See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/266736429

Rotational Dynamics

Chapter · January 2014

CITATIONS READS

0 3,683

1 author:

V. V. Sidorenko
Russian Academy of Sciences
59 PUBLICATIONS   297 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

9th Humboldt Colloquium on Celestial Mechanics View project

Dynamics of Tethered Satellite Systems View project

All content following this page was uploaded by V. V. Sidorenko on 11 October 2014.

The user has requested enhancement of the downloaded file.

ROTATIONAL DYNAMICS
Vladislav Sidorenko
Keldysh Institute of Applied Mathematics, Moscow, RUSSIA
Keywords: rotational motion, gravity torque, Euler’s angles, Euler’s equations, Andoyer’s variables,
“action-angle” variables, spin-orbit coupling, resonances, Cassini’s laws

Contents

1. Introduction. Main assumptions

2. Kinematics of rotational motion
2.1. Reference frames used in studies of rotational motion
2.2. Euler angles
2.3. Euler’s kinematical equations
2.4. Singularities accompanying the use of Euler angles
3. Rotational dynamics: Euler’s formalism
3.1. The relation between angular momentum and angular velocity
3.2. Tensor of inertia and ellipsoid of inertia
3.3. Euler’s dynamical equations
4. Rotational dynamics: Lagrangian formalism
5. Rotational dynamics: Hamiltonian formalism
5.1. Andoyer’s variables
5.2. Hamiltonian of rotational motion
5.3. Modified Andoyer’s variables
5.4. “Action-angle” variables
6. Euler-Poinsot motion: torque-free rotation of the rigid body
6.1. Motivation. Equations of motion
6.2. The torque-free rotations in the triaxial case
6.3. The torque-free rotations in the axisymmetric case
7. Torques applied to celestial body
7.1. The gravity torque
7.2. The other torques applied to celestial bodies
8. Perturbed Euler-Poinsot motion in the gravity field
8.1. The rotational motion of the body when the ellipsoid of inertia is nearly a sphere
8.2. Fast rotations of the body in gravity field
9. Spin-orbit coupling
9.1. Planar motions
9.2. Investigation of the resonant planar motions: various strategies
9.3. The application of averaging to reveal the secular effects
9.4. Moon-like resonant rotations
9.5. Mercury-like resonant rotations
9.6. Generalization: “spatial” resonant rotations
9.7. The origin of the resonant rotational motions
10. Rotational dynamics in the case of the motion in an evolving orbit
10.1 Cassini’s laws
10.2 The evolution of the orbit as a source of chaos in rotational dynamics
11. Conclusion

1
Summary

This chapter provides a short introduction into the main dynamical problems related to the rotational
motion of celestial bodies. We start by considering various ways to characterize this motion and to derive
the equations of motion. Although the main attention is given to the influence of the gravity torque on the
rotational motion, the role of other torques is also briefly discussed. In an elementary way, we establish
the key property of the non-resonant, slightly perturbed, rotational motion of a celestial body (under the
action of gravity torque only) - the precession of the angular momentum vector around the normal to the
orbital plane. The resonant spin-orbit coupling is considered as well.

1. Introduction. Main assumptions

Since any real celestial body is not a material point, a complete theory of its motion should consider not
only the orbital dynamics, but also the rotation of this body around its mass center O . The main
properties of the rotational motion are discussed in the next sections. For further reading we can
recommend the textbooks by V.V.Beletsky (2001), C.D. Murray and S.F.Dermott (1999) and the reviews
from the volume “Dynamics of extended celestial bodies and rings” published in a series “Lecture notes
in Physics” under the editorship of J. Souchay (2006).
The rotational motion of the celestial bodies is usually studied within a “restricted” model, which
is based on the assumption that the rotation does not influence the orbital motion. If this model is
accepted, the orbital motion (or, more exactly, the motion of the mass center) is supposed to be known –
it can be modeled, for example, by considering the celestial body as a point mass.
The “restricted” model is accurate enough when the size of the body is much smaller than the
distance to the center of the celestial body (a star or a planet) around which the orbital motion occurs. If
the body is orbiting an object of substantially greater mass with more or less spherically symmetric
internal structure, then a further simplification is possible: the gravity field of this object is approximated
by the gravity field of the attracting center O* . In this case the “restricted” model is equivalent to the

assumption that the body’s mass center O moves in a Keplerian orbit around O* .
Sometimes the assumptions of the “restricted” problem are too restrictive. As an example we can
mention the studies on the dynamics of binary asteroids where the analysis of the rotational motion
beyond the scopes of the “restricted” problem is needed.
Another important assumption is that we will consider the celestial body as non-deformable (i.e.,
the distances between any two points of the body keep their values). Often enough the term “rigid body”
is used to specify this approximation. Due to the necessity of explaining the tidal phenomena, the

2
rotational dynamics of deformable bodies is actively investigated too. Despite the progress achieved, the
theory of the rotation of deformable bodies remains complicated and will not be discussed here.

2. Kinematics of rotational motion

2.1. Reference frames used in studies of rotational motion. To characterize the rotational
motion of a body we need two Cartesian reference frames with the origin in the mass center O . One
reference frame is fixed in the body – we will denote it as O . The rotational motion leads to a change
in the orientation of the fixed reference frame O with respect to the second reference frame, the
choice of which depends on the specific features of the problem under consideration. Often enough it is
convenient to introduce the “inertial” reference frame Oxyz with the axes preserving their orientation in
the absolute space (the quotation is applied because the translational motion of the origin is not required
to be uniform). Since we will usually suppose that the mass center O moves in a non-evolving Keplerian
orbit, we can orient the axis Oz of the inertial reference frame along the normal to the orbital plane (in the
direction of the angular momentum of the orbital motion with respect to the attracting center O* ) and the

axis Ox along the direction to the pericentre from O* ; in that case the axis Oy is tangent to the orbit when

the body moves through the pericentre. If the orbit is circular, the axis Ox can be directed along the line
passing through the attracting center O* and the arbitrary point of the orbit.
Sometimes the rotational motion of the body is considered with respect to the so-called orbital
reference frame OxO yO zO defined in the following way: the axis OzO is oriented along the radius-vector

R of the mass center O ( R  O*O ); the axis OyO is perpendicular to the osculating plane of orbital

motion and the axis OxO forms an acute angle with the direction of the body’s motion along its orbit.
2.2. Euler angles. In the XVIII century the famous mathematician Leonard Euler established that
the rigid body with a fixed point can be moved from one position to any other by only one rotation. This
statement provides the following opportunity to define the orientation of the body: we specify the rotation
which allows to achieve a current orientation of the fixed reference frame with respect to, for example,
the inertial reference frame from a position where the orientations of these reference frames coincide.
The set of all rotations is a group (under the operation of composition) denoted as SO(3) . To
parameterize this group three parameters are needed. One of the possible parameterizations is to represent
an element of SO(3) as a product of three elementary rotations about the axes with pre-defined
orientation. In particular such parametrization can be performed by means of the so-called Euler’s angles
 , , (which are called the precession angle, the nutation angle and the proper rotation angle,
respectively) corresponding to a sequence of rotations about the axes Oz , ON and O (Fig. 1).
3
 Fig.1. Euler’s angles used to define the orientation of the body-fixed
reference frame with respect to the inertial reference frame.
In studies concerning the rotational dynamics it is frequently necessary to write down the
components of a vector in the reference frame under consideration, once they are known in some other
frame. To relate the components of the vector in the different reference frames, a transition matrix of the
following form is used:

 v x   a x a x a x   v 
v   a  
 y   y a y a y   v  .
v  a  v 
 z   z a z a z   
Here v x , v y , v z and v , v , v denote the component of the vector v in the reference frames Oxyz and

O respectively. To obtain the inverse transformation the transposed matrix should be used.
The elements of the transition matrix are functions of the angles used to define the orientation of
the body:

 a x a x a x 
 
 a y a y a y   R3 ( ) R1 ( ) R3 ( ) ,
a a z a z 
 z 

where R1 () and R3 () are the matrices defining the elementary rotations around the axis of Cartesian
reference frame:

1 0 0   cos   sin  0

R1 ( )   0 cos   sin   , R3 ( )   sin 
 cos  0  .
 0 sin  cos    0 1 
  0

By elementary calculations one obtains

a x  cos  cos  sin  cos sin , a x   cos  sin  sin  cos cos ,
a x  sin  sin  ,
4
 a y  sin  cos  cos  cos sin , a y   sin  sin  cos  cos  cos ,
a y   cos  sin  ,

az  sin sin , az  cos sin , az  cos.

2.3. Euler’s kinematical equations. To describe how the body changes its orientation, we
introduce a vector quantity known as an “angular velocity”. It is a pseudovector which specifies the
angular speed of the body and the direction of the instantaneous axis of rotation in the motion around the
mass center O . Denoting an angular velocity as ω , we write it down as the sum of three terms
corresponding to the elementary rotations:

ω   e z   e N   e . (2.1)

Here e z and e denote the unit vectors of the axis Oz and O respectively, the unit vector e N is directed

along the line of nodes ON (Fig. 1). In scalar form the relation (2.1) gives us
   cos   sin  sin , (2.2)

   sin   sin  cos ,

     cos.

Resolving (2.2) with respect to  , , we obtain the classical Euler’s kinematical equations:


1
sin 
 sin   cos  , (2.3)

   cos   sin ,
    ctg   sin   cos  .

2.4. Singularities accompanying the use of Euler angles. As one can see, Euler’s kinematical
equations (2.3) become singular at sin   0 . This singularity (very unpleasant for numerical studies) is
not connected with something special in rotational motion. It is an artifact of the rotation group SO(3)
parametrization by means of the Euler angles. To avoid this kind of singularity, the other parametrizations
of SO(3) can be applied (for example, by means of quaternions).

3. Rotational dynamics: Euler’s formalism

3.1. The relation between angular momentum and angular velocity. Euler’s approach to the
rotational dynamics of celestial bodies is based on the angular momentum equation
dG
M, (3.1)
dt

5
written in the “inertial” reference frame Oxyz . In the equation (3.1) G denotes the angular momentum of
the body with respect to the mass center O , M is the total torque (with respect to O ) of all forces applied
to this body.
To compute G we should sum up the angular momenta of all elements of the body:

G    (r  v )dV , (3.2)
V

where r and v denote the radius vector and the velocity of the infinitesimal volume element dV with
respect to the mass center,  characterizes the local density of the matter inside the body and the symbol
“  ” is used to denote the vector product. Taking into account the relation
v  ωr ,
we rewrite (3.2) as follows:

G    (r  (ω  r ))dV    [r 2ω  (r, ω)r]dV . (3.3)

V V

Here and below the notation (, ) is applied for the scalar product in R 3 .
As one can see from (3.3) G depends linearly on the angular velocity ω . To write down the
relation between these quantities in a more concise way we introduce an operator : R3  R3 , defined by
the formula

   [ E3r 2  rrT ]dV . (3.4)

V

In formula (3.4) E3 is the 3  3 identity matrix and the dyadic product of vectors is used:

 a1b1 a1b2 a1bn 

a b a b a2bn 
ab   2 1
T 2 2
, a  Rm , b  Rn .
 
 
 amb1 amb2 ambn 

The relation between G and ω takes now the remarkable form

G  ω. (3.5)
3.2. Tensor of inertia and ellipsoid of inertia. The formula (3.5) is valid regardless of the
reference frame where the components of G and ω are provided. For the matrix representation of the
operator (which depends on the choice of the reference frame) the term “tensor of inertia” is used. We
can write down the tensor of inertia both in the reference frame Oxyz (which preserves the orientation)
and in the rotating body-fixed reference frame O , but only in the last case the coefficients of the
matrix do not vary with time.

6
 The eigenvectors of give us the directions of the so called principal central axes of inertia: if the
body rotates around such an axis, then G is parallel to ω . In general there exist three mutually
perpendicular principal axes of inertia (fixed in the body!). It allows us to introduce the body-fixed
reference frame O in a way which simplifies the structure of the equations of motion – we will
suppose below that the axes O , O , O are directed along the principal axes of inertia. In this
reference frame the tensor of inertia is given by the diagonal matrix:
 diag( A, B, C ) ,

A   ( 2   2 )  dV , B   ( 2   2 )  dV ,
V V

C   ( 2   2 )  dV .
V

The quantities A, B, C are called the principal central moments of inertia.

The relation
(r, r)  1

defines in R 3 the quadratic surface which is called the ellipsoid of inertia (or, more precisely, the
ellipsoid of inertia corresponding to the mass center O ). It is easy to prove that the ellipsoid of inertia is
rigidly connected to the body: if the body orientation varies in the inertial space, then the orientation of
the ellipsoid of inertia varies in the same way. Taking it into account, one can characterize the rotational
motion of the body in terms of its inertia ellipsoid motion (See Sec. 6.2).
Often enough some kind of resemblance exists between the shapes of the body and of its inertia
ellipsoid. For example let us consider a homogeneous body bounded by the tri-axial ellipsoid (i.e., by the
ellipsoid with the different semi-principal axes). The directions of its longest, intermediate and shortest
principal axes coincide with the directions of the corresponding inertia ellipsoid principal axes at the
mass center O .
3.3. Euler’s dynamical equations. In the rotating reference frame O the angular momentum
equation (3.1) takes the form
d G
 ωG  M . (3.6)
dt
Here the prime indicates that the components of the differentiated vector should be expressed in the frame
O :
T
d G  dG dG dG 
 , ,  .
dt  dt dt dt 

7
Substituting (3.5) into (3.6) and taking into account that in the reference frame O the matrix has
constant coefficients, we obtain
d ω
 ω ω  M
dt
or (in scalar form)
d
A  (C  B)  M  , (3.7)
dt
d
B  ( A  C )   M ,
dt
d
C  ( B  A)   M  .
dt
The equations (3.7) are called “Euler’s dynamical equations”.
If the components of the torque M are the known functions of the variables  ,  ,  , , ,

(and, may be, of the time t ) then Euler’s dynamical equations (3.7) and Euler’s kinematical equations
(2.3) form a closed system of differential equations describing the rotational motion of the celestial body.

4. Rotational dynamics: Lagrangian formalism

Sometimes it is more convenient to derive the equations of the rotational motion in the form of
Lagrange’s equations. Let us choose the Euler’s angles  , , as generalized coordinates. The
Lagrangian equations can be written as
d  L  L d  L  L
    0,    0,
dt     dt    
d  L  L
dt       0,

where the Lagrangian function L is given by
L  Tr   .

In the last formula Tr denotes the kinetic energy of the rotational motion,  is the potential of the body in
the field of external forces (we suppose for simplicity that the non potential forces are absent). Using the
relations (2.2) it is not difficult to write down the expression for Tr as a function of the Lagrangian

variables  , , , , , :

8
 1 1
Tr  ( A2  B2  C2 )   A( cos   sin  sin ) 2
2 2
 B(  sin   sin  cos ) 2  C ( 2   sin  ) 2  .

5. Rotational dynamics: Hamiltonian formalism

5.1 Andoyer’s variables. In most of the modern theories dealing with the dynamics of the
celestial bodies the canonical sets of variables are used to describe the rotational motion. The main
advantage of the canonical variables is an opportunity to apply various perturbation techniques based on
canonical transformations.
A very convenient set of canonical variables for studies of rotational motion was proposed by the
French astronomer and mathematician Henri Andoyer. In order to define the Andoyer’s variables we
introduce the auxiliary Cartesian reference frame OxG yG zG related to the angular momentum G : the axis

OzG of this reference frame is oriented along the angular momentum, the axis OxG lies in the plane Oxy ,

and the angle between the axes Oz and OyG does not exceed  2 (Fig. 2). The intersections of the plane

OxG yG (which is normal to G ) with the planes Oxy and O provide two lines of nodes ON G and

ON E , respectively; we denote the angle between the planes OxG yG and Oxy as I and the angle between

the planes OxG yG and O as J .

Fig.2. Andoyer angles.

The canonical set of Andoyer’s variables consists of three angle variables l , g , h and their
conjugated momenta L, G, H . The angle variables are defined as follows:

9
  h is an angle between the axis Ox and the axis OxG (which is directed along the lines of

nodes N G ). This angle is measured in the plane Oxy .

 g is an angle between the axis OxG and the ascending node of the equator (i.e., of the plane

O ) with respect to the plane OxG yG .

 l is an angle between the ascending node of the equator with respect to OxG yG and the axis O .
Conjugated momenta are given by formulas
G | G |, L  G cos J , H  G cos I .
5.2 Hamiltonian of rotational motion. It is not difficult to write down the components of the
angular velocity ω in terms of Andoyer’s variables:
T
 G sin J sin l G sin J cos l G cos J  (5.1)
ω , ,  .
 A B C 
Here

L G 2  L2
cos J  , sin J  .
G G
Substituting (5.1) into the formula for the kinetic energy of the rotational motion we obtain the
Hamiltonian of the free body motion (i.e., for the motion in the absence of an external torque)

1 1 2 2  sin 2 l cos2 l  L2 (5.2)

 (ω, ω)  (G  L )   
B  2C
free .
2 2  A
The Hamiltonian for the rotation of the rigid body in the potential field of external forces has form
 free , (5.3)

where  as in the previous section denotes the potential of the forces applied to the body. If the
expression for  is known we can obtain the equations of rotational motion in terms of Andoyer’s
variables:

d  d 
( L, G, H )   , (l , g , h )  .
dt (l , g , h) dt ( L, G, H )
5.3 Modified Andoyer’s variables. It is important to note that it may happen that some of the
Andoyer’ variables are undefined:
 In the case of the body rotation around the axis O ( J  0 or J   ), the undefined variables
are the angles l and g ;

 In the case of the angular momentum vector oriented normally to the plane Oxy ( I  0 or I  
), the undefined variables are the angles g and h ;

10
  In the case when the axes Oz , O and the angular momentum G are aligned
( I  0(mod  ), J  0(mod  ) ), all the angles h, g , l are undefined.
The transformation to non-singular variables allows to avoid such problems. As an example, we
present the set of non-singular variables used to study the properties of motion in the case when J  0
and I  0 simultaneously:

P  G, p1  2(G  H ) cosh, p2  2(G  L) cos l ,

  h  g  l , q1   2(G  H ) sin h, q2   2(G  L) sin l.
After the canonical transformation
( L, G, H , l , g, h) ( P, p1, p2 , , q1, q2 ),
the Hamiltonian (5.3) takes the form
 p 2  q22   q22 p22  1  p22  q22 
 2 P  2   
4   A B  2C  2 
P

 ( P, p1 , p2 , , q1 , q2 ).

5.4. “Action-angle” variables. Typically the phase space of a physically meaningful

integrable Hamiltonian system with n degrees of freedom is foliated (at least locally) by invariant n -
dimensional tori. In this case a set of canonical “action-angle” variables (I, φ) can be introduced, where
the angles φ  (1, ,n ) provide the coordinates on these tori, while the conjugate momenta (“action”

variables) I  ( I1, , I n ) are used to “enumerate” them. In terms of the “action-angle” variables the
Hamiltonian of the system depends only on the “action” variables (but, as the case may be, not
depending on all of them):  (I) . It means that the behavior of the system can be described in a
very simple way: some of the variables vary as linear functions of time, the other (in particular, all
“action” variables) preserve their initial values. The formal conditions at which the “action-angle”
variables exist are provided by the Lioville-Arnold Theorem.
In the case A  B (i.e, in the case of the dynamical symmetry of the body) the Hamiltonian (5.2)
depends only on the conjugated momenta L and G . In such case the Andoyer’s variables are “action-
angle” variables and the application of perturbation techniques is simplified. If the symmetry is absent,
but the triaxiality is small, the Hamiltonian (written in terms of Andoyer’s variables) can be split into a
sum of “axisymmetrical” leading term and a perturbation due to the triaxiality. Then the last term is
considered as a part of the perturbing potential  .
In the general case, the “action-angle” variables for the torque-free rotation of the rigid body were
introduced by Yury Sadov. In the paper, published in 1970, Sadov constructed the transformation
11
 ( L, G, H , l , g , h) ( I l , I g , I h , l ,  g , h ),

where
1
Il 
2  Ldl, I g  G, Ih  H .

In practice, the straightforward application of the “action-angle” variables for studies of the Euler-Poinsot
perturbed motion results in rather complicated expressions for the right hand parts of the equations of
motion. For this reason the other variants of the variables (both canonical and non-canonical) are often
used to simplify the analytical investigation of the rotational dynamics.

6. Euler-Poinsot motion: torque-free rotation of the rigid body

6.1. Motivation. Equations of motion. In this section we would like to discuss the properties
of the rotational motion in absence of any perturbation (the so-called Euler-Poinsot case in the rigid body
dynamics). There are at least two reasons why it is worth doing this task:
 If the external torques are sufficiently small then the Euler-Poinsot motion can be used as a
rough model when the rotational dynamics of the celestial body is considered over a short time
interval.
 The Euler-Poinsot motion often provides a basis for various perturbation techniques to reveal the
secular effects of the body dynamics over a long time interval.
In absence of any external torques, the angular momentum G does not vary, so that the Andoyer’s
variables G, H , h preserve their initial values. The behavior of the angles l , g, J  arccos( L / G) is
governed by the equations
dJ 1 1
 G sin J sin l cos l    , (6.1)
dt  A B
dg  sin 2 l cos2 l 
 G  ,
dt  A B 
dl  1 sin 2 l cos2 l 
 G cos J   
B 
.
dt C A

Here G should be considered as a parameter.

6.2. The torque-free rotations in the triaxial case. The term “triaxial case” means that the
three moments of inertia are different from each other.
A  B, B  C, C  A.
The explicit formulas for the variation of the angles l , g , J in the torque free motion of the triaxial body
can be found in the textbooks on classical mechanics. Since these formulas are rather complicated (they

12
are written in terms of the higher transcendental functions), it is important to mention an opportunity to
classify the unperturbed motions geometrically. To start we present in Fig. 3 the inertia ellipsoid
A 2  B 2  C 2  1
with several polhodes (a polhode is a curve consisting of the points where the inertia ellipsoid is
intersected by an instantaneous axis of rotation in the motion under consideration). This ellipsoid
corresponds to a body with principal moments of inertia satisfying the inequalities
A  B  C.
The axis O is the longest axis of the inertia ellipsoid and the axis O is its shorter axis. If the polhode
surrounds the axis O , the motion is called “complex long axis mode” (complex LAM). If the polhode
surrounds the axis O , the motion is called “complex short axis mode” (complex SAM). Rotations about
the axis O or the axis O are called simple LAM and SAM, respectively. Although these terms are not
introduced in classical textbooks on rigid body dynamics, they are widely used in publications on the
rotational motion of celestial bodies. To determine the mode for given values of the kinetic energy Tr and

angular momentum G , the quantity

2 BTr
w
G2
can be used. If w  (1, B / A) , then we have complex LAM, while in the case w  ( B /C,1) we have
complex SAM.

Fig. 3. The inertia ellipsoid and polhodes.

With rare exceptions the simple SAM or close to the simple SAM rotations are typical for
examples of weakly perturbed Euler-Poinsot motion provided by celestial bodies. For this reason we will
pay in the following the attention mainly to SAM. The case of LAM can be treated mutatis mutandis.
13
 In complex SAM the quantity w and the maximum value  of the angle between the short axis of
the inertia ellipsoid and the angular momentum vector are related in the following way:
B  B
w()   1   sin 2 .
C  C
The integration of the equations (6.1) for this type of motion can be reduced to the integration of a single
differential equation:

dl G (C  B)( B  wA)
 U (l , ),
dt B AC
where

AC B  B  2
U (l , )    w     1 cos l
(C  B )( B  wA)  A  A 
B B B 
       1 cos2 l
A C A  w w (  )

Let g (t ), J (t ), l (t ) denote the solution of (6.1) corresponding to complex SAM motion. Then the
functions g (t ) and l (t ) can be written as
g (t )  g t  g1 (t ), l (t )  l t  l1 (t ) , (6.2)

where the quantities g and l are called the frequencies of the Euler-Poinsot motion and they are given

by the formulas:
G   C   ( , k ) 
g  1  1   ,
C   A  K (k ) 
  G  (C  B )( B  wA)
l     .
2 K (k )  B  AC
Here K (k ) and ( , k ) denote the complete elliptic integrals of the first and third kind, respectively, and
their parameters are

C ( B  A) ( B  A)( B  wC )
 , k .
A(C  B) ( B  C )( B  wA)

In general the frequencies g and l are incommensurable. The functions g1 (t ) and l1 (t ) in (6.2) are

(Tl / 2) -periodic functions of time, where Tl  2 / | l | . The variation of the angle J in complex SAM is

also described by a (Tl / 2) -periodic function of time:

14
 B  B  2
  w     1 cos l
J  arccos   A 
A
.
 B B  B  2
      1 cos l
 A C  A 
6.3. The torque-free rotations in the axisymmetric case. It is the case when the moment of
inertia has the same value for any axis lying in the equatorial plane (i.e., in the plane O ) and passing
through the mass center. Setting A  B in (6.1) we obtain the equations describing the torque-free motion
in the axisymmetric case:
dJ dg G dl 1 1
 0,  ,  G cos J    . (6.3)
dt dt A dt C A
As it follows from (6.3), in this case the motion can be represented as a superposition of two rotations.
The symmetry axis of the body (or, more exactly, the symmetry axis of the ellipsoid of inertia) rotates at
a constant angular velocity g  G / A around the angular momentum vector forming a constant nutation

angle J . The body itself rotates around its symmetry axis at a constant angular velocity
1 1
l  G cos J   .
 C A
Such a motion is called regular precession.

7. Torques applied to celestial bodies

7.1. The gravity torque. The interaction of the body with the field of the attracting center O* is
characterized by the potential
 dV
 g    , (7.1)
V
|R r |

where  is the gravity parameter, R is the radius vector of the body mass center O with respect to O*

and r denotes the radius vector of the infinitesimal volume element dV with respect to the body mass
center.
Since we assume that the size of the body is substantially smaller than the distance to the
attracting center we can expand the integrand in (7.1) as a series with respect to powers of x /R , y /R ,
z /R . Neglecting the terms of third and higher order we obtain
m  3
g    3
tr  (e R , e R ) . (7.2)
R 2R 2R3

15
Here m is the mass of the body, denotes its tensor of inertia with respect to the mass center O and
eR  R | R | .

The expression (7.2) allows us to compute the gravity torque M g applied to the body. Let σ

denotes the infinitesimal rotation of the body – the rotation around the axis e  σ /| σ | on the angle

  | σ |  1 . The work of the torque M g on the rotation σ and the change of the gravitation potential

 g due to this rotation are related in the following way:

(M g , σ)   g . (7.3)

The first two terms in (7.2) are independent from the body orientation. So they are insignificant for the
computation of  g and our attention should be concentrated on the third term. We have evidently

3 3
3
(e R , e R )  3
[ A(e R , e )2  B(e R , e )2  C (e R , e )2 ] . (7.4)
2R 2R
The rotation σ changes the directions of the axes O , O, O :

e  e  σ  e , e  e  σ  e , e  e  σ  e , (7.5)

where e , e , e are the unit vectors of the corresponding axes after the rotation. Substituting (7.5) into

(7.2) and taking into account (7.4) we obtain:

3
  g   g   (e R  e R , σ ) .
R3
Since (7.3) is valid for any infinitesimal rotation we arrive at the conclusion that
3
Mg  eR  eR .
R3
It is useful to find the components of the gravity torque M g in the body reference frames. Simple

calculations give us the result:

3
M g  (C  B)(e R , e )(e R , e ),
R3
3
M g  ( A  C )(e R , e )(e R , e ),
R3
3
M g  ( B  A)(e R , e )(e R , e ).
R3
To complete the discussion of the expression for the gravity torque, we recall that in the case of
the body motion in a Keplerian orbit one has

  a  n (1  e cos )
2 3 3
 n 2
   . (7.6)
R3 R (1  e2 )3
16
Here a and e denote the semimajor axis and the eccentricity of the orbit,  denotes the true anomaly of
the body and n denotes the mean motion (the mean motion is defined by the formula n  2 / P , where
P is the period of the body motion in orbit).
7.2. Other torques applied to celestial bodies. Many aspects of the rotational motion
exhibited by the real celestial bodies can be understood by neglecting all external torques except the
gravity torque M g . Nevertheless, sometimes a more accurate approach is needed. Without going into the

details, we mention only several examples:

 The long term evolution of planets and exoplanets spin states depends substantially on the
so-called tidal torques;
 It is believed that the radiative torque due to infrared emission of absorbed sunlight is large
enough to alter the spin states of asteroids (Yarkovsky–O'Keefe–Radzievskii–Paddack effect, or
YORP effect);
 The anisotropic ice sublimation on the comet nucleus surface (induced by solar radiation
heating) results in a reactive torque, modifying the spin state of the nucleus.

8. Perturbed Euler-Poinsot motion in the gravity field

As it was mentioned in Sec. 7 we will focus our attention on the rotational motion of celestial
bodies under the influence of only the gravity torque. In general there are two cases when this influence is
weak and the rotational motion can be treated as a perturbed Euler-Poinsot motion (Sec. 6). The first case
takes place when the ellipsoid of inertia of the celestial body is nearly a sphere:
A  B, B  C
The second case corresponds to the fast rotation, when the absolute value of the body angular velocity ω
is substantially greater than the value of the mean motion n .
8.1. The rotational motion of the body when the ellipsoid of inertia is nearly a sphere. To
simplify the description of the main dynamical effects, we assume that the body moves in a circular orbit
( e  0 ) and that its ellipsoid of inertia is axisymmetric (let us take for definiteness A  B ). Under these
assumptions the Hamiltonian (5.3) takes the form

1  G2  1 1  2  3 2
      L   n (C  A)(e R , e )2 . (8.1)
2  A  C A  2
Here
(e R , e )  sin I cos J sin(h   )
 sin J cos I cos g sin(h   )  sin g cos(h   ) ,   nt   0
It is worth mentioning that even in this reduced form the discussed dynamical problem is non-integrable.
17
 Since the angle l is absent in the expression for the Hamiltonian (8.1), the corresponding
conjugate variable L preserves its value. It means that formally we deal with a two-degrees of freedom
Hamiltonian system, which depends on L as a parameter. After the change of variables
h h  ,
the Hamiltonian of the perturbed rotational motion becomes autonomous and can be rewritten as
 0  1 , (8.2)

where
G2 3
0  nH , 1  n 2 (C  A)(e R , e ) 2 .
2A 2
As a consequence of the assumption that the ellipsoid of inertia is nearly spherical we have (in general)
| 1 |  | 0 |.

The Hamiltonian 0 in (8.2) is integrable and satisfies the condition of isoenergetic non-degeneracy:

2 0 2 0  0

G 2
GH G
2 0 2 0  0
0
H G H 2 H
 0  0
0
G H

It allows to apply KAM-theory to establish the properties of the perturbed motion described by the
Hamiltonian (8.2). In particular, if the nonsphericity of the inertia ellipsoid is small enough, the quantities
G and H remain forever near their initial values. This means that the angles
H L
I  arccos , J  arccos
G G
preserve approximately their values in any motion of the body. Therefore the perturbed motion of the
celestial body under the above assumptions can be described as a combination of two precessions: the
precession of the axis O around the angular momentum vector G on approximately the same angular
distance J and the precession of the angular momentum vector G around the normal to the orbital plane
on approximately the same angular distance I . The rates of the these precessions are
G
g  (8.3)
A
and
 3(C  A)n 2
h   (1  3cos2 J ) cos I , (8.4)
H
1
G

18
respectively. The designation  is used for averaging over the variables g and h .

The described qualitative features of the rotational motion as a superposition of two precessions is
valid independently of any condition on the variation of the variables g and h . Obviously, if g and h
vary with commensurable rates then the resonance phenomena take place, but the scale of these

| AC |
phenomena is rather small (of order  1 or smaller).
A
More information on resonant and non-resonant rotations of the body with close values of the
central principal moments of inertia can be found in the literature.
8.2. Fast rotations of the body in gravity field. For simplicity we assume as before that the
body moves in a circular orbit ( e  0 ) and that A  B . First we need to rewrite the dynamical equations
in a form convenient for the investigations of the fast rotations. With this aim we introduce the
dimensionless variables
G H G
G ,H ,  0 t.
G0 G0 A

Here G0 denotes the initial value of the angular momentum. The equations of motion can be written as

dG  dH 
 ,  , (8.5)
d g d h
dg  dh 
 ,  ,
d G d H
where
 00 (G)   01 (H )   2 11 (G, H , g, h ) , (8.6)

G2 3CA nA
00  , 01  H , 11  ( e R , e ) 2 ,   .
2 2 A G0
Actually the parameter  in the expression for the Hamiltonian (8.6) equals the ratio of the mean motion
n to the rate of the body precession g  G0 / A around the angular momentum vector G . Since in the

case of the fast body rotation

n  g ,

then  can be considered as a small parameter.

Within the framework of KAM-theory the Hamiltonian (8.6) corresponds to the case of proper
degeneracy. An attempt to prove the existence of invariant tori was undertaken by V.V. Beletsky and
A.P. Torjevskii in 1972. Although the proposed proof is not technically perfect, there is no doubt that the
quantities G and H remain close to their initial values. This means that the angles I and J preserve

19
(approximately) their values too, and therefore we obtain a result similar to that in Sec. 8.1: the perturbed
motion of a rapidly rotating axisymmetric body in a gravity field can be described as a superposition of
two precessions: the precession of the axis O around the angular momentum vector G and the
precession of the angular momentum vector G around the normal to the orbital plane; the rates of
precessions are provided by the formulas (8.3) and (8.4).
Nevertheless we should note an important difference: in the system (8.5) the angle variables vary
with the rates of different order:
dg dh
1, ~ .
d d
This allows to apply the averaging over the fast variable g only with the aim to investigate the properties

of the motion on a time interval of order  1 . The averaged motion equations correspond to a one-degree
of freedom Hamiltonian system with a Hamiltonian which depends on G as a parameter

(G, H , h )   01 (H )   2 11 ( G, H , g , h ) . (8.7)
g

Since (8.7) is integrable, it can be studied in details analytically.

In the case of fast rotation of a celestial body with different principal moments of inertia, there is a
slow precession of the angular momentum vector around the normal to the orbital plane, but its rate
depends in a complicated way on the parameters of the Euler-Poinsot motion (Sec. 6).

9. Resonant spin-orbit coupling

The resonant spin-orbit coupling corresponds to the rotation of a celestial body with a rate which
is commensurate with the orbital mean motion n . Almost all really observed examples of resonant spin-
orbit coupling correspond to the case of a body rotation with a rate equal to n . Such a spin state is
peculiar of the Moon, Mars’s natural satellites Phobos and Deimos, as well as of many satellites of the
giant planets. The only known example of a different resonant spin state is provided by Mercury, which
rotates with an angular velocity equal to 3n / 2 .
The aim of this section is to discuss the main properties of the Moon-like and Mercury-like
resonant spin states on the basis of a simple dynamical model.
9.1. Planar motions. If the approximation (7.2) is used for the computation of the potential
energy of the body in the gravity field, the equations of motions in terms of the modified Andoyer’s
variables (Sec. 5.3), namely
d  d 
( P, p1 , p2 )   , (, q1, q2 )  ,
dt (, q1 , q2 ) dt ( P, p1, p2 )
possess a 2-dimensional integral manifold defined by the relation

20
 p1  p2  q1  q2  0 .
The behavior of the variables P and  on this manifold is governed by the equations
d  P dP 3
 ,   3 ( B  A)sin 2(   ) . (9.1)
dt C dt 2R
The phase trajectories laying on correspond to the motions of the rigid body with the axis O
permanently directed along the normal to the orbital plane (and, as a consequence, with the angular
velocity ω permanently normal to the orbital plane too). These motions are called “planar”, since all
points of the body move parallel to the orbital plane (Fig. 4).

Fig. 4. The geometry of planar motion.

The investigations of the planar motions reveal to some extent the most general properties of the
rotational dynamics of the celestial bodies (in particular those that relate to stability, resonance
phenomena, chaos, etc). Among the specialists this area of activity is known as “spin-orbit problem”.
Using the relation (7.6) we rewrite down the equations (9.1) in a form which is more convenient
for the analytical treatment, namely
d2
(2 )   F ( , e)sin 2(   )  0 , (9.2)
d 2
where
3( B  A) (1  e cos )3
 , F ( , e)  .
C (1  e2 )3
As independent variable in (9.2) we take the mean anomaly   nt . The true anomaly  in (9.1) and
(9.2) should be considered as a known function of the mean anomaly.
The value of the parameter  in (9.2) depends on the shape and internal structure of the body. For
a celestial body with a nearly spherically symmetric structure, one has
  1 .

21
 Some specialists prefer to deal with the equation of the planar oscillations in the form proposed by
V.V. Beletsky:
d 2 d
(1  e cos )  2e sin   sin  cos   2e sin . (9.3)
d 2
d
Here      denotes the angle between the local vertical and the axis O of the body-fixed reference
frame. The Beletsky equation (9.3) was obtained from (9.2) by using the true anomaly as independent
variable. We recall the relation between the true anomaly  and the mean anomaly  in Keplerian
motion:
d (1  e cos )2
 .
d (1  e2 )3/2
In the case of a motion in a circular orbit ( e  0 ) the equation (9.3) reduces to the integrable
equation of a mathematical pendulum:
d 2 (2 )
  sin 2  0 .
d 2
In the general case ( e  0 ) the equations (9.1) are non-integrable and chaotic motions might take
place. The standard example of a real body chaotic rotational dynamics is provided by the Saturn’s
satellite Hyperion.
Remark. It is worth noting that the equations (9.2) and (9.3) are derived and valid in the
“restricted” setting of the rotational motion problem introduced at the beginning of this Chapter.
9.2. Investigation of the resonant planar motions: various strategies. Below we will discuss
the resonant spin-orbit coupling of nearly spherically symmetric celestial bodies (   1 ). For our aims it
is enough to present the general ideas in a qualitative way.
Nevertheless we should mention that the integrability of the equations (9.2) in the case of the
motion in a circular orbit provides an opportunity to apply a perturbation technique for the investigation
of the resonant rotational motions in a near-circular orbit ( e  1 ) without any restriction on the relations
between the moments of inertia.
9.3. The application of averaging to reveal the secular effects. In order to avoid the
separated investigation of Moon-like and Mercury-like regimes, we consider the more general situation of
a body motion close to a permanent rotation with the resonant angular velocity resulting in exactly k
rotations during two revolutions around the attracting center O* (at k  2 we have the Moon-like

rotation, while k  3 gives us the Mercury-like resonance).

To characterize the deviation of the motion from the above mentioned permanent resonant rotation
we introduce the critical angle

22
 k
  .
2
Bearing in mind the application of the perturbation technique it is convenient to describe the behavior of
the angle  by means of the equations in Hamiltonian form (which can be easily obtained from (9.2)) :
dp  d 
 ,  ,
d  d p
where
p2 1
  1 ( , , e), 1   F ( , e) cos(k  2  2 ).
2 4
To study the long term oscillations of the critical angle, an averaging over the mean anomaly is applied.
The averaged Hamiltonian takes the form
p2 
  H (k , e) cos 2 ,
2 4
where H (k , e)  X k3,22 (e) (the designation X nj ,m (e) is used in the theory of elliptic motion for Hansen

coefficients). Since the quantities H (k , e) characterize the “strength” of the resonance, their properties
were carefully studied. In particular, a convenient expression for H (k , e) in terms of Bessel functions
J n ( x ) was derived:

1 n 
2  n
H ( k , e)  a (k , e) J n (ke) z |n k | . (9.4)
3e n 
Here

an | n  k |3  | n  k | 2k | n  k | 1  e 2  k ,

1  1  e2 .
z
e
With (9.4), it is easy to obtain
H (k , e) ~ e|k 2| at e  0 (k  0).
If the eccentricity e is small then the Taylor expansion for H (k , e) can be useful:
1 3 11 5 313 7 3355 9
H ( 1, e)  e  e  e  e  ,
48 768 30720 442368
1 1 5 5 143 7 9097 9
H (1, e)   e  e3  e  e  e  ,
2 16 384 18432 1474560
5 13 35 6 5 8 49 10
H (2, e)  1  e2  e4  e  e  e  ,
2 16 288 576 3600

23
 7 123 3 489 5 1763 7 13527 9
H (3, e)  e  e  e  e  e 
2 16 128 2048 163840
The graphs of H (k , e) as a function of e for some values of k are presented in Fig. 5. At k  2 the

quantity H (k , e) equals 0 at certain ek*  (0,1) ,

H (k , e)  0 at e  (0, ek* )
and
H (k , e)  0 at e  (ek* ,1).

Fig. 5. Graphs of H (k , e) ( k  1,2,3,4,5,6,8,10,14 ).

9.4. Moon-like resonant rotations (k=2). If the eccentricity e of the orbit satisfies the
inequality
e  e2*  0.6819384366...

then the averaged equations

dp  d 
 ,  (9.5)
d  d p
admit the stable stationary solution   0 (mod  ). It corresponds to the motion in which the same part of
the body is oriented to the attracting center both at apocentre and pericentre passages. The long axis of the
ellipsoid of inertia at these times is directed along the radius-vector of the body mass center O (Fig. 6,a)

24
 
In the case e  e2* the stable solution is   (mod  ). In this motion the body is also turned by
2
the same part to attracting center at apocentre and pericentre, but the long axis of the inertia ellipsoid is
directed along the normal to the radius-vector (Fig. 6,b) .
The equations (9.5) allow to study the stability of the Moon-like resonant motion with respect to
“in-plane” perturbations (i.e., the perturbed motion is also planar). The analysis of the stability with
respect to “out-of-plane” perturbations is much more complicated.

Fig. 6. Orientation of the inertia ellipsoid (qualitatively) in the case of 1:1

resonance between the orbital and rotational motion of the celestial body.
9.5. Mercury-like resonant rotations (k=3). In the case e  e3*  0.781685126... the stable
(with respect to “in-plane” perturbations) solution of the averaged equations corresponds to the rotational
motion in which the long axis of the inertia ellipsoid is directed along the radius-vector at pericentre and
normal to it when the body passes through apocentre (Fig. 7). At every moment the orientation is
opposite to the orientation possessed by the body in a previous orbit. For e  e3* in a stable rotational
motion the long axis is normal to the radius-vector at pericentre and directed along it at apocentre.

Fig. 7. Orientation of the inertia ellipsoid (qualitatively) in the case of 3:2

resonance between rotational and orbital motion ( e  e3* ).

25
 9.6. Generalization: “spatial” resonant rotations. As it was stressed at the beginning of this
section we considered the simplest case of resonant spin-orbit coupling, when the body rotates around the
normal to the orbital plane. More complicated resonant motions exist also. In particular, we would like to
mention two curious classes of motions described by Likins in 1965 as “conic” and “hyperboloid”
precessions, respectively. Both classes exist only for an axisymmetric body moving in a circular orbit. In
the case of a “conic” precession, the axis of symmetry preserves its orientation in the plane defined by the
normal to the orbit and the radius-vector corresponding to the mass center O . Therefore when the body
moves in an orbit, its axis of symmetry slides along the surface of the cone. In an “hyperboloid”
precession, the axis of symmetry is fixed in the plane normal to the radius-vector of the mass center and
slides along the surface of the hyperboloid.
9.7. The origin of the resonant rotational motions. The capture into resonant rotation was
investigated by many specialists. It is evident that tidal torques play a crucial role in this process.
Nevertheless, due to the absence of a reliable theory describing the tidal phenomena on cosmological time
scales all conclusions have a rather hypothetical character.

10. Rotational dynamics in the case of the motion in an evolving orbit

10.1 Cassini’s laws. The celestial bodies move in orbits which evolve under the action of various
perturbations. The evolution of the orbit complicates substantially the analysis of the rotational motion.
Without going into the details, we present some general results related to the rotational dynamics in an
evolving orbit.
We start from the laws of the Moon’s rotational motion established at the end of the XVII century
by the Italian astronomer Giovanni Domenico Cassini. We recall that the Moon’s orbit is inclined to the
ecliptic (the mean inclination equals approximately 5.1o ) and the line of nodes in ecliptic has a retrograde
motion with a period of 18.6 years (i.e., the plane of the Moon’s osculating orbit rotates around the
normal to the ecliptic at the Earth-Moon baricenter). Cassini’s laws can be formulated as follows:
First law. The Moon rotates uniformly around the axis fixed in its body; the period of rotation
coincides with the period of the Moon’s orbital motion around the Earth.
Second law. The Moon’s obliquity (the angle between the ecliptic and the equatorial plane of the
Moon) preserves a constant value (  1o32' ).
Third law. The ascending node of the Moon’s equator on the ecliptic always coincides with the
descending node of the Moon’s orbit (i.e., the plane of the Moon’s orbit, the plane of the ecliptic and the
plane of the Moon’s equator intersect on the same line).

26
 Cassini formulated his laws only on the basis of observations. Lately they were derived in a
rigorous way. Actually they characterize the general properties of the synchronous rotational motion in an
evolving orbit and after appropriate modifications they can be applied to other natural satellites.
10.2 The evolution of the orbit as a source of chaos in rotational dynamics. In Sec. 8 we
presented a nice picture of close to quasi-periodic motions when the external torques are small enough. In
the case of an evolving orbit, “secular” resonances become possible when one of the frequencies of the
body’s precessional motion is commensurate with certain frequency, characterizing the evolution of the
osculating elements. If the capture into such a resonance and the escape from it take place over and over
again, it gives rise to the chaotization of the rotational motion. This mechanism was proposed to explain
the chaotic variation of Mars obliquity.

11. Conclusion

The rotational dynamics as a part of celestial mechanics has a long history. Nevertheless new
problems and new directions of investigation continue to appear. In particular, we can mention the
intensive studies on the dynamics of gyrostats (or close to gyrostats bodies) as possible basis for modeling
the celestial bodies with a complicated internal structure. The application of Lie-Poisson integrators for
the numerical analysis of the long term evolution of the rotational motion is actively discussed. The recent
discovery of the numerous planetary systems in the vicinity of other stars stimulated the speculations
about the possible rotational states of the planets in these systems with special emphasis on the
habitability conditions. Therefore there are no doubts that we will see a lot of interesting new
achievements in the future.

Related chapters: The gravitational two-body problem, Classical Hamiltonian perturbation theory

Glossary

Angular velocity: a quantity used to characterize how fast and in what direction the rigid body is turning.
Euler angles: three angles used to describe the orientation of the rigid body.
Euler’s dynamical equations: the differential equations describing the evolution of the angular velocity
in the rotational motion of the rigid body.
Euler’s kinematical equations: the differential equations used to describe the variation of the rigid body
orientation due to the rotational motion.

27
Gravity torque: it appears due to the fact that some parts of the celestial body are closer to the attracting
center, while the other are farther; the emerging difference in the attractive force results in the mechanical
torque applied to the body.
Precession: a movement of the celestial body rotational axis in which this axis traces out a cone.
Rigid body: an idealization of a solid body of finite size in which the deformation is neglected.
Tensor of inertia: a tensor object used in the computation of the kinetic energy and/or angular
momentum for the rigid body rotational motion with a given value of the angular velocity.

Bibliography

Arnold, V.I., Kozlov, V.V., Neishtadt, A.I.: Mathematical Aspects of Classical and Celestial Mechanics,
3rd Edition, Springer New York (2006) [An outstanding monograph on Classical Mechanics, Dynamical
Systems and Celestial Mechanics]

Atobe, K., Ida, S.: Obliquity evolution of extrasolar terrestrial planets, Icarus, 188, 1-17 (2007) [This
paper is devoted to the rotational dynamics of extrasolar planets]

Beletsky, V.V.: Motion of an Artificial Satellite about its Center of Mass. Israel program for scientific
translations, Jerusalem (1966) [An important early monograph on the rotational motion of artificial
celestial bodies]

Beletsky, V.V.: Resonance rotation of celestial bodies and Cassini’s laws, Celest. Mech., 6, 356-378
(1972) [In this paper the Cassini’s laws are obtained in a rigorous way]

Beletsky, V.V.: Essays on the motion of celestial bodies, Birkhauser Verlag, Basel-Boston-Berlin (2001)
[An informal introduction into Celestial Mechanics and Spaceflight Dynamics. Large part of the book is
devoted to the rotational motion of celestial bodies]

Beletsky, V.V., Torjevskii, A.P.: Rapid-rotation stability of axisymmetric satellite in the gravitational
field, Soviet Physics Doklady, 17, 214-217 (1972) [An application of the KAM-theory for the stability
analysis of the fast rotational motion in the gravity field]

Benettin, G., Guzzo, M., Marini, V.: Adiabatic chaos in the spin-orbit problem, Celest. Mech. Dyn.
Astron., 101, 203-224 (2008) [This paper presents an analysis of large scale chaotic rotational motions in
the case of a spin orbit resonance]

28
Breiter, S., Nesvorny, D., Vokrouhlicky, D.: Efficient Lie-Poisson integrator for secular spin dynamics of
rigid bodies, Astron. J., 130, 1267-1277 (2005) [This paper presents a numerical integration method
allowing very long time steps]

Burov, A.A.: Non-integrability of planar oscillation equation for satellite in elliptical orbit. Vestnik
Moskov. Univ. Ser. Mat. Mekh. 1, 71–73 (1984) (in Russian) [This paper presents first rigorous results
related to non-integrability in rotational dynamics of celestial bodies]

Celletti, A.: Analysis of resonances in the spin-orbit problem in celestial mechanics. Part I: The
synchronous resonance, J. Applied Math. and Physics (ZAMP) 41,174–204 (1990); Part II: Higher order
resonances and some numerical experiments, J. Applied Math. and Physics (ZAMP) 41, 453–479 (1990)
[The author presents rigorous results on the stability of the rotational motion in the case of its resonance
with an orbital motion]

Chernousko, F.L.: The motion of the satellite around its mass center under the action of the gravity
torque. Prikladnaya Matematika i Mekhanika, 27, 473-483 (1963) (in Russian) [This paper presents an
analysis of secular effects in the case of the fast rotations]

Chernousko, F.L.: Resonance phenomena in the motion of a satellite relative to its mass center. USSR
Comp. Math. and Math. Phys. 3, 699–713 (1964) (in Russian) [A useful paper to understand the main
properties of the spin-orbit resonance]

Correia, A.C.M., Laskar, J.: Mercury’s capture into the 3/2 spin-orbit resonance as a result of its chaotic
dynamics, Nature, 429, 848-850 (2004) [In this paper the chaotic evolution of Mercury's orbit is taken
into account to provide a realistic explanation of spin-orbit resonance formation]

Dobrovolskis, A.R.: Spin states and climates of eccentric exoplanets, Icarus, 192, 1-23 (2007) [This paper
is devoted to the rotational dynamics of extrasolar planets with a special emphasis on the habitability
conditions]

Elipe, A., Gurfil, P., Tangren, W., Efroimsky, M.: The Serret-Andoyer formalism in rigid-body dynamics:
I. Symmetries and perturbations, Regul. Chaotic Dyn. 12, 389-425 (2007) [This paper presents a very
detailed discussion of Andoyer’s variables and “action-angles” in the rotational dynamics]

Elipe, A., Lanchares, V.: Exact solution of a triaxial gyrostat with one rotor, Celest. Mech. Dyn. Astron.
101, 49-68 (2008) [In this paper the attitude dynamics of a triaxial gyrostat under no external torques is
described analytically in terms of elliptic functions]
29
Efroimsky, M.: Body tides near spin-orbit resonances, Celest. Mech. Dyn. Astron. 112, 283-330 (2012)
[This paper provides the critical analysis of the models used to study the influence of tides on the
rotational motion of celestial bodies]

Fukushima, T.: New canonical variables for orbital and rotational motions, Celest. Mech. Dyn. Astron.
60, 57-68 (1994) [An alternative approach to describe the orbital and rotational motion]

Getino, J., Ferrandiz, J.: A rigorous Hamiltonian approach to the rotation of elastic bodies, Celest. Mech.
Dyn. Astron. 58, 277-295 (1994) [This paper demonstrates how the rotational motion of deformable
objects can be studied]

Goldreich, P., Peale, S.J.: Spin-orbit coupling in the solar system, Astron. J., 71, 425-438 (1966) [The
seminal paper on the spin-orbit resonances]

Goldreich, P., Peale, S.J.: The dynamics of planetary rotations, Annual Rev. Astron. Astrophys., 6, 287-
320 (1968) [An important early review paper on the rotational dynamics in the Solar System]

Hellstrom, C., Mikkola, S.: Satellite attitude dynamics and estimation with the implicit midpoint method,
New Astronomy, 14, 467-477 (2009) [This paper presents a numerical integration method allowing very
long time steps]

Julian, W.H.: Free precession of the comet Halley nucleus. Nature 326, 57–58 (1987) [Useful example of
qualitative approach to the rotational dynamics of a real celestial body]

Kinoshita, H.: Analytical expansions of torque-free motions for short and long axis modes. Celest. Mech.
Dyn. Astron. 53, 365–375 (1992) [This paper presents some properties of torque-free rotational motion
which are absent in the ordinary textbooks]

Kouprianov,V.V., Shevchenko, I.I.: Rotational dynamics of planetary satellites: a survey of regular and
chaotic behavior. Icarus, 176, 224–234 (2005) [In this paper the stability diagram for the 1:1 spin-orbit
resonance are constructed]

Likins, P.W.: Stability of symmetrical satellite in attitude fixed in an orbiting reference frame. J.
Astronaut. Sci. 12, 18-24 (1965) [A classical paper on precessional motion under the influence of the
gravity torque]

Lutze, F.H., Jr., Abbit, M.W., Jr.: Rotational locks for near-symmetric satellites. Celest. Mech. 1, 31–35
(1969) [A useful paper to understand the main properties of the spin-orbit resonance]
30
Maciejewski, A.J.: Reduction, relative equilibria and potential in the two rigid bodies problem. Celest.
Mech. Dyn. Astron. 63, 1–28 (1995) [This paper presents the analysis of the rotational motion beyond the
scopes of the “restricted” problem]

Maciejewski, A.J., Przybylska, M.: Non-Integrability of the Problem of a Rigid Satellite in Gravitational and
Magnetic Fields. Celest. Mech. Dyn. Astron. 87, 317–351 (2003) [To prove the non-integrability the
authors apply the modified variant of Ziglin theory]

Markeev, A.P.: About the rotations of the near-axisymmetric satellite in elliptic orbit at resonance of the
Mercury type. Prikladnaya Matematika i Mekhanika, 72, 707-720 (2008) (in Russian) [This paper
presents the stability analysis of the rotational motion in the case of the 3:2 spin-orbit resonance
(Mercury-like resonant rotation)]

Murray, C.D., Dermott, S.F.: Solar system dynamics, Cambridge University Press (1999) [A modern
textbook on Celestial Mechanics. Various aspects of rotational dynamics are discussed in Chapters 4, 5
and 9]

Neishtadt, A.I., Scheeres, D.J., Sidorenko, V.V., Vasiliev, A.A.: Evolution of comet nucleus rotation.
Icarus, 157, 205–218 (2002) [The authors apply the averaging procedure to reveal the secular effects due
to the ice sublimation in rotational motion of comet nucleus]

Noyelles, B.: Expression of Cassini’s third law for Callisto, and theory of its rotation. Icarus, 202, 225-
239 (2009) [In this paper the author applies the perturbation theory to establish the main properties of the
natural satellites rotational motion]

Petrov, A.L., Sazonov, V.V., Sarychev, V.A.: Stability of the near-axisymmetric satellite periodic
oscillations in the plane of elliptic orbit. Izv. AN SSSR. Mech. Tverd. Tela 41–50 (1983) (in Russian) [In
this paper the Floquet theory is applied for the stability analysis of the rotational motion in the case of the
1:1 spin-orbit resonance]

Rubincam, D.P.: Radiative Spin-up and Spin-down of Small Asteroids. Icarus, 148, 2-11 (2000) [The
pioneering paper on the YORP-effect]

Sadov, Yu.A.: The Action-Angles Variables in the Euler-Poinsot Problem, Prikladnaya Matematika i
Mekhanika, 34, 962-964 (1970) (in Russian. English translation: Journal of Applied Mathematics and
Mechanics, 34, 922-925 (1970)) [Seminal paper on “action-angle” variables in the rotational dynamics]

31
Sadov, S.: Stability of resonance rotation of a satellite with respect to its mass center in the orbit plane.
Kosmicheskie Issledovaniya, 44, 170-181 (2006) (in Russian. English translation: Cosmic Res. 44, 160–
171 (2006)) [An advanced analysis of the spin orbit resonance conditions]

Sansaturio, M.E., Vigueras, A.: Translatory-rotatory motion of a gyrostat in a Newtonian force field,
Celest. Mech. 41, 297-311 (1988) [This paper can used as an introduction into the gyrostat dynamics in
the gravity field]

Sidi, M.J.: Spacecraft dynamics and control: a practical engineering approach, Cambridge University
Press (1997) [A textbook where the application of quaternions in rotational dynamics is discussed]

Sidorenko, V.V., Scheeres, D.J., Byram, S.M.: On the rotation of comet Borrelly’s nucleus, Celest. Mech.
Dyn. Astron., 102, 133–147 (2008) [This paper presents a set of non-canonical variables for studies of
secular effects in rotational motion]

Souchay, J. (editor): Dynamics of extended celestial bodies and rings, Lect. Notes. Phys., 682 (2006)
[This volume provides a discussion of rotational motion from different points of view]

Torjevskii, A.P.: Motion of artificial satellites around mass center and resonances, Astronautica Acta, 14,
241-259 (1969) [This paper presents a generalization of the spin-orbit problem]

Touma, J., Wisdom, J.: The chaotic obliquity of Mars, Science, 259, 1294-1297 (1993) [This paper shows
the possibility of the rotational motion chaotization due the slow evolution of the orbit]

Touma, J., Wisdom, J.: Lie-Poisson integrators for rigid body dynamics in the solar system, Astron. J.,
107, 1189-1202 (1994) [This paper presents a numerical integration method allowing very long time
steps]

Touma, J., Wisdom, J. Nonlinear core-mantle coupling, Astron. J., 122, 1030-1050 (2001) [This paper
shows how the rotational motion of a planet with a liquid core can be studied]

Whittaker, E.T.: A Treatise on the Analytical Dynamics of Particles and Rigid Bodies, 2nd edn.
University Press, Cambridge, UK (1917) [A classical textbook on classical mechanics]

Wisdom, J., Peale, S.J., Mignard, F.: The chaotic rotation of Hyperion, Icarus, 58, 137-152 (1984) [The
fundamental paper on chaos in rotational dynamics of celestial bodies]

32
 Zlatoustov, V.A., Markeev, A.P.: Stability of planar oscillations of a satellite in an elliptic orbit, Celest.
Mech., 7, 31-45 (1973) [Here the authors apply KAM-theory for the stability analysis of the rotational
motion]

Biographical sketch
Vladislav Sidorenko (born in 1961 in Krasnoyarsk, Russia) received his Master degree in 1984 at the
Moscow Institute of Physics and Technology. Since 1987 he has been working at the Keldysh Institute of
Applied Mathematics, where he defended his Ph.D. Thesis (prepared under supervision of Prof.
V.A.Sarychev) in 1988 and Habilitation Thesis in 1997. He is a professor of Moscow Institute of Physics
and Technology also. His research activity belongs to Celestial Mechanics (rotational dynamics of
celestial bodies, mean-motion resonances), Spaceflight Mechanics (mathematical simulation of space
debris dynamics) and General theory of non-linear oscillations. He is the author of about 100 scientific
publications.

33

View publication stats