1:
2: \documentstyle[preprint,aps,graphics]{revtex}
3: \tightenlines
4: \title {Coherent processing of a light pulse stored in a medium of four-level atoms}
5: \author{A. Raczy\'nski$^{1}$
6: \footnote{email: raczyn@phys.uni.torun.pl}, J. Zaremba$^{1}$ and
7: S. Zieli\'nska-Kaniasty$^{2}$}
8: \address{$^{1}$Instytut Fizyki, Uniwersytet Miko\l aja Kopernika,
9: ul.Grudzi\c{a}dzka 5,
10: 87-100 Toru\'n, Poland,\\$^{2}$ Instytut Matematyki i Fizyki,
11: Akademia Techniczo-Rolnicza, Al. Prof. S. Kaliskiego 7, 85-796
12: Bydgoszcz, Poland.}
13: \begin{document}
14:
15: \maketitle
16:
17: \begin{abstract}
18: It is demonstrated that the properties of light stored in a
19: four-level atomic system can be modified by an additional control
20: interaction present during the storage stage. By choosing the
21: pulse area of this interaction one can in particular continuously
22: switch between two channels into which light is released.
23: \\
24: PACS numbers: 42.50.Gy, 03.67.-a
25: \end{abstract}
26: \newpage
27: In recent years a number of nonlinear optical phenomena in weak
28: fields have been intensively investigated. They are connected with
29: a coherent excitation of an atomic medium, the optical properties
30: of which can then be drastically modified. An important example is
31: the electromagnetically induced transparency \cite{w1}, possibly
32: additionally controlled in time, which is manifested as light
33: slow-down or even its storage and a controlled release
34: \cite{w2,w3,w4,w5,w6,w7}. The simplest realization of those
35: effects occurs in atomic systems with three active states in a
36: $\Lambda$ configuration. Extending such a configuration by adding
37: a coherently coupled fourth level opens new possibilities of an
38: external control of such processes \cite{w8,w9,w10,w11} . In our
39: previous paper \cite{w12} we have shown that in a double $\Lambda$
40: system it is possible to change the light frequency of the stored
41: light or even to release two pulses of different frequencies by
42: applying two control fields, properly chosen and delayed in time.
43:
44: Light storing in the form of atomic coherences joins the
45: advantages of the efficiency of light as an information carrier
46: and of an atomic medium as an information store. Thus it might be
47: a question of a practical importance how to modify in a controlled
48: way the properties of the released light by processing the atomic
49: medium during the storage stage. In the present work we
50: investigate two possibilities of controlling the released pulse or
51: pulses by modifying the atomic coherence due to the stopped light.
52: In the case (a) the lower, initially empty state of a typical
53: $\Lambda$ system is additionally coupled to a fourth state by
54: another laser. In the case (b) the initially occupied state of the
55: $\Lambda$ system is coupled by some kind of an effective
56: interaction (e.g., magnetic or two-photon electric coupling) with
57: another state of the same parity. We show that the Rabi
58: oscillations due to the new interactions modify in a coherent way
59: the properties of the released light.
60:
61: We consider a quasi one-dimensional medium of four-level atoms
62: with three lower metastable states $|b>$ and $|c>$ and $|d>$ and
63: an upper state $|a>$ (Fig.\ref{fig1}). The position of an atom is
64: described by the variable $z$, which is considered continuous. The
65: states $b,a$ and $c$ constitute a typical $\Lambda$ system with
66: the weak signal field 1 and a strong control field 2. In the case
67: (a) the state $c$ is additionally coupled with another state $d$
68: by a laser field 4. In the case (b) the state $a$ is coupled with
69: a fourth state $d$ by a weak signal field 3 while the states $b$
70: and $d$ are coupled by some effective coupling 4. The interaction
71: Hamiltonian in the case (a) is $V=-\hat{d}\sum_{j=1,2,4}
72: \epsilon_{j}\cos\phi_{j}$, with $\phi_{j}=\omega_{j} t-k_{j} z$,
73: $\epsilon_{j}=\epsilon_{j}(z,t)$ being slowly varying envelopes
74: and $\hat{d}$ - the dipole moment operator; we have assumed that
75: all the fields have the same linear polarization. In the case (b)
76: the Hamiltonian reads $V=-\hat{d}\sum_{j=1,2,3}
77: \epsilon_{j}\cos\phi_{j}+ [(iU)|b><d|-(iU)|d><b|] \cos \phi_{4}$,
78: where for numerical reasons we have made a physically
79: insignificant assumption that the matrix element $iU$ of the
80: effective interaction 4 is imaginary. The matrix elements of the
81: dipole moment $d_{1}=(\hat{d})_{ab}$, $d_{2}=(\hat{d})_{ac}$,
82: $d_{3}=(\hat{d})_{ad}$, $d_{4}=(\hat{d})_{cd}$ are taken real.
