1: \documentclass[aps,prl,twocolumn]{revtex4}
2: \usepackage{epsfig}
3:
4: \begin{document}
5:
6: \title{
7: Coherent population trapping in ruby crystal at room temperature
8: }
9:
10: \author{
11: Roman Kolesov
12: }
13:
14: \address{
15: Department of Physics and Institute for Quantum Studies,
16: Texas A\&M University,\\
17: College Station, Texas~~77843-4242\\
18: % Institute of Applied Physics, RAS, 603600, Nizhny Novgorod, Russia
19: }
20: \date{\today}
21:
22:
23: \begin{abstract}
24: Observation of coherent population trapping (CPT) at ground-state Zeeman sublevels
25: of $Cr^{3+}$-ion in ruby is reported. The experiments are performed
26: at room temperature by using both nanosecond optical pulses and nanosecond trains
27: of ultrashort pulses. In both cases sharp drops in the resonantly induced
28: fluorescence are detected as the external magnetic field is varied.
29: Theoretical analysis of CPT in a transient regime due to pulsed action of optical
30: pulses is presented.
31: \end{abstract}
32: \maketitle
33:
34: In last few decades very much attention has been paid to studying optical interference phenomena
35: originating from the atomic coherence excited
36: by laser light while interacting with multi-level atomic media. These
37: phenomena include coherent population trapping (CPT) \cite{cpt_ref}, electromagnetically
38: induced transparency (EIT) \cite{eit_ref}, lasing without population inversion (LWI)
39: \cite{lwi_ref}, slow light \cite{slow}, and many others. The interest to this field
40: arises from a number of potential applications of these phenomena. Among
41: them are metrology, quantum information storage and processing, ultra-sensitive
42: magnetometry, development of lasers in vacuum ultraviolet (VUV), X-ray, and $\gamma$-ray
43: ranges, etc. However, CPT and related phenomena are mostly being studied in gaseous media
44: while for practical purposes solid-state materials are preferable. The advantages of solids
45: are high atomic density, absence of atomic diffusion, compactness, and robustness. At the
46: same time, decoherence processes in solid media are severe. Nevertheless, several
47: demonstrations of EIT \cite{Pr_YSO,EIT}, slow light \cite{Pr_YSO},
48: and non-degenerate four-wave mixing \cite{hakuta} in solids have been reported by several groups.
49: However, all those works were performed at temperatures close to the liquid helium one. This fact,
50: in turn, makes real solid-state applications of the above phenomena questionable.
51:
52: In this Letter I report the first, to my knowledge, experimental observation
53: of CPT resonances in crystalline solid at room temperature. Prior to that, I have to
54: discuss some general issues in order to identify the class of solid-state materials
55: in which the phenomena of EIT/CPT can be observed at room temperature. I will consider only
56: optical crystals doped with either rare-earth or transition metal ions possessing
57: discrete electronic states in the bandgap of the host
58: material. The electronic transitions between those states, both optical and Zeeman and/or
59: hyperfine, can be very narrow even at room temperature. The class of so-called S-state
60: ions doped into wide-bandgap optical crystals is of particular interest. It includes
61: $Cr^{3+}$, $Eu^{2+}$, $Fe^{3+}$, $Mn^{4+}$, $Gd^{3+}$, and several others. The most important
62: feature of these ions is that, when incorporated into a crystal, their first excited electronic
63: state lies several thousands $cm^{-1}$ above the ground one. Consequently,
64: the strongest relaxation process between the ground-state Zeeman or hyperfine sublevels,
65: owing to resonant inelastic scattering of phonons at low-energy electronic transition
66: (Orbach relaxation, \cite{Orbach}) is suppressed. As a result, the decoherence rate
67: at the transitions between Zeeman/hyperfine sublevels can be as low as a few $MHz$
68: even at room temperature. This fact is confirmed by a vast number of electron paramagnetic
69: resonance (EPR) measurements in which the ions mentioned above serve as paramagnetic probes.
70: Some of them, namely, $Cr^{3+}$, $Eu^{2+}$, and $Gd^{3+}$, offer very favorable optical
71: properties as well. In the following discussion, the case of ruby, i.e. $Cr^{3+}:Al_2O_3$,
72: is considered.
73:
74: Another important feature of doped solids is that the population decay from
75: the excited electronic state is much slower than decoherence at Zeeman/hyperfine
76: transition. This is exactly opposite to the case of alkali-metal
77: vapors in which most of experimental work on EIT/CPT is being performed. Therefore, one cannot
78: expect to populate the coherent superposition of ground-state sublevels, which
79: does not interact with the laser field under the conditions of CPT (so-called, ``dark'' state),
80: by decay of the excited optical level, as is the case in the alkali-metal experiments.
81: The only way of exciting ground-state coherence in that situation is to remove as much population
82: as possible from the ``bright'' state (the coherent superposition of sublevels orthogonal to the
83: ``dark'' one) by transferring it to the upper optical level. This should be done faster
84: than the coherence decays, i.e. laser excitation has to be pulsed.
85:
86: \begin{figure}
87: \center{
88: \includegraphics[width=6cm]{growth.eps}
89: \caption{\label{fig:growth} Temporal evolution of the upper optical state population
90: under pulsed excitation. The inset shows the model of a three-level atomic medium interacting
91: with light pulse.}
92: }
93: \end{figure}
94:
95:
96: Let us consider the following theoretical model. A three-level medium
97: is illuminated with a resonant laser pulse interacting with
98: both $1\leftrightarrow 3$ and $2\leftrightarrow 3$ optical transitions (see inset of Fig.\ref{fig:growth}).
99: In the rest of the paper we will be interested in the amount of population transferred into
100: level $3$ by a laser pulse or pulse train as a function of the energy separation between ground-state Zeeman sublevels
101: $1$ and $2$ which can be varied by means of external magnetic field. The following notations are introduced: $\Gamma$
102: is the homogeneous halfwidth of optical transitions, $\gamma$ is the Zeeman decoherence rate,
103: $\Omega$ is the laser field time-dependent Rabi frequency (dipole matrix elements of both optical
104: transitions are assumed equal). The initial populations of the two ground-state sublevels
105: are $\rho_{11}=\rho_{22}=1/2$, $\rho_{33}=0$. Assume also, that the pulse duration is longer
106: than $\Gamma^{-1},\gamma^{-1}$. This assumption somewhat contradicts the last statement
107: in the previous paragraph, but it is relevant under the experimental conditions described below.
108: Algebraically solving the density matrix equations for all the coherences, one comes to
109: the following linear equation for the upper state population:
110: \begin{equation}
111: \label{eq:pop3}
112: \frac{\rho_{33}}{dt}=P\left(1-\frac{P\left(P+\gamma\right)}{\delta^2+\left(P+\gamma\right)^2}\right)
113: \left(1-3\rho_{33}\right),
114: \end{equation}
115: where $P=2\Omega^2/\Gamma$ and $\delta$ is the frequency separation between the ground-state
116: sublevels. The second term in brackets represents the effect of Zeeman coherence on the
117: rate at which level $3$ is populated. Laser pulse starts acting on the medium at $t=0$.
118: As $t\rightarrow\infty$, all three levels become equally populated, as expected. However,
119: if at some instant $T$ the laser pulse is shut off, the final population of level $3$ is
120: smaller for $\delta=0$ than it is for $\delta\neq 0$ because of its lower growth rate
121: as illustrated in Fig.\ref{fig:growth}. The width of this transient CPT resonance
122: is determined by $P+\gamma$, i.e. by Zeeman decoherence rate and power broadening.
123:
124: Similar effects can be observed under the illumination of the medium with a short (nanosecond)
125: train of ultrashort (picosecond) pulses. In addition to CPT resonance at zero magnetic field,
126: one should expect drops in the upper state population as the Zeeman sublevel separation
127: becomes equal to the multiple of the pulse repetition rate in the train \cite{train_kochar}.
128: It worths mentioning that transient coherent effects in ruby crystal at room temperature
129: were first studied in 1974 \cite{magnetization}. In that work precessing magnetization
130: was induced by a train of ultrashort pulses. Significant enhancement in magnetization was
131: observed as the splitting between Zeeman sublevels of $Cr^{3+}$-ion became multiple of
132: the pulse repetiotion rate. However, no effect of Zeeman coherence on the absorptive
133: properties of the crystal was studied.
134:
135: \begin{figure}
136: \center{
137: \includegraphics[width=5cm]{levels.eps}
138: \caption{\label{fig:levels} Energy levels of $Cr^{3+}$-ion. Optical transitions allowed for
139: $\sigma$-polarization in ${\bf B}\perp c$ geometry are indicated.}
140: }
141: \end{figure}
142:
143: To understand how the above theoretical description applies to ruby, let us consider
144: the energy levels of $Cr^{3+}$ in that crystal (see Fig.\ref{fig:levels}).
145: Its ground state is described by the following spin Hamiltonian:
146: \begin{equation}
147: \label{eq:hamiltonian}
148: H=g_{||}B_{||}S_z+g_{\perp}{\bf B}_{\perp}{\bf S}_{\perp}+D\left(S_z^2-S\left(S+1\right)/3\right),
149: \end{equation}
150: where $g_{||}=1.982$ and $g_{\perp}=1.987$ are the ground-state $g$-factors in the directions parallel
151: and perpendicular to the crystal $c$-axis respectively, $B_{||}$ and ${\bf B}_{\perp}$ are
152: the corresponding magnetic field components, $2D=-11.47\;GHz$ is the zero-field
153: splitting of the ground state, and $S=3/2$ is the spin of $Cr^{3+}$. The excited
154: optical state is split into two spin-doublets separated by $29\;cm^{-1}$ giving rise
155: to $R_1$ ($694.3\;nm$) and $R_2$ ($692.9\;nm$) optical lines of ruby. The width of both
156: optical transitions is $11\;cm^{-1}$ while the oscillator strength is $7.5\times 10^{-7}$.
157:
158: The situation when ${\bf B}\perp c$-axis is of particular interest. The selection rules for
159: $\sigma^{\pm}$-polarizations of light propagating along the c-axis are summarized in Tables
160: \ref{tab:selection1} and \ref{tab:selection2} \cite{selection}. The matrix elements are normalized so that
161: the sum of their squares for each of the $\sigma^{\pm}$-polarizations is unity. One can easily see that there are a number
162: of Raman processes connecting levels $G_1$ and $G_2$ and $G_3$ and $G_4$. Since
163: the optical linewidth is greater than the separation between the ground-state Zeeman sublevels,
164: all $\Lambda$-systems will give their contributions to the Zeeman coherence excitation.
165: Therefore, if one uses a linearly polarized laser light, the net contribution of all
166: Raman processes to the coherence excitation at either $G_1\leftrightarrow G_2$ or
167: $G_3\leftrightarrow G_4$ transition is exactly zero. For example, at $R_2$-line the coherence
168: between levels $G_1$ and $G_2$ due to Raman process going though $E_3$ for $\sigma^{+}$-polarization
169: is exactly compensated by the one going though $E_4$ for $\sigma^{-}$-polarization.
170: The situation is the same for $R_1$-line. However, there is no such difficulty for circularly polarized light.
171: \begin{table}[t]
172: \begin{center}
173: \begin{tabular}{|c|cccc|}
174: \hline
175: State & $G_1$ & $G_2$ & $G_3$ & $G_4$ \\
176: \hline
177: $E_1$ & $\pm 0.25$ & $0.25$ & $-0.31$ & $\pm 0.31$ \\
178: $E_2$ & $0.25$ & $\pm 0.25$ & $\pm 0.31$ & $-0.31$ \\
179:
180: \hline
181: \end{tabular}
182: \end{center}
183: \caption{\label{tab:selection1}Selection rules for $\sigma^{+}$ (upper sign) and
184: $\sigma^{-}$ (lower sign) polarizations interacting with $R_1$-line.}
185: \end{table}\begin{table}[t]
186: \begin{center}
187: \begin{tabular}{|c|cc|cc|}
188: \hline
189: State & \multicolumn{2}{|c|}{$G_1$} & \multicolumn{2}{|c|}{$G_2$} \\
190: & $\sigma^{+}$ & $\sigma^{-}$ & $\sigma^{+}$ & $\sigma^{-}$ \\
191: \hline
192: $E_3$ & $0.43$ & $-0.07$ & $-0.43$ & $-0.07$ \\
193: $E_4$ & $-0.07$ & $-0.43$ & $0.07$ & $-0.43$ \\
194:
195: \hline
196: \end{tabular}
197: \end{center}
198: \caption{\label{tab:selection2}Selection rules for $\sigma^{\pm}$ polarizations interacting with $R_2$-line. Transitions from
199: the lower ground-state doublet to levels $E_3$ and $E_4$ are forbidden.}
200: \end{table}
201:
202: The experiments were performed on a $3\times 3\times 5\; mm^3$ dilute ruby crystal
203: ($0.002\%$ of $Cr^{3+}$) with $c$-axis being perpendicular to one of $3\times 5\;mm^2$ sides
204: (Scientific Materials Corp.) A homemade tunable Ti:Sapphire laser pumped
205: by the doubled $Nd:YLF$ pulsed laser (Photonics Industries, model GM-30) could operate
206: both in a long pulse regime (pulse duration $\approx 30\;ns$ FWHM) and in a mode-locked regime
207: delivering $\approx 30\;ns$ pulse trains with repetition rate $\approx 260\;MHz$.
208: The laser was mode-locked by using a cryptocyanine saturable absorber dissolved
209: in ethanol. The typical
210: output energy per pulse or per train was several tenths of a $mJ$ at a $100\;Hz$ repetition rate. The laser polarization
211: was linear. The typical shapes of pulses in both regimes are shown on the insets of
212: Figs.\ref{fig:long_exp} and \ref{fig:train_exp}. The shape of the pulse train
213: was taken with a $350\;ps$-risetime photodiode and a $500\;MHz$-bandwidth
214: oscilloscope. The duration of
215: each individual pulse in the train was not measured.
216: The crystal was placed into a varying magnetic field with ${\bf B}\perp c$-axis and
217: illuminated by laser pulses sent through
218: a $\lambda/4$-plate and focused by an $f=10\;cm$ lens. Presumably, the laser spot diameter
219: at the sample was $\sim 100\;\mu m$. The fluorescence was detected as a function
220: of slowly varying periodic magnetic field. The sweeping frequency of the magnetic field was $0.17\;Hz$.
221:
222: %\begin{figure}
223: %\center{
224: %\includegraphics[width=6cm]{pulses.eps}
225: %\caption{\label{fig:pulses} Typical shapes of laser pulses used in the experiment: a) multi-mode, no mode-locking and
226: %b) mode-locked with saturable absorber. Trace b) is taken with $350\;ps$ photodiode and $500\;MHz$-bandwidth oscilloscope.}
227: %}
228: %\end{figure}
229:
230: The experimental results for the case of long pulse excitation for both $R_1$ and $R_2$ lines
231: are shown in Fig.\ref{fig:long_exp}. The spectra were averaged over several hundreds of magnetic field sweeps.
232: In the case of $R_1$-line, there are two features present in the spectrum:
233: the broad one with the plateau on the bottom and a sharp one of much smaller amplitude. The smaller one
234: corresponds to less than $0.5\%$ decrease in the fluorescence amplitude and has a width of $18\;G$. It corresponds
235: to CPT resonance due to $G_1\leftrightarrow E_1,E_2 \leftrightarrow G_2$ $\Lambda$-system, thus its width
236: in terms of frequency units is $\approx 100\;MHz$ since $g$-factor for $G_1\leftrightarrow G_2$ transition
237: is $\approx 4$. The broad drop in fluorescence signal has an amplitude $\approx 7-8\%$. Its FWHM is close to
238: $\approx 950\;G$. The presence of plateau at the bottom of that feature indicates that this is a CPT resonance
239: in the $G_3\leftrightarrow E_1,E_2 \leftrightarrow G_4$ $\Lambda$-system with levels $G_3$ and $G_4$
240: being split very slowly as the magnetic field increases. Its width recalculated into frequency
241: domain is $\approx 40-50\;MHz$, i.e. this CPT resonance is, in fact, much sharper than the one discussed
242: above. The possible explanation for this fact is as follows. The main contribution to the EPR linewidth
243: of $Cr^{3+}$ in ruby is known to originate from magnetic dipole-dipole hyperfine interaction
244: with neighboring $^{27}Al$ nuclei \cite{broadening}. Thus, the magnetic-allowed Zeeman transition
245: $G_1\leftrightarrow G_2$ should be broadened much stronger than the magnetic-forbidden $G_3\leftrightarrow G_4$.
246: This also explains more than an order of magnitude difference in the amplitudes of those two CPT resonances, since
247: in $\Lambda$-system with longer-lived Zeeman coherence CPT should be more pronounced.
248:
249: For the $R_2$-line only one CPT resonance, represented by a sharp feature at the bottom of a smooth
250: background, is observed. It has the width of $18\;G$, i.e. the same as the one observed at $R_1$-line,
251: thus it corresponds to $G_1\leftrightarrow E_3,E_4 \leftrightarrow G_2$ $\Lambda$-system. The decrease
252: in the fluorescence is $3-4\%$ at the center of the resonance. The fact that this resonance is much stronger
253: is easily understandable, since 1) there is no background fluorescence excited from the other pair of
254: levels, $G_3$ and $G_4$, and 2) the transitions $G_1,G_2\leftrightarrow E_3,E_4$ are 3 times stronger
255: than $G_1,G_2\leftrightarrow E_1,E_2$. The smooth background originates, probably, from level
256: mixing in high magnetic fields and consequent change in the optical selection rules. That was confirmed
257: by removing the $\lambda/4$-plate, i.e. by making laser light linearly polarized. Under this condition all
258: three CPT resonances, discussed above, disappeared, as they were supposed to, while the smooth feature in question remained.
259:
260: \begin{figure}
261: \center{
262: \includegraphics[width=8cm]{long_exp.eps}
263: \caption{\label{fig:long_exp} CPT resonances in zero magnetic field: a) trace for $R_1$-line and b) trace for $R_2$-line.
264: Narrow peaks are shown in detail on insets. See text for details.}
265: }
266: \end{figure}
267:
268: \begin{figure}
269: \center{
270: \includegraphics[width=8cm]{train_exp.eps}
271: \caption{\label{fig:train_exp} CPT resonances with mode-locked laser: a) trace for $R_1$-line and b) trace for $R_2$-line.
272: See text for details.}
273: }
274: \end{figure}
275:
276: The results for trains of ultrashort pulses are shown in Fig.\ref{fig:train_exp}. In the spectrum obtained for
277: $R_1$-line additional resonances appear in pairs. They are not equidistant because of nonlinear dependence of the
278: levels $G_3$ and $G_4$ separation on the magnetic field strength. However, if recalculated into
279: frequency domain, their positions exactly correspond to the harmonics of laser repetition rate ($260\;MHz$).
280: The first three of them are shown on the inset of Fig.\ref{fig:train_exp}a. Their widths (FWHM) are in the range $35-38\;MHz$
281: and their positions, according to the fit, are $265\;MHz$, $533\;MHz$, and $804\;MHz$. The widths are in
282: rather good agreement with the result obtained for the $G_3\leftrightarrow G_4$ transition
283: in a long pulse regime. It was not possible to observe multiple CPT resonances corresponding to
284: $G_1\leftrightarrow E_3,E_4 \leftrightarrow G_2$ $\Lambda$-system with appreciable signal-to-noise ratio
285: due to very small amplitude of the CPT effect ($<0.5\%$) and rather unstable operation of laser
286: in a mode-locked regime. However, the signature of these resonances is the increased quasi-noise
287: at the bottom of the broad feature.
288:
289: For $R_2$-line almost 30 CPT resonances are indicated. In order to improve the signal-to-noise ratio, the original
290: spectrum was flipped with respect to the zero magnetic field and the two sets of data were combined together.
291: That explains the fact that the plot shown in Fig.\ref{fig:train_exp}b is completely symmetric. The magnetic field
292: separation between each pair of resonances is $\approx 45\;G$. This value corresponds to $252\;MHz$ of frequency
293: separation which is somewhat close to $260\;MHz$ of laser repetition rate. Due to rather poor signal-to-noise
294: ratio and rater low spectral resolution it is not possible to determine the linewidth of resonances exactly,
295: but it seems to correspond rather well to the value of $18\;G$ obtained in a long pulse regime.
296:
297: In summary, the first observation of CPT in a crystal at room temperature is reported. Possible applications of
298: the reported phenomena include room-temperature all-optical analogue of EPR spectroscopy in low magnetic fields,
299: suppression of excited state absorption in laser crystals \cite{ESA}, alternative technique of mode-locking
300: in lasers \cite{mode_lock}, etc.
301:
302: The author is grateful to F.Vagizov, P.Hemmer, Y.Rostovtsev, E.Kuznetsova, and O.Kocharovskaya, for stimulating discussions and to
303: S.Lissotchenko for technical assistance. The work was supported by NSF and AFOSR.
304:
305: \begin{thebibliography}{1}
306:
307: \bibitem{cpt_ref} E. Arimondo, in {\it Progress in Optics}, edited by E. Wolf
308: (Elsevier, Amsterdam, 1996), Vol. 35, pp. 257-354
309:
310: \bibitem{eit_ref} J.P. Marangos, J. Mod. Opt. {\bf 45}, 471 (1998); S.E.Harris, Phys. Today {\bf 50}, 36 (1997)
311:
312: \bibitem{lwi_ref} J. Mompart and R. Corbalan, J. Opt. B: Quantum Semiclassical Opt. {\bf 2}, R7 (2000)
313:
314: \bibitem{slow} A.B. Matsko et al., Adv. At., Mol., Opt. Phys. {\bf 46} 191 (2001)
315:
316: \bibitem{Pr_YSO} A.V. Turukhin et al., Phys. Rev. Lett. {\bf 88}, art. no. 023602 (2002)
317:
318: \bibitem{EIT} Y. Zhao, C.K. Wu, B.S. Ham, M.K. Kim, and E. Awad, Phys. Rev. Lett. {\bf 79} 641 (1997);
319: C.J. Wei and N.B. Manson, Phys. Rev. A {\bf 60} 2540 (1999); P.R. Hemmer,
320: A.V. Turukhin, M.S. Shahriar, and J.A. Musser, Opt. Lett. {\bf 26} 361 (2001);
321: M. Phillips and H.L. Wang, Phys. Rev. Lett. {\bf 89} art. no. 186401 (2002);
322: G.B. Serapiglia, E. Paspalakis, C. Sirtori, K.L. Vodopyanov, and C.C. Phillips, Phys. Rev. Lett. {\bf 84} 1019 (2000)
323:
324: \bibitem{hakuta} J.Q. Liang, M. Katsuragava, F.L. Kien, and K. Hakuta, Phys. Rev. Lett. {\bf 85} 2474 (2000)
325:
326: \bibitem{Orbach} R. Orbach, Proc. Roy. Soc. (London) {\bf A264} 458 (1961)
327:
328: \bibitem{train_kochar} O.A. Kocharovskaya and Ya.I. Khanin, Zh. Eksp. Teor. Fiz. {\bf 90} 1610 (1986)
329: [Sov. Phys. JETP {\bf 63} 945 (1986)]; V.A. Sautenkov et al., Phys. Rev. A, to be published
330:
331: \bibitem{magnetization} Y. Fukuda, Y. Takagi, and T. Hashi, Phys. Lett. {\bf 48A} 183 (1974)
332:
333: \bibitem{selection} S. Sugano and Y. Tanabe, J. Phys. Soc. Jpn. {\bf 13} 880 (1958);
334: M.O. Schweika-Kresimon, J. Gutschank, and D. Suter, Phys. Rev. A {\bf 66} art. no. 043816 (2002)
335:
336: \bibitem{broadening} N. Laurance, E.C. McIrvine, and J. Lambe, J. Phys. Chem. Solids, {\bf 23} 515 (1962);
337: R.F. Wenzel and Y.W. Kim, Phys. Rev. {\bf 140} A1592 (1965)
338:
339: \bibitem{ESA} E. Kuznetsova, R. Kolesov, and O. Kocharovskaya, Phys. Rev. A {\bf 70} art. no. 043801 (2004);
340: R. Kolesov, E. Kuznetsova, and O. Kocharovskaya, {\it ibid.} to be published
341:
342: \bibitem{mode_lock} O.A. Kocharovskaya, Ya.I. Khanin, and V.B. Tsaregradskii, Kvantovaya Elektron. {\bf 12} 1227 (1985)
343: [Sov. J. Quantum Electron. {\bf 15} 810 (1985)]
344:
345: \end{thebibliography}
346:
347: \end{document}
348: