This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

Fast electromagnetic response of a thin film of resonant atoms with permanent dipole

J.-G. Caputo1 E. Kazantseva1,2 and A. Maimistov3
(July 27, 2025)
Abstract

We consider the propagation of extremely short pulses through a dielectric thin film containing resonant atoms (two level atoms) with permanent dipole. Assuming that the film width is less than the field wave length, we can solve the wave equation and reduce the problem to a system of generalized Bloch equations describing the resonant atoms. We compute the stationary solutions for a constant irradiation of the film. Superimposing a small amplitude linear wave we compute the reflection and transmission coefficients. From these, one can then deduce the different parameters of the model. We believe this technique could be used in experiments to obtain the medium atomic and relaxation parameters.

1:Laboratoire de Mathématiques, INSA de Rouen,
B.P. 8, 76131 Mont-Saint-Aignan cedex, France
E-mail: caputo@insa-rouen.fr
2 Center for Engineering Science Advanced Research,
Computer Science and Mathematics Division,
Oak Ridge National Laboratory,
Oak Ridge, TN 37831, USA
E-mail: kazantsevaev@ornl.gov
3: Department of Solid State Physics and Nanosystems,
Moscow Engineering Physics Institute,
Kashirskoe sh. 31, Moscow 115409, Russia
E-mail: amaimistov@gmail.com

PACS numbers : 42.25.Bs, 42.65.Pc, 42.81.Dp

Keywords: thin film, two level atoms, permanent dipole, refractive phenomena

1 Introduction

One of the famous models describing field-matter interaction is the Maxwell-Bloch system [1]. It corresponds to an ensemble of two-level atoms whose states alternate due to the electromagnetic field. In a simple approach the electromagnetic field is assumed to be scalar and the operator of dipole transition to only have non-diagonal matrix elements. This model was the base for the description of many coherent nonlinear effects, the coherent pulse propagation (self-induced transparency [2]) and the coherent transient effects (optical nutations, free induction decay, quantum beats, superradiance, photon echo). A detailed review of these phenomena may be found in [3, 4] and in the book [1].

The Maxwell-Bloch system can be extended further than the model of two-level atoms discussed above. In particular one can generalize the resonant atomic model. The model of three-level atom now attracts great interest, because it describes electromagnetic induced transparency, slow light propagation in (three-level) atomic vapor [5] and the coherent population transfer [6]. The coherent interaction of electromagnetic pulses with quantum dots [7, 8] can be described by such a generalized Maxwell-Bloch system. Another generalization of the two-level model is to take into account diagonal matrix elements of the dipole transition operator [12]. In this medium steady state one-half cycle pulses were obtained in the sharp line limit [13]. A new kind of steady state pulse was found, characterized by an algebraic decay of the electric field.

The reduced Maxwell-Bloch equation in the sharp line limit have a zero-curvature representation[14, 15]. A number of results related with the complete integrability of this reduced system have been obtained in [16, 17, 18]. The numerical simulation of the propagation of extremely short pulses [19] shows the existence of an extraordinary breather with non zero pulse square. An analytical expression for this breather has been found in [14, 15] and reproduced in [20] using the Darboux transformation method. Recently Zabolotzki[21] developed the inverse scattering method to find the solution for the isotropic limit of the general model of a two-component electromagnetic field interacting with two-level atoms with a permanent dipole moment without invoking the slowly varying envelope approximation.

The influence of the permanent dipole on parametric processes was studied in [9, 10]. Recently Weifeng et al studied the generation of attosecond pulses in a two-level system with permanent dipole moment[11]. For this, higher harmonics are generated and the spectrum can be extended to the X-ray range. The quantum interference of both even and odd harmonics results in the generation of higher intensity attosecond pulses.

The propagation of ultrashort pulses propagation through a thin film containing resonant atoms located at the interface between two dielectrics is also described by the Maxwell-Bloch equations. However, if the width of the film is less than a wave length [22], the atomic system is compressed into a ”point”. An insightful comment was made in [23, 24], concerning the critical role of the local Lorentz field. This local field induces a nonlinearity so that the thin film of two-level atoms acts as a nonlinear Fabry-Perot resonator. One then expects optical bistability for this device. There are many generalizations of the thin film model, where three (or more) level atoms, two-photon transition between resonant levels and non-resonant nonlinearity were studied. Here we consider a thin film corresponding to two-level atoms with permanent dipole moment [25]. There the author analyzed numerically the pulse propagation through the film taking into account the local field. He showed that a dense film irradiated by a one-circle pulse emits a short response with a delay much longer than the characteristic cooperative time of the atom ensemble. Contrary to [25] we assume that the atoms of the film are rare so that the local field can be neglected.

Recently [28] we introduced a general formalism to describe the interaction of a (linear polarized) electromagnetic pulse with a medium. With this we studied a ferroelectric film which was described by a Duffing oscillator naturally giving a double well potential. Here we generalize this approach to the case of a layer of resonant atoms described by the generalized Bloch equations taking into account the permanent dipole moment. An important feature of the model is that there is a clear separation between the medium and the surrounding vacuum so that no dispersion relation can be written. Instead we obtain a scattering problem where the reflection and transmission coefficients need to be calculated.
After presenting the model in section 2, we compute its equilibria and their stability in section 3. In section 4 we assume an additional periodic modulation around a fixed background field. This enables to do a spectroscopy study of the film so that both the dipole parameter and the coupling parameter can be extracted from the reflection curve. Section 5 concludes the article.

2 The model

2.1 One dimensional wave propagation

Refer to caption

Figure 1: Schematic of a incoming electromagnetic pulse e0e_{0} on a thin dielectric film containing resonant atoms.

Following the formalism of [28] we assume that an electromagnetic wave is incident from the left x<0x<0 on a medium whose position is given by the function I(x)I(x). The configuration is shown in Fig. 1. Denoting by subscripts the partial derivatives, the equations are

ettexx\displaystyle e_{tt}-e_{xx} =\displaystyle= g(x,t),\displaystyle g(x,t), (1)
g(x,t)\displaystyle g(x,t) =\displaystyle= γpttI(x),\displaystyle-\gamma p_{tt}I(x), (2)

where pp is the polarization in the medium. Let the dielectric susceptibilities of surrounding mediums be the same. That eliminates the Fresnel refraction. The boundary conditions are

e(t,x=±)=0,et(t,x=±)=0,e(t,x=\pm\infty)=0,~~e_{t}(t,x=\pm\infty)=0, (3)

and the initial conditions are e=e0(xt)e=e_{0}(x-t) following the scattering problem. Then we have

e(t=0,x)=e0(x),et(t=0,x)=xe0(x).e(t=0,x)=e_{0}(x),~~e_{t}(t=0,x)=-\partial_{x}e_{0}(x). (4)

We assume that the initial pulse is located at the left far from the film.

Using the general procedure for solving the wave equation (see [27]), we showed in [28] that the solution of this problem is

e(x,t)=e0(xt)+12+0tg(y,τ)[θ(xyt+τ)θ(xy+tτ)]𝑑τ𝑑y,e(x,t)=e_{0}(x-t)+\frac{1}{2}\int\limits_{-\infty}^{+\infty}\int\limits_{0}^{t}g(y,\tau)\left[\theta(x-y-t+\tau)-\theta(x-y+t-\tau)\right]d\tau dy, (5)

where we use the step-function

θ(z)=zδ(x)𝑑x,\theta(z)=\int\limits_{-\infty}^{z}\delta(x)dx,

which can also be written as

θ(z)={1z>0,1/2z=0,0z<0,\theta(z)=\left\{\begin{array}[]{lr}1&z>0,\\ 1/2&z=0,\\ 0&z<0,\end{array}\right.

When the medium is reduced to a single thin film placed at x=ax=a,

g(x,t)=γpttI(x),whereI(x)=δ(xa).g(x,t)=-\gamma p_{tt}I(x),~~~{\rm where}~~~I(x)=\delta(x-a).

The field at x=ax=a is given by

e(a,t)\displaystyle e(a,t) =\displaystyle= e0(at)+γ20tpt(a,τ)[δ(τt)δ(τt)]𝑑τ\displaystyle e_{0}(a-t)+\frac{\gamma}{2}\int\limits_{0}^{t}p_{t}(a,\tau)\left[\delta(\tau-t)-\delta(\tau-t)\right]d\tau
=\displaystyle= e0(at)γpt(a,t).\displaystyle e_{0}(a-t)-\gamma p_{t}(a,t).

With this general formalism, one can address the question of what happens for a medium represented by an ensemble of resonant atoms. These could be molecules, quantum dots, or two or three level atoms. Here we will restrict ourselves to the two-level atoms with permanent dipole embedded into a thin film.

2.2 The case of a thin film of resonant atoms

Let the plane electromagnetic wave interact with the atoms or molecules characterized by the operator of the dipole transition between resonant energy levels and let this operator have both non-diagonal and diagonal matrix elements [12]. In the two-level approximation the Hamiltonian of the considered model can be written as[15]

H^=ω02(1001)(d11Ed12Ed21Ed22E),\hat{H}=\frac{\hbar\omega_{0}}{2}\left(\begin{array}[]{cc}-1&0\\ 0&1\end{array}\right)-\left(\begin{array}[]{cc}d_{11}E&d_{12}E\\ d_{21}E&d_{22}E\end{array}\right),

where EE is the amplitude of the electric field of the electromagnetic wave and ω0\omega_{0} is the frequency difference between the levels. The polarization of the medium is

P(x,y,t)=lδ(xa)nAp(t),P(x,y,t)=l\delta(x-a)n_{A}p(t), (6)

where ll is the film thickness and where nAn_{A} is the volume density of the atoms. We assumed a homogeneous film and neglected the transverse yy dependence of PP. The atomic polarizability pp is given by the expression

p=trρ^d^=ρ11d11+ρ22d22+ρ12d21+ρ21d12p={\rm tr}\hat{\rho}\hat{d}=\rho_{11}d_{11}+\rho_{22}d_{22}+\rho_{12}d_{21}+\rho_{21}d_{12} (7)
=12(d11+d22)+12(d11d22)(ρ11ρ22)+ρ12d21+ρ21d12,=\frac{1}{2}\left(d_{11}+d_{22}\right)+\frac{1}{2}\left(d_{11}-d_{22}\right)(\rho_{11}-\rho_{22})+\rho_{12}d_{21}+\rho_{21}d_{12},

where ρ^\hat{\rho} is the density matrix and where we used the constraint ρ11+ρ22=1\rho_{11}+\rho_{22}=1. In the expression above, the first term corresponds to the constant polarizability of the molecules. The average of this quantity over all atoms must be zero because we assume no polarization of the medium in the absence of an electromagnetic field. Note that all relaxation processes are neglected because we assume short electromagnetic pulses. We also neglect the dipole-dipole interaction, ie we assume the atoms carrying dipoles are rare in the film. The evolution of the elements of ρ^\hat{\rho} is given by the Heisenberg equation iρ^/t=H^ρ^ρ^H^i\hbar\partial\hat{\rho}/\partial t=\hat{H}\hat{\rho}-\hat{\rho}\hat{H} and yields the Bloch equations.

So the total system of equations describing the interaction of an electromagnetic wave with a collection of identical atoms in a film placed at x=ax=a is [12]

2Ex21c22Et2=4πnAc2lδ(xa)2t2(12(d22d11)r3+d12r1),\frac{\partial^{2}E}{\partial x^{2}}-\frac{1}{c^{2}}\frac{\partial^{2}E}{\partial t^{2}}=\frac{4\pi n_{A}}{c^{2}}l\delta(x-a)\frac{\partial^{2}}{\partial t^{2}}\left(\frac{1}{2}\left(d_{22}-d_{11}\right)r_{3}+d_{12}r_{1}\right), (8)
r1t=[ω0+(d11d22)E/]r2,\frac{\partial r_{1}}{\partial t}=-\left[\omega_{0}+\left(d_{11}-d_{22}\right)E/\hbar\right]r_{2}, (9)
r2t=[ω0+(d11d22)E/]r1+2(d12E/)r3,\frac{\partial r_{2}}{\partial t}=\left[\omega_{0}+\left(d_{11}-d_{22}\right)E/\hbar\right]r_{1}+2(d_{12}E/\hbar)r_{3}, (10)
r3t=2(d12E/)r2,\frac{\partial r_{3}}{\partial t}=-2(d_{12}E/\hbar)r_{2}, (11)

where r1,2,3r_{1,2,3} are the components of the Bloch vector

r1=ρ12+ρ21,r2=i(ρ12ρ21),r3=ρ22ρ11.r_{1}=\rho_{12}+\rho_{21},\quad r_{2}=-i(\rho_{12}-\rho_{21}),\quad r_{3}=\rho_{22}-\rho_{11}. (12)

In equations (9)-(11) the field EE is taken in the film, i.e., at x=ax=a. This system differs from the well-known Maxwell-Bloch equations [29, 30] by the terms containing the parameter (d11d22)\left(d_{11}-d_{22}\right).

We normalize time, space and electric field as

τ=ω0t,ζ=ω0x/c,e=2d12E/ω0.\tau=\omega_{0}t,\quad\zeta=\omega_{0}x/c,\quad e=2d_{12}E/\hbar\omega_{0}.

We also introduce the parameter μ\mu which measures the strength of the permanent dipole:

μ=d11d222d12.\mu={d_{11}-d_{22}\over 2d_{12}}. (13)

Then the Maxwell-Bloch equations (8)-(11) take the form:

r1,τ=(1+μe)r2,r2,τ=(1+μe)r1+er3,r3,τ=er2,r_{1,\tau}=-(1+\mu e)r_{2},\quad r_{2,\tau}=(1+\mu e)r_{1}+er_{3},\quad r_{3,\tau}=-er_{2}, (14)
e,ζζe,ττ=α(r1μr3),ττδ(ζζ0),e_{,\zeta\zeta}-e_{,\tau\tau}=\alpha(r_{1}-\mu r_{3})_{,\tau\tau}\delta(\zeta-\zeta_{0}), (15)

where ζ0=ω0a/c\zeta_{0}=\omega_{0}a/c, and

α=8πnAlωcd122ω0.\alpha=8\pi n_{A}l{\omega\over c}{d_{12}^{2}\over\hbar\omega_{0}}. (16)

A rough estimation of this parameter taking d12d_{12}\approx 1 Debye =3.3361030=3.336~10^{-30} gives α=nAl0.00795\alpha=n_{A}l0.00795 so that for a film thickness of l=1μml=1\mu m and a volume density 107<nA<10910^{7}<n_{A}<10^{9} we get 0.1<α<100.1<\alpha<10. If the film is thinner, l=1nml=1nm we can take the same range of α\alpha with densities 1010<nA<101210^{10}<n_{A}<10^{12}. This is still compatible with our hypothesis of dilute impurities in the medium so that we can neglect the dipole-dipole interaction.

The system can be further simplified by noting from equation (14) that (r1μr3)τ=r2(r_{1}-\mu r_{3})_{\tau}=-r_{2}. We then get

e,ττe,ζζ=αr2,τδ(ζζ0),e_{,\tau\tau}-e_{,\zeta\zeta}=\alpha r_{2,\tau}\delta(\zeta-\zeta_{0}), (17)
r1,τ=(1+μe)r2,r2,τ=(1+μe)r1+er3,r3,τ=er2.r_{1,\tau}=-(1+\mu e)r_{2},\quad r_{2,\tau}=(1+\mu e)r_{1}+er_{3},\quad r_{3,\tau}=-er_{2}. (18)

This is the model that we will analyze in detail in this article.

We will assume the general initial conditions where the medium is initially at rest so that e=0,e=0, e,ζ=e,τ=0,r1=r2=0,e_{,\zeta}=e_{,\tau}=0,r_{1}=r_{2}=0, r3=1r_{3}=-1, at τ\tau\longrightarrow-\infty. From the Bloch equations (14) we obtain (r12+r22+r32),τ=0.\left(r_{1}^{2}+r_{2}^{2}+r_{3}^{2}\right)_{,\tau}=0. and using the initial conditions we get the value of this integral of motion

r12+r22+r32=1.r_{1}^{2}+r_{2}^{2}+r_{3}^{2}=1. (19)

We consider the effect of an electromagnetic pulse impinging on the thin film placed at ζ0=0\zeta_{0}=0. For that we can use the general result (5) to solve the wave equation (17). Taking the right part of (17) as a function g(y,τ)g(y,\tau) under the integral in (5) we obtain the following expression

e(0,τ)=e0(τ)+α0τr2(τ)δ(ττ)𝑑τ=e0(τ)+α2r2(τ).e(0,\tau)=e_{0}(-\tau)+\alpha\int\limits_{0}^{\tau}r_{2}(\tau^{\prime})\delta(\tau-\tau^{\prime})d\tau^{\prime}=e_{0}(-\tau)+\frac{\alpha}{2}r_{2}(\tau). (20)

which represents the strength of the electrical field in the film. Now, using this expression we can write the modified Bloch equations as

r1.τ\displaystyle r_{1.\tau} =\displaystyle= (1+μe0)r2(αμ/2)r22,\displaystyle-(1+\mu e_{0})r_{2}-(\alpha\mu/2)r_{2}^{2}, (21)
r2,τ\displaystyle r_{2,\tau} =\displaystyle= (1+μe0)r1+e0r3+(αμ/2)r1r2+(α/2)r2r3,\displaystyle(1+\mu e_{0})r_{1}+e_{0}r_{3}+(\alpha\mu/2)r_{1}r_{2}+(\alpha/2)r_{2}r_{3}, (22)
r3,τ\displaystyle r_{3,\tau} =\displaystyle= e0r2(α/2)r22.\displaystyle-e_{0}r_{2}-(\alpha/2)r_{2}^{2}. (23)

These are the correct equations describing the resonant responses of two-level atoms of a thin film to an ultra-short electromagnetic pulse. It is worth noting that the electromagnetic wave is incident normally on the film. Second, all atoms of the film are identical. Finally note that we did not assume any limitation on the time duration of the electromagnetic pulse. It may be a half period pulse, i.e. an electromagnetic spike or a quasiharmonic wave.

Finally note that since the motion occurs on the Bloch sphere, the system (23) has the constraint r12+r22+r32=1r_{1}^{2}+r_{2}^{2}+r_{3}^{2}=1. It is then natural to write it in the reduced coordinates (m,ϕ)(m,\phi) such that

r1=1m2cosϕ,r2=1m2sinϕ,r3=m.r_{1}=\sqrt{1-m^{2}}\cos\phi,~~~~r_{2}=\sqrt{1-m^{2}}\sin\phi,r_{3}=m. (24)

The system is then

mτ=e01m2sinϕα2(1m2)sin2ϕ,\displaystyle m_{\tau}=-e_{0}\sqrt{1-m^{2}}\sin\phi-{\alpha\over 2}(1-m^{2})\sin^{2}\phi, (25)
ϕτ=(μ+m1m2cosϕ)(e0+α21m2sinϕ)+1.\displaystyle\phi_{\tau}=(\mu+{m\over\sqrt{1-m^{2}}}\cos\phi)(e_{0}+{\alpha\over 2}\sqrt{1-m^{2}}\sin\phi)+1. (26)

3 Equilibrium states

We now study the stationary points of the system (23). Initially before the wave reaches it, the film is at rest. The electromagnetic field e0e_{0} shifts the film state to a new equilibrium. The system then relaxes back to its original state after the wave has passed. We will examine these new transient states and their stability.

For e0e_{0} constant, the system of equations (21-23) has the fixed point (r1,0,r3)(r_{1}^{\ast},0,r_{3}^{\ast}) where r1,3r_{1,3}^{\ast} satisfy

(1+μe0)r1+e0r3=0.(1+\mu e_{0})r_{1}^{\ast}+e_{0}r_{3}^{\ast}=0. (27)

Assuming r20r_{2}^{\ast}\neq 0 leads to a contradiction. From (19) it follows that

r12+r32=1r_{1}^{\ast 2}+r_{3}^{\ast 2}=1 (28)

Combining these two equations we obtain

r3=±[1+e02(1+μe0)2]1/2,r1=e01+μe0r3.r_{3}^{\ast}=\pm[1+{e_{0}^{2}\over(1+\mu e_{0})^{2}}]^{-1/2},~~r_{1}^{\ast}=-{e_{0}\over 1+\mu e_{0}}r_{3}^{\ast}. (29)

This fixed point corresponds to a stationary polarization and population induced in the two-level atoms by the incident constant field e0e_{0}. When e0e_{0}\rightarrow\infty

r3±μ1+μ2,r1r3μ.r_{3}^{\ast}\rightarrow\pm{\mu\over\sqrt{1+\mu^{2}}},~~~r_{1}^{\ast}\rightarrow-{r_{3}^{\ast}\over\mu}. (30)

To understand geometrically the position of the fixed points, we can parametrize the Bloch sphere with r2=0r_{2}^{\ast}=0 by

r1=sinθ,r3=cosθ,r_{1}^{\ast}=\sin\theta,~~r_{3}^{\ast}=\cos\theta,

yielding the relation

tanθ=e01+μe0.\tan\theta=-{e_{0}\over 1+\mu e_{0}}. (31)

The dependence of the values r1,3r_{1,3}^{\ast}  on the electric field e0e_{0} is monotonic so there is no bistability. The two fixed points are shown in Fig. 2 for μ=±0.1,±1\mu=\pm 0.1,~\pm 1. For μ\mu small (left panel) the fixed points start from (0,±1)(0,\pm 1) and rotate following (sinθ,cosθ)(\sin\theta,\cos\theta) with θ\theta given by (31). For μ=1\mu=-1 and e0=1e_{0}=1 θ=π/2+nπ\theta=\pi/2+n\pi so we obtain (±1,0)(\pm 1,0). Since r3r_{3} is the difference in the population of the levels, these are equally populated. When μ\mu is large and positive (top right panel) the fixed points do not change very much when the electric field is increased. On the contrary when μ\mu is large and negative (bottom right panel) increasing the electric field shifts the fixed points from (0,±1)(0,\pm 1) to (±1,0)(\pm 1,0) for e0=1e_{0}=1 and about (±1/2,±1/2)(\pm 1/\sqrt{2},\pm 1/\sqrt{2}) for e0=5e_{0}=5. For this last value of the parameter we can have a strong population inversion.

To study the stability of the fixed point we linearize the modified Bloch equations. We set r1=r1+δr1r_{1}=r_{1}^{\ast}+\delta r_{1}, r2=0+δr2r_{2}=0+\delta r_{2}, r3=r3+δr3r_{3}=r_{3}^{\ast}+\delta r_{3}. The linearized modified Bloch equations read

δr1τ\displaystyle{\delta r_{1}}_{\tau} =\displaystyle= (1+μe0)δr2,\displaystyle-(1+\mu e_{0})\delta r_{2},
δr2τ\displaystyle{\delta r_{2}}_{\tau} =\displaystyle= (1+μe0)δr1+e0δr3+(αμ/2)r1δr2+(α/2)r3δr2,\displaystyle(1+\mu e_{0})\delta r_{1}+e_{0}\delta r_{3}+(\alpha\mu/2)r_{1}^{\ast}\delta r_{2}+(\alpha/2)r_{3}^{\ast}\delta r_{2}, (32)
δr3τ\displaystyle{\delta r_{3}}_{\tau} =\displaystyle= e0δr2,\displaystyle-e_{0}\delta r_{2},

where as usual the τ\tau subscript indicates the derivative. We introduce the field modified frequency

Ω2=(1+μe0)2+e02,\Omega^{2}=(1+\mu e_{0})^{2}+e_{0}^{2}, (33)

and ratio

b=α4(μr1+r3)=α4r31+μe0=±α4Ω.b={\alpha\over 4}(\mu r_{1}^{\ast}+r_{3}^{\ast})={\alpha\over 4}{r_{3}^{\ast}\over 1+\mu e_{0}}=\pm{\alpha\over 4\Omega}. (34)

From (32) we can get the equation for δr2\delta r_{2}

δr2ττ2bδr2,τ+Ω2δr2=0.{\delta r_{2}}_{\tau\tau}-2b{\delta r_{2}}_{,\tau}+\Omega^{2}\delta r_{2}=0. (35)

The characteristic equation is

λ22bλ+Ω2=0,\lambda^{2}-2b\lambda+\Omega^{2}=0,

whose roots are

λ1,2=b±iΩ2b2=±α4Ω±iΩ1α216Ω4\lambda_{1,2}=b\pm i\sqrt{\Omega^{2}-b^{2}}=\pm{\alpha\over 4\Omega}\pm i\Omega\sqrt{1-{\alpha^{2}\over 16\Omega^{4}}}

Then if b<0b<0 and Ω2>b2\Omega^{2}>b^{2} the fixed point is stable. Depending on α\alpha two cases occur. Consider first a small α\alpha so that Ω2>b2\Omega^{2}>b^{2}. Then if μ>0\mu>0, b>0b>0 if r3>0r_{3}^{\ast}>0. Then the fixed point such that r3>0r_{3}^{\ast}>0 is unstable. Conversely the fixed point such that r3<0r_{3}^{\ast}<0 is stable. If 1+μe0<01+\mu e_{0}<0 which occurs for μ<0\mu<0 and large fields e0e_{0} the situation is reversed. The fixed point such that r3>0r_{3}^{\ast}>0 is stable while the one such that r3<0r_{3}^{\ast}<0 is unstable. The oscillation frequency of the orbit as it approaches the fixed point is given by the imaginary part of λ\lambda

ω=Ω1α216Ω4\omega^{\ast}=\Omega\sqrt{1-{\alpha^{2}\over 16\Omega^{4}}} (36)

.

When Ω2>b2\Omega^{2}>b^{2} corresponding to a large α\alpha, there is no imaginary part of λ\lambda. Even for very large α\alpha the stability remains unchanged because for the first term b=α4Ωb={\alpha\over 4\Omega} will always dominate the second one Ωα216Ω41\Omega\sqrt{{\alpha^{2}\over 16\Omega^{4}}-1}. Therefore the fixed point such that r3<0r_{3}^{\ast}<0 is stable (resp. unstable) for μ>0\mu>0 (resp. 1+μe0<01+\mu e_{0}<0). In this case there are no oscillations around the fixed point.

Refer to caption Refer to caption

Refer to caption Refer to caption

Figure 2: Position of the fixed points (r1,0,r3)(r_{1}^{\ast},0,r_{3}^{\ast}) on the Bloch sphere for four different values of the dipole parameter μ\mu, top from left to right μ=0.1,1\mu=0.1,~1 and bottom from left to right μ=0.1,1\mu=-0.1,~-1. The shades correspond to different amplitudes of the electric field e0e_{0}, empty is for e0=0.1e_{0}=0.1 , half-empty is for e0=1e_{0}=1 and full is for e0=5e_{0}=5. The stable fixed points (see text) are indicated by the letter s.

To conclude, for all values of α\alpha, the stability is shown in Fig. 3 in the parameter space (μ,e0)(\mu,e_{0}).

Refer to caption

Figure 3: Stability diagram in the (μ,e0)(\mu,e_{0}) plane indicating which fixed point is stable.

3.1 Numerical results

To illustrate the previous analysis, we have solved numerically the system of ordinary differential equations (21-23). We used the Runge-Kutta 4-5 Dopri5[34] as solver. For all the runs presented, we choose an incoming pulse

e0(t)=e0012[tanh(tt1w)tanh(tt2w)],e_{0}(-t)=e_{00}{1\over 2}[\tanh({t-t_{1}\over w})-\tanh({t-t_{2}\over w})], (37)

where t1=64t_{1}=-64, t2=5t_{2}=-5 and w=0.2w=0.2. In the following we will abusively name the amplitude e00e_{00} e0e_{0}. We first consider μ=1\mu=1. Fig. 4 shows the evolution of the Bloch vector components as a function of time, with from left to right e0=0.1,1e_{0}=0.1,1 and 2. As expected the system reaches the stable equilibrium state given by (29) with r3<0r_{3}^{*}<0. The oscillations are given by the frequency ω1.1,2.2\omega^{*}\approx 1.1,2.2 and 3.6 from left to right. These plots correspond to the upper right panel of Fig. 2.

Refer to caption Refer to caption Refer to caption

Figure 4: Time evolution of the components of the Bloch vector (r1,r2,r3)(r_{1},r_{2},r_{3}) respectively in continuous line, long dash and short dash (red, green and black online). From the left panel to the right panel the amplitude of the field is increased from e0=0.1e_{0}=0.1 (left), e0=1e_{0}=1 (middle) to e0=2e_{0}=2 (right). The dipole parameter is μ=1\mu=1 and the coupling coefficient is α=1\alpha=1.

The case μ=1\mu=-1 is shown in Fig. 5. When e0e_{0} is small, we have an equilibrium very close to the one for μ=1\mu=1 as expected from the bottom right panel of Fig. 2. When e0=1e_{0}=1 we obtain the situation where r3=0r_{3}^{*}=0 and r1=±1r_{1}^{*}=\pm 1. Note the typical frequency ω=1\omega^{*}=1 as in the left panel. For a larger field e0=2e_{0}=2, shown in the right panel we obtain as expected an equilibrium r3>0r_{3}^{*}>0. Note the relatively small oscillation frequency ω1.7\omega^{*}\approx 1.7 compared for the one for the same e0e_{0} and μ=1\mu=1 (right panel of Fig. 4).

Refer to caption Refer to caption Refer to caption

Figure 5: Same as for Fig. 4 except that the dipole parameter is μ=1\mu=-1.

When α\alpha is large, the fixed point does not change but the eigenvalue of the Jacobian is now real. We then get no oscillations as the system reaches the equilibrium as shown in the right panel of Fig. 6 for which e0=0.4,μ=1e_{0}=0.4,\mu=1 and α=10\alpha=10. Reducing α\alpha to 1 restores the oscillations of frequency ω\omega^{*} as shown in the left panel of Fig. 6.

Refer to caption Refer to caption

Figure 6: Time evolution of the components of the Bloch vector (r1,r2,r3)(r_{1},r_{2},r_{3}) respectively in continuous line, long dash and short dash (red, green and black online) for two values of the coupling parameter α=1\alpha=1 for the left plot and α=10\alpha=10 for the right plot. The amplitude of the electric field is e0=0.4e_{0}=0.4. All the other parameters are as in Fig. 4.

4 Spectroscopic analysis of the film

We now assume a constant field e0e_{0} applied to the film together with a small harmonic wave. It is then possible to examine how these waves of different frequencies get scattered by the film. Using this spectroscopic analysis of the film, we will see that one can recover the atomic dipolar parameter μ\mu and the coupling parameter α\alpha.

For a constant e0e_{0}, the system (14,17) becomes

δeττδeζζ\displaystyle\delta e_{\tau\tau}-\delta e_{\zeta\zeta} =\displaystyle= αδD(x)δr2,τ,\displaystyle\alpha\delta_{D}\left(x\right)\delta r_{2,\tau}, (38)
δr1,τ\displaystyle\delta r_{1,\tau} =\displaystyle=- (1+μe0)δr2,\displaystyle(1+\mu e_{0})\delta r_{2}, (39)
δr2,τ\displaystyle\delta r_{2,\tau} =\displaystyle= (1+μe0)δr1+e0δr3+(μr1+r3)δe,\displaystyle(1+\mu e_{0})\delta r_{1}+e_{0}\delta r_{3}+(\mu r_{1}^{\ast}+r_{3}^{\ast})\delta e, (40)
δr3,τ\displaystyle\delta r_{3,\tau} =\displaystyle= e0δr2.\displaystyle-e_{0}\delta r_{2}. (41)

where we temporarily use δD\delta_{D} to define the Dirac delta function.

We separate time and space

δe=E(ζ)eiωτ,δr1,2,3=R1,2,3eiωτ,\delta e=E(\zeta)e^{-i\omega\tau},~~~~\delta r_{1,2,3}=R_{1,2,3}e^{-i\omega\tau}, (42)

and obtain the system

Eζζ+ω2E\displaystyle E_{\zeta\zeta}+\omega^{2}E =\displaystyle= iωαδ(ζ)R2,\displaystyle i\omega\alpha\delta\left(\zeta\right)R_{2}, (43)
iωR1\displaystyle i\omega R_{1} =\displaystyle= (1+μe0)R2,\displaystyle(1+\mu e_{0})R_{2}, (44)
iωR2\displaystyle-i\omega R_{2} =\displaystyle= (1+μe0)R1+e0R3+(μr1+r3)E,\displaystyle(1+\mu e_{0})R_{1}+e_{0}R_{3}+(\mu r_{1}^{\ast}+r_{3}^{\ast})E, (45)
iωR3\displaystyle i\omega R_{3} =\displaystyle= e0R2.\displaystyle e_{0}R_{2}. (46)

From the second and fourth equations we get

R1=i(1+μe0)ωR2,R3=ie0ωR2.R_{1}=-i\frac{(1+\mu e_{0})}{\omega}R_{2},~~~~\ R_{3}=-i\frac{e_{0}}{\omega}R_{2}. (47)

Plugging these expression into (43) results in

R2=i(μr1+r3)ωEω2Ω2,R_{2}=i\frac{(\mu r_{1}^{\ast}+r_{3}^{\ast})\omega E}{\omega^{2}-\Omega^{2}}, (48)

where Ω\Omega is given by (33). The substitution of the previous equation into (38) leads to the non standard eigenvalue problem

Exx+ω2[1+α(μr1+r3)δ(x)ω2Ω2]E=0.\displaystyle E_{xx}+\omega^{2}\left[1+\frac{\alpha(\mu r_{1}^{\ast}+r_{3}^{\ast})\delta(x)}{\omega^{2}-\Omega^{2}}\right]E=0. (49)

As usual we assume a scattering experiment so that the field is given by

E\displaystyle E =\displaystyle= {Aeiωx+Beiωx,x<0,Ceiωx,x>0.\displaystyle\left\{\begin{array}[]{l l}Ae^{i\omega x}+Be^{-i\omega x},&x<0,\\ Ce^{i\omega x},&x>0.\end{array}\right. (52)

At the film x=0x=0 the field is continuous and its gradient satisfies the jump condition

{E(0+)=E(0)=E(0)Ex(0+)Ex(0)+βE(0)=0.\displaystyle\left\{\begin{array}[]{l l}E(0+)=E(0-)=E(0)\\ E_{x}(0+)-E_{x}(0-)+\beta E(0)=0.\end{array}\right. (55)

where

β=α(μr1+r3)ω2ω2Ω2\beta=\frac{\alpha(\mu r_{1}^{\ast}+r_{3}^{\ast})\omega^{2}}{\omega^{2}-\Omega^{2}}

Writing the two conditions (55) using the left and right fields results in

{A+B=Ciω(CA+B)+βC=0.\left\{\begin{array}[]{l l}A+B=C\\ i\omega(C-A+B)+\beta C=0.\end{array}\right. (56)

The solution is

B=iβ2ω11+iβ2ωA,C=11+iβ2ωAB=\frac{i\beta}{2\omega}\frac{1}{1+\frac{i\beta}{2\omega}}A,~~~~C=\frac{1}{1+\frac{i\beta}{2\omega}}A

The fixed point is such that (1+μe0)r1+e0r3=0(1+\mu e_{0})r_{1}^{\ast}+e_{0}r_{3}^{\ast}=0, so that μr1+r3=r1e0\mu r_{1}^{\ast}+r_{3}^{\ast}=-\frac{r_{1}^{\ast}}{e_{0}}, so the reflection and transmission coefficients are

R=BA=iαω(r1/e0)2(ω2Ω2)+iαω(r1/e0)R={B\over A}=-\frac{i\alpha\omega(r_{1}^{\ast}/e_{0})}{2(\omega^{2}-\Omega^{2})+i\alpha\omega(r_{1}^{\ast}/e_{0})} (57)
T=CA=(ω2Ω2)(ω2Ω2)+iαω(r1/2e0)T={C\over A}=\frac{(\omega^{2}-\Omega^{2})}{(\omega^{2}-\Omega^{2})+i\alpha\omega(r_{1}^{\ast}/2e_{0})} (58)

where the reflection and transmission coefficients satisfy |R|2+|T|2=1|R|^{2}+|T|^{2}=1 From (29) we have

r1/e0=±1/Ω,r_{1}^{\ast}/e_{0}=\pm 1/\Omega, (59)

so that the transmission coefficient is

T=2Ω(ω2Ω2)2Ω(ω2Ω2)±iαω.T=\frac{2\Omega(\omega^{2}-\Omega^{2})}{2\Omega(\omega^{2}-\Omega^{2})\pm i\alpha\omega}. (60)

The modulus squared of the transmission and reflection coefficients are then

|T|2=4Ω2(ω2Ω2)24Ω2(ω2Ω2)2+α2ω2,|T|^{2}=\frac{4\Omega^{2}(\omega^{2}-\Omega^{2})^{2}}{4\Omega^{2}(\omega^{2}-\Omega^{2})^{2}+\alpha^{2}\omega^{2}}, (61)
|R|2=α2ω24Ω2(ω2Ω2)2+α2ω2,|R|^{2}=\frac{\alpha^{2}\omega^{2}}{4\Omega^{2}(\omega^{2}-\Omega^{2})^{2}+\alpha^{2}\omega^{2}}, (62)

Let us examine how |R|2|R|^{2} depends on the different parameters. It attains its maximum 1 for ω=Ω\omega=\Omega and decays at infinity as 1ω2{1\over\omega^{2}}. The half-width of the resonance, such that |R|2(ωh)=1/2|R|^{2}(\omega_{h})=1/2 can be easily obtained. Solving the quadratic equation for ωh2\omega_{h}^{2} we get

ωh2=Ω2±α21+α216Ω4+α28Ω2Ω2±α2,\omega_{h}^{2}=\Omega^{2}\pm{\alpha\over 2}\sqrt{1+{\alpha^{2}\over 16\Omega^{4}}}+{\alpha^{2}\over 8\Omega^{2}}\approx\Omega^{2}\pm{\alpha\over 2}, (63)

for large Ω\Omega. Therefore the half-width is

ωhΩ±α4Ω.\omega_{h}\approx\Omega\pm{\alpha\over 4\Omega}. (64)

Fig. 7 shows the square of the modulus |R|2|R|^{2} for a fixed Ω\Omega (e0=1,μ=1e_{0}=1,~\mu=1). As expected the half-width is proportional to α\alpha. Notice how the resonance becomes asymmetric for large α\alpha indicating that all higher order terms in (63) should be considered.

Refer to caption

Figure 7: Square of the modulus |R|2|R|^{2} of the reflection coefficient as a function of the frequency ω\omega for three different values of the coupling coefficient α=0.2\alpha=0.2 (continuous line, red online) , 1 (long dashed line, green online) and 5 (short dashed line, blue online). The parameters are μ=1,e0=1\mu=1,e_{0}=1.

When μ\mu is varied Ω\Omega varies so |R|2|R|^{2} will be shifted. Fig. 8 shows |R|2|R|^{2} for μ=1,0\mu=-1,0 and 1. As expected for μ<0\mu<0 (resp. μ>0\mu>0) the resonance is shifted towards low (resp. high) frequencies. For μ=1\mu=-1, the higher order terms in (63) should be taken into account and the resonance is not symmetric. When μ=1\mu=1, the higher order terms can dropped and the resonance curve becomes symmetric.

Refer to caption

Figure 8: Square of the modulus |R|2|R|^{2} of the reflection coefficient as a function of the frequency ω\omega for three different values of the dipole parameter μ=1,0.1\mu=-1,~0.1 and 1 as indicated in the plot. The parameters are e0=1,α=1e_{0}=1,~\alpha=1.

The amplitude of the incoming pulse e0e_{0} will also change the resonance by shifting Ω\Omega. Fig. 9 shows |R|2|R|^{2} for e0=0.1,1e_{0}=0.1,1 and 2. As expected, for e0=0.1e_{0}=0.1 the resonance is asymmetric. It becomes symmetric for e01e_{0}\geq 1.

Refer to caption

Figure 9: Square of the modulus |R|2|R|^{2} of the reflection coefficient as a function of the frequency ω\omega for three different values of the field amplitude e0=0.1,1e_{0}=0.1,~1 and 2. The parameters are μ=1,α=1\mu=1,~\alpha=1.

The reflection curve R(ω)R(\omega) allows to estimate the atomic parameter μ\mu and coupling parameter α\alpha through the formulas (33) and (63). The position of the resonance gives Ω\Omega and from there one can compute μ\mu. We have

μ=1e0±Ω2e02e0.\mu=-{1\over e_{0}}\pm{\sqrt{\Omega^{2}-e_{0}^{2}}\over e_{0}}. (65)

The coupling parameter α\alpha can then be obtained from (63).

5 Conclusion

We consider the electromagnetic pulse propagation thought the thin film containing the two-level atoms with taking account both non-diagonal and diagonal matrix elements of the operator of the dipole transition between resonant energy levels. The exact solution of the wave equation allows to derive the modified Bloch equations. Thus the problem of propagating extremely short (one- or few-cycle) pulses through a thin film is reduced to the analysis of a system of nonlinear ordinary differential equations on the Bloch sphere.

In the presence of a constant background field the equilibrium state of the system film/field is changed. These new equilibrium states are the fixed points of the modified Bloch equations. They depend on the field amplitude, the difference of diagonal elements of the operator of the dipole transition (dipole parameter) and the coupling constant. The stability analysis of these fixed points indicates which states will be attained by the system. For the stable states the population difference is negative or positive depending on the sign of the dipole parameter. When the film is illuminated by an electromagnetic field it reaches the new stable state with a typical relaxation time which depends on the field amplitude, the dipole parameter and the coupling constant. Using our estimate, experimentalists could measure the dipole parameter.

In the last part of the article we considered that in addition to a constant background, the thin film is irradiated by a small harmonic field of frequency ω\omega. This spectroscopy analysis yields the reflection and transmission coefficients of the film as a function of ω\omega. As expected the film is completely opaque i.e. its reflection coefficient is equal to 1 for the field modified transition frequency Ω\Omega given by (33). We found that the width of the resonant curve is proportional to the coupling constant. The position of the resonance depends on the dipole parameter and the ground field.

The modern progress in nano-technology allows to produce thin films of different features and the investigation of such features is attractive. Our study shows that the fast electromagnetic response of a thin film could be used in experiments to measure intrinsic parameters of generalized atoms (quantum dots, meta-atoms, molecules …) and their coupling parameter to the field.

Acknowledgments

One of the authors (A.I.M.) is grateful to the Laboratoire de Mathématiques, INSA de Rouen for hospitality and support. Elena Kazantseva thanks the Region Haute-Normandie for a Post-doctoral grant. Jean-Guy Caputo thanks the Centre de Ressources Informatiques de Haute-Normandie for access to computing ressources. The authors express their gratitude to S.O. Elyutin for enlightening discussions.

References

  • [1] L. Allen, J.H. Eberly, Optical Resonance and Two-Level Atoms, Wiley, N.Y. 1975).
  • [2] S.L. McCall, E.L. Hahn, Phys.Rev.Lett. 18, 908 (1967); Phys.Rev.183, 457 (1969)
  • [3] R.L.Shoemaker Ann.Rev.Phys.Chem. 30, 239-270 (1979)
  • [4] A.I. Maimistov, A.M. Basharov, S.O. Elyutin, Yu.M. Sklyarov, Phys. Report 191, 1 (1990).
  • [5] M.Fleischhauer, A.Imamoglu, and J.P. Marangos Rev.Mod. Phys. 77, 633-673 (2005)
  • [6] K. Bergmann, H. Theuer, and B. W. Shore Rev.Mod.Phys. 70, 1003-1025 (1998)
  • [7] S.O. Elyutin, E.V. Kazantseva, and A.I. Maimistov, Opt. Spectrosc. 90, 439-445 (2001)
  • [8] A.I. Maimistov and S.O. Elyutin, Opt. Spectrosc. 93, 257-262 (2002).
  • [9] J.P. Lavoine, C. Hoerner, A.A. Volleys, Phys.Rev. A 44, 5947 (1991)
  • [10] R. Bavli, Y.B. Band, Phys.Rev. A 43, 5039 (1991),
  • [11] Weifeng Yang, Shangqing Gong, Ruxin Li and Zhizhan Xu, Phys.Lett. A362, No. 1, 37-41 (2007)
  • [12] L.W. Casperson, Phys. Rev. A,57 , 609-621 (1998).
  • [13] A.I. Maimistov, J.-G. Caputo, Physica D 189, 107 (2004).
  • [14] M. Agrotis, N.M. Ercolani, S.A. Glasgow, J.V. Moloney, Physica D 138. 134-162 (2000)
  • [15] J. G. Caputo and A. I. Maimistov, Phys. Lett. A 296, 34-42, (2002).
  • [16] M. Agrotis Physica D183, 141-158 (2003)
  • [17] S.A. Glasgow, M.A. Agrotis, and N.M. Ercolani Physica D212, 82-99 (2005)
  • [18] A.A. Zabolotski, Optics and Spectroscopy 95, 751-759 (2003).
  • [19] S. O. Elyutin JETP 101, 11 (2005)
  • [20] N. V. Ustinov, Proceedings SPIE, Vol. 6725 ICONO 2007: Nonlinear Space-Time Dynamics, Yuri Kivshar; Nikolay Rosanov, Editors, 67250F (2007); Breather-like pulses in a medium with the permanent dipole moment, arXiv: nlin.SI/0512056
  • [21] A. A. Zabolotskii, J. Exp. Theor. Phys. 106, No. 5, 846-857 (2008)
  • [22] V. I. Rupasov and V. I. Yudson, Sov. J. Quantum Electronics, 12, 415, (1982).
  • [23] , M. G. Benedict, A. I. Zaitsev, V. A. Malyshev, and E.D. Trifonov, Opt. Spectrosk. 66, 726-728 (1989)
  • [24] M. G. Benedict, V. A. Malishev, E.D. Trifonov and A. I. Zaitsev, Phys. Rev A 43, 3845, (1991).
  • [25] S.O. Elyutin, Propagation of a videopulse through a thin layer of two-level dipolar atoms, J. Phys. B: At. Mol. Opt. Phys. 40, 2533-2550 (2007)
  • [26] A.I. Maimistov, I.R. Gabitov, Nonlinear optical effects in artificial materials, Eur. Phys. J. Special Topics 147, 1, 265-286 (2007) in ”Nonlinear aves in complex systems: energy flow and geometry” (Springer, 2007)
  • [27] A. N. Tikhonov and A. A. Samarski, Equations of Mathematical Physics (Dover, New York, 1983).
  • [28] J.-G. Caputo, E. V. Kazantseva, A.I. Maimistov, Electromagnetically induced switching of ferroelectric thin films, Phys. Rev. B 75, 014113 (2007)
  • [29] G.L.Lamb, Jr., Rev.Mod Phys. 43, 99- (1971).
  • [30] R.K. Bullough, P.M. Jack, P.W. Kitchenside, R. Saudders , Phys.Scr. 20, 364- (1979).
  • [31] F.T. Arecchi, E. Courtens, Phys.Rev. A 2, 1730 (1970)
  • [32] N.E. Rehler, J.H. Eberly, Phys.Rev. A 3, 1735 (1971).
  • [33] M.J. Ablowitz, H. Segur, Solitons and the Inverse Scattering Transformation, SIAM, Philadelphia, 1981.
  • [34] E. Hairer, S. P. Norsett and G. Wanner. Solving ordinary differential equations I (Springer-Verlag, 1987).