Physics Letters A 372 (2008) 2984–2988

www.elsevier.com/locate/pla

Fractional Heisenberg equation

Vasily E. Tarasov
Skobeltsyn Institute of Nuclear Physics, Moscow State University, Moscow 119992, Russia
Received 10 July 2007; received in revised form 9 January 2008; accepted 11 January 2008
Available online 29 January 2008
Communicated by A.P. Fordy

Abstract
Fractional derivative can be defined as a fractional power of derivative. The commutator (i/h̄)[H, ·], which is used in the Heisenberg equation,
is a derivation on a set of observables. A derivation is a map that satisfies the Leibnitz rule. In this Letter, we consider a fractional derivative
on a set of quantum observables as a fractional power of the commutator (i/h̄)[H, ·]. As a result, we obtain a fractional generalization of the
Heisenberg equation. The fractional Heisenberg equation is exactly solved for the Hamiltonians of free particle and harmonic oscillator. The
suggested Heisenberg equation generalize a notion of quantum Hamiltonian systems to describe quantum dissipative processes.
© 2008 Elsevier B.V. All rights reserved.

PACS: 03.65.-w; 03.65.Ca; 45.10.Hj; 03.65.Db

Keywords: Heisenberg equation; Fractional derivative; Fractional equation

1. Introduction Fractional derivative can be defined as a fractional power

of derivative (see for example [9]). It is known that the in-
finitesimal generator (i/h̄)[H, ·], which is used in the Heisen-
The fractional calculus has a long history from 1695, when
berg equation, is a derivation of observables. A derivation is
the derivative of order 1/2 has been described by Leibniz [1,2].
a linear map D, which satisfies the Leibnitz rule D(AB) =
The theory of derivatives and integrals of non-integer order goes
(DA)B + A(DB) for all operators A and B. In this Letter,
back to Leibnitz, Liouville, Grunwald, Letnikov and Riemann.
we consider a fractional derivative on a set of observables as
There are many books about fractional calculus and fractional
a fractional power of derivative. As a result, we obtain a frac-
differential equations [1–4]. Derivatives of fractional order, and
tional generalization of the Heisenberg equation. It allows us to
fractional differential equations have found many applications
generalize a notion of quantum Hamiltonian systems. Note that
in recent studies in physics (see for example [5–8] and refer-
fractional generalization of classical Hamiltonian systems has
ences therein).
been suggested in [10]. The suggested fractional Heisenberg
In the quantum kinematics, the observables are given by self-
equation is exactly solved for the Hamiltonians of free particle
adjoint operators. The dynamical description of system is given
and harmonic oscillator. A quantum system that is presented by
by a superoperator, which is a rule that assigns to each opera-
fractional Heisenberg equation can be considered as a dissipa-
tor exactly one operator. The natural description of the motion
tive system [11]. Fractional derivatives can be used as a possible
is in terms of the infinitesimal change of the system. The equa-
approach to describe an interaction between the system and an
tion of motion for quantum observable is called the Heisenberg
environment. Note that it is possible to consider quantum dy-
equation. For Hamiltonian systems, the infinitesimal superop-
namics with low-level fractionality by some generalization of
erator of the Heisenberg equation is defined by some form of
method suggested in [12] (see also [13]).
derivation.
In Section 2, the fractional power of derivative and the frac-
tional Heisenberg (FH) equation are suggested. In Section 3,
E-mail address: tarasov@theory.sinp.msu.ru. the properties of time evolution, which is described by the frac-
0375-9601/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.physleta.2008.01.037
 V.E. Tarasov / Physics Letters A 372 (2008) 2984–2988 2985

tional equation, are considered. In Section 4, the FH equation called the generating superoperator, or infinitesimal generator,
for free particle is solved. In Section 5 the solution of FH equa- of the semi-group {Φt , t  0}.
tion with harmonic oscillator Hamiltonian is derived. Let us consider the Cauchy problem for fractional Heisen-
berg equation (3) in which the initial condition is given by A0 .
2. Fractional derivative and Heisenberg equation Then its solution can be presented in the form
(α)
Let us consider a set of quantum observables in the Heisen- At (α) = Φt A0 ,
berg picture. A superoperator L is a rule that assigns to each (α)
where the superoperators Φt , t > 0, form a semi-group
operator A exactly one operator L(A). (About superoperator which will be called the fractional semi-group. The super-
formalism see for example [14–16] and references therein.) For operator (L− α
H ) is infinitesimal generator of the semi-group
Hamiltonian H , let L−H be the superoperator given by (α)
{Φt , t  0}.
1
L−
HA = (H A − AH ). 3. Properties of fractional time evolution
i h̄
The operator differential equation Let us consider some properties of time evolution described
(α)
d by a fractional semi-group {Φt , t  0}.
At = −L− H At (1) (α)
dt (1) The superoperators Φt can be constructed in terms
is called the Heisenberg equation for Hamiltonian systems. of the superoperators Φt by the Bochner–Phillips formula
The time evolution of a Hamiltonian system is induced by the [17,18,20]
Hamiltonian H . ∞
It is interesting to obtain a fractional generalization of (α)
Φt = ds fα (t, s)Φs (t > 0), (5)
Eq. (1). We will consider here concept of fractional power
for L− −
H . If LH is a closed linear superoperator with an every-
0

where dense domain D(L− −

H ), having a resolvent R(z, LH ) =
where fα (t, s) is defined by
− −1
(zLI − LH ) on the negative half-axis, then there exists 
a+i∞
[19–21] the superoperator 1  
fα (t, s) = dz exp sz − tzα , (6)
2πi
∞ a−i∞
sin πα
−(L−
H) =
α
dz zα−1 R(−z, L− −
H )LH , (2) where a, t > 0, s  0, and 0 < α < 1. The branch of zα is so
π
0 taken that Re(zα ) > 0 for Re(z) > 0. This branch is a one-
defined on D(L− for 0 < α < 1. The superoperator (L− α valued function in the z-plane cut along the negative real axis.
H) H ) is
a fractional power of the Lie left superoperator. The convergence of this integral is obviously in virtue of the
As a result, we obtain the equation convergence factor exp(−tzα ). By denoting the path of in-
tegration in (6) to the union of two paths r exp(−iθ), and
d
At = −(L− α
(3) r exp(+iθ), where r ∈ (0, ∞), and π/2  θ  π , we obtain
H ) At ,
dt ∞
where t and H /h̄ are dimensionless variables. This is the frac- 1  
fα (t, s) = dr exp sr cos θ − tr α cos(αθ )
tional Heisenberg equation. π
0
Note that this equation cannot be presented in the form  
× sin sr sin θ − tr α sin(αθ ) + θ . (7)
d i
At = −L−
Hnew At = [Hnew , At ] (4) If we have a solution At of the Heisenberg equation (1), then
dt h̄
formula (5) gives the solution
with some operator Hnew . Therefore, quantum systems de-
scribed by (3) are not Hamiltonian systems. The systems will be ∞
called the fractional Hamiltonian systems (FHS). Usual Hamil- At (α) = ds fα (t, s)As (t > 0), (8)
tonian systems can be considered as a special case of FHS. 0
If we consider the Cauchy problem for equation (1) in which of fractional equation (3).
the initial condition is given at the time t = 0 by A0 , then its (2) In quantum theory, the most important is the class of real
solution can be written in the form superoperators. Let A∗ be an adjoint of A. A quantum observ-
At = Φt A0 . able is a self-adjoint operator. If Φt is a real superoperator and
A is a self-adjoint operator A∗ = A, then the operator At = Φt A
The one-parameter superoperators Φt , t  0 have the properties is self-adjoint, i.e.,
Φt Φs = Φt+s (t, s > 0), Φ0 = LI , (Φt A)∗ = Φt A.
where LI is unit superoperator (LI A = A). Then the set of Φt , A superoperator as a map from a set of observables into itself
t  0, is called the semi-group. Then the superoperator L−H is should be real. All possible dynamics, i.e., temporal evolutions
2986 V.E. Tarasov / Physics Letters A 372 (2008) 2984–2988

d 1 d
of quantum observables, should be described by real superoper- Qt = Pt , Pt = 0.
ators. Therefore the following statement is very important. If Φt dt m dt
(α) The well-known solutions of these equations are
is a real superoperator, then Φt is real. It can be proved by us-
ing the Bochner–Phillips formula and Eq. (7). t
(3) An adjoint superoperator of Φt is a superoperator Φ̄t , Qt = Q0 + P0 , Pt = P0 . (9)
m
such that
    Using these solutions and the Bochner–Phillips formula, we
Tr (Φ̄t A)∗ B = Tr A∗ Φt (B) . will obtain the solutions of fractional Heisenberg equations

Let us give the basic statement regarding the adjoint superoper- d 1 d

Qt = − α (L−
P2
)α Qt , Pt = 0. (10)
ator. If Φ̄t is an adjoint superoperator of Φt , then the superop- dt m dt
erator
Note that (L− P2
)α = L−
P 2α
. The solutions of fractional equa-
∞ tions (10) have the forms
(α)
Φ̄t = ds fα (t, s)Φ̄s (t > 0),
∞
(α)
0
Qt (α) = Φt Q0 = ds fα (t, s)Qs , Pt (α) = P0 ,
(α)
is an adjoint superoperator of Φt .
We prove this statement by 0
using the Bochner–Phillips formula: where Qs is given by (9). Then
∞
 (α) ∗    1
Tr Φ̄t A B = ds fα (t, s) Tr (Φ̄s A)∗ B Qt = Q0 + gα (t)P0 , Pt = P0 ,
m
0 where
∞
  ∞
= ds fα (t, s) Tr A∗ Φs (B)
gα (t) = ds fα (t, s)s.
0
  0
= Tr A∗ Φt (B) .
(α)
If α = 1/2, then we have
Let {Φ̄t , t > 0} be a semi-group, such that the density matrix
∞
operator ρt = Φ̄t ρ0 is described by the von Neumann equation t 1 t2
ds √ e−t /4s = .
2
g1/2 (t) = √
d 1 2 π s 2
ρt = [H, ρt ]. 0
dt i h̄
Then
(α)
Then the semi-group {Φ̄t , t > 0} describes the evolution of t2
the density operator Qt = Q0 − P0 , Pt = P0 . (11)
2m
(α)
ρt (α) = Φ̄t ρ0 These equations describe a fractional free motion for α = 1/2.
For the operators (11), the average values and dispersions have
by the fractional equation the form
d t2
ρt = −(−L− α
H ) ρt . Qt  = x0 − p0 , Pt  = p0 ,
dt 2m
This is the fractional von Neumann equation. and
(4) Let Φt , t > 0, be a positive one-parameter superoperator,  
i.e., Φt A  0 for A  0. Using the Bochner–Phillips formula h̄2 b2 h̄2 t 4
DP (t) = 2 , DQ (t) = 1+ 2 4 .
and the property fα (t, s)  0, (s > 0), it is easy to prove that 2b 2 m b
(α) (α)
the superoperators Φt are also positive, i.e., Φt A  0 for Here we use the coordinate representation and the pure state
A  0.
 
(5) It is known that Φ̄t is a real superoperator if Φt is real. √ −1/2 (x − x0 )2 i
(α) (α)
Analogously, if Φt is a real superoperator, then Φ̄t is real. Ψ (x) = x|Ψ  = (b π ) exp − + p0 x .
2b h̄
(12)
4. Fractional free particle
The average value and dispersion are defined by the well-known
equations
Let us consider the Hamiltonian H = P 2 /2m, where P is di-
 
mensionless variable, and m−1 has the action dimension. Then At  = Tr |Ψ Ψ |At = Ψ |At |Ψ ,
the Heisenberg equation (1) describes a free one-dimensional

particle. For A = Q, and A = P , Eq. (1) gives DA (t) = A2t − At 2 = Ψ |A2t |Ψ  − Ψ |At |Ψ 2 .
 V.E. Tarasov / Physics Letters A 372 (2008) 2984–2988 2987

5. Fractional Heisenberg equation for harmonic oscillator These functions can be presented through the Macdonald func-
tion (see [22], Section 2.5.37.1.) such that
Let us consider the Hamiltonian 
 2 1/4   2

ωt +πi/8 +πi/4 ωt
1 2 mω2 2 C1/2 (t) = e K−1/2 2e
H= P + Q , (13) 4π 4
2m 2 
 
ωt 2
where t and P are dimensionless variables. Then Eq. (1) de- + e−πi/8 K−1/2 2e−πi/4 ,
scribes a harmonic oscillator. For A = Q, and A = P , Eq. (1) 4

gives  2 1/4   
ωt ωt 2
d 1 d S1/2 (t) = i e+πi/8 K−1/2 2e+πi/4
Qt = Pt , Pt = −mω2 Qt . 4π 4
dt m dt 
 
ωt 2
The well-known solutions of these equations are − e−πi/8 K−1/2 2e−πi/4 , (19)
4
1
Qt = Q0 cos(ωt) + P0 sin(ωt), where ω > 0, and Kα (z) is the Macdonald function [1,2], which
mω is also called the modified Bessel function of the third kind.
Pt = P0 cos(ωt) − mωQ0 sin(ωt). (14) Using (12), we get the average values
Using these solutions and the Bochner–Phillips formula, we 1
will obtain the solutions of fractional Heisenberg equations Qt  = x0 Cα (t) + p0 Sα (t),
mω
d d Pt  = p0 Cα (t) − mωx0 Sα (t),
Qt = −(L− α
H ) Qt , Pt = −(L− α
H ) Pt , (15)
dt dt and the dispersions
where H is defined by (13). The solutions of fractional equa-
h̄2 2 b2 m2 ω2 2
tions (15) have the forms DP (t) = C (t) + Sα (t),
2b2 α 2
∞ b2 h̄2
Qt (α) = Φt Q0 =
(α)
ds fα (t, s)Qs , DQ (t) = Cα2 (t) + 2 2 2 Sα2 (t).
2 2b m ω
0 As a result, the fractional harmonic oscillator is a simple dis-
∞
sipative systems. The dispersion of the wave packet is defined
(α)
Pt (α) = Φt P0 = ds fα (t, s)Ps . (16) by these equations. The solutions are characterized by the frac-
0 tional damping of in the fractional harmonic oscillator. The
dumping is described by the modified Bessel function (19) of
Substitution of (14) into (16) gives the third kind. An important property of the evolution described
1 by the fractional equations are that the solutions have power-
Qt = Q0 Cα (t) + P0 Sα (t), (17) like tails.
mω
Pt = P0 Cα (t) − mωQ0 Sα (t), (18)
6. Conclusion
where
∞ In this Letter, we consider derivatives of noninteger order as
fractional powers of derivative. We derive a fractional general-
Cα (t) = ds fα (t, s) cos(ωs),
ization of the Heisenberg equation and a generalization of quan-
0 tum Hamiltonian system to describe open quantum systems.
∞ A quantum system that is presented by fractional Heisenberg
Sα (t) = ds fα (t, s) sin(ωs). equation can be considered as a dissipative systems. Fractional
0 derivatives can be used as an approach to describe an interaction
between the quantum system and an environment. This inter-
Eqs. (17) and (18) describe solutions of the fractional Heisen- pretation caused by following reasons. Using the properties
berg equations (15) for quantum harmonic oscillator.
If α = 1/2, then ∞
fα (t, s) = 1, fα (t, s)  0 (for all s > 0),
∞
t cos(ωs) 2
C1/2 (t) = √ ds 3/2 e−t /4s , 0
2 π s we can assume that fα (t, s) is a density of probability distribu-
0
tion. Then the Bochner–Phillips formula (5) can be considered
∞
t sin(ωs) −t 2 /4s as a smoothing of Hamiltonian evolution Φt with respect to
S1/2 (t) = √ ds e . time s > 0. This smoothing can be considered as an influence of
2 π s 3/2
0 the environment on the system. As a result, the parameter alpha
2988 V.E. Tarasov / Physics Letters A 372 (2008) 2984–2988

can be used as a simple model of the interaction between the [7] G.M. Zaslavsky, Phys. Rep. 371 (2002) 461.
system and the environment. Note that the quantum Markovian [8] E.W. Montroll, M.F. Shlesinger, The wonderful world of random walks,
in: J. Lebowitz, E. Montroll (Eds.), Studies in Statistical Mechanics,
equations, which is also called the Lindblad equations [23], can
vol. 11, North-Holland, Amsterdam, 1984, pp. 1–121.
be generalized by suggested approach to describe completely [9] V.E. Tarasov, Int. J. Math. 18 (3) (2007) 281.
positive evolution of dissipative and open quantum systems. [10] V.E. Tarasov, J. Phys. A 38 (2005) 5929.
[11] U. Weiss, Quantum Dissipative Systems, World Scientific, Singapore,
References 1993.
[12] V.E. Tarasov, G.M. Zaslavsky, Physica A 368 (2) (2006) 399.
[13] A. Tofighi, H.N. Pour, Physica A 374 (1) (2007) 41.
[1] K.B. Oldham, J. Spanier, The Fractional Calculus: Theory and Appli- [14] V.E. Tarasov, J. Phys. A 37 (2004) 3241.
cations of Differentiation and Integration to Arbitrary Order, Academic [15] V.E. Tarasov, Phys. Rev. E 66 (2002) 056116.
Press, New York, 1974. [16] V.E. Tarasov, Phys. Lett. A 288 (2001) 173.
[2] S.G. Samko, A.A. Kilbas, O.I. Marichev, Fractional Integrals and Deriva- [17] S. Bochner, Proc. Natl. Acad. Sci. USA 35 (1949) 369.
tives Theory and Applications, Gordon and Breach, New York, 1993. [18] R.S. Phillips, Pacific J. Math. 2 (1952) 343.
[3] I. Podlubny, Fractional Differential Equations, Academic Press, San [19] V. Balakrishnan, Pacific J. Math. 10 (1960) 419.
Diego, 1999. [20] K. Yosida, Functional Analysis, Springer, Berlin, 1965.
[4] A.A. Kilbas, H.M. Srivastava, J.J. Trujillo, Theory and Application of [21] S.G. Krein, Linear differential Equations in Banach Space, Translations of
Fractional Differential Equations, Elsevier, Amsterdam, 2006. Mathematical Monographs, vol. 29, Amer. Math. Soc., 1971.
[5] G.M. Zaslavsky, Hamiltonian Chaos and Fractional Dynamics, Oxford [22] A.P. Prudnikov, Yu.A. Brychkov, O.I. Marichev, Elementary Functions,
Univ. Press, Oxford, 2005. Integrals and Series, vol. 1, Gordon and Breach, New York, 1986.
[6] A. Carpinteri, F. Mainardi (Eds.), Fractals and Fractional Calculus in Con- [23] G. Lindblad, Commun. Math. Phys. 48 (1976) 119.
tinuum Mechanics, Springer, Wien, 1997.