1: \documentstyle[aps,12pt]{revtex}
2: \draft
3: \title{Spatial evolution of short laser pulses under coherent
4: population trapping}
5: \author{V.G.Arkhipkin}
6: \address{Institute of Physics, Russian Academy of Sciences, 660036
7: Krasnoyarsk ; Krasnoyarsk State Technical University, 660074
8: Krasnoyarsk, Russia}
9: \author{I.V.Timofeev}
10: \address{Krasnoyarsk state university,660041 Krasnoyarsk, pr. Svobodny, 79}
11: \begin{document}
12: \maketitle
13:
14: \begin{abstract}
15:
16: Spatial and temporal evolution is studied of two powerful short
17: laser pulses having different wavelengths and interacting with a
18: dense three-level Lambda-type optical medium under coherent
19: population trapping. A general case of unequal oscillator
20: strengths of the transitions is considered. Durations of the probe
21: pulse and the coupling pulse $T_{1,2}$ ($T_2>T_1$) are assumed to
22: be shorter than any of the relevant atomic relaxation times. We
23: propose analytical and numerical solutions of a self-consistent
24: set of coupled Schr\"{o}dinger equations and reduced wave
25: equations in the adiabatic limit with the account of the first
26: non-adiabatic correction. The adiabaticity criterion is also
27: discussed with the account of the pulse propagation. The dynamics
28: of propagation is found to be strongly dependent on the ratio of
29: the transition oscillator strengths. It is shown that envelopes of
30: the pulses slightly change throughout the medium length at the
31: initial stage of propagation. This distance can be large compared
32: to the one-photon resonant absorption length. Eventually, the
33: probe pulse is completely reemitted into the coupling pulse during
34: propagation. The effect of localization of the atomic coherence
35: has been observed similar to the one predicted by Fleischhauer and
36: Lukin (PRL, {\bf 84}, 5094 (2000).
37:
38: \end{abstract}
39:
40: PACS numbers: 42.50.Hz, 42.50.Rh, 42.65.Ky
41:
42: \section{INTRODUCTION}
43:
44: The electromagnetically induced transparency (EIT) can be used to
45: make optically thick media transparent to resonant laser radiation
46: \cite{Harr}. The EIT is the result of various quantum interference
47: effects such as nonlinear interference \cite{Raut}, coherent
48: population trapping (CPT) \cite{Agap,Arim}, adiabatic population
49: transfer (APT) \cite{Berg}. Optical characteristics of the matter
50: undergo drastic changes under those effects to such an extent that
51: they can now be manipulated. A lot of interesting applications
52: based on that have been proposed and experimentally realized (see,
53: e.g. \cite{Agap,Arim,Berg,Yam,Hau,Kash,Phillips,Liu}).
54:
55: Interesting and unusual phenomena caused by the above indicated
56: effects can be observed when laser pulses propagate in a resonant
57: three-level medium. Propagation of pulses under EIT conditions was
58: studied for example in \cite{1,2,3,4,5,6,7,7a}. As a rule, the
59: situations are considered when both pulses have identical forms
60: and their duration is longer than the relaxation time of the
61: intermediate resonant state (matched pulses \cite{1}; dressed
62: field pulses \cite{2,3}) or when the duration of the coupling
63: radiation considerably exceeds that of the probe radiation
64: (adiabatons \cite{5,6,Kasapi}). A theoretical study of certain
65: features of spatial evolution under APT conditions is presented in
66: \cite{6,7,8}. Propagation of soliton-like pulses in a three-level
67: system is studied in \cite{Hioe}. A three-level system with equal
68: oscillator strengths is under consideration in all the above
69: mentioned studies, whereas in actual fact the transition
70: oscillator strengths most often are different.
71:
72: In this paper, the spatial and temporal evolution is studied of
73: two overlapping short laser pulses propagating in a resonant
74: optically thick medium that consists of three-level
75: $\Lambda$-atoms. Pulses of such configuration are widely used to
76: enhance the efficiency of nonlinear generation processes
77: \cite{9a,9b}. The two pulses are assumed to have identical shapes
78: but different durations $(T_2>T_1)$ as shown in Fig.1. The pulse
79: durations are much shorter than any of the times of relaxation in
80: the medium (short pulses). It is also assumed that the pulse
81: envelopes satisfy the adiabaticity criterion \cite{3,10,11}:
82:
83: \begin {eqnarray} \label {e1}
84: (G_2\dot{G_1}-G_1\dot{G_2})/G^3\ll1,
85: \end {eqnarray}
86: %
87: where $G_{1,2}$ are the Rabi frequencies of the respective
88: fields, $G=\sqrt{G_1^2+G_2^2}$; the dot refers to time
89: derivatives. Condition (\ref{e1}) is easy to satisfy by making one
90: of the pulse amplitudes or both of them large even for short
91: pulses. This will induce a strong coherence at the Raman
92: transition resulting in the effect of CPT. The latter considerably
93: decreases the absorption of the propagating resonant pulses. The
94: dynamics of propagation of such pulses is studied here without
95: restriction of the relationship between oscillator strengths of
96: transitions.
97:
98: Our theoretical model involves a set of coupled Schr\"{o}dinger
99: equations and a set of reduced wave equations, allowing thus a
100: simultaneous description of temporal and spatial evolution of the
101: atomic system and the radiation. The equations are analyzed in
102: approximation (\ref {e1}) with the account of the first
103: non-adiabatic correction. It will be shown that the dynamics of
104: propagation strongly depends on the oscillator strengths ratio.
105: Also the spatial and temporal behavior is analyzed of the atomic
106: Raman coherence. A possibility of localizing the atomic coherence
107: spatially has been established. The results obtained are compared
108: to the results reported in \cite{Lukin}.
109:
110: The paper is organized as follows. In Section II, we describe the
111: model and present the basic equations. Section III contains
112: self-consistent solutions of those equations in the adiabatic
113: limit and describes the temporal behaviour of the level
114: populations and the atomic Raman coherence in the optically thin
115: medium. The spatial evolution of pulses in the optically thick
116: medium is described in Section IV for various oscillator strength
117: ratios. In Section IV, we also discuss the adiabaticity criterion
118: and demonstrate the effect of spatial localization of the atomic
119: coherence. Finally we summarize the results obtained.
120:
121: \section{BASIC EQUATIONS}
122:
123: The three-level system under consideration is shown in Fig.1
124: together with the temporal configuration of the pulses as they
125: enter the medium. The pulses travel along the same direction, $z$.
126: States $|0\rangle$, $|1\rangle$ and $|2\rangle$ are connected by
127: laser pulses $E_1=1/2\,E_1(t)\ exp[-i(\omega_1 t-k_1 z)]+c.c.$ and
128: $E_2=1/2\,E_2(t)\ exp[-i(\omega_2 t-k_2 z)]+c.c.$, respectively.
129: In our further consideration we shall refer to the first pulse
130: $E_1$ as the probe and to the second pulse $E_2$ as the coupling
131: pulse. The Rabi frequency of probe pulse is comparable
132: with that of the coupling pulse. The pulses are sent simultaneously
133: into an atom. The pulse durations $T_{1,2}$ $(T_2>T_1)$ are
134: assumed to be much less than any of the relaxation times of atoms.
135: The transition $|0\rangle-|2\rangle$ is electric dipole forbidden.
136: The intermediate state $|1\rangle$ is in a one-photon resonance
137: with each field interacting only with the respective transition.
138:
139: The following standard set of equations describes the spatial and
140: temporal dynamics of the probability amplitudes of atomic states
141: $b_{0,1,2}$ and slowly varying Rabi frequencies
142: $G_1=d_{10}E_1(t)/2\hbar$, $G_2=d_{21}E_2(t)/2\hbar$ in the
143: local-time coordinate system $\tau = t-z/c$:
144:
145: \begin {eqnarray} \label {eq1}
146: \frac {\partial b_0}{\partial \tau}& =& iG_1^*b_1\exp {(-ik_1 z)},\nonumber \\
147: \frac {\partial b _ 2} {\partial \tau}& =& iG_2^{*} b_1\exp {(-ik_2z)}, \nonumber \\
148: \frac {\partial b _ 1} {\partial \tau}& =& iG_1b_0\exp {(ik_1 z)} +
149: iG_2b_2\exp {(ik_2 z)}.
150: \end {eqnarray}
151:
152: \begin {eqnarray} \label {eq2}
153: \frac {\partial G_{1}} {\partial z}& =& iK_1 b_1b_0^*\exp {(ik_1 z)},\nonumber \\
154: \frac {\partial G_{2}} {\partial z}& =& iK_2 b_1b_2^*\exp{(ik_2z)}.
155: \end {eqnarray}
156:
157: Here we assumed zero one-photon detunings. $K_{1,2} =
158: {\pi\omega_1\left|d_{10,12}\right|^2}N/\hbar c$ are the
159: propagation coefficients, $N$ is the atomic concentration,
160: $d_{10,12}$ are the dipole transition matrix elements,
161: $\omega_{1,2}$, $k_{1,2}$ are the frequencies and the wave numbers
162: of the interacting waves in vacuum, $c$ is the light velocity in
163: vacuum. All atoms are assumed to be initially in the ground state
164: $|0\rangle$: $b_0(-\infty,z)=1$, $b_{1,2}(-\infty,z)=0$. We use
165: Gaussian pulses at the medium entrance $z=0$ for the purpose of
166: numerical simulation: $G_1(\tau)=G_1^0 \exp(-\tau^2ln2/T_1^2)$,
167: $G_2(\tau)=G_2^0 \exp[-\tau^2ln2/T_2^2]$.
168:
169: In terms of $a_0=b_0\exp{(i k_1 z)}$, $a_2=b_2\exp{(i k_2 z)}$,
170: $a_1=i b_1$, equations ({\ref{eq1}}) and ({\ref{eq2}}) can be
171: written as:
172:
173: \begin {eqnarray} \label {equ1}
174: \frac {\partial a_0} {\partial\tau}& =& G_1^{*} a_1, \nonumber \\
175: \frac {\partial a_2}{\partial\tau}& =& G_2^{*}a_1, \nonumber \\
176: \quad\frac{\partial a_1}{\partial\tau}& =& - G_1 a_0 - G_2 a_2.
177: \end {eqnarray}
178: %
179: \begin {eqnarray} \label {equ2}
180: \frac {\partial G_{1}} {\partial z}& =& -K_1 a_1 a_0^{*},\nonumber \\
181: \frac {\partial G_{2}} {\partial z}& =& -K_2 a_1 a_2^{*}.
182: \end {eqnarray}
183:
184: The coupled equations ({\ref{equ1}}) and ({\ref{equ2}}) give a
185: complete semiclassical description of the resonant
186: different-wavelengths propagation problem we are dealing with.
187:
188: \section{TEMPORAL DYNAMICS OF LEVEL POPULATIONS AND RAMAN COHERENCE
189: IN THE ADIABATIC APPROXIMATION (OPTICALLY THIN MEDIUM)}
190:
191: In this section we study the temporal dynamics of populations and
192: the atomic coherence in the given time-dependent field, assuming
193: that the medium is optically thin. Based on that, $G_{1,2}$ will
194: not depend on the coordinate $z$. One can show that condition
195: (\ref{e1}) for Gaussian pulses reduces to $G_2^0 T_1\gg1$ when
196: $T_2/T_1>\sqrt{2}$. With the first non-adiabatic correction, the
197: solution of Eq. ({\ref{equ1}}) takes on the form:
198:
199: \begin {eqnarray} \label {eq3}
200: a_0 &\simeq& \frac{G_2(\tau)}{G(\tau)}, \quad
201: a_2\simeq-\frac{G_1(\tau)}{G(\tau)}, \nonumber \\
202: a_1 &\simeq& \frac {1}{G_1}\frac{\partial(G_2/G)}
203: {\partial\tau}\simeq
204: -\frac{1}{G_2}\frac{\partial(G_1/G)}{\partial\tau},
205: \end {eqnarray}
206: %
207: where $G(\tau)=\sqrt{G_1^2(\tau)+G_2^2(\tau)}$.
208:
209: The solutions for probability amplitudes are convenient to be
210: represented as:
211:
212: \begin {eqnarray} \label {eq3a}
213: a_0=\cos{\theta(\tau)}, \quad a_2=-\sin{\theta(\tau)},
214: \end {eqnarray}
215: %
216: where the mixing angle $\theta(\tau)$ is defined as
217: $\tan{\theta(\tau)}=G_1(\tau)/G_2(\tau)$ (we shall discuss its
218: meaning later on).
219:
220: The expression for $a_1$ can be reduced to:
221:
222: \begin {eqnarray} \label {eq4}
223: a_1=(G_2\dot{G_1}-G_1\dot{G_2})/G^3=\dot{\theta}/G.
224: \end {eqnarray}
225: %
226: In adiabatic limit (\ref {e1}) $|a_1|=|\dot{\theta}/G|\ll1$
227: ($\dot{\theta}=\partial{\theta}/\partial{\tau}$), i.e. the
228: population of the intermediate state $|1\rangle$ is close to zero
229: all the time during the interaction with pulses. This also implies
230: that the resonant absorption of the light pulses is weak
231: (electromagnetically induced transparency) and the population is
232: mainly distributed between the initial $|0\rangle$ and the final
233: $|2\rangle$ states:
234:
235: \begin {eqnarray} \label {eq5}
236: |a_0|^2+|a_2|^2\simeq 1.
237: \end {eqnarray}
238: %
239: Equality (\ref {eq5}) reflects the fact that atoms are trapped in
240: the CPT state:
241: $a_{cpt}=(G_2/G)a_0-(G_1/G)a_2=a_0\cos{\theta}-a_2\sin{\theta}=1$.
242: This effect is responsible for the decrease in the resonant
243: absorption of propagating pulses. Also the Raman coherence
244: $\rho_{20}=a_0a_2^*$ occurs:
245:
246: \begin {eqnarray} \label {eq5a}
247: \rho_{20}=-\frac{1}{2}\sin{(2\theta)} \quad \mbox{or} \quad
248: \rho_{20}=-\frac{G_1G_2}{G_1^2+G_2^2}.
249: \end {eqnarray}
250:
251: Obviously, the maximum coherence (in absolute value)
252: $|\rho_{20}|=1/2$ is reached when $\theta=\pi/4$ ($G_1^0=G_2^0$).
253: Figure 2 shows the temporal behavior of level populations
254: $|a_{0,2}(\tau)|^2$, the atomic Raman coherence
255: $|\rho_{20}(\tau)|$ and the mixing angle $\theta(\tau)$ for
256: Gaussian pulses in the optically thin medium.
257:
258: The above results can be interpreted in terms of the
259: three-dimensional vector model where vector variables
260: $\vec{a}=(a_0,a_2,a_1)$ and $\vec{G}=(G_2,G_1,0)$ (the torque
261: vector) are introduced. Using these variables, we can rewrite (\ref
262: {equ1}) as
263:
264: \begin {eqnarray} \label {eq7}
265: \dot {\vec {a}} = \vec {G} \times\vec {a},
266: \end {eqnarray}
267: %
268: where the sign '$\times$' means the vector product. The solution
269: of equation (\ref {eq7}) is vector $\vec{a}=(G_2/G,-G_1/G,
270: \dot{\theta}/G)$. Components of the vector $\vec{a}$ coincide with
271: adiabatic solution (\ref {eq3}). Figure 3 illustrates the dynamics
272: of vectors $\vec{G}$ and $\vec{a}$ in the three-dimensional vector
273: model. The torque vector $\vec{G}$ moves in the
274: $\vec{e}_1$-$\vec{e}_2$ plane. And vector $\vec{a}$ having a small
275: angle with respect to vector $\vec{G}$ ($|\dot{\theta}/G| \ll 1$)
276: follows it. Thus one can see an absolute analogy with the adiabatic
277: following in the case of a light pulse interacting with a
278: two-level atom \cite{13}. Such a simple picture can be observed
279: only in optically thin media. In optically dense media, $G_{1,2}$
280: and hence $\theta$ become dependent on the $z$ coordinate.
281:
282: \section{SPATIAL EVOLUTION OF INTERACTING PULSES IN OPTICALLY DENSE MEDIA}
283:
284: \subsection{General case:
285: unequal oscillator strengths $(K_1\neq K_2)$}
286:
287: The condition $K_1\neq K_2$ means that the probability of the
288: $|0\rangle-|1\rangle$ transition is not equal to that of the
289: $|2\rangle-|1\rangle$ transition. We note that in the ideal
290: adiabatic limit $a_1=0$, and the pulses would not change their
291: shape as they propagate in a medium that is optically thick for
292: each of the pulses (see (\ref {equ2})). However this is not the
293: case. The non-adiabatic correction has to be introduced for real
294: situations, which results in the induced dipole moments at the
295: transitions $|1\rangle-|0\rangle$ and $|1\rangle-|2\rangle$ , and
296: in the change of both pulses traveling in the medium. In order to
297: attribute this effect to propagation of the interacting pulses in
298: an optically dense medium, it is necessary to solve Eqs. (\ref
299: {equ1}) and (\ref {equ2}) in a self-consistent way.
300:
301: Use (\ref {eq3a}) and (\ref {eq4}) to rewrite field equations
302: (\ref {equ2}) in the following form
303:
304: \begin {eqnarray} \label {eq8}
305: \frac{\partial G_1} {\partial z} =
306: -K_1\frac{\dot{\theta}}{G}\cos{\theta}, \qquad
307: \frac{\partial G_2}{\partial z} =
308: K_2 \frac{\dot{\theta}}{G}\sin{\theta}.
309: \end {eqnarray}
310:
311: From (\ref {eq8}), one can show that
312:
313: \begin {eqnarray} \label {eq8a}
314: K_2 G_1^2(\tau,z) &+& K_1 G_2^2(\tau, z) \nonumber \\
315: &=& K_2 G_1^2(\tau,z=0)+K_1G_2^2(\tau,z=0),
316: \end {eqnarray}
317: %
318: i.e. $K_2 G_1^2(\tau,z)+K_1 G_2^2(\tau, z)$ does not depend on the
319: $z$ coordinate.
320:
321: Equation (\ref {eq8a}) describes the Manley-Raw relation, i.e.
322: the law of conservation of the total energy density during
323: propagation under CPT conditions.
324:
325: Using the definitions of $\theta$ and $G$, we obtain the following
326: expressions for $G_{1,2}(\tau,z)$
327:
328: \begin {eqnarray} \label {eq3b}
329: G_1(\tau,z)=G(\tau,z)\sin{\theta(\tau,z)}, \nonumber \\
330: G_2(\tau,z)=G(\tau,z)\cos {\theta(\tau,z)}.
331: \end {eqnarray}
332:
333: Substitution (\ref {eq3b}) into the Manley-Raw relation gives the
334: following expression for $G(\tau,z)$:
335:
336: \begin {eqnarray} \label {eq10}
337: G^2(\tau,z)=G^2_0(\tau)\frac{K_2\sin^2(\theta_0(\tau))+
338: K_1\cos^2(\theta_0(\tau))}{K_2\sin^2(\theta(\tau,z))+
339: K_1\cos^2(\theta(\tau,z))},
340: \end {eqnarray}
341: %
342: where $G_0^2(\tau)=G_{1}^2(\tau,0)+G_{2}^2(\tau,0)$ and
343: $\theta_0(\tau)=\theta_0(\tau,0)$ are the functions at the medium
344: entrance $z=0$.
345:
346: So the dynamics of level populations as well as the atomic Raman
347: coherence and the evolution of the pulse shape are completely
348: determined by the function $\theta(\tau,z)$ that depends on both
349: the time $\tau$ and the coordinate $z$. Differentiating
350: $\tan{\theta}=G_1/G_2$ on the $z$ and using (\ref {eq8}) we obtain
351: the following equation for $\theta(\tau,z)$:
352:
353: \begin {eqnarray} \label {eq11}
354: \frac{\partial\theta}{\partial\tau}+\frac{G^2(\tau,z)}{K(\theta)}
355: \frac{\partial\theta}{\partial z}=0,
356: \end {eqnarray}
357: %
358: where $K(\theta(\tau,z))=K_1\cos^2{(\theta(\tau,z))}+
359: K_2\sin^2{(\theta(\tau,z))}$.
360:
361: Eq.(\ref {eq11}) is similar to the equations describing nonlinear
362: waves with the sharpening of the wave front during propagation
363: \cite{14}. The parameter $u=G^2/K$ can be treated as the
364: "nonlinear" velocity. The nonlinear velocity can be described as
365: $u(\tau,z)=A(\tau)/K^2(\theta(\tau,z))$ where the first factor
366: $A(\tau)=
367: G_0^2(\tau)[K_2\sin^2{(\theta_0(\tau))}+K_1\cos^2{(\theta_0(\tau))}]$
368: is independent of the $z$ coordinate, and the second factor
369: $K(\theta(\tau,z))$ is maintained along a characteristic of
370: Eq.(\ref {eq11}) $\theta(\tau,z)=const$. This allows to write down
371: the characteristic curve equation in an evident form:
372:
373: \begin{eqnarray} \label {eq12}
374: z=\frac{1}{K^2(\theta_0)}\int_{\tau_0}^{\tau}A(\tau')d\tau'
375: \end{eqnarray}
376: %
377: Here $\tau_0$ is the time at which the characteristic curve goes
378: out of the medium boundary.
379:
380: The solution for $\theta(\tau,z)$ has the form:
381:
382: \begin{eqnarray} \label {eq13}
383: \theta(\tau,z)=\theta_0(\tau_0,0)
384: \end{eqnarray}
385: %
386: Here $\tau_0$ is to be determined from Eq.(\ref {eq12}).
387:
388: \subsection{The case of equal oscillator strengths $(K_1=K_2)$}
389:
390: In the case of equal oscillator strengths $K_1=K_2\equiv K$ the
391: $G(z,\tau)$ function is not subject to changes during propagation
392: ($G(\tau,z)=G_0(\tau)$ -- the Manley-Raw relation for this case).
393: Therefore Eq. (\ref {eq11}) substantially simplifies:
394:
395: \begin {eqnarray} \label {eq14}
396: \frac{\partial\theta}{\partial\tau}+\frac{G^2_0(\tau)}{K}
397: \frac{\partial\theta}{\partial z}=0.
398: \end {eqnarray}
399: %
400: The solution of Eq. (\ref {eq14}) can be written in the following
401: form:
402: %
403: \begin {eqnarray} \label {eq15}
404: \theta(\tau,z)=\theta_0(Z^{-1}(Z(\tau)-z),0),
405: \end {eqnarray}
406: %
407: where $Z(\tau)=K^{-1}\int_{-\infty}^{\tau}d\tau^{'}
408: G^2(0,\tau^{'})$, $Z^{-1}(\tau)$ is the inverse function of
409: $Z(\tau)$.
410:
411: It is not difficult to show that in this case
412: $G=\sqrt{G_1^2(\tau,z)+G_2^2(\tau,z)}$ coincides with the
413: definition of dressed field pulses \cite{2,3}: $G_{-}=a_0 G_2 -
414: a_2 G_1$. Thus the pulses in our case can be identified as dressed
415: field pulses (only at $K_1=K_2$). It would be interesting to note
416: that other combination $G_{+}=a_0 G_1+a_2 G_2\equiv 0$ (see also
417: \cite{2}). The concept of dressed field pulses cannot be applied
418: to the case of $K_1\neq K_2$, but in both cases the Manley-Raw
419: relation remains valid.
420:
421: \subsection{The adiabaticity criterion}
422:
423: The above results were obtained on the assumption that the
424: adiabaticity criterion (\ref {e1}) (or $|\dot{\theta}/G|\ll1$)
425: remains valid during propagation of pulses. However it is not
426: necessary the case. Therefore we investigate the adiabaticity
427: condition with the account of propagation. Differentiating (\ref
428: {eq13}) on $\tau$, we can write the following expression for the
429: adiabaticity criterion:
430:
431: \begin {eqnarray} \label {eq16}
432: \frac{\dot{\theta}(\tau,z)}{G(\tau,z)} &=&
433: \frac{\partial\theta_0}{\partial\tau_0}
434: \frac{G(\tau,z)}{G^2_0(\tau_0)}
435: \frac{K(\theta(\tau,z))}{K(\theta_0(\tau_0))} \nonumber \\
436: &\times& [1+\sin{(2\theta_0(\tau_0))}\frac{2(K_2-K_1)z}{G^2_0(\tau_0)}
437: \frac{\partial{\theta_0}}{\partial{\tau_0}}]^{-1}
438: \ll1
439: \end {eqnarray}
440: %
441: As follows from (\ref {eq16}), the adiabaticity condition is
442: destroyed ($\dot{\theta}/G \rightarrow \infty$) when the factor in
443: the square brackets tends to zero. Since
444: $\sin{(2\theta_0(\tau_0))}>0$ in the entire range of change of
445: $\theta$ ($0\leq\theta\leq\pi/4$), relation (\ref{eq16}) is not
446: fulfilled under the following conditions
447:
448: \begin {eqnarray} \label {eq18}
449: \frac{2(K_1-K_2)z}{G^2_0(\tau_0)}\frac{\partial{\theta_0}}
450: {\partial{\tau_0}}\sin{(2\theta_0(\tau_0))}=1, \nonumber \\
451: (K_1-K_2)\frac{\partial{\theta_0}}{\partial{\tau_0}}>0,
452: \end {eqnarray}
453:
454: Evidently, condition (\ref {eq18}) is not fulfilled at $K_1=K_2$,
455: and the adiabaticity criterion holds throughout the propagation
456: process in this case. It should be mentioned that numerical
457: analysis of Eq. (\ref{eq18}) reveals that, generally speaking,
458: there is a range of change of the $q=K_1/K_2$ parameter when the
459: first condition in (\ref{eq18}) is not fulfilled either: $q_{min}<
460: q < q_{max}$, $q_{min}<1$, $q_{max}>1$. The values $q_{max}$ and
461: $q_{min}$ depend on the ratio $a = T_2/T_1$ and the shape of
462: pulses. For example, the $q$ parameter for Gaussian pulses with
463: $T_2/T_1=3$ may vary within the limit $0.40<q<1.35$, as shown by
464: the numerical simulation.
465:
466: In this case for the thick medium the adiabaticity criterion has
467: the form:
468: %
469: \begin {eqnarray} \label {eq19}
470: \frac{\partial\theta_0}{\partial\tau_0}\frac{G(\tau,z)}{G^2_0(\tau_0)}\ll1.
471: \end {eqnarray}
472: %
473: It can be readily shown that condition (\ref{eq19}) for Gaussian
474: pulses is fulfilled when $T_2/T_1>\sqrt{3}$ and $G_2^0T_1\gg1$
475: (compare with the case of the thin medium).
476:
477: For Gaussian pulses ($T_2>T_1$) and the initial conditions
478: $a_0(-\infty)=1$, $a_2(-\infty)=0$, we have $\dot{\theta_0}>0$ for
479: $-\infty<\tau_0<0$. With $K_1>K_2$ ($q>q_{max}$) the adiabaticity
480: criterion begins to break down at the leading edge: the mixing
481: angle front becomes steeper. Condition (\ref {eq16}) gets
482: destroyed at the trailing edge (where $\dot{\theta_0}<0$) when
483: $K_1<K_2$ ($q<q_{min}$). Let’s introduce a critical length $z_c$ at
484: which the adiabaticity condition (\ref{eq16}) cannot be satisfied.
485: Using (\ref {eq18}), one can obtain the following simple
486: estimation for $z_c$ for the case $K_1 \neq K_2$:
487:
488: \begin {eqnarray} \label {eq18a}
489: z_c \simeq \frac{G_0^2 T_1}{2|K_1-K_2|}.
490: \end {eqnarray}
491:
492: Let us now consider the experimental parameters: $N = 10^{15}
493: cm^{-3}$; $1/\lambda_{1,2} = 10000, 20000 (cm^{-1})$;
494: $1/\gamma_{1,2} = 10, 50 ns$ are the relaxation times of levels
495: $|1\rangle$ and $|2\rangle$, respectively; $T_1 = 0.1 ns$;
496: $G_{1,2}^0T_1 = 20$; $a = T_2/T_1 = 3$; $fg_{10} = 0.1$ is the
497: oscillator strength of the probe transition. From (\ref{eq18a}) we
498: obtain $z_c = 10^5 z_0 \approx 2 cm$, where $z_0$ is the linear
499: absorption length. This estimation agrees with the results
500: presented in Figs. 5-7.
501:
502: Using the characteristic equation (\ref{eq12}) it is easy to find
503: the area where the adiabaticity criterion breaks down.
504: Characteristic curves for various values of $q=K_1/K_2$ are shown
505: in Fig~4. The thickening of the characteristic curves means the
506: sharpening of the mixing angle front at a certain medium depth. At
507: the point of intersection (see Fig.~4~b,c), $\dot{\theta}
508: \rightarrow \infty$, condition (\ref{eq16}) collapses. However
509: there are no such points in the case when $q_{min}<q<q_{max}$ (see
510: Fig.4a), i.e. here the adiabaticity criterion (\ref{eq16}) is
511: maintained during propagation of pulses.
512:
513: The above discussion is illustrated by Fig.5 where the temporal
514: behavior of the mixing angle $\theta$ is presented for
515: $q=K_1/K_2=1,0.25,4$ and for different normalized propagation
516: lengths. On the Fig.5a-b the analytical results are presented
517: (formulae (\ref {eq13}) and (\ref {eq15})). Figure 5c shows the
518: numerical solution for $\theta$ at $K_1>K_2$.
519:
520: One can see that the evolution of the $\theta$ parameter at
521: $K_1>K_2$ is different from that at $K_1<K_2$. At $K_1>K_2$ the
522: adiabaticity condition fails for all values of $z$, beginning from
523: the critical lengths $z_c$ defined by Eq. (\ref {eq18a}). Here,
524: the analytical theory does not apply at the very late stage of
525: nonlinear wave propagation. In the case of $K_1<K_2$ ,
526: nonadiabaticity develops at the trailing edge (the front becomes
527: steeper at a certain propagation length), but it does not go
528: deeper into the medium. A good agreement between the analytical
529: and numerical solutions for $\theta$ at $K_1\leq K_2$ is observed
530: over the entire propagation length. This leads us to conclude that
531: the interaction adiabaticity is fairly sensitive to the ratio of
532: the oscillator strengths of the transitions interacting with the
533: pulses.
534:
535: \subsection{Discussion of results}
536:
537: The solutions obtained have been used to analyze the temporal
538: dynamics and spatial behavior of the propagating EIT pulses and
539: the atomic coherence for various oscillator strength ratios.
540: Figure 6 illustrates the temporal and spatial evolution of
541: normalized Rabi frequencies of the both pulses
542: $g_{1,2}(\tau)=G_{1,2}/\sqrt{{G_1^0}^2+{G_2^0}^2}$ as they
543: propagate inside the optically thick medium. The temporal
544: evolution of pulses can be seen to depend on the ratio between the
545: transition oscillator strengths. In the case of $K_1\leq K_2$ both
546: pulses undergo reshaping as they propagate in the medium (Fig.6a
547: and b). The probe pulse is gradually depleted and the coupling gets
548: stronger. Note that the pulse shape at the initial stage of
549: propagation shows very little change along the length of the
550: medium, which may exceed the linear absorption length. Complete
551: reemitting of the probe pulse into the coupling one during
552: propagation is possible. Using the Eq. (\ref {eq12}) one can
553: obtain the following expression for the maximal distance $z_m$ on
554: which the probe pulse propagates into the medium:
555:
556: \begin{eqnarray} \label {eq20}
557: z_m=\frac{1}{K^2(\theta_0(-\infty))}
558: \int_{-\infty}^{\infty}A(\tau)d\tau.
559: \end{eqnarray}
560: %
561: For the Gaussian pulses we have
562:
563: \begin{eqnarray} \label {eq21}
564: z_m = \sqrt{\frac{\pi}{2ln{2}}}\,(G_2^0)^2\; T_1
565: \frac{K_2(G_1^0/G_2^0)^2+K_1a}{K_1^2}.
566: \end{eqnarray}
567: %
568: Here $a=T_2/T_1$.
569:
570: For the same parameters as in the previous sub-section, we have
571: $z_m = 4.8 \cdot 10^5 z_0 \approx 9 cm$ for $K_1/K_2 = 1$ and
572: $z_m = 8.4 \cdot 10^5 z_0 \approx 16 cm$ for $K_1/K_2 = 1/4$.
573: These values agree with the results shown in Figs. 5-7.
574:
575: An interesting feature of the spatial distribution of the probe
576: pulse is illustrated in Fig. 6a,b: in some areas of the medium, the
577: field in the tail of the pulse is different from zero. This is
578: believed to be the result of the spatial compression of the probe
579: caused by the slowing down of the group velocity of the probe pulse
580: \cite{Lukin}.
581:
582: In the case of $K_1>K_2$, the adiabaticity condition is maintained
583: over the $z_c$ range, which can also be much longer than the
584: length of the linear absorption. The leading edge of the probe
585: pulse undergoes gradual depletion and the pulse amplitudes display
586: only small changes in that range. Outside that range, the pulse
587: splits into several peaks. This occurs due to the nonadiabaticity
588: of interaction (see Fig. 6c where numerical simulations are
589: presented for $g_{1,2}(\tau)$. Here the first three curves
590: correspond to the adiabatic interaction).
591:
592: The pulse evolution, described above, is determined by spatial and
593: temporal behaviour of the atomic Raman coherence $\rho_{20}$ shown
594: in Fig.7 (see also \cite{7a}). Figures 7a,b reveal an unusual
595: spatial and temporal behavior of the atomic coherence, which we
596: interpret as the slowing down, stopping and localization of the
597: atomic coherence in the medium. The probe pulse is transferred
598: into and stored in the collective atomic excitation under the
599: control of the coupling. One can say that a phase grating is
600: created in the atomic medium. The phase grating is preserved
601: throughout the entire period of relaxation of the atomic
602: coherence. During that period of time, the information stored in
603: the atomic excitations can be transferred back to the radiation
604: using another coupling pulse of the same or of a different
605: frequency.
606:
607: We find these effects to be similar to the ones predicted and
608: demonstrated in \cite{Phillips,Liu,Lukin}, but for some
609: differences as indicated below. In our case, both the coupling
610: pulse and the probe pulse are strong. Unlike \cite{Lukin}, we used
611: the boundary conditions for the pulse envelopes yielding the time
612: distribution of pulses at the medium boundary, $z=0$. We believe,
613: this condition is more natural than the one used in \cite{Lukin}
614: where the authors use the probe pulse distribution in the medium
615: at fixed time as the initial condition. Also they do not take into
616: consideration evolution of the coupling pulse. The effect takes
617: place in the case $K_1\leq K_2$ and is not observed in the case of
618: $K_1>K_2$.
619:
620: \section{CONCLUSION}
621:
622: The propagation of two short overlapping pulses with durations
623: $T_2>T_1$ in optically thick three-level media under CPT
624: conditions has been studied for the general case of unequal
625: transition oscillator strengths. An analytical solution has been
626: obtained for the set of reduced wave equations under the adiabatic
627: following condition. Also it has been shown how spatial evolution
628: of pulses depends on the oscillator strength ratio.
629:
630: The condition of adiabaticity provided at the medium entrance
631: preserves for any value of propagation lengths if $K_1\leq K_2$
632: ($q<q_{max}$) and breaks down at $K_1>K_2$ ($q>q_{max}$). In the
633: range $q<q_{max}$, the probe pulse is completely depleted and
634: reemitted into the coupling pulse during propagation. This is not
635: possible in the case of $q>q_{max}$. It has been established that
636: to provide for the adiabaticity condition in an optically thin
637: medium, the restriction $T_2/T_1> \sqrt{2}$ has to be ensured,
638: whereas in a thick medium $T_2/T_1> \sqrt{3}$.
639:
640: We also have studied the spatial behavior of the atomic coherence
641: $\rho_{20}$, which plays a significant role, for example, in
642: nonlinear mixing processes. It has been found that a strong
643: coherence can be maintained over a length equal to several
644: hundreds of thousand of one-photon absorption lengths during
645: propagation. The effect of localization of the atomic coherence is
646: demonstrated.
647:
648: %\acknowledgements
649:
650: \begin{references}
651:
652: \bibitem{Harr} S.E.Harris, Physics Today {\bf 50}, No.7, 36 (1997).
653:
654: \bibitem{Raut} S.G.Rautian, A.M.Shalagin, {\it Kinetic Problem of Nonlinear
655: Spectroscopy} (North-Holland: Amsterdam, 1991); A.K.Popov, {\it
656: Vvedenie v nelineinuyu spectroscopiyu} (Novosibirsk: Nauka, 1983).
657:
658: \bibitem{Agap} B.D.Agap'ev, M.B.Gornyi, B.G.Matisov, Yu.V.Rozhdestvensky,
659: Usp. Fiz. Nauk {\bf 163}, 1 (1993).
660:
661: \bibitem{Arim} E.Arimondo, in {\it Progress in Optics} edited by
662: E.Wolf (Elsevier, Science, Amsterdam, 1996), {\bf 35}, p. 257.
663:
664: \bibitem{Berg} K.Bergman, H.Theuer, B.W.Shore, Rev.Mod.Phys., {\bf 70},
665: 1003 (1998).
666:
667: \bibitem{Yam} S.E.Harris, Y.Yamamoto, Phys. Rev. Lett.,
668: {\bf 81}, 4611 (1998).
669:
670: \bibitem{Hau} L.V.Hau, S.E.Harris, Z.Dutton, C.H.Behroozi. Nature (London)
671: {\bf 397}, 594 (1999).
672:
673: \bibitem{Kash} M.M.Kash, V.A.Sautenkov, A.S.Zibrov {\it et al.},
674: Phys. Rev. Lett., {\bf 82}, 5229 (1999).
675:
676: \bibitem{Phillips} D.F.Phillips, A.Fleischhauer, A.Mair, R.L.Walsworth,
677: M.D.Lukin, Phys.Rev.Lett. {\bf 86}, 783 (2001).
678:
679: \bibitem{Liu} C.Liu, Z.Dutton, C.H.Behroozi, L.V.Hau, Nature (London)
680: {\bf 409}, 490 (2001).
681:
682: \bibitem{1} S.E.Harris, Phys. Rev. Lett., {\bf 72}, 52 (1994).
683:
684: \bibitem{2} J.H.Eberly, M.L.Pons, H.R.Haq, Phys. Rev. Lett. {\bf 72},
685: 56 (1994).
686:
687: \bibitem{3} J.H.Eberly, Quantum Semiclass. Opt. {\bf 7}, 373 (1995).
688:
689: \bibitem{4} J.H.Eberly, A.Rahman, R. Grobe, Phys. Rev. Lett. {\bf 76}, 3687 (1996).
690:
691: \bibitem{5} R.Grobe, F.T.Hioe, J.H.Eberly, Phys. Rev. Lett. {\bf 73},
692: 3183 (1994).
693:
694: \bibitem{6} M.Fleischhauer, A.S.Manka, Phys. Rev. A {\bf 54}, 794 (1996).
695:
696: \bibitem{7} V.G.Arkhipkin, D.V.Manushkin, V.P.Timofeev.
697: Quantum Electronics, {\bf 28}, 1055 (1998).
698:
699: \bibitem{7a} V.G.Arkhipkin, I.V.Timofeev.
700: in {\it Proceedings of International Conference on Laser Physics
701: and Spectroscopy, 1999}, edited by V.L.Derbov, L.A.Melnikov,
702: V.P.Ryabukho, SPIE {\bf 4002}, p. 45.
703:
704: \bibitem{Kasapi} A.Kasapi, M.Jain, G.Y.Yin, S.E.Harris, Phys. Rev. Lett.
705: {\bf 74}, 2447 (1995).
706:
707: \bibitem{8} G.G.Grigoryan, Y.T.Pashayan, in {\it Proceedings of
708: International Conference on New Trends in Atomic and Molecular
709: Spectroscopy, 1999}, edited by G.G.Gursadian, A.V.Karmenyan, SPIE
710: {\bf 4060}, p. 21.
711:
712: \bibitem{Hioe} F.T.Hioe, R.Grobe, Phys. Rev. Lett.
713: {\bf 73}, 2559 (1994).
714:
715: \bibitem{9a} M.Jain, X.Hia, G.Y.Yin, A.J.Merriam, S.E.Harris,
716: Phys. Rev. Lett. {\bf 77}, 21 (1996); A.J.Merriam, S.J.Sharpe,
717: H.Xia, D.Manuszak, G.Y.Yin, S.E.Harris, Opt. Lett. {\bf 24}, 9
718: (1999).
719:
720: \bibitem{9b} V.G.Arkhipkin, D.V.Manushkin, S.A.Myslivets,
721: A.K.Popov. Quantum Electronics, {\bf 28}, 637 (1998).
722:
723: \bibitem{10} J.R.Kuklinski, U.Gaubats, F.T.Hioe, K.Bergman,
724: Phys. Rev. A. {\bf 40}, 6471 (1989).
725:
726: \bibitem{11} U.Gaubats, P.Rudecki, S.Schiemann and
727: K.Bergman, J.Chem Phys. {\bf 92}, 5363 (1990).
728:
729: \bibitem{Lukin} M.Fleischhauer, M.D.Lukin,
730: Phys. Rev. Lett. {\bf 84}, 5094 (2000).
731:
732: \bibitem{13} L.Allen, J. Eberly. {\it Optical resonance and two-level atoms}
733: (New York, 1975).
734:
735: \bibitem{14} R.Z.Sagdeev, D.A.Usikov, G.M.Zaslavsky, {\it Nonlinear Physics}
736: (Harwood, Chur, Swtzerland, 1988).
737: \end{references}
738:
739: \newpage
740: \centerline{FIGURES}
741:
742: FIG. 1. (a) The three-level Lambda-type system coupled by two
743: resonant pulses with Rabi frequencies $G_1$ and $G_2$. (b) The
744: shapes of probe $G_1$ and coupling $G_2$ pulses at the medium
745: entrance.
746:
747: FIG. 2. The time evolution of (a) level populations
748: $|a_{0,2}(\tau)|^2$ and the atomic Raman coherence
749: $|\rho_{20}(\tau)|$, (b) the mixing angle $\theta(\tau)$ in an
750: optically thin medium for the Gaussian pulses. $G_1^0 T_1=G_2^0
751: T_1=20$, (a) $T_2/T_1=3$; (b) $T_2/T_1=\sqrt{2}$ (A), $\sqrt{3}$
752: (B), $3$ (C), $10$ (D).
753:
754: FIG. 3. The vector model of adiabatic interaction of two short
755: pulses with a three-level Lambda-type system.
756:
757: FIG. 4. The characteristic curves for the Eq.(\ref{eq11}): (a)
758: $K_1/K_2=1$, (b) $K_1/K_2=0.25$, (c) $K_1/K_2=4$.
759:
760:
761: FIG. 5. The time evolution of the parameter $\theta$ for different
762: relationships between $K_1$ and $K_2$ at different propagation
763: lengths. (a) $K_1=K_2$, (b) $K_1/K_2=0.25$, (c) $K_1/K_2=4$.
764: $G_{1,2}^0 T_1=20$, $T_2/T_1=3$. Here and in all the other figures
765: the time $\tau$ is measured in the units of the pulse duration
766: $T_1$, and the propagation length $z$ of pulses in the medium is
767: measured in the units of the length of linear absorption of the
768: probe radiation determined in accordance with the Beer's low. In
769: (c) the numerical solution for the case $K_1/K_2=4$ is presented.
770:
771: FIG. 6. The time evolution of the normalized Rabi frequencies
772: $g_{1,2}=G_{1,2}/\sqrt{{G_1^0}^2+{G_2^0}^2}$ of the probe and
773: coupling pulses for different relationships between $K_1$ and
774: $K_2$ at different propagation lengths. (a) $K_1=K_2$, (b)
775: $K_1/K_2=0.25$, (c) $K_1/K_2=4$. $g_{1,2}^0=20$, $T_2/T_1=3$.
776:
777: FIG. 7. The time evolution of the atomic Raman coherence for
778: different relationships between $K_1$ and $K_2$ at different
779: propagation lengths. (a) $K_1=K_2$, (b) $K_1/K_2=0.25$, (c)
780: $K_1/K_2=4$. $G_{1,2}^0T_1 =20$, $T_2/T_1=3$.
781:
782: \end{document}
783: