Annals of Physics 350 (2014) 605614

Contents lists available at ScienceDirect

Annals of Physics
journal homepage: www.elsevier.com/locate/aop

Classical states of an electric dipole in an

external magnetic field: Complete solution for
the center of mass and trapped states
Boris Atenas, Luis A. del Pino, Sergio Curilef
Departamento de Fsica, Universidad Catlica del Norte, Avenida Angamos 0610, Antofagasta, Chile

highlights

Bound states without turning points.

Lagrangian Formulation for an electric dipole in a magnetic field.
Motion of the center of mass and trapped states.
Constants of motion: pseudomomentum and energy.

article

info

Article history:
Received 29 April 2014
Accepted 11 August 2014
Available online 20 August 2014
Keywords:
General physics
Electric dipole
Lagrangian formulation

abstract
We study the classical behavior of an electric dipole in the presence
of a uniform magnetic field. Using the Lagrangian formulation, we
obtain the equations of motion, whose solutions are represented
in terms of Jacobi functions. We also identify two constants of
motion, namely, the energy E and a pseudomomentum C . We obtain
a relation between the constants that allows us to suggest the
existence of a type of bound states without turning points, which
are called trapped states. These results are consistent with and
complementary to previous results.
2014 Elsevier Inc. All rights reserved.

1. Introduction
In the present day, many specialists study the world at the molecular scale. Nanotechnology is
slowly exploring molecular rotors, and applications of this concept are extensive. Using electric fields,

Corresponding author.
E-mail addresses: scurilef@ucn.cl, scurilef@gmail.com (S. Curilef).

http://dx.doi.org/10.1016/j.aop.2014.08.007
0003-4916/ 2014 Elsevier Inc. All rights reserved.

606

B. Atenas et al. / Annals of Physics 350 (2014) 605614

molecules can change in orientation and/or remain controlled [13]. Molecular-level devices can be
obtained from the conversion of energy into controlled motion; nevertheless, it is difficult to repeat
this process using a mechanical molecular motor, although it is common in biological systems. For
the time being, it is expected that the physical principles at the scale of a molecular engine can be
identified by applying rotor dynamics in two dimensions. These rotors are modeled as electric dipoles
in electric or magnetic fields.
The primary goal of the present work is to describe the motion of a classical electric dipole in the
presence of an external magnetic field, perpendicular to the dipoles plane of motion. This system
has been approached from various perspective by several authors [47]. However, the trajectory of
the center of mass and the conditions for the existence of trapped states in terms of the constants of
motion have not been fully studied. In this article, we describe in detail the solution of the equations
of motion in the coordinates of the relative motion and the center of mass, which we derive from the
Lagrangian formulation of the problem. The relation between the constants of motion, which permits
the existence of trapped states, is established.
As previously discussed, a model of rigid and non-rigid dipoles is considered, constraining the
motion of the center of mass to a direction that is perpendicular to the magnetic field [7]. The motion
of the relative coordinate into the plane is defined by the direction of the magnetic field and a direction
perpendicular to the motion of the center of mass. It is possible to show that for certain values of the
characteristic parameters defined in the problem, there is a functional relation between two constants
of motion that allows the existence of trapped states [4]; this relation has not yet been analytically
established. In other words, an interval of values is found for the constant of motion where solutions
are possible and its trend of these solutions is well defined for certain limiting values. These states are
called classical bound states embedded in a continuum. The quantum analogue is also discussed [7].
In addition, equations and constants of motion are found for the model of an electric dipole
in an external magnetic field, and a preliminary discussion of the existence of trapped states is
introduced [4]. In addition, a model of two interacting particles is discussed [6], and special trajectories
are found in this model for several initial conditions of the velocity, direction, charges and values of the
magnetic field. The distance between particles may vary, but the conditions constrain the motion to a
plane perpendicular to the field and to a fixed distance between particles. Furthermore, the classical
dynamics of two interacting particles becomes an interesting problem where the challenge is to find
solutions that are fully analytical [47].
These solutions could significantly impact the future of the applications and construction to
molecular motors, as they describe the overall behavior of a dipole from a classical perspective. This
paper is organized as follows: In Section 2 we present the theoretical basis of the system, deriving the
equations and constants of motions. In Section 3 the solution of the equation of motion is obtained
for the center of mass coordinates. In Section 4 we address the conditions that lead to trapped states,
as mentioned above. Finally, in Section 5, we offer some concluding remarks.
2. Basic definitions and equations of motion
In the present model [4], we consider two charges in the presence of a uniform magnetic field. The
as follows: B = A.
We assign to the particle
magnetic field is obtained from a vector potential A,
1(2) the charge e1 (e2 ), the position
r1 (
r2 ), the velocity
r 1 (
r 2 ) and the mass m1 (m2 ). The Lagrangian
formulation leads to the following expression:
L(
r1 ,
r2 ;
r 1 ,
r 2 ) =

1
2

m1
r 1 +

1
2

m2
r 2

e1
c

(r1 ) r 1
A

e2
c

(r2 ) r 2
A

e1 e2

|r2 r1 |

(1)

where is the dielectric constant of the medium in which the motion of charges occurs. We define
using the symmetric gauge as follows:
the vector potential A

(ri ) =
A

1
2

ri ,
B

for i = 1, 2,

(2)

B. Atenas et al. / Annals of Physics 350 (2014) 605614

607

is the uniform magnetic field. Now, we consider the following change of variables:
where B
r = r2 r1 ,
=
R

(3)

m1
r1 + m2
r2
m1 + m2

(4)

is the position of the center of mass. Now, if we substitute

where
r is the relative position and R
Eqs. (2)(4) into Eq. (1), we obtain the following function:

, r ; R, r ) =
L(R

1
2

+
MR

e1 e2
1 2
e1 + e2
R R
r

B
2
|r |
2c

e2 m1 e1 m2
2cM

R r + r R
B

(e1 m22 + e2 m21 )

2cM 2

B
r
r ,

(5)

where = m1 m2 /M is the reduced mass and M = m1 + m2 is the total mass of the dipole. Hereafter,
we consider a rigid dipole composed of an internal coupling that holds the two charges together and
ensures that the Coulomb interaction between the charges is constant. Then, one of the particles
carries charge +e, whereas the other carries charge e, the masses are equal m1 = m2 and a is the
fixed length of the dipole. Thus, the Lagrangian is

, r ; R, r ) =
L(R

1
2

+
MR

1 2
e2
e
R r + B r R .
r +
+
B
2
a 2c

(6)

The energy of the system

, r ; R, r ) = PR R + pr r L(R, r ; R, r ),
E (R

(7)

L
r

r =
where p
is the relative conjugate momentum and PR =
center of mass. It is additionally found, for the energy, to be:

, r ; R, r ) =
E (R

1
2

+
MR

L
,
R

the conjugate momentum for the

e2
1 2
r
.
2
a

(8)

The other constant of motion that appears from the analysis is called pseudomomentum [4,6,8],

R + ec A r ,
C = PR + qA

(9)

where q = e1 + e2 is the total charge and ec = e1

m2
M

e2 mM1 is the coupling charge, A R =

R,

1
B
2c

r = 1 B r and PR = M R + e B r . Thus, the relation between the conjugate momentum of the

A
2c
2c
center of mass and the relative coordinates, for q = 0 is given by
C = PR +

e
2c

r
B

(10)

by considering the definition of the conjugate momentum for the center of mass, the constant of
motion C yields

+
C = M R

e
c

r .
B

(11)

By combining Eqs. (8) and (11), we can express the energy in terms of only the relative variable and,
the relation between E and C :

E =

2
1 2
1
e
2 e
r +
C B r .
2
2M
c
ka

(12)

Up to this point, this result corresponds to the general motion of the rigid dipole in a constant magnetic
field.

608

B. Atenas et al. / Annals of Physics 350 (2014) 605614

then
Now, if we restrict the motion of the particles to a plane perpendicular to the magnetic field B,
the pseudomomentum C becomes perpendicular to the magnetic field. This choice allows us to define
unit vectors in suitable directions; these are, e C , parallel to the vector C and e B , parallel to the vector
Additionally, we can define a third unit vector
B.
e = e C e B

(13)

to constitute a complete set of orthonormal vectors, namely, a basis.

in the above basis, are

The expressions for
r and R,
= cos e + sin e C ,

(14)

= sin e + ( cos )eC ,

(15)

eB
where is the angle between the vectors
r and e , = ar , = Ra , dtd = c dd , c = Mc
, is the
cyclotron frequency. The pseudomomentum is also defined as a dimensionless constant as follows:
= |C |/M c a. If we define Eq. (8) in terms of the dimensionless units introduced before, then

2
2
d

1 d

= + c ,
d
d

(16)

where = M
, = 22e 3 , = 22E2 and by considering Eqs. (14) and (15), we obtain
c
c M a
c a M

2 2 cos + 1 + 2 c .

(17)

= 0, we
By taking the time derivative of the previous equation and dismissing the trivial solution
obtain
+ 2 sin = 0,

with 2 = ,

(18)

where the time derivative is in terms of the dimensionless time . We emphasize that Eq. (18) coincides with the equation of motion of a nonlinear pendulum whose general solution is [9]:

= sgn0 k [ 0 ] + sn1 (k0 | ),

= 2 arcsin[sn(( )| )]sgn(cn(( )| )),
= 2sgn0 k dn(( )| ),

02 +4 2 k20

(19)

where = 1k , k =
, k0 = sin 20 , 0 , 0 are the initial angular velocity and the initial
2
orientation of the dipole, respectively, and the sgn(x) function is defined as
sgn(x) =

x 0,
x < 0.

(20)

Furthermore, we must take into account some additional definitions; these are the Jacobi functions
[10]
sn(x|k) = sin(am(x|k)),
cn(x|k) = cos(am(x|k)),

(21)

dn(x|k) = (1 k2 sn2 (x|k))

and am(x|k) is the inverse of an incomplete elliptic function [10] of the first kind, with
am(x|k)

x=
0

(1 k2 sin2 ( ))

(22)

B. Atenas et al. / Annals of Physics 350 (2014) 605614

609

The set of equations (19) is valid for any value of the parameter and the value of this parameter
classifies the motion of the dipole into two possible states: if 1, the dipole has sufficient energy
to rotate and, if > 1, the dipole oscillates around equilibrium. For the latter case, it is suitable, from
the numerical point of view, to reformulate the set of equations (19) as a function of the parameter k,
which takes values in accordance with 0 k 1:

= sgn0 [ 0 ] + sn1

k ,
k

k0

= 2 arcsin[ksn(|k)],
= 2sgn0 k cn(|k).

(23)

The angle and the angular velocity are periodic functions with the following period:

T =

2 K ( )/
4K (1/ )/

1,
> 1,

(24)

where K (x) is the complete elliptic integral of the first kind.

3. Motion of the center of mass
The law of motion that satisfies the position of the center of mass can be found by integrating Eq.
(15), with the aid of Eqs. (19) and (23), and is written as follows:

1 = 2 sn(| )cn(| ),
2 = 1 + 2sn2 (| ),

(25)
(26)

where 1 and 2 are the components of the velocity of the center of mass in the directions e and e C
respectively. Eq. (25) is easily integrable, so
d(dn(x| ))
dx
d1
d

= 2 sn(x| )cn(x| ),

sgn0 d(dn(| ))

= 4

E0 2
sgn0

1 = 10 4

E0 2

(27)

[dn(| ) dn(0 | )],

where 10 is the initial condition for 1 , and the parameters 0 and E0 are: 0 = sn1 (k0 | ) and
2

E0 = 0 + 4 2 k20 , respectively. The solution of Eq. (26) can also be obtained from the primitive of

the function sn2 (x| ) [11] as 12 [x E (am(x| ))], where E (x| ) is an incomplete elliptic function of
the second kind:
sgn0

2
2 = 20 + 2
( 1)( 0 ) + 2 { 0 + E (am(0 | )) E (am(| ))} . (28)

E0 2
Eqs. (27) and (28) represent the general laws of motion for the position of the center of mass regardless
whether the dipole possesses enough energy to rotate, namely, for any value of . To clarify the
formulation, we rephrase Eqs. (27) and (28) for > 1 as a function of k as follows:

1 = 10 2
2 = 20 +

k sgn0

sgn0

[cn(|k) cn(0 |k)],

[( 1)( 0 ) + 2{ 0 + E (am(0 |k)) E (am(|k))}] .

(29)

610

B. Atenas et al. / Annals of Physics 350 (2014) 605614

Consider two sets of initial conditions that correspond to limiting cases of the integrals of motion.
Case 1: (Fixed dipole): 0 = 0, 0 = 0, k = 0
The set of equations (29) becomes the set

1 = 10 ,
2 = 20 + ( 1) ( 0 ) ,

(30)

which represents a free motion of the fixed dipole in the direction defined by C . We can
distinguish two possibilities depending on the value of . This is, if > 1, the motion is
parallel to C . If < 1, the motion is antiparallel to C , which is the same trajectory observed
by Curilef et al. [5] and Troncoso et al. [6]. For a direct derivation of this last sentence, see
Appendix. The resting system corresponds to the marginal case where = 1.
Certainly, the marginal case is quite unexpected. Charges are at rest in the relative
coordinate. Therefore, only the Coulomb force is present. By contrast, the center of mass
moves perpendicular to the direction of the constant of motion C . Indeed, such a motion of
the center of mass is additionally perpendicular to the relative vector between particles, thus
both particles move on straight lines with the same velocity. This fact permits the appearing,
by the Lorentz force, of an electric force which cancels out the Coulomb interaction between
charges. Besides, all vectors involved in this motion are perpendicular among them and
coincide with the definition of the basis given by Eq. (13). Alternatively, we remark that
the combination of the parameters corresponds to a motion that occurs in the minimum of
the effective potential that can be derived from Eq. (12). Thus, the system with both charges
constrained by the fixed distance naturally move, on straight lines with the same velocity,
with an energy level that exactly coincides with the minimum of the effective potential.

Case 2: (Rotating dipole): = 0, 0 = 0 eB

In this case, the chosen basis becomes meaningless because the vectors eC and e are not
defined. The new basis is 0 , eB 0 and eB , with 0 = 0. Thus, k =
the set of equations (19) become the set

0
2

, = 0 and

= 0 ,
= 0 ,

(31)

which represents a rotation with constant angular velocity for the relative variable that is
one of the special trajectories identified in [6].
Now, by inserting the set of equations given in Eq. (31) into Eq. (15) and performing the subsequent
integration, we can write:

= 0 +

(0 ) .

(32)

In the limit 0 0, Eq. (32) corresponds to Case 1. In other words, the center of mass moves freely
in the direction perpendicular to the constant orientation of the dipole.
4. Trapped states
Trapped states exhibit the property of having a mean velocity equal to zero [4,7]. For the variable
of the center of mass, Eqs. (27) and (28) mean:

1 ( ) 10 = 0,
2 ( ) 20 = 0,
= 2 sgn0 k( ) + 0 .

(33)

B. Atenas et al. / Annals of Physics 350 (2014) 605614

611

Fig. 1. The pseudomomentum (from Eq. (11)) is depicted as a function of from Eq. (34).

The first equation of (33) is immediately satisfied because of the periodicity of Jacobi functions. The
second is satisfied only if
=12
=12

K ( ) E ( )

1,

2 K ( )
K (k) E (k)
K (k)

> 1,

(34)

where E (x) is the complete elliptic integral of the second kind. The condition for the existence of
trapped states is included in Eq. (34). As concluded previously in [4], trapped states can only exist if
1, which is the same result as in (34).
Case 1: 1 represents rotating dipoles Eq. (34) has nonnegative solutions (Fig. 1) if = 0, = 0.
Eq. (32) represents a trapped motion. In the spatial reference system centered on the initial
position of the center of mass, the previous equation is a curve that forms a sort of cardioid
(see Fig. 2).
Case 2: > 1
Eq. (34) may possess positive solutions if 0 k < kmax , kmax 0.9. By analyzing Eq. (34)
in terms of k 0 and preserving the terms of order k2 ; 1 k2 and Eqs. (29) can be
converted into the following:

1 10 2
2 20

sgn0

sgn0
2

k(cos cos 0 ),
k2 (sin(2) sin(20 )).

(35)

The approximation for k 0 that leads to Eq. (35) defines a Lissajous figure with = 2
centered in a suitable system of reference. However, if we repeat the calculation numerically
for higher values of k, surprisingly, we obtain a nearly identical figure. As stated above, the
orbits for all values of the parameter k are very similar.
In order to show a graphical example for the functions obtained from Eq. (35), we define
the following normalized coordinates for the center of mass

1 10
,
Max(1 10 )
2 20
=
,
Max(2 20 )

1N =
2N

(36)

612

B. Atenas et al. / Annals of Physics 350 (2014) 605614

Fig. 2. The normalized orbit of the center of mass. K = 0, = 0, which form a cardioid.

Fig. 3. The normalized path of the center of mass (of the dipole) for the initial conditions given by 0 = 0. Orbits for two values
of k are compared. The path for small k (namely, k = 0.01) is represented by a solid line, and the case k = 0.85 is represented
by red points. It is apparent that the two curves nearly coincide.

where Max(x) means the maximum value of the variables x. Thus, in Fig. 3 we depict 1N
versus 2N , with 0 = 0.
5. Concluding remarks
In summary, the present study supports and generalizes previous analyzes discussed in the
literature, which can be considered as particular cases of the present analysis.
The Lagrangian function is first explicitly written in terms of particle coordinates, but after a suitable change of variables we write the Lagrangian in terms of center of mass and relative coordinates.
All subsequent analysis is performed using these coordinates. Thus, the motion of the center of mass
is essentially negligible and trapped states are only briefly mentioned. Here, the equations of motion
are integrated directly and precisely, establishing the conditions for the existence of trapped states.
These conditions are in agreement with previous results [4] and establish the relation between the
values of constants of motion C and E that enables the existence of trapped states.

B. Atenas et al. / Annals of Physics 350 (2014) 605614

613

The integration is performed in terms of special functions, such as Jacobi functions [10] that have
been previously defined. We show a complete analytical solution for the problem, classifying separately the motion of the center of mass and the relative motion. In one hand, we derive the two
components for the trajectory of the motion of the center of mass. Two limiting cases that emphasize
particular initial conditions are shown, these are specifically the fixed dipole and the rotating dipole.
First, it is interesting to note that the system, under certain conditions, can move against the sense of
the pseudomomentum. As expected, the motion in the pseudomomentum sense is possible too. In the
latter case, a rotation with constant angular velocity is obtained for the relative variable previously
identified in [6].
On the other hand, the relevant observation on the existence of a kind of special states is made
that we have called: trapped states. This is a family of states, which are appointed in the special
motion of the relative coordinate. The main property of having a mean velocity equal to zero without
turning points is the fingerprint of this kind of interesting states. Here, we again identify two limiting
cases. The first limiting case, a cardioid-like curve is shaped (see Fig. 2) in the spatial reference system
centered on the initial position of the center of mass. Another limiting and interesting case, which
is graphically illustrated in Fig. 3, is the identical behavior exhibited by the normalized paths of the
center of mass for several values of k. A comparison between the small values of k and k = 0.85 is
presented in Fig. 3, where it can be seen that the curves are very similar.
Certainly, the present study of the behavior of the electric dipole in presence of a magnetic field
is classical. The behavior strongly depends on the initial conditions. If we slightly modify the initial
conditions for the system, we no longer obtain this special family of states. We think this is a relevant
contribution that we expect to follow studying by several perspectives. As said before, for the time
being, it is expected that some physical principles at the scale of a molecular engine can be identified
by applying rotor modeled as electric dipoles in external magnetic fields.
Acknowledgments
We acknowledge partial financial support from CONICYT-UCN PSD-065. In addition, one of us (B.A.)
acknowledges the financial support from CONICyT/Beca Magster Nacional 2014, Folio 22140054.
Appendix
The constant of motion, C , whose dependence on the conjugate momentum of the center of mass
and the relative coordinates is given, in Eq. (11), by

+
C = M R

e
c

r ,
B

(37)

whose compound in one of the elements of the basis (13), e C , yields

e
r e C ,
B
c
which can be written in terms of the dimensionless variables, defined in the text, as follows:

e C +
C e C = M R

= C + e B e e C .

(38)

(39)

Now, the physical meaning of this quantity imposes

= C + 1 0.

(40)

Therefore,
1 = C 1.

(41)

Thus, 1 C < ; hence, if 1 C < 0, the dipole moves against the sense of the pseudomomentum that seems not to be intuitive. If C = 0, we obtain the marginal case, the resting dipole. If
C > 0, the dipole has a motion parallel to the pseudomomentum.

614

B. Atenas et al. / Annals of Physics 350 (2014) 605614

References
[1] V.V. Volchkov, M.N. Khimich, M. Ya Melnikov, B.M. Uzhinov, Hihg Energy Chem. 47 (2013) 224229.
[2] Y. Yoshida, Y. Shimizu, T. Yajima, G. Maruta, S. Takeda, Y. Nakano, T. Hiramatsu, H. Kageyama, H. Yamochi, G. Saito,
Chemistry-A European J. 19 (2013) 1231312324.
[3] N. Ogiwara, T. Yanagibashi, Y. Hikichi, M. Nishikawa, J. Kamiya, K. Wada, Vacuum 98 (2013) 1821.
[4] P. Troncoso, S. Curilef, European J. Phys. 27 (2006) 13151322.
[5] S. Curilef, F. Claro, Amer. J. Phys. 65 (1997) 244250.
[6] M.A. Escobar-Ruiz, A.V. Turbiner, J. Math. Phys. 54 (2013) 022901.
[7] D.L. Pursey, N.A. Sveshnikov, A.M. Shirokov, Theoret. Math. Phys. 117 (1998) 12621273.
[8] L.P. Gorkov, I.E. Dzyaloshinskii, Zh. Eksp. Teor. Fiz. 53 (1967) 717722.
[9] O. Karlheinz, Eur. J. Phys. 32 (2011) 479490.
[10] P.F. Byrd, M.D. Friedmann, Handbook of Elliptical Integrals for Engineers and Scientists, Springer, New York, 1971;
M. Abramowitz, I.A. Stegun, Handbook of Mathematical Functions with Formulas Graphs and Mathematical Tables, Dover,
New York, 1972;
H. Hancock, Elliptic Integrals, Dover, New York, 1958.
[11] Hebert Bristol Dwight, Table of Integrals and Other Mathematical Data, Chicago, New York, 1957.