83: Resonant conditions concerning all the couplings are assumed.
84:
85: The evolution equation $i\hbar \dot{\rho}=[H,\rho]$ for the
86: density matrix $\rho=\rho(z,t)$ for an atom at position $z$, after
87: making the rotating-wave approximation, transforming-off the
88: rapidly oscillating factors: $\rho_{ab}=\sigma_{ab}
89: \exp(-i\phi_{1})$, $\rho_{ac}=\sigma_{ac} \exp(-i\phi_{2})$,
90: $\rho_{bc}=\sigma_{bc} \exp[i(\phi_{1}-\phi_{2})]$,
91: $\rho_{db}=\sigma_{db} \exp(-i\phi_{4})$, $\rho_{dc}=\sigma_{dc}
92: \exp[i(\phi_{2}-\phi_{3})]$, $\rho_{ad}=\sigma_{ad}
93: \exp[i(\phi_{3})]$, $\rho_{ii}=\sigma_{ii}$, and after adding
94: relaxation terms describing the spontaneous emission within the
95: system, takes the form in the more complicated case (b)
96:
97: \begin{eqnarray}
98: i \dot{\sigma}_{aa}&=&
99: -\frac{1}{2\hbar}\epsilon_{1}d_{1}(\sigma_{ba}-\sigma_{ab})
100: +\frac{1}{2}\Omega_{2}(\sigma_{ca}-\sigma_{ac})
101: -\frac{1}{2\hbar}\epsilon_{3}d_{3}(\sigma_{da}-\sigma_{ad})
102: -i(\Gamma^{a}_{b}+\Gamma^{a}_{c}+\Gamma^{a}_{d})
103: \sigma_{aa},\nonumber\\
104: i \dot{\sigma}_{bb}&=&
105: -\frac{1}{2\hbar}\epsilon_{1}d_{1}(\sigma_{ab}-\sigma_{ba})
106: +\frac{1}{2\hbar} iU (\sigma_{bd}+\sigma_{db})
107: +i \Gamma^{a}_{b} \sigma_{aa} ,\nonumber\\
108: i \dot{\sigma}_{cc}&=&
109: \frac{1}{2}\Omega_{2}(\sigma_{ac}-\sigma_{ca})
110: +i\Gamma^{a}_{c} \sigma_{aa},\nonumber\\
111: i \dot{\sigma}_{dd}&=&
112: -\frac{1}{2\hbar}\epsilon_{3}d_{3}(\sigma_{ad}-\sigma_{da})
113: -\frac{1}{2\hbar} iU (\sigma_{bd}+\sigma_{db})+i \Gamma^{a}_{d}
114: \sigma_{aa},\nonumber\\
115: i \dot{\sigma}_{ab}&=&-\frac{1}{2\hbar}\epsilon_{1} d_{1} (\sigma_{bb}-
116: \sigma_{aa})+\frac{1}{2}\Omega_{2}\sigma_{cb}
117: -\frac{1}{2\hbar}\epsilon_{3}d_{3}\sigma_{db}
118: +\frac{1}{2\hbar}iU\sigma_{ad}
119: - \frac{i}{2}(\Gamma^{a}_{b}+\Gamma^{a}_{c}+\Gamma^{a}_{d})
120: \sigma_{ab},\\
121: i \dot{\sigma}_{ac}&=&-\frac{1}{2\hbar}\epsilon_{1} d_{1}
122: \sigma_{bc}+
123: \frac{1}{2} \Omega_{2}(\sigma_{cc}-\sigma_{aa})
124: - \frac{1}{2\hbar} \epsilon_{3}d_{3}\sigma_{dc}-
125: \frac{i}{2}(\Gamma^{a}_{b}+\Gamma^{a}_{c}+\Gamma^{a}_{d})
126: \sigma_{ac},\nonumber\\
127: i \dot{\sigma}_{ad}&=&-\frac{1}{2\hbar}\epsilon_{1} d_{1}
128: \sigma_{bd}
129: +\frac{1}{2}\Omega_{2}\sigma_{cd}
130: -\frac{1}{2\hbar} \epsilon_{3}d_{3}(\sigma_{dd}-\sigma_{aa})
131: -\frac{1}{2\hbar}iU\sigma_{ab}
132: -\frac{i}{2}(\Gamma^{a}_{b}+\Gamma^{a}_{c}+\Gamma^{a}_{d})
133: \sigma_{ad},\nonumber\\
134: i \dot{\sigma}_{bc}&=&-\frac{1}{2\hbar}\epsilon_{1} d_{1}
135: \sigma_{ac}
136: -\frac{1}{2} \Omega_{2}\sigma_{ba}
137: +\frac{1}{2\hbar}iU\sigma_{dc},\nonumber\\
138: i \dot{\sigma}_{bd}&=&
139: -\frac{1}{2\hbar}\epsilon_{1}d_{1}\sigma_{ad}
140: +\frac{1}{2\hbar}\epsilon_{3}d_{3}\sigma_{ba} + \frac{1}{2\hbar}
141: iU(\sigma_{dd}-\sigma_{bb}),\nonumber\\
142: i \dot{\sigma}_{cd}&=&
143: \frac{1}{2} \Omega_{2}\sigma_{ad}
144: +\frac{1}{2\hbar}\epsilon_{3}d_{3}\sigma_{ca}- \frac{1}{2}
145: iU\sigma_{cb},\nonumber
146: \end{eqnarray}
147: where $\Gamma^{a}_{b}$ is the decay rate of the state $|a>$ to
148: $|b>$, etc., and $\Omega_{2}=-\epsilon_{2}d_{2}/\hbar$ is the Rabi
149: frequency corresponding to the driving field 2.
150:
151: The corresponding equations for the case (a) are obtained from the
152: above set (1) by interchanging the indices $b\leftrightarrow c$,
153: $1\leftrightarrow 2$, by setting $\epsilon_{3}=0$, $d_{3}=0$,
154: $\Gamma^{a}_{d}=0$, by replacing the effective coupling $iU$ by
155: $=-\epsilon_{4} d_{4}$ (when multiplied by $\sigma_{dc}$,
156: $\sigma_{da}$, $\sigma_{ac}$, $\sigma_{db}$, $\sigma_{bc}$,
157: $\sigma_{cc}$ and $\sigma_{dd}$) and by changing the sign of
158: $\omega_{4}$.
159:
160: The propagation equations for the signal field 1 (in the case (a))
161: and for both signal fields (1) and (3) (in the case (b)), in the
162: slowly varying envelope approximation and in the conditions of the
163: resonance read
164: \begin{eqnarray}
165: \frac{\partial \epsilon_{1}}{\partial z}+\frac{1}{c}
166: \frac{\partial \epsilon_{1}}{\partial t}=-i N
167: d_{1}\frac{\omega_{1}}{\epsilon_{0}c} \sigma_{ba},\nonumber\\
168: \frac{\partial \epsilon_{3}}{\partial z}+\frac{1}{c}
169: \frac{\partial \epsilon_{3}}{\partial t}=-i N
170: d_{3}\frac{\omega_{3}}{\epsilon_{0}c} \sigma_{da},
171: \end{eqnarray}
172: where $N$ is the atom density and $\epsilon_{0}$ is the vacuum
173: electric permittivity. Similarly as in earlier papers, we have
174: neglected propagation effects for the driving fields, i.e. we take
175: $\epsilon_{2,4}=\epsilon_{2,4}(t)$.
176:
177: Eqs (1) and (2) have been solved numerically in the moving window
178: frame of reference: $t'=t-z/c,z'=z$. Switching the driving field 2
179: on and/or off was modeled by a hyperbolic tangent, while the
180: additional pulse 4 was taken rectangular. The initial probe pulse
181: was taken as the sine square shape
182: \\$\epsilon_{1}(0,t)=\epsilon_{10}
183: \sin^{2}[\pi(t-\tau_{1})/(\tau_{2}-\tau_{1})]\Theta(t-\tau_{1})
184: \Theta(\tau_{2}-t)$,
185: while the initial condition for the atomic part was
186: $\sigma_{bb}(z,0)=1$, with other matrix elements equal to zero.
187:
188: We have performed model computations for data being of realistic
189: orders of magnitude, however without making attempt to imitate any
190: real atom. The atomic energies were $E_{a}$= -0.10 a.u.,
191: $E_{b}=$-0.20 a.u. $E_{c}=$-0.18 a.u. with $E_{d}=$ -0.22 a.u (in
192: the case (a)) and $E_{d}=E_{b}+10^{-7}$ a.u. in the case (b) (the
193: latter value is of order of a magnetic energy splitting). The
194: relaxation rates for the spontaneous emission from the level
195: $E_{a}$ to $E_{b}$, $E_{c}$ and in the case (b) also to $E_{d}$
196: were taken equal to $2.4\times 10^{-9}$ a.u., from which the
197: dipole moments have been calculated. The dipole moment for the
198: electric transition $c\leftrightarrow d$ was taken $-2.74\times
199: 10^{-1}$ a.u., which corresponded to a negligible width of the
200: level $E_{c}$. The length of the atomic sample was
201: $2.5\times10^{7}$ a.u. (1.3 mm) in the case (a) and $3\times
202: 10^{7}$ a.u. (1.6 mm) in the case (b) and its density $3\times
203: 10^{-13} $ a.u. $(2\times 10^{12}$ cm$^{-3}$). The initial signal
204: pulse length was $10^{11}$ a.u. (2.4 $\mu$s) and
205: $\epsilon_{10}=10^{-10}$ a.u. (which corresponded to the power
206: density of 3.5$\times10^{-4}$ Wcm$^{-2}$); the maximum value of
207: the amplitude of the control field 2 was $1.2 \times10^{-9}$ a.u.
208: (50 mWcm$^{-2}$). The values of the $b-d$ effective coupling $U$
209: in the case (b) were of order of $10^{-10}$ a.u. while in the case
210: (a) we took a coupling with $\epsilon_{4}=2\times 10^{-9}$ a.u.
211:
212: In Fig.\ref{fig2} we show the released part of the pulse 1 as a
213: function of the local time $t'$ for different values of the area
214: of the pulse 4 (case(a)). The pulse can be lowered, completely
215: damped or its sign reversed depending on the final phase of the
216: Rabi oscillations between the levels c and d. Of course the final
217: results do not depend on particular time instants of switching the
218: interaction 4 on and off, provided that the pulse arrived after
219: the signal pulse 1 has been stored and before the release stage
220: has started. The presence of the Rabi oscillations becomes clearly
221: visible in the situation in which the control pulses 2 and 4
222: partially overlap. In this case, with the latter pulse being now
223: by an order of magnitude stronger than before, the restored pulse
224: is constructed of parts freed in those intervals of the Rabi
225: period in which the coherence $\sigma_{bc}$ differs significantly
226: from zero. The Rabi oscillations between the levels c and d are
227: thus imposed on the leaving signal pulse (see Fig.\ref{fig3}).
228:
229: The Rabi oscillations due to the additional control field may be
230: used not only to destroy in a reversible way the atomic coherence
231: $\sigma_{ab}$ necessary to release the pulse 1 (case (a)) but also
232: to create a new coherence $\sigma_{ad}$ which can be converted
233: into a new pulse 3 (case (b)). In Fig.\ref{fig4} we show the
234: shapes of the two signal pulses 1 and 3 for different values of
235: the area of the control pulse 4, switched on and off in the
236: storage stage. If the area is a multiple of $\pi$ only the pulse 1
237: is released, with its sign being changed in the case of an odd
238: multiple. If the area is an odd multiple of $\frac{\pi}{2}$ only
239: the pulse 3 appears, alternatively with a changed sign. For pulse
240: areas being not a particular multiple of $\frac{\pi}{2}$ both
241: pulses 1 and 3 are released, their heights being under control.
242:
243: As in the previous papers the problem can be analyzed in terms of
244: dark state polaritons. Such an analysis allows one to describe the
245: whole process of light storing in a single $\Lambda$ system in
246: terms of a shape preserving solution of the Maxwell-Bloch
247: equations (Eqs (1,2)), the components of which, i.e. the signal
248: field and the atomic coherence, adiabatically turn one into
249: another. In the case of a four-level system the evolution could
250: not in general be fully adiabatic, which means that bright-state
251: polaritons must appear at some stage of the process and are later
252: damped \cite{w12}. Thus the dark-state polaritons at the initial
253: and final stages are not identical.
254:
255: The approach of Ref. \cite{w12} generalized in our case (b) leads
256: to the following results. One can attempt to solve Eqs (1) and (2)
257: perturbatively (as concerns signal fields), in an adiabatic and
258: relaxationless approximation. The stage of light stopping occurs
259: as in the case of a three-level system: the polariton solution
260: \begin{equation}
261: \Psi=\frac{\Omega_{2}\epsilon_{1}+\frac{2\omega_{1}N
262: d_{1}}{\epsilon_{0}}\sigma_{bc}}
263: {\sqrt{\Omega_{2}^{2}+\frac{2\omega_{1}Nd_{1}^{2}}
264: {\hbar\epsilon_{0}}}}[-\frac{d_{2}}{|d_{2}|}]
265: \end{equation}
266: describes an adiabatic conversion of the pulse 1 into the
267: coherence $\sigma_{bc}$ and the sign correction guarantees that
268: for large $|\Omega_{2}|$ we get $\psi=\epsilon_{1}$ (we assume
269: that $\epsilon_{2}>0$). After pulse stopping, say at time instant
270: $t_{1}$, the control pulse 4 is switched on and is present up to
271: the time instant $t_{2}$. As a consequence the density matrix
272: evolves and at $t=t_{2}$ we obtain in the case (b)
273: $\sigma_{bb}=\cos^{2}\theta$, $\sigma_{dd}=\sin^{2} \theta$,
274: $\sigma_{bd}=-\sin\theta \cos\theta$,
275: $\sigma_{bc}=\sigma_{bc}(t_{1}) \cos\theta$,
276: $\sigma_{dc}=-\sigma_{bc}(t_{1})\sin\theta$, where
277: $\theta=\frac{U(t_{2}-t_{1})}{2\hbar}$ is the pulse area.
278:
279: At time instant $t_{3}$ ($t_{3}>t_{2}$) the control field 2 is
280: switched on in order to release the trapped pulse. In the assumed
281: approximations $\sigma_{bb}$, $\sigma_{dd}$ and $\sigma_{bd}$ do
282: not change any more. The pulses 1 and 3 satisfy the equations
283: \begin{equation}
284: (\frac{\partial}{\partial t}+c\frac{\partial}{\partial z})
285: \epsilon_{j}=\frac{1}{\Omega_{2}}\frac{\partial}{\partial t}
286: \frac{1}{\Omega_{2}} \sum_{k} M_{jk} \epsilon_{k},
287: \end{equation}
288: where $j,k=1,3$ and
289: $M_{11}=-\frac{2N\omega_{1}d_{1}^{2}}{\epsilon_{0}
290: \hbar}\cos^{2}\theta$,
291: $M_{13}=\frac{2N\omega_{1}d_{1}d_{3}}{\epsilon_{0}
292: \hbar}\sin\theta\cos\theta$,
293: $M_{31}=\frac{2N\omega_{3}d_{3}d_{1}}{\epsilon_{0}
294: \hbar}\sin\theta\cos\theta$,
295: $M_{33}=-\frac{2N\omega_{3}d_{3}^{2}}{\epsilon_{0}\hbar }\sin^{2}
296: \theta$.
297:
298: Eqs (4) can be decoupled by a linear transformation. One of the
299: solutions can be shown, similarly as in previous papers, to be a
300: shape-preserving solution traveling with a time-dependent velocity
301: \begin{equation}
302: v(t)=c\frac{1}{1+\frac{2N(d_{1}^{2}\omega_{1}\cos^{2}\theta
303: +d_{3}^{2}\omega_{3}\sin^{2}\theta)}{\hbar\epsilon_{0}\Omega_{2}^{2}}}.
304: \end{equation}
305: (The other solution is zero due to the initial conditions at
306: $t=t_{2}$.) The polariton, being a combination of two fields and
307: two coherences, has the form
308: \begin{eqnarray}
309: \Psi=\sqrt{d_{1}^{2}\omega_{1}\cos^{2}\theta+d_{3}^{2}\omega_{3}
310: \sin^{2}\theta} \frac{\sqrt{\omega_{1}})}
311: {\sqrt{1+\frac{2N}{\epsilon_{0}\hbar \Omega_{2}^{2}}
312: (d_{1}^{2}\omega_{1}\cos^{2}\theta+d_{3}^{2}\omega_{3}\sin^{2}\theta)}}
313: \frac{d_{1}}{|d_{1}|} \nonumber\\
314: \times
315: [\frac{d_{1}\epsilon_{1}\cos\theta-d_{3}\epsilon_{3}\sin\theta}
316: {d_{1}^{2}\omega_{1}\cos^{2}\theta+d_{3}^{2}\omega_{3}\sin^{2}\theta}
317: +\frac{2N}{\epsilon_{0}\Omega_{2}}(\sigma_{bc}\cos\theta
318: -\sigma_{dc}\sin\theta)].
319: \end{eqnarray}
320: The solution (6) has been normalized so that it is equal to the
321: solution (3) at $t=t_{2}$, i.e.
322: $\Psi(t_{2})=\sqrt{\frac{2N\hbar\omega_{1}}{\epsilon_{0}}}
323: [\sigma_{bc}(t_{2})\cos\theta-\sigma_{dc}(t_{2})\sin\theta]$.
324: However, the final form of the polariton is, again for large
325: $\epsilon_{2}$, a combination of the signal fields with the
326: coefficients different from $\cos\theta$ and $\sin\theta$, except
327: in the case of $d_{1}=d_{3}$ and $\omega_{1}=\omega_{3}$. In
328: particular for $\theta=-\frac{\pi}{2}$ one finds that
329: $\Psi\rightarrow \sqrt{\frac{\omega_{1}}{\omega{3}}}
330: \epsilon_{3}\frac{d_{1}d_{3}}{|d_{1}d_{3}|}$ instead of
331: $\epsilon_{3}\frac{d_{1}d_{3}}{|d_{1}d_{3}|}$. As described in
332: detail in our previous paper \cite{w12}, this means that the
333: evolution cannot be fully adiabatic and bright state polaritons
334: (which are later damped) must be invoked. In the case (a) the
335: nonadiabatic element of the evolution is even more conspicuous:
336: only the part of the transformed coherence, namely that
337: proportional to $\cos\theta$ is active in the release stage and
338: turns adiabatically, being a shape-preserving solution, into
339: $\cos\theta \epsilon_{1}$. The other part of the excitation,
340: namely that proportional to $\sin\theta$, "survives" the release
341: stage inside the medium unless relaxations in the states $b$ and
342: $d$ are taken into account.
343:
344: In summary, we have demonstrated that a modification of the atomic
345: coherence due to a stopped light pulse can be used as a new way of
346: changing the properties of the released light. One can in
347: particular release two pulses of different frequencies or
348: polarizations, with their envelopes being regulated in a
349: continuous way. This may serve as a kind of a switch which allows
350: one e.g., to continuously steer the information by sending it to
351: particular channels or to temporarily hide it.
352:
353:
354: \acknowledgments{This work is a part of a program of the
355: National Laboratory of AMO Physics in Toru\'n, Poland}
356:
357: \newpage
358: \begin{references}
359: \bibitem{w1}
360: S. E. Harris, Phys. Today {\bf 507} (1997) 36.
361: \bibitem{w2}
362: M. Fleischhauer and M. D. Lukin, Phys. Rev. Lett. {\bf 84} (2000)
363: 5094.
364: \bibitem {w3}
365: M. Fleischhauer, S. F. Yelin and M. D. Lukin, Opt. Commun. {\bf
366: 179} (2000) 395.
367: \bibitem{w4}
368: O. Kocharovskaya, Y. Rostovtsev and M. O. Scully, Phys. Rev. Lett.
369: {\bf 86} (2001) 628.
370: \bibitem{w5}
371: D. F. Phillips, A. Fleischhauer, A. Mair, R. L. Walsworth and M.
372: D. Lukin, Phys. Rev. Lett. {\bf 86} (2001) 783.
373: \bibitem{w6}
374: C. Liu, Z. Dutton, C. H. Behroozi and L. V. Hau, Nature {\bf 409}
375: (2001) 490.
376: \bibitem{w7}
377: A. B. Matsko, O. Kocharovskaya, Y. Rostovtsev, G. R. Welch, A. S.
378: Zibrov, and M. O. Scully, Adv. in At. Mol. Opt. Phys. {\bf 46},
379: (2001) 191.
380: \bibitem{w8}
381: M. D. Lukin, S. F. Yelin, M. Fleischhauer and M. O. Scully, Phys.
382: Rev. A {\bf 60} (1999) 3225.
383: \bibitem {w9}
384: E. A. Korsunsky and D. V. Kosachiov, Phys. Rev. A. {\bf 60} (1999)
385: 4996.
386: \bibitem{w10}
387: M. D. Lukin, P. R. Hemmer and M. O. Scully, Adv. At. Mol. Opt.
388: Phys. {\bf 42} (2000) 347.
389: \bibitem{w11}
390: A. S. Zibrov, A. B. Matsko, O. Kocharovskaya, Y. V. Rostovtsev, G.
391: R. Welch and M. O. Scully, Phys. Rev. Lett. {\bf 88} (2002)
392: 103601.
393: \bibitem{w12}
394: A. Raczy\'nski and J. Zaremba, Opt. Commun. {\bf 209} (2002) 149
395: \end{references}
396: \newpage
397: \noindent
398:
399: \begin{figure}
400: %\includegraphics{f1.eps}
401: \caption{Level and coupling schemes; the indices 1 and 3 refer to
402: signal fields and 2 and 4 - to control fields.} \label{fig1}
403: \end{figure}
404:
405: \begin{figure}
406: %\includegraphics{f2.eps}
407: \caption{The field amplitude of the released pulse 1 as a function
408: of the local time $t'$ for different pulse areas of the control
409: field 4 in the case (a): curve 1: 0, curve 2: $\frac{\pi}{6}$,
410: curve 3: $\frac{\pi}{4}$, curve 4: $\frac{\pi}{3}$, curve 5:
411: $\frac{\pi}{2}$, curve 6: $\frac{3\pi}{4}$, curve 7: $\pi$,}
412: \label{fig2}
413: \end{figure}
414:
415: \begin{figure}
416: %\includegraphics{f3.eps}
417: \caption{The field amplitude of the released pulse 1 as a function
418: of the local time $t'$ in the case (a) for overlapping pulses 2
419: and 4: curve 1: part of the pulse transmitted before light
420: storing, curve 1a: the released pulse in the absence of the
421: additional coupling field, curve 1b: the released pulse in the
422: presence of the additional coupling field, curve 2: the control
423: field 2, curve 4: the additional coupling field 4. The values of
424: the fields 2 and 4 have been reduced by the factors of 40 and 20,
425: respectively. } \label{fig3}
426: \end{figure}
427:
428: \begin{figure}
429: %\includegraphics{f4.eps}
430: \caption{The amplitudes (in $10^{-11}$ a.u.) of the signal fields
431: 1: full line, and 3: dashed line, as functions of the local time
432: $t'$ (in $10^{11}$ a.u.) in the case (b) for different values of
433: the pulse area of the interaction 4.} \label{fig4}
434: \end{figure}
435:
436:
437:
438: \end{document}
439: