Research Article Volume 5 Issue 3
1Physics, Rajamangala, University of Technology Thanyaburi, Thailand
1Department of Physics, Tezpur University, India
Correspondence: Surarit Pepore, Physics, Rajamangala, University of Technology Thanyaburi, Thailand, Tel 0925202842
Received: May 30, 2021 | Published: October 19, 2021
Citation: Pepore S. Propagators for a particle in a time-dependent linear potential and a free particle. Phys Astron Int J. 2021;5(3):83-88. DOI: 10.15406/paij.2021.05.00238
In this paper, the propagators for a particle moving in a time-dependent linear potential and a free particle with linear damping are calculated by the application of the integrals of the motion of a quantum system. The propagator for a charged harmonic oscillator is derived from the Feynman path integrals and the propagator for a damped harmonic oscillator is evaluated by the Schwinger method. The relation between the integrals of the motion, Feynman path integrals, and Schwinger method are also described.
Keywords: propagators, free particle, harmonic oscillator, integrals of the motion, feynman path integrals, schwinger method
In quantum mechanics and quantum field theory, the propagator or Green function is represented as the transition probability amplitude for a particle to travel from initial space-time configuration to final space-time configuration. The standard method in calculating the propagator is Feynman path integral.1 In 2006, S.Pepore and et al.2 applied the Feynman path integral method to Calculate the propagator for a harmonic oscillator with time-dependent mass and frequency. The one aim of this paper is using the path integral method to derive the propagator for a charged harmonic oscillator in time-dependent electric field.
The another method in calculating the propagator is the Schwinger method.3 This method was first formulated by Schwinger in 1951 for solving the gauge invariance and vacuum polarization in QED. In 2015, the Schwinger method was used to derive the propagator for time-dependent harmonics oscillators by S.Pepore and B.Sukbot.4–6 The one purposes of this article is applying the Schwinger method to calculate the propagator for a damped harmonic oscillator.
In 1975, V.V. Dodonov, I.A. Malkin, and V.I. Man’ko7 presented the connection between the integrals of the motion of a quantum system and its propagator that is the eigenfunction of the integrals of the motion describing
initial points of the system trajectory in the phase space. In 2018, S. Pepore applied the integrals of the motion to calculate the propagators for time-dependent harmonic oscillators.8,9 The one aim of this article is applying the integrals of the motion to derive the propagators for a particle moving in a time-dependent linear potential and a free particle with linear damping. The organization of this paper are as follows. In Sec.2, the propagator for a particle in a time-dependent linear potential is derived. In Sec.3, the propagator for a free particle with linear damping is obtained with the aid of the integrals of the motion. In Sec 4, the Feynman path integrals is applied to evaluate the propagator for a charged harmonic oscillator in time-dependent electric field. In Sec.5, the procedures of the Schwinger method are described. In Sec.6, the propagator for a damped harmonic oscillator are derived by the Schwinger method. Finally, the conclusion is presented in Sec.8.
In this section, we will calculate the propagator for a particle moving in a
time-dependent linear potential described by the Hamiltonian operator.6
ˆH(t)=ˆp22m−ktˆx (1)
Where k is a constant and t is time.
The classical equation of motion for this system is
m¨x−kt=0 (2)
The classical paths in the phase space under the initial conditions x(0)=x0 and P(0)=P0 are given by
x(t)=x0+tmp0+kt36m (3)
p(t)=p0+kt22. (4)
Now we consider the systems of Eqs.(3) and (4) as an algebraic system for unknown initial position x0 and initial momentum p0 . The variables x,p, and t are taken as the parameters. The solution of this system can be written as the operator in Hilbert space as
ˆx0(ˆx,ˆp,t)=ˆx−tmˆp+kt33m (5)
ˆp0(ˆx,ˆp,t)=ˆp−kt22. (6)
The operators ˆx0 and ˆp0 are the integrals of the motion because theirs satisfy equation of
dˆIdt=∂ˆI∂t+iћ[ˆH,ˆI]=0, (7)
Where ˆI may be ˆx0 and ˆp0 . Then these operatos must satisfy equations for the Green function or propagator,2,5,6
ˆx0(x)K(x,x′,t)=ˆx(x′)K(x,x′,t) (8)
ˆp0(x)K(x,x′,t)=−ˆp(x′)K(x,x′,t), (9)
where the operators on the left-hand sides of the equations act on variables x , and on the right-hand sides, on x′ .
Now we write Eqs.(8) and (9) explicitly,
(x+itћm∂∂x+kt33m)K(x,x′,t)=x′K(x,x′,t), (10)
(−iћ∂∂x−kt22)K(x,x′,t)=iћ∂K(x,x′,t)∂x′ . (11)
By modifying Eqs.(10) and (11), the system of equation for deriving the propagator are
∂K(x,x′,t)∂x=[im(x−x′)ћt+ikt23ћ]K(x,x′,t), (12)
∂K(x,x′,t)∂x′=[−im(x−x′)ћt+ikt26ћ]K(x,x′,t). (13)
Now one can integrate Eq. (12) with respect to the variable x to obtain
K(x,x′,t)=C(x′,t)expexp[iћ(m(x−x′)22ћ+kt23x)], (14)
Substituting Eq.(14) into Eq.(13), we obtain the differential equation for C(x′,t) as
∂C(x′,t)∂x′=(ikt26ћ)C(x′,t). (15)
Solving Eq.(15), the function C(x′,t) can be expressed as
C(x′,t)=C(t)expexp[ikt26ћx′]. (16)
So, the propagator in Eq.(14) can be written as
K(x,x′,t)=C(t)expexp[iћ(m(x−x′)22t+kt23x+kt26x′)]. (17)
To obtain C(t), we must substitute the propagator of Eq. (17) into the Schrodinger’s equation
iћ∂K(x,x′,t)∂t=(−ћ22m∂2K(x,x′,t)∂x2−ktxK(x,x′,t)). (18)
After some algebra, we obtain an equation
dC(t)dt=−(12t+ik2t418ћm)C(t). (19)
Equation (19) can be simply integrated with respect to time t ,and one obtains
C(t)=C√texpexp(−iћk2t590m), (20)
where C is a constant. Substituting Eq.(20) into Eq.(17) and applying the initial condition
limt→0+K(x,x′,t)=δ(x−x′), (21)
we obtain
C=√m2πiћ. (22)
So, the propagator for a particle moving in time-dependent linear potential is
K(x,x′,t)=√m2πiћtexpexp[iћ(m(x−x′)22t+kt23x+kt26x′−k2t590m)], (23)
which is the same form as the result of S.Pepore and B.Sukbot calculated by the Schwinger method.6
This section is the calculation of the propagator for a free particle with linear damping by the application of integrals of motion operators. Considering the motion of a free particle with constant mass in a linear damping which has the damping coefficient β , the Hamiltonian operator of this system can be written as10
ˆH(t)=e−γtˆp22m, (24)
Where γ=βm.
The Hamilton equation of motion for position and momentum are11
˙x=pme−γt,˙p=0. (25)
The classical paths in the phase space under the initial conditions x(0)=x0 and p(0)=p0 are given by
x(t)=x0+(1−e−γt)mγp0, (26)
p(t)=p0. (27)
Now we rewrite the systems of Eqs. (26) and (27) in terms of the initial position operator ˆx0 and initial momentum operator ˆp0 5as
ˆx0(ˆx,ˆp,t)=ˆx−[1−e−γtmγ]ˆp, (28)
ˆp0(ˆx,ˆp,t)=ˆp. (29)
The operators ˆx0 and ˆp0 are the integrals of the motion because theirs satisfy Eq.(7). Then these operators must satisfy Eqs.(8), (9), and (x+iћ(1−e−γtmγ)∂∂x)K(x,x′,t)=x′K(x,x′,t), (30)
−iћ∂K(x,x′,t)∂x=iћ∂K(x,x′,t)∂x′. (31)
By modifying Eqs.(30) and (31), the system of equation for calculating the propagator are
∂K(x,x′,t)∂x=[imγ(x−x′)ћ(1−e−γt)]K(x,x′,t), (32)
∂K(x,x′,t)∂x′=−[imγ(x−x′)ћ(1−e−γt)]K(x,x′,t). (33)
Now one can integrate Eq.(32) with respect to the variable x to obtain
K(x,x′,t)=C(x′,t)expexp[iћ(mγ1−e−γt(x22−xx′))]. (34)
Substituting Eq.(34) into Eq.(33), we obtain the differential equation for C(x′,t) as
∂C(x′,t)∂x′=iћ(mγ1−e−γt)x′C(x′,t). (35)
Solving Eq.(35), the function C(x′,t) can be express as
C(x′,t)=C(t)exp[i2ћ(mγ1−e−γt)x′2]. (36)
So, the propagator in Eq.(34) can be written as
K(x,x′,t)=C(t)exp[imγ(x−x′)22ћ(1−e−γt)]. (37)
To find C(t), we must substitute the propagator of Eq.(37) into the Schrodinger’s equation
iћ∂K(x,x′,t)∂t=−ћ22me−γt∂2K(x,x′,t)∂x2. (38)
After some algebra, we obtain an equation
dC(t)dt=−[γe−γt2(1−e−γt)]C(t). (39)
Equation (39) can be simply integrated with respect to time t , and one obtains
C(t)=C√1−e−γt, (40)
where C is a constant. Substituting Eq.(40) into Eq.(37) and applying the initial condition of Eq.(21), we obtain
C=√m2πiћ. (41)
So, the propagator for a free particle with linear damping is
K(x,x′,t)=√mγ2πiћ(1−e−γt)exp(imγ(x−x′)22ћ(1−e−γt)). (42)
The aim of this section is to derive the propagator for a charged harmonic oscillator in time-dependent electric field by Feynman path integral method.1 Considering the motion of a charged harmonic oscillator which has mass m and positive charge q moving in time-dependent electric field EcosΩt, the Lagrangian of this system can be written as
L=12m˙x2−12mω2x2−qEcosΩtx . (43)
By using the Euler-Lagrange equation for the Lagrangian in Eq.(43), the equation of motion can be written as
¨x+ω2x+qEcosΩt=0. (44)
The general solution of Eq.(44) is
x(t)=Acosωt+Bsinωt+qE(Ω2−ω2)cosΩt, (45)
Where A and B are constants. The constants A and B in Eq.(45) can be determined by imposing the boundary conditions of x(t′)=x′ and x(t'')=x'' . The classical path that connects the point of (x',t') and (x'',t'') can be written as
xcl(t)=[sinω(t−t′)sinωT]x''−[sinω(t−t'')sinωT]x′
−qE(Ω2−ω2)sinωT[cosΩt''sinω(t+t′)−cosΩt′sinω(t+t'')−sinωTcosΩt]. (46)
The action can be calculated from the time-integration of the Lagrangian from t′ to t"
S(x'',t'';x′,t′)=t''∫t′L(˙x,x,t)dt. (47)
For the action of our system, the Lagrangian in Eq.(43) is substituted into Eq.(47), and then integrated by parts of the first term on the right hand side of Eq.(43) and using the equation of motion in Eq.(44).
The classical action can be written as
Scl(x'',t'';x′,t′)=m2(x''cl˙x''cl−x'cl˙x'cl). (48)
Substituting the classical paths of Eq.(46) into Eq.(48), the classical action becomes
Scl(x'',t'';x′,t′)=mω2cotωT(x''2+x′2)−mωsinωTx''x′+mqE2(Ω2−ω2)sinωT[ωcosΩt''(sin2ωt′−cosω(t''+t′))+ωcosΩt′(cos2ωt′−sinω(t''+t′))+ωcotωT(cosΩt′sin2ωt′−cosΩt''sinω(t''+t′))+ωcosΩTcosΩt'']x''
+mqE2(Ω2−ω2)sinωT[ωcosΩt''(cos2ωt′−sin2ωt′cotωT−cosω(t''+t′)cscωT)−ωcosΩt′(cosω(t''+t′)−sinω(t''+t′)cotωT−cos2ωt'')+ΩsinΩt′sinωT+ωcosωTcosΩt′−ΩsinΩt'']x′
+mq2E22(Ω2−ω2)2sin2ωT
[ωcos2Ωt''(2cosωt''sinωt′cosω(t''+t′)−sin2ωt′cos2ωt′)−
ωcos2Ωt′(2sinωt''cosωt′cosω(t''+t′)−sin2ωt''cos2ωt′)+
ΩsinωT((sinΩt''+sinΩt′)cosΩt′sinω(t''+t′)−sinΩt''cosΩt′sin2ωt''−sinΩt′cosΩt''sin2ωt′)−
ωcosΩt′cosΩt''((sin2ωt''−sin2ωt′)cosω(t''+t′)−(cos2ωt''+cos2ωt′)sinωT)
−Ω2(sin2Ωt''+sin2Ωt′)sin2ωT
(49)
The quadratic Lagrangian propagator can be separated into a pure function of time F(t'',t′) and the exponential function of classical action Scl(x'',t'';x′,t′) as suggested in Ref.1
K(x'',t'';x′,t′)=F(t'',t′)eiScl(x'',t'';x′,t′)/ћ . (50)
Calculation of the function F(t'',t′) presented by Pauli,12 Morette,13 or Jones and Papadoupoulos14 can be performed by the semi classical approximation of path integral formula
F(t'',t′)=√12πiћ|∂2Scl∂x′∂x''| . (51)
By substituting the classical action of Eq.(49) into Eq.(51), the pre-exponential factor can be obtained as
F(t'',t′)=√mω2πiћsinωT . (52)
From Eqs.(49),(50) and (52), the propagator for a charged harmonic oscillator in time-dependent electric field can be expressed by
K(x'',t'';x′,t′)=√mω2πiћsinωTeiScl(x'',t'';x′,t′)/ћ . (53)
Begin by considering a time-dependent Hamiltonian operator ˆH(t) , the propagator is defined by
K(x,x′;t)=〈x|ˆTexp(−iћt∫0ˆH(t)dt)|x′〉, (54)
where ˆT is the time-ordering operator and |x〉,|x′〉 are the eigenvectors of the position operator ˆx (in the Schrodinger picture) with eigenvalues x and x′ , respectively.
The differential equation for the propagator in Eq.(54) can be written as
iћ∂K(x,x′;t)∂t=〈x|ˆHˆTexp(−iћt∫0ˆH(t)dt)|x′〉. (55)
Applying the relation between the operators in the Heisenberg and Schrodinger pictures, we obtain the equation for the propagator in the Heisenberg picture
iћ∂K(x,x′;t)∂t=〈x(t)|ˆH(ˆx(t),ˆp(t))|x′(0)〉, (56)
where |x(t)〉 and |x′(0)〉 are the eigenvectors of the operators ˆx(t) and ˆx(0), respectively, with the corresponding eigenvalues x and x′ . Besides, ˆx(t) and ˆp(t) satisfy the Heisenberg equations
iћdˆx(t)dt=[ˆx(t),ˆH],iћdˆp(t)dt=[ˆp(t),ˆH]. (57)
The main idea of the Schwinger method consists in the following steps.
(1). The first step is solving the Heisenberg equations for ˆx(t) and ˆp(t), and writing the solution for ˆp(t), only in terms of the operators ˆx(t) and ˆx(0).
(2). The next step is substituting the solutions obtained in step (1) into the expression for ˆH(ˆx(t),ˆp(t)) in Eq.(56) and employing the commutator [ˆx(0),ˆx(t)] to rewrite each term of ˆH(t) in a time ordered form with all operators ˆx(t) to the left and all operators ˆx(0). to the right. The time ordered Hamiltonian can be defined as ˆHord(ˆx(t),ˆx(0)).
(3). After this ordering, Eq.(56) can be written in the form
iћ∂K(x,x′;t)∂t=H(x,x′;t)K(x,x′;t), (58)
with H(x,x′;t) being an ordinary function defined as
H(x,x′;t)=〈x(t)|ˆHord(ˆx(t),ˆx(0))|x′(0)〉〈x(t)|x′(0)〉. (59)
Integrating Eq.(58) over , the propagator takes the form
K(x,x′;t)=C(x,x′)expexp{−iћt∫0H(x,x′;t)dt}, (60)
where C(x,x′) is an integration constant.
(4). The last step is the calculating of C(x,x′) This is obtained by using the following conditions
−iћ∂K(x,x′;t)∂x=〈x(t)|ˆp(t)|x′(0)〉, (61)
iћ∂K(x,x′;t)∂x′=〈x(t)|ˆp(0)|x′(0)〉, (62)
and the initial condition
limt→0+K(x,x′;t)=δ(x−x′). (63)
The Hamiltonian for a damped harmonic oscillator is described by8
H(t)=e−rtp22m+12mω2ertx2, (64)
Where r is the damping constant coefficient.
The equation of motion corresponding to the Hamiltonian in Eq. (64) is
¨x+r˙x+ω2x=0. (65)
The classical solution of Eq.(65) can be written in the form
x(t)=e−rt2(cosΩt+r2ΩsinΩt)x′+(e−rt2sinΩtmΩ)p′, (66)
where we impose the initial conditions x′=x(0) and p′=p(0).
The reduced frequency Ω in Eq.(66) is defined by Ω=√ω2−r24 . The reduced frequency Ω is real when ω2−r24>0. That is, we will be concerned with the under-damped case.
By solving the Heisenberg equation in Eq. (57), the position operators ˆx(t) can be written similarly to Eq.(66) as
ˆx(t)=e−rt2(cosΩt+r2ΩsinΩt)ˆx(0)+(e−rt2sinΩtmΩ)ˆp(0). (67)
The momentum operator ˆp(t)=m(t)ertˆ˙x(t) can be written by using Eq.(67) as
ˆp(t)=−(mω2ert2sinΩtΩ)ˆx(0)+ert2(cosΩt−rsinΩt2Ω)ˆp(0). (68)
By using Eq.(67), we can eliminate ˆp(0) from Eq.(68) by
ˆp(t)=mert(ΩcotΩt−r2)ˆx(t)−(mΩert2cscΩt)ˆx(0). (69)
Substituting ˆx(t) and ˆp(t) into the Hamiltonian operator
ˆH(t)=e−rtˆp22m+12mω2ertˆx2 with the aid of
[ˆx(0),ˆx(t)]=iћsinΩtmΩe−rt/2, (70)
the ordered Hamiltonian operator can be expressed as
ˆHord(t)=mert2(Ω2csc2Ωt−rΩcotΩt+r22)ˆx2(t)
−mΩert2(ΩcscΩtcotΩt−r2cscΩt)ˆx(t)ˆx(0)+12mΩ2csc2Ωtˆx2(0)
−iћ2(ΩcotΩt−r2)
. (71)
Applying Eqs.(58)-(60), the propagator takes the form
K(x,x′;t)=C(x,x′)exp[−iћt∫0{12mert(Ω2csc2Ωt−rΩcotΩt+r22)
+12mΩ2csc2Ωtx′2−mΩert2(ΩcscΩtcotΩt−r2cscΩt)xx′
−iћ2(ΩcotΩt−r2)}dt].
(72)
Now, we will integrate each term of Eq.(72) with respect to time. The first term of Eq.(72) can be integrated as
−im2ћx2t∫0ert(Ω2csc2Ωt−rΩcotΩt+r22)dt=imΩ2ћertcotΩtx2−imr4ћertx2. (73)
The second term of Eq.(72) can be calculated by
−imΩ22ћx′2t∫0csc2Ωtdt=imΩ2ћcotΩtx′2. (74)
The third term of Eq.(72) can be derived by
imΩћxx′t∫0ert2(ΩcscΩtcotΩt−r2cscΩt)dt=−imΩћert2cscΩtxx′. (75)
Finally, integrating the last term of Eq.(72), the result is
−t∫0(Ω2cotΩt−r4)dt=−12lnln(sinΩt)+rt4. (76)
Combining the results of Eqs.(73)-(76), the propagator can be written as
K(x,x′;t)=C(x,x′)√ert2sinΩtexp(−imr4ћertx2)
×exp[imΩ2ћsinΩt(ertcosΩtx2+cosΩtx′2−2ert2xx′)]
. (77)
The final step is deriving the function C(x,x′). Substituting Eq.(77) into Eq.(62), its can be obtained that
iћ∂C(x,x′)∂x′=−mr2x′C(x,x′). (78)
The solution of Eq.(78) can be written as
C(x,x′)=C(x)expexp(imr4ћx′2), (79)
Where C(x) is a position function
The propagator in Eq.(77) can be expressed as
K(x,x′;t)=C(x)√ert2sinΩtexp(−imr4ћ(ertx2−x′2))
×expexp[imΩ2ћsinΩt(ertcosΩtx2+cosΩtx′2−2ert2xx′)].
(80)
The next step is calculating C(x). Substituting Eq.(80) into Eq.(61), the result is
∂C(x)∂x=0, (81)
which implies that C(x). is a constant independent of x .
After applying Eq.(63), it can be obtained that
C=√mΩ2πiћ. (82)
So, the propagator for a damped harmonic oscillator can be written as
K(x,x′;t)=√mΩert22πiћsinΩtexp(−imr4ћ(ertx2−x′2))
×exp[imΩ2ћsinΩt(ertcosΩtx2+cosΩtx′2−2ert2xx′)]
. (83)
This propagator is the same as the result of S.Pepore,8 found by applying the integrals of motion of a quantum systems.
In this paper we have successfully calculated the exact propagators for time-dependent Hamiltonian systems. The method for deriving the propagators with the helping of integrals of motion of quantum systems presented in this paper can be successfully applied in solving a time-dependent linear potential and a free particle with linear damping problems. This method has the important steps in finding the constant of motion x0 and p0 and implying that the propagator K(x,x′,t) is the eigen functions of the operators ˆx0(x) and ˆp0(x) . The exact propagator for a charged harmonic oscillator in time-dependent electric field was calculated by the Feynman path integral method. The crucial result in our calculation is to derive the classical action as mentioned in E.(49). The propagator for a damped harmonic oscillator has calculated by the Schwinger method. The important step in the Schwinger formalism is to find the solution of the Heisenberg equation in Eq.(67) and to express the Hamiltonian operator in an appropriate order with the aid of the commutator in Eq.(70). The advantage of the Schwinger method in this paper is that it requires only fundamental operator algebra and some basic integration. In fact, the application of the integrals of the motion method has many common features with the Schwinger method, but the Schwinger method requires the operators ˆx(t) and ˆp(t) in deriving the matrix element of Hamiltonian operator in calculating the propagator in Eq.(72). In the Feynman path integrals, the pre-exponential function C(t) comes from sum over all fluctuating paths that depend on calculation of the functional integration while in the integrals of the motion method this term appears from solving the Schrodinger equation of propagator. In the Schwinger formalism, the pre-exponential function C(t) arises from the commutation relation of [ˆx(t),ˆx(0)]. These different points of view may show the connection between classical mechanics and quantum mechanics.
Finally, we have presented simple techniques in calculating the propagator. It is preferable to have many methods in deriving the propagators in the field of time-dependent Hamiltonian systems and the Feynman path integrals, Schwinger method, and integrals of the motion method are effective and appropriate techniques.
None.
The author declares there is no conflict of interest.
©2021 Pepore, et al. This is an open access article distributed under the terms of the, which permits unrestricted use, distribution, and build upon your work non-commercially.