathbf {X}|^2+1}{2}p_z - \langle \mathbf {X},\mathbf {P} \rangle z, \end{gather*} and, in addition, we have the relation \begin{align} \label {relation} |\eta | = \frac {1}{2}(|\mathbf {X}|^2 + 1)|\mathbf {P}| = \frac {|\mathbf {P}|}{1 - \xi _0}. \end{align} \begin{align} \label {ham*K*c*xi*eta} \widetilde {K}_c (\xi ,\eta ) = |\eta | f(\xi ,\eta ) - 1, \end{align} \begin{gather*} f(\xi ,\eta ) := 1 + (\eta _1 \xi _2 - \eta _2 \xi _1)(1-\xi _0) + (a g_1^2 + b g_2^2 + \frac {1}{2} g_3^2)(1 - \xi _0) - (c + \frac {1}{2})(1 - \xi _0), \\ g_k := g_k(\xi ,\eta ) := \eta _k(1 - \xi _0) + \xi _k \eta _0,\quad k=1,2,3. \end{gather*} By shifting and squaring the Hamiltonian ( 22 ) we obtain the new smooth Hamiltonian \(Q(\xi ,\eta )\) on a subset of \(T^*S^3\), \begin{align} \label {Hamiltonian*Q} Q(\xi ,\eta ) = \frac {1}{2}|\eta |^2 f(\xi ,\eta )^2. \end{align} \begin{align} \label {points*near*origin} |\mathbf {P}|=|\eta |(1-\xi _0) < \varepsilon . \end{align} \begin{align} \label {liouville*v*f} X = \sum _{i = 0}^{3} \eta _i \frac {\partial }{\partial \eta _i}, \end{align} is transverse to \(Q^{-1}(\frac {1}{2})\) over points \((\xi ,\eta )\) satisfying ( 24 ). \begin{align} \label {to*show} dQ(X) | _{Q^{-1}(\frac {1}{2})} > 0. \end{align} \begin{align*} dQ(X) &= |\eta |^2 f(\xi ,\eta )^2 + |\eta |^2 f(\xi ,\eta ) \sum _{i=0}^{3} \frac {\partial f}{\partial \eta _i} (\xi ,\eta ) \eta _i \\ &= 2 Q + |\eta |^2 f(\xi ,\eta ) (1-\xi _0) ( \eta _1 \xi _2 - \eta _2 \xi _1 + 2a g_1^2 + 2b g_2^2 + g_3^2 ). \end{align*} In order to prove ( 26 ), we first show that we can choose \(\varepsilon > 0\) so small such that \begin{align} \label {bound*lower*f} |f(\xi ,\eta )| \geq \frac {1}{2}. \end{align} \begin{align*} |f(\xi ,\eta )| &= \left | 1 + (\eta _1 \xi _2 - \eta _2 \xi _1)(1-\xi _0) + (a g_1^2 + b g_2^2 + \frac {1}{2} g_3^2)(1 - \xi _0) - (c + \frac {1}{2})(1 - \xi _0) \right | \\ &= \left | 1 + ( b g_2^2 + \frac {1}{2} g_3^2 - (c + \frac {1}{2}) )(1-\xi _0) + (\eta _1 \xi _2 - \eta _2 \xi _1)(1-\xi _0) + a g_1^2(1 - \xi _0) \right | \\ &\geq 1 - |\eta _1 \xi _2 - \eta _2 \xi _1|(1-\xi _0) - |a|g_1^2(1-\xi _0) \\ &\geq 1 - |\eta _1 \xi _2 - \eta _2 \xi _1|(1-\xi _0) - g_1^2(1-\xi _0). \end{align*} Furthermore, \(|\eta _1 \xi _2 - \eta _2 \xi _1| \leq |\eta ||\xi |\), and because \(|\xi |=1\), we have in view of ( 24 ), \begin{align} \label {inequality*epsilon} |\eta _1 \xi _2 - \eta _2 \xi _1|(1-\xi _0) \leq |\eta |(1-\xi _0) < \varepsilon . \end{align} \[ |f(\xi ,\eta )| \geq 1 - \varepsilon - g_1^2(1-\xi _0). \] If \(\varepsilon \) approaches 0, then \(\xi _0 \to 1\), which means that we can choose \(\varepsilon \) so small such that ( 27 ) holds. By using the level set condition \(Q^{-1}(\frac {1}{2})\) together with the lower bound ( 27 ) for \(|f(\xi ,\eta )|\), we find \[ \frac {1}{2} = Q(\xi ,\eta ) = \frac {1}{2} |\eta |^2 f(\xi ,\eta )^2 \geq \frac {1}{2} |\eta |^2 \frac {1}{2}, \] which gives an upper bound for \(|\eta |\), i.e., \[ dQ(X) \geq 2 Q - |\eta |^2 \left | f(\xi ,\eta ) \right | \left | (1-\xi _0) \left ( \eta _1 \xi _2 - \eta _2 \xi _1 + 2a g_1^2 + 2b g_2^2 + g_3^2 \right ) \right |. \] Notice that by ( 29 ) we obtain \[|\eta ||\eta ||f(\xi ,\eta )| \leq 2 \sqrt { 2Q(\xi ,\eta ) } = 2 \sqrt {2 \frac {1}{2}} = 2, \] which implies, together with ( 28 ), \begin{align*} dQ(X) &\geq 1 -2 \left ( \left | (1-\xi _0) (\eta _1 \xi _2 - \eta _2 \xi _1) \right | + \left | (1-\xi _0) (2a g_1^2 + 2b g_2^2 + g_3^2 ) \right | \right ) \\ &\geq 1 -2 \varepsilon \left (1 + |2a g_1^2 + 2b g_2^2 + g_3^2| \right ). \end{align*} Since the latter term can be bounded by some constant \(A\) on a compact set away from the origin, we obtain \[dQ(X) \geq 1 - 2\varepsilon (1 + A). \] Now we choose \(\varepsilon \) sufficiently small such that \(dQ(X)>0\), which proves ( 26 ). We have seen that for \(c < H(L_{1/2})\) the bounded component \(\Sigma _c^b\) of the energy level set can be Moser-regularized to form a compact 5-dimensional manifold \(\widetilde {\Sigma }_c^b \subset T^*S^3\) which is diffeomorphic to \(S^*S^3\). Since the Liouville vector field ( 10 ) on \(T^*\mathbb {R}^3\) and the natural one ( 25 ) on \(T^*S^3\) coincide after Moser regularization, we obtain a Liouville vector field that is defined near the whole regularized level set, and in fact, it is the natural one. By the transversality results from Proposition 3.4 and Proposition 3.9 we obtain that the natural Liouville vector field on \(T^*S^3\) is transverse to \(\widetilde {\Sigma }_c^b\), which means that \(\widetilde {\Sigma }_c^b\) is fiberwise starshaped, and moreover, \(\widetilde {\Sigma }_c^b \cong (S^*S^3,\xi _{st})\). For the planar problem, one can of course perform the same computation to obtain the same result. But since the planar problem corresponds to the restriction of the spatial system to the fixed point set of the symplectic symmetry \(\sigma \) from ( 6 ), the transversality result in the planar case follows immediately. This consequence is based on a general construction from [5] . Namely, if a energy level set \(\Sigma \) is of contact type and the entire system has a symplectic symmetry \(\sigma \), such as ( 6 ), then the restriction of the contact form on \(\Sigma \) to \(\Sigma | _{\text {Fix}(\sigma | _{\Sigma })}\) is a contact form on \(\Sigma | _{\text {Fix}(\sigma | _{\Sigma })}\). Therefore, we have the same result in the planar problem, in which we denote the Moser-regularized compact 3-dimensional manifold by \(\widetilde {\Sigma }_c^b|_{\text {Fix}(\sigma )} \cong S^* S^2 \subset T^*S^2\). This completes the proof of Theorem 1.1 . [1] Albers P., Frauenfelder U., van Koert O., Paternain G.: Contact geometry of the restricted three-body problem. Comm. Pure Appl. Math. 65(2), 229–263 (2012) https://doi.org/10.1002/cpa.21380 [2] Almeida A.J.C., Peixinho N., Correia A.C.M.: Neptune Trojans and Plutinos: colors, sizes, dynamics, and their possible collisions. Astron. Astrophys. 508, 1021–1030 (2009) https://doi.org/10.1051/0004-6361/200911943 [3] Alvarez-Ramirez M., Vidal C.: Dynamical aspects of an equilateral restricted four-body problem. Math. Probl. Eng. 2009(181360) (2009) https://doi.org/10.1155/2009/181360 [4] Aydin C.: From Babylonian lunar observations to Floquet multipliers and Conley–Zehnder Indices. J. Math. Phys. 64(8) (2023) https://doi.org/10.1063/5.0156959 [5] Aydin C.: Symplectic splitting of Hamiltonian structures and reduced monodromy matrices. Linear Algebra Appl. 674, 426–441 (2023) https://doi.org/10.1016/j.laa.2023.06.008 [6] Aydin C.: The Conley–Zehnder indices of the spatial Hill three-body problem. Celest. Mech. Dyn. Astron. 135(32) (2023) https://doi.org/10.1007/s10569-023-10134-7 [7] Aydin C.: The linear symmetries of Hill's lunar problem. Arch. Math. (Basel) 120(3), 321–330 (2023) https://doi.org/10.1007/s00013-022-01822-1 [8] Baltagiannis A.N., Papadakis K.E.: Equilibrium points and their stability in the restricted four-body problem. Internat. J. Bifur. Chaos 21(8), 2179–2193 (2011) https://doi.org/10.1142/S0218127411029707 [9] Baltagiannis A.N., Papadakis K.E.: Periodic solutions in the Sun–Jupiter–Trojan Asteroid–Spacecraft system. Planet. Space Sci. 75, 148–157 (2013) https://doi.org/10.1016/j.pss.2012.11.006 [10] Birkhoff G.D.: The restricted problem of three bodies. Rend. Circ. Mat. Palermo 39, 265–334 (1915) https://doi.org/10.1007/BF03015982 [11] Burgos-García J., Celletti A., Gales C., Gidea M., Lam W.T.: Hill four-body problem with oblate bodies: an application to the Sun-Jupiter-Hektor-Skamandrios system. J. Nonlinear Sci. 30(6), 2925–2970 (2020) https://doi.org/10.1007/s00332-020-09640-x [12] Burgos-García J., Gidea M.: Hill's approximation in a restricted four-body problem. Celest. Mech. Dyn. Astron. 122, 117-141 (2015) https://doi.org/10.1007/s10569-015-9612-9 [13] Burgos-García J., Delgado J.: Periodic orbits in the restricted four-body problem with two equal masses. Astrophys. Space Sci. 345, 247–263 (2012) https://doi.org/10.1007/s10509-012-1118-2 [14] Cho W., Jung H., Kim G.: The contact geometry of the spatial circular restricted 3-body problem. Abh. Math. Semin. Univ. Hambg. 90, 161–181 (2020) https://doi.org/10.1007/s12188-020-00222-y [15] Colin V., Dehornoy P., Hryniewicz U., Rechtman A.: Generic properties of 3-dimensional Reeb flows: Birkhoff sections and entropy. Comment. Math. Helv. 99(3), 557–611 (2024) https://doi.org/10.4171/CMH/573 [16] Colin V., Dehornoy P., Rechtman A.: On the existence of supporting broken book decompositions for contact forms in dimension 3. Invent. math. 231, 1489–1539 (2023) https://doi.org/10.1007/s00222-022-01160-7 [17] Connors M. Stacey G., Brasser R., Wiegert P.: A survey of orbits of co-orbitals of Mars. Planet. Space Sci. 53, 617–624 (2005) https://doi.org/10.1016/j.pss.2004.12.004 [18] Cristofaro-Gardiner D., Hryniewicz U., Hutchings M., Liu H.: Contact three-manifolds with exactly two simple Reeb orbits. Geom. Topol. 27(9), 3801–3831 (2023) https://doi.org/10.2140/gt.2023.27.3801 [19] Cronin J., Richards P.B., Russell L.H.: Some periodic solutions of a four-body problem. Icarus 3, 423–428 (1964) https://doi.org/10.1016/0019-1035(64)90003-X [20] Floer A.: Morse theory for Lagrangian intersection. J. Differential Geom. 28(3), 513–547 (1988) https://doi.org/10.4310/jdg/1214442477 [21] Frauenfelder U., van Koert O.: The Restricted Three-Body Problem and Holomorphic Curves. Pathw. Math., Birkhäuser (2018) [22] Gabern F., Jorba A.: A restricted four-body model for the dynamics near the Lagrangian points of the Sun-Jupiter system. Discrete Contin. Dyn. Syst. Series B 1(2), 143–182(2001) https://doi.org/10.3934/dcdsb.2001.1.143 [23] Geiges, H.: An Introduction to Contact Topology. Cambridge studies in advanced mathematics 109, Cambridge University Press (2008) [24] Hill G.W.: Researches in the Lunar Theory. Am. J. Math. 1(3), 245–260 (1878) https://doi.org/10.2307/2369430 [25] Hofer H.: Pseudoholomorphic curves in symplectizations with applications to the Weinstein conjecture in dimension three. Invent. Math. 114, 515–563 (1993) https://doi.org/10.1007/BF01232679 [26] Hofer H., Wysocki K., Zehnder E.: The Dynamics on Three-Dimensional Strictly Convex Energy Surfaces. Ann. of Math. (2) 148(1), 197–289 (1998) https://doi.org/10.2307/120994 [27] Irie K.: Strong closing property of contact forms and action selecting functors. J. Fixed Point Theory Appl. 26(15) (2024) https://doi.org/10.1007/s11784-024-01102-1 [28] Kumari R., Kushvah B.S.: Stability regions of equilibrium points in restricted four-body problem with oblateness effects. Astrophys. Space Sci. 349, 693–704 (2014) https://doi.org/10.1007/s10509-013-1689-6 [29] Lagrange J.L.: Essai sur le problème des trois corps. Oeuvres 6, 229–331 (1772) [30] Lamb J.S.W., Roberts J.A.G.: Time-reversal symmetry in dynamical systems: A survey. Physica D 112, 1–39 (1998) https://doi.org/10.1016/S0167-2789(97)00199-1 [31] Michalodimitrakis M.: The circular restricted four-body problem. Astrophys. Space Sci. 75, 289–305 (1981) https://doi.org/10.1007/BF00648643 [32] Moreno A., Aydin C., van Koert O., Frauenfelder U., Koh D.: Bifurcation graphs for the CR3BP via symplectic methods. To appear in The Journal of the Astronautical Sciences (2024) [33] Moreno A., van Koert O.: Global hypersurfaces of section in the spatial restricted three-body problem. Nonlinearity 35(6), 2920–2970 (2022) https://doi.org/10.1088/1361-6544/ac692b [34] Moser J.: Regularization of kepler's problem and the averaging method on a manifold. Comm. Pure Appl. Math. 23(4), 609–636 (1970) https://doi.org/10.1002/cpa.3160230406 [35] Moulton F.R.: On a class of particular solutions of the problem of four bodies. Trans. Am. Math. Soc. 1(1), 17–29 (1900) [36] Murray C.D., Cooper N.J., Evans M.W., Beurle K.: S/2004 S 5: A new co-orbital companion for Dione. Icarus 179(1), 222–234 (2005) https://doi.org/10.1016/j.icarus.2005.06.009 [37] Nicholson S.B.: The Trojan Asteroids. Publ. Astr. Soc. Pacific Leaflets 8(381), 239–246 (1961) [38] Olkin C., Vincent M., Adam C. et al.: Mission Design and Concept of Operations for the Lucy Mission. Space. Sci. Rev. 220(47) (2024) https://doi.org/10.1007/s11214-024-01082-1 [39] Poincaré H.: Les méthodes nouvelles de la mécanique céleste. I–III, Gauthiers-Villars, Paris (1892) [40] Poincaré H.: Sur un théorème de géométrie. Rend. Circ. Mat. Palermo 33, 375–407 (1912) https://doi.org/10.1007/BF03015314 [41] Santana-Ros T., Micheli M., Faggioli L. et al.: Orbital stability analysis and photometric characterization of the second Earth Trojan asteroid 2020 \(\text {XL}*5\). Nat. Commun. 13(447) (2022) https://doi.org/10.1038/s41467-022-27988-4 [42] Scheeres D.J.: The Restricted Hill Four-Body Problem with Applications to the Earth–Moon–Sun System. Celest. Mech. Dyn. Astron. 70, 75-98 (1998) https://doi.org/10.1023/A:1026498608950 [43] Schwarz R., Dvorak R., Süli A., Erdi B.: Survey of the stability region of hypothetical habitable Trojan planets. Astron. Astrophys. 474(3), 1023–1029 (2007) https://doi.org/10.1051/0004-6361:20077994 [44] Stacey, R.G., Connors M.: A centenary survey of orbits of co-orbitals of Jupiter. Planet. Space Sci. 56, 358–367 (2008) https://doi.org/10.1016/j.pss.2007.11.002 [45] Steves B.A., Roy A.E.: Some special restricted four-body problems—I. Modelling the Caledonian problem. Planet. Space Sci. 46, 1465–1474 (1998) https://doi.org/10.1016/S0032-0633(98)00077-4 [46] Taubes C.H.: The Seiberg–Witten equations and the Weinstein conjecture. Geom. Topol. 11(4), 2117-2202 (2007) https://doi.org/10.2140/gt.2007.11.2117 [47] Wintner A.: On Hill's Periodic Lunar Orbit. Amer. J. Math. 60(4), 937–948 (1938) https://doi.org/10.2307/2371269 1. Introduction
Astronomical significance.
Hill's approximation.
Why we care about contact property.
Main result.
Theorem 1.1.
For any given \(\mu \in [0,\frac {1}{2}]\) it holds that \(\seteqnumber{0}{}{0}\) Organization of the paper.
2. Hill's approximation in the spatial equilateral circular R4BP
2.1. Hamiltonian.
Figure 1. Equilateral circular restricted four-body problem. Left: Case of \(m_1 > m_2 > m_3\). Right: Case of \(m_2 = m_3\) in a rotating frame of reference; \(B_2\) and \(B_3\) are located symmetrically with respect to \(B_1\)..
Furthermore, after applying a rotation in the \(xy\)-plane, the system is equivalent with the system characterized by the Hamiltonian
Figure 2. The quantities \(a\) (red), \(b\) (green), \(\lambda _1\), \(\lambda _2\) (both blue) and \(d\) (black).. 2.2. Linear symmetries.
Together with the previous linear symplectic symmetries, they form the group \(\mathbb {Z}_2 \times \mathbb {Z}_2 \times \mathbb {Z}_2\). If one restrict the system to Fix(\(\sigma \)), linear anti-symplectic symmetries for the planar problem are given by
that together with the linear symplectic ones \(\{ \pm \text {id} \}\) form a Klein-four group \(\mathbb {Z}_2 \times \mathbb {Z}_2\). These symmetries show that it is not possible to say which of the two primaries at infinity we are moving to or away from. Remark 2.1.
In [7] it shown that the Hill 3BP (\(\mu = 0\)) has two special properties.
2.3. Lagrange points and Hill's region.
Figure 3. Hill's region (gray shaded domains) for planar problem \(\{z=0\}\) for \(\mu =0.2\). White domains correspond to forbidden regions. Red dots indicate \(L_{1/2}\); blue dots indicate \(L_{3/4}\). Right: For \(c < H(L_{1/2})\). Left: For \(H(L_{1/2}) < c < H(L_{3/4})\). In the Hill 3BP (\(\mu =0\)), when \(L_{3/4}\) are sent to infinity, below the critical value the Hill's region consists of one bounded component and two unbounded components.. 3. Contact property - Proof of Theorem 1.1
3.1. Basic notations.
Definition 3.1.
Let \(M\) be a smooth manifold of odd dimension \(2n +1\). A “contact form” on \(M\) is a 1-form \(\alpha \in \Omega ^1(M)\) such that \(\alpha \wedge (d\alpha )^{\wedge n} \neq 0\). Given a contact form \(\alpha \), the hyperplane field \(\xi = \text {ker}\alpha \subset TM\) is oriented by \(d\alpha \), and this oriented codimension-1 field is called the “contact structure”. The pair \((M,\xi )\) is called “contact manifold”. The “Reeb vector field” \(R_{\alpha }\) is the unique vector field defined by the equations \(d \alpha (R_{\alpha }, \cdot ) = 0\) and \(\alpha (R) = 1\), whose flow is called “Reeb flow”. Definition 3.2.
A “Liouville vector field” \(X\) on a symplectic manifold \((M,\omega )\) is a vector field satisfying \(\mathscr {L} _X \omega = \omega \), where \(\mathscr {L}\) denotes the Lie derivative, i.e., the Lie derivative along \(X\) preserves \(\omega \). Example 3.3.
The cotangent bundle \(T^* Q\) of a smooth manifold \(Q\) of dimension \(n\) is endowed with the so-called “Liouville one-form”. In local coordinates \((q_1,...,q_n)\) on \(Q\) and dual coordinates \((p_1,...,p_n)\) on the fibers of \(T^* Q\), the Liouville one-form is defined by \(\lambda _{can} = \sum _{i=1} ^n p_i dq_i\). Since the standard symplectic form is characterized by \(\omega _{can} = d \lambda _{can} = \sum _{i=1}^n dp_i \wedge dq_i\), the “natural Liouville vector field” \(X\) on \(T^* Q\) associated to \(\lambda _{can}\) is defined by \(\iota _X \omega _{can} = \lambda _{can}\). In local coordinates, 3.2. Proof of transversality in non-regularized case.
Proof.
Proof.
Proof.
Proof.
Proof of Proposition 3.4 .
3.3. Moser-regularized energy level set and proof of transversality near the origin.
Proposition 3.9.
For \(\varepsilon > 0 \) small enough, the natural Liouville vector field on \(T^*S^3\) given by \(\seteqnumber{0}{}{24}\) Proof of Proposition 3.9 .
Acknowledgement.
References
1