1: \documentstyle[preprint,pra,floats,aps,psfig]{revtex}
2: \renewcommand{\baselinestretch}{1.5}
3: %\textwidth=16cm
4: \oddsidemargin=1cm
5: \evensidemargin=1cm
6: \textheight=24cm
7: %\topmargin=-1cm
8:
9: \begin{document}
10:
11: \title{Coherent control of nuclear spin isomers of\\ molecules:
12: The role of molecular motion\thanks{The work was presented at The
13: VIIth International Symposium on Magnetic Field and Spin Effects
14: in Chemistry and Related Phenomena, 15-20 July 2001, Japan}}
15: \author{P.L.~Chapovsky\thanks{E-mail: chapovsky@iae.nsk.su}}
16:
17: \address{Institute of Automation and Electrometry,\\
18: Russian Academy of Sciences, 630090 Novosibirsk, Russia}
19:
20: \date{\today}
21:
22: \draft
23: \maketitle
24:
25: \begin{abstract}
26: Molecular center-of-mass motion is taken into account in the
27: theory of coherent control of nuclear spin isomers of molecules.
28: It is shown that infrared radiation resonant to the molecular
29: rovibrational transition can substantially enrich nuclear spin
30: isomers and speed up their conversion rate.
31: \end{abstract}
32:
33: \vspace{2cm}
34: \pacs{}
35:
36: \section{Introduction}
37:
38: Symmetrical molecules exist in the form of nuclear spin isomers
39: \cite{Landau81}. Most famous are the ortho and para isomers of H$_2$.
40: \cite{Farkas35}. On the other hand, spin isomers of molecules heavier
41: than hydrogen are almost unknown. This is because their investigations
42: were not possible for a long time due to the lack of enrichment methods.
43: Now such methods start to appear that has advanced the field significantly
44: (see the review \cite{Chap99ARPC}). Yet one needs more efficient
45: enrichment methods in order to perform investigations
46: of spin isomers of various molecules. It will also stimulate
47: practical applications of spin isomers.
48:
49: It was proposed recently to use strong electromagnetic fields for
50: the spin isomer enrichment. In the first version of the method
51: an enrichment is produced by the light-induced change of molecular
52: level populations \cite{Ilichov98CPL,Shalagin99JETPL}. In the second
53: method enrichment is due to combined action of the light-induced
54: level populations change, level energy shift and
55: light-induced coherences \cite{Chap01PRA2,Chap01JPB}.
56: The latter allows to refer to this method as coherent control of spin isomers.
57:
58: Analysis of the isomer coherent control \cite{Chap01PRA2,Chap01JPB} was performed
59: using simplified model of molecules interacting with a strong electromagnetic
60: field. Molecules were assumed to be at rest and have nondegenerate
61: states. The purpose of the present paper is to develop more general model
62: which will account molecular center-of-mass motion. This can be essential
63: because the process takes place in the gas phase at room temperature where
64: light-molecular interaction is known to be affected by the molecular motion.
65:
66: \section{Qualitative picture}
67:
68: Coherent control of spin isomers has been analysed for the case of
69: microwave excitation of molecular rotational transitions \cite{Chap01PRA2}
70: and for infrared excitation of
71: rovibrational transitions \cite{Chap01JPB}. Although each of these two
72: schemes has some advantages we will account molecular motion only for
73: the case of rovibrational excitation because this scheme allows to
74: achieve larger enrichment. Note also, that molecular motion is more
75: important for infrared excitation than for microwave excitation because
76: the latter has smaller Doppler broadening.
77:
78: Let us start with the qualitative description of the phenomenon. We assume
79: that the molecules have two nuclear spin states, ortho and para.
80: This is the case, for example,
81: for the CH$_3$F molecules \cite{Landau81,Chap99ARPC}. Nuclear spin states are different
82: by the symmetry of nuclear spin wave function of equivalent nuclei in the molecule.
83: Long wavelength electromagnetic radiation cannot change the nuclear spin state
84: directly. An important point is that the spin conversion in polyatomic
85: molecules is governed by intramolecular mixing of close (accidentally resonant)
86: ortho and para states (see Ref.~\cite{Chap99ARPC} for more details).
87: Interaction of an external radiation with these particular states can affect
88: the spin conversion.
89:
90: The level scheme for rovibrational excitation is shown
91: in Fig.~\ref{fig1}. Similar to Ref.~\cite{Chap01JPB} we assume that each
92: of the two vibrational states
93: (ground and excited) has one ortho-para level pair mixed by intramolecular
94: perturbation. The infrared radiation excites the rovibrational transition $m-n$
95: of the ortho molecules. By assumption, collisions cannot change spin state
96: of the test molecules directly, i.e., collisional cross-section of
97: ortho-para transitions vanish. Suppose that the molecule is placed
98: initially in the ortho subspace of the ground vibrational state.
99: Due to collisions the test molecule will undergo fast rotational
100: relaxation {\it inside} the ortho subspace. This will proceed until the molecule
101: jumps to the state $m'$, which is mixed with the para state $k'$
102: by the intramolecular perturbation $\hat V'$ (see Fig.~\ref{fig1}),
103: or to the state $n$ which is mixed with the para state $k$ by the
104: combined action of the external field and intramolecular perturbation
105: $\hat V$. Admixture of a para state
106: implies that the next collision can move the molecule to another para
107: states and thus localize it inside the para subspace. The spin conversion
108: occurs.
109:
110: Vibrational relaxation in this enrichment scheme plays an important role.
111: After ortho-para transition in upper vibrational state the molecule relaxes
112: to the ground vibrational state. Consequently there is no back conversion through
113: the $m-k$ channel in excited vibrational state but back conversion
114: occurs through the unperturbed $m'-k'$ channel. Steady state concentrations
115: of the spin isomers is determined by the ortho-to-para
116: conversion rate in the excited vibrational state and the para-to-ortho
117: conversion rate in the ground vibrational state. It is essential that conversion
118: through the $m-k$ channel can be substantially enhanced by electromagnetic
119: radiation.
120:
121: It is clear that molecular center-of-mass motion affects the phenomenon
122: because it makes the ortho-para mixing in upper vibrational state
123: velocity selective. One obvious effect is that the
124: radiation does not interact with all molecules in the state $n$.
125: Molecular motion influences also the level energy shift and coherences produced
126: in the molecule by external radiation. One needs quantitative account of
127: these effects in order to understand the real efficiency of the method.
128:
129: \section{Rate equations}
130:
131: Quantitative analysis of the problem can be performed with the help of
132: kinetic equation for the density matrix, $\hat\rho$.
133: The molecular Hamiltonian consists of four terms,
134: \begin{equation}
135: \hat H = \hat H_0 + \hbar\hat G + \hbar\hat V + \hbar\hat V'.
136: \label{H}
137: \end{equation}
138: The main part, $\hat H_0$, has the eigen ortho and para states
139: shown in Fig.~\ref{fig1}. $\hbar\hat G$ describes
140: the molecular interactions with the external radiation that will be
141: taken in the form of monochromatic travelling wave,
142: \begin{equation}
143: \hat G = - ({\bf E}_0\hat{\bf d}/\hbar)\cos(\omega_Lt-\bf k\bf r),
144: \label{g}
145: \end{equation}
146: where ${\bf E}_0$; $\omega_L$ and $\bf k$ are the amplitude, frequency
147: and wave vector of the electromagnetic wave, respectively; $\hat {\bf d}$
148: is the operator of the molecular electric dipole moment.
149: $\hat V$ and $\hat V'$ are the intramolecular perturbations
150: that mix the ortho and para states in excited and ground vibrational
151: states, respectively. These perturbations originate from the nuclear
152: spin-spin, or spin-rotation interactions
153: \cite{Curl67JCP,Chap91PRA,Guskov95JETP,Ilisca98PRA}.
154:
155: In the representation of the eigen states of $\hat H_0$ kinetic equation reads
156: \cite{Rautian79},
157: \begin{equation}
158: \partial\text{\boldmath$\rho$}/\partial t +{\bf v}\cdot\nabla\text{\boldmath$\rho$} =
159: {\bf S} - i [{\bf G}+{\bf V}+{\bf V'},\text{\boldmath$\rho$}],
160: \label{r1}
161: \end{equation}
162: where ${\bf S}$ is the collision integral; ${\bf v}$ is
163: the molecular center-of-mass velocity. We have considered molecular
164: motion classically in this equation and neglected spontaneous
165: decay which is slow for molecular rovibrational transitions in comparison with
166: collisional transitions. Account of molecular motion makes the matrices
167: {\boldmath$\rho$} and ${\bf S}$ velocity dependent. The matrices
168: ${\bf G}$, ${\bf V}$, and ${\bf V'}$ remain to be velocity independent.
169:
170: Kinetic equation for the total concentration of molecules in one spin state,
171: e.g., ortho can be obtained from Eq.~(\ref{r1}),
172: \begin{equation}
173: \partial\rho_o/\partial t =
174: 2Re\int i(\rho_{mk}V_{km}+\rho_{m'k'}V'_{k'm'})d{\bf v};\ \ \
175: \rho_o = \sum_j\int\rho(j,{\bf v})d{\bf v};\ \ \ j\in {\text{ortho}}.
176: \label{ro}
177: \end{equation}
178: Here $j$ indicates the rotational states of the molecule. These sates
179: are assumed to have no degeneracy. In Eq.~(\ref{ro})
180: a uniform spatial distribution of molecular density has been assumed.
181: Collision integral did not enter into Eq.~(\ref{ro}) because by assumption
182: collisions do not change the molecular spin state, i.e.,
183: $\sum_j\int S_{jj}d{\bf v}=0$, if $j\in$~ortho, or $j\in$~para. {\bf G} did
184: not enter into Eq.~(\ref{ro}) also because the matrix elements of {\bf G}
185: off-diagonal in nuclear spin states vanish.
186:
187: The off-diagonal matrix elements $\rho_{mk}$ and $\rho_{m'k'}$ can be found
188: using the perturbation theory. We assume further the perturbations $\hat V$
189: and $\hat V'$ being small and consider zeroth- and first-order terms of the
190: density matrix,
191: \begin{equation}
192: \text{\boldmath$\rho$} = \text{\boldmath$\rho$}^{(0)} +
193: \text{\boldmath$\rho$}^{(1)}.
194: \label{split}
195: \end{equation}
196:
197: Collisions in our system will be described by the model standard
198: in the theory of light-molecular interaction. The off-diagonal elements
199: of ${\bf S}$ have only decay terms,
200: \begin{equation}
201: S_{jj'} = -\Gamma\rho_{jj'}; \ \ \ j\neq j'.
202: \label{S_off}
203: \end{equation}
204: The decoherence rates,
205: $\Gamma$, were taken equal for all off-diagonal elements of collision
206: integral. Relaxation of the off-diagonal elements of the density matrix
207: is determined in molecular systems mainly by the level-population decay
208: which is thus assumed to be $j$-independent. Note that there
209: are opposite examples, e.g., \cite{Trappeniers79JCP}.
210:
211: The diagonal elements of ${\bf S}$ can be expressed
212: through the kernel of collision integral, $A$, as follows,
213: \begin{equation}
214: S(\alpha) = \sum_{a_1j_1}\int A(\alpha|\alpha_1)\rho(\alpha_1)d{\bf v_1} -
215: \rho(\alpha)\sum_{a_1j_1}\int A(\alpha_1|\alpha)d{\bf v_1};\ \ \
216: \alpha\equiv \{a,j,{\bf v} \}.
217: \label{S}
218: \end{equation}
219: Here $a=e,g$ denotes the excited and ground vibrational states, respectively.
220: The nuclear spin index is omitted in Eq.~(\ref{S}) because there are
221: no collisional ortho-para transitions and the kernels $A$ for ortho and para
222: states are assumed to be the same. We will consider further the model of strong
223: collisions with the following collision kernel,
224: \begin{equation}
225: A(\alpha|\alpha_1) = \nu_rw(j)\delta_{aa_1}\delta({\bf v}-{\bf v_1}) +
226: \nu_tw(j)f({\bf v})\delta_{aa_1} + \nu_vw(j)f({\bf v})\delta_{a_1e}\delta_{ag}.
227: \label{A}
228: \end{equation}
229: Here $w(j)$ is the Boltzmann distribution of rotational state populations,
230: \begin{equation}
231: w(j)=Z^{-1}\exp(-E_j/kT),
232: \label{B}
233: \end{equation}
234: where $Z$ is the rotational partition function for one vibrational state;
235: $E_j$ is the rotational energy; $T$ is the
236: gas temperature; $k_B$ is the Boltzmann constant. We assume the same function
237: $w(j)$ for the two vibrational states of both isomers. $f({\bf v})$ is the
238: Maxwell distribution,
239: \begin{equation}
240: f({\bf v}) =\frac{1}{\pi^{3/2}v_0^3}\exp(-\frac{{\bf v}^2}{v^2_0});\ \ \
241: v_0=\sqrt{2k_BT/M},
242: \label{M}
243: \end{equation}
244: where $M$ is the molecular mass.
245:
246: In Eq.~(\ref{A}) the three relaxation rates were introduced, rotational
247: relaxation ($\nu_r$) without change of molecular velocity (velocity
248: persistence effect) and vibrational state; translational relaxation
249: ($\nu_t$) that equilibrate also rotational
250: distribution but do not change the vibrational state;
251: vibrational relaxation ($\nu_v$) which is accompanied by
252: velocity and rotational relaxations. The introduction of different relaxation
253: rates for different degrees of freedom
254: makes the model of strong collisions more accurate and flexible. It allows
255: to adjust the model to the particular experimental conditions.
256: Because of its simplicity model of strong collisions is often used in
257: laser physics, see, e.g., \cite{Dykhne80JETP,Mironenko81IANS}.
258:
259: \section{Zeroth-order theory}
260:
261: For the zeroth-order term of the density matrix one has the following kinetic
262: equation,
263: \begin{equation}
264: \partial\text{\boldmath$\rho$}^{(0)}/\partial t +
265: {\bf v}\cdot\nabla\text{\boldmath$\rho$}^{(0)} =
266: {\bf S}^{(0)} - i [{\bf G},\text{\boldmath$\rho$}^{(0)}].
267: \label{ro0}
268: \end{equation}
269: Electromagnetic field excites only ortho molecules. Consequently, para
270: molecules remain at equilibrium in the zeroth-order perturbation theory.
271: For the population of para molecules one has,
272: \begin{equation}
273: \rho^{(0)}_p(g,j,{\bf v}) =
274: (N-\rho^{(0)}_o) w(j)f({\bf v});\ \ \ \rho^{(0)}_p(e,j,{\bf v}) = 0,
275: \label{0rp}
276: \end{equation}
277: where $N$ is the total concentration of the test molecules;
278: $\rho^{(0)}_o$ is the concentration of ortho molecules.
279:
280: Zeroth-order theory for ortho molecules
281: is in fact a familiar phenomenon of light-molecular interaction.
282: There are numerous publications
283: on this subject including many textbooks. Our zeroth-order theory
284: is close to the approach \cite{Mironenko81IANS}, although the relaxation
285: scheme given by Eq.~(\ref{A}) is different from \cite{Mironenko81IANS}.
286: Eqs.~(\ref{S_off}),(\ref{S}),(\ref{A}), and (\ref{ro0}) allow to
287: deduce the system of equations for the stationary populations of ortho molecules,
288: \begin{eqnarray}
289: \nu_e\rho^{(0)}_o(e,j,{\bf v}) & = & \nu_r w(j)\rho^{(0)}_o(e,{\bf v})
290: + \nu_t w(j)f({\bf v})\rho^{(0)}_o(e)
291: + \rho^{(0)}_o p(\Omega,{\bf v})\,\delta_{jm}; \nonumber \\
292: \nu_g\rho^{(0)}_o(g,j,{\bf v}) & = & \nu_r w(j)\rho^{(0)}_o(g,{\bf v})
293: + \nu_t w(j)f({\bf v})\rho^{(0)}_o(g) + \nonumber \\
294: &&\nu_v w(j)f({\bf v})\rho^{(0)}_o(e) -
295: \rho^{(0)}_o p(\Omega,{\bf v})\,\delta_{jn}; \nonumber \\
296: \rho^{(0)}_o p(\Omega,{\bf v}) & = & \frac{2\Gamma |G|^2}{\Gamma^2 + (\Omega-{\bf kv})^2}
297: \left[\rho^{(0)}_o(g,n,{\bf v}) - \rho^{(0)}_o(e,m,{\bf v})\right],
298: \label{sys0}
299: \end{eqnarray}
300: where $\nu_e=\nu_r+\nu_t+\nu_v$; $\nu_g=\nu_r+\nu_t$;
301: $\rho^{(0)}_o(e)$ and $\rho^{(0)}_o(g)$ are the concentrations
302: of ortho molecules in excited and ground vibrational states;
303: $\rho^{(0)}_o(a,{\bf v})=\sum_j\rho^{(0)}_o(a,j,{\bf v});\ \ a=e,g$;
304: and $p(\Omega,{\bf v})$ is the probability of optical excitation.
305: Eq.~(\ref{sys0}) is written in the rotating wave approximation.
306: Nonzero matrix element of ${\bf G}$ is given by,
307: \begin{equation}
308: G_{mn} = - Ge^{i({\bf kr}-\Omega t)};\ \
309: G\equiv E_0 \overline{d}_{mn}/2\hbar,
310: \label{rwa}
311: \end{equation}
312: where $\Omega=\omega_L - \omega_{mn}$ is the radiation frequency
313: detuning from the absorption line center, $\omega_{mn}$;
314: the bar over a symbol indicates a time-independent factor;
315: $|G|$ is the Rabi frequency.
316:
317: Solution of Eqs.~(\ref{sys0}) has no difficulty. Integration of
318: Eqs.~(\ref{sys0}) over {\bf v} and summation over $j$ gives
319: the system of algebraic equations for the distributions
320: $\rho^{(0)}_o(a,j,{\bf v})$, $\rho^{(0)}_o(a,{\bf v})$, and $\rho^{(0)}_o(a)$
321: that can be easily solved. The level populations
322: can be expressed as the sum of three terms, equilibrium in $j$ and ${\bf v}$,
323: equilibrium in $j$, and nonequilibrium in $j$ and ${\bf v}$,
324: \begin{eqnarray}
325: \rho^{(0)}_o(e,j,{\bf v})& = & \rho^{(0)}_o\{p(\Omega)\tau_{ev}w(j)f({\bf v})+
326: p(\Omega,{\bf v})[\tau_{et}w(j)+\tau_e\delta_{jm}]\} \nonumber \\
327: \rho^{(0)}_o(g,j,{\bf v})& = & \rho^{(0)}_o\{w(j)f({\bf v})[1-
328: p(\Omega)\tau_{gv}]-p(\Omega,{\bf v})[\tau_{gt}w(j)+\tau_g\delta_{jn}]\}.
329: \label{sol0}
330: \end{eqnarray}
331: here $p(\Omega)=\int p(\Omega,{\bf v})d{\bf v}$ and,
332: \begin{equation}
333: p(\Omega,{\bf v}) = \frac{w(n)\Gamma^2_Bf({\bf v})}
334: {\tau_1[\Gamma^2_B+(\Omega-{\bf kv})^2][1+1/\kappa+Y(\Omega)\tau_2/\tau_1]}.
335: \label{pov}
336: \end{equation}
337: In this equation, $\Gamma_B=\Gamma\sqrt{1+\kappa}$ is the power dependent
338: homogeneous linewidth (Bennett width) and $\kappa=2|G|^2\tau_1\Gamma^{-1}$
339: is the saturation parameter. The relaxation times are expressed through
340: the collisional parameters of the kernel $A$ from Eq.~(\ref{A}),
341: \begin{eqnarray}
342: \tau_{ev} & = & \nu^{-1}_v-(\nu_v+\nu_t)^{-1};\ \ \
343: \tau_{et}=(\nu_v+\nu_t)^{-1}-\nu_e^{-1};\ \ \ \tau_e=\nu_e^{-1} \nonumber \\
344: \tau_{gv} & = & \nu^{-1}_v-\nu^{-1}_t;\ \ \ \tau_{gt}=\nu^{-1}_t -\nu^{-1}_g ;\ \ \
345: \tau_g=\nu^{-1}_g;\ \ \ \nonumber \\
346: \tau_1 & = & \tau_{et}w(m)+\tau_{gt}w(n)+\tau_e+\tau_g;\ \ \
347: \tau_2=\tau_{ev}w(m)+\tau_{gv}w(n).
348: \label{tau}
349: \end{eqnarray}
350: The function,
351: \begin{equation}
352: Y(\Omega) = \sqrt{\pi}yReW(x+iy);\ \ \ x=\Omega/kv_0;\ \ \ y=\Gamma_B/kv_0,
353: \label{Y}
354: \end{equation}
355: is the Foigt profile of the absorption coefficient. $W(z)$ is the probability
356: integral,
357: \begin{equation}
358: W(z) = e^{-z^2}\left[1+\frac{2i}{\sqrt{\pi}}\int^z_0e^{u^2}du\right].
359: \label{W}
360: \end{equation}
361:
362: Similar to the calculation of the diagonal matrix elements of
363: {\boldmath$\rho$}$^{(0)}_o$, one can find the off-diagonal density matrix
364: element, $\rho^{(0)}_o(n|m;{\bf v})$ which amplitude is equal to,
365: \begin{equation}
366: \overline\rho^{(0)}_o(n|m;{\bf v}) = -i\rho^{(0)}_o
367: \frac{p(\Omega,{\bf v})}{2G\Gamma}[\Gamma-i(\Omega-{\bf kv})].
368: \label{ro_off}
369: \end{equation}
370:
371: \section{First order theory}
372:
373: In zeroth-order perturbation theory, one neglects perturbations
374: $\hat V$ and $\hat V'$. It implies that there are no coherences
375: between ortho and para states, $\rho^{(0)}_{mk}=0$;
376: $\rho^{(0)}_{m'k'}=0$. Consequently, one has,
377: \begin{equation}
378: \partial\rho_o/\partial t = 2Re \int
379: i(\rho^{(1)}_{mk}V_{km}+\rho^{(1)}_{m'k'}V'_{k'm'})d{\bf v},
380: \label{ro2}
381: \end{equation}
382: instead of Eq.~(\ref{ro}). Thus the spin conversion appears
383: in the second order approximation.
384:
385: The first order correction to the
386: density matrix, {\boldmath$\rho$}$^{(1)}$, is determined by the equation,
387: \begin{equation}
388: \partial\text{\boldmath$\rho$}^{(1)}/\partial t +
389: {\bf v}\cdot\nabla\text{\boldmath$\rho$}^{(1)} = {\bf S}^{(1)}
390: - i [{\bf G},\text{\boldmath$\rho$}^{(1)}]
391: - i [{\bf V} + {\bf V'},\text{\boldmath$\rho$}^{(0)}].
392: \label{ro1}
393: \end{equation}
394: For $\rho^{(1)}_{m'k'}$ one has from this equation,
395: \begin{equation}
396: \rho^{(1)}_{m'k'} = \frac{-iV'_{m'k'}}{\Gamma + i\omega'}
397: [\rho^{(0)}_p(g,k',{\bf v})-\rho^{(0)}_o(g,m',{\bf v})],
398: \label{1rho}
399: \end{equation}
400: where $\omega'\equiv\omega_{m'k'}$. The density matrix element
401: $\rho^{(1)}_{mk}$ can be obtained from equations
402: which are deduced from Eq.~(\ref{ro1}),
403: \begin{eqnarray}
404: \partial\rho^{(1)}_{mk}/\partial t + {\bf v}\cdot\nabla\rho^{(1)}_{mk} +
405: \Gamma\rho^{(1)}_{mk} + iG_{mn}\rho^{(1)}_{nk}
406: & = & iV_{mk} \rho^{(0)}_o(e,m,{\bf v}); \nonumber \\
407: \partial\rho^{(1)}_{nk}/\partial t + {\bf v}\cdot\nabla\rho^{(1)}_{nk} +
408: \Gamma\rho^{(1)}_{nk} + iG_{nm}\rho^{(1)}_{mk}
409: & = & iV_{mk} \rho^{(0)}_o(n|m;{\bf v}).
410: \label{sys1}
411: \end{eqnarray}
412: Substitutions,
413: $V_{mk}=\overline{V}e^{i\omega t},\ \ (\omega\equiv\omega_{mk});\ \
414: \rho^{(1)}_{mk} = \overline{\rho}^{(1)}_{mk}e^{i\omega t};\ \
415: \rho^{(1)}_{nk} = \overline{\rho}^{(1)}_{nk}
416: e^{i[(\Omega-\omega)t-{\bf kv}]}$,
417: transform Eqs.~(\ref{sys1}) to algebraic equations from which one
418: finds $\rho^{(1)}_{mk}$. Using this result for $\rho^{(1)}_{mk}$ and
419: $\rho^{(1)}_{m'k'}$
420: from Eq.~(\ref{1rho}) one has the rate equation (\ref{ro2}) in the form,
421: \begin{eqnarray}
422: \frac{\partial\rho_o}{\partial t} & = & \frac{2\Gamma|V'|^2}{\Gamma^2+\omega'^2}
423: \left[\rho^{(0)}_p(g,k')-\rho^{(0)}_o(g,m') \right] - \nonumber \\
424: &&2|V|^2 Re\int\frac{[\Gamma+i(\Omega+\omega-{\bf kv})]\rho^{(0)}_o(e,m,{\bf v}) +
425: iG\overline\rho^{(0)}_o(n|m;{\bf v})}{(\Gamma+i\omega)
426: [\Gamma+i(\Omega+\omega-{\bf kv})]+|G|^2}d{\bf v}.
427: \label{ro3}
428: \end{eqnarray}
429: This rate equation has the same general structure as in
430: \cite{Chap01JPB}. On the other hand the contributions from the level population in
431: upper state, $\rho^{(0)}_o(e,m,{\bf v})$, and from the light-induced coherence,
432: $\rho^{(0)}_o(n|m;{\bf v})$, are more complicated now due to their
433: velocity dependencies.
434:
435: \section{Enrichment and conversion}
436:
437: Similar to Ref.~\cite{Chap01JPB}, the rate equation (\ref{ro3}) is
438: presented in the form,
439: \begin{equation}
440: \partial\rho_o/\partial t = N\gamma'_{op} - \rho_o\gamma;\ \ \
441: \gamma \equiv \gamma'_{op} + \gamma'_{po} - \gamma'_n + \gamma_n + \gamma_{coh}.
442: \label{final}
443: \end{equation}
444: We have neglected in the right-hand side of Eq.~(\ref{final}) the small
445: difference between $\rho^{(0)}_o$ and the total concentration of
446: ortho molecules $\rho_o$. In Eq.~(\ref{final}) the following partial conversion
447: rates were introduced. The field independent rates,
448: \begin{equation}
449: \gamma'_{op}=\frac{2\Gamma|V'|^2}{\Gamma^2+\omega'^2}w(k');\ \ \
450: \gamma'_{po}=\frac{2\Gamma|V'|^2}{\Gamma^2+\omega'^2}w(m').
451: \label{g'}
452: \end{equation}
453: $\gamma_{free}\equiv \gamma'_{op} + \gamma'_{po}$, is
454: the isomer conversion rate without an external field. The rates $\gamma'_{op}$ and
455: $\gamma'_{po}$ coincide with the analogous rates in \cite{Chap01JPB}.
456: This is not surprising because intramolecular ortho-para state mixing is
457: velocity independent. The field dependent contribution to the
458: conversion rate,
459: \begin{equation}
460: \gamma'_n = \gamma'_{po}p(\Omega)w(m')(\tau_{gv}+\tau_{gt}),
461: \label{g'n}
462: \end{equation}
463: is also similar to that in \cite{Chap01JPB} except for the new definition of
464: relaxation parameters. This rate enters to the total rate $\gamma$ in
465: Eq.~(\ref{final}) with minus sign because it is caused by the light-induced
466: population depletion of the ground vibrational state of ortho molecules.
467: The last two contributions to $\gamma$ are more complicated now.
468: The ``noncoherent'' term appears due to the level population,
469: $\rho^{(0)}_o(e,m,{\bf v})$, in Eq.~(\ref{ro3}),
470: \begin{equation}
471: \gamma_n = 2|V|^2 Re\int\frac{[\Gamma+i(\Omega+\omega-{\bf kv})]
472: [p(\Omega)\tau_{ev}w(m)f({\bf v})+p(\Omega,{\bf v})(\tau_{et}w(m)+\tau_e)]}
473: {(\Gamma+i\omega)[\Gamma+i(\Omega+\omega-{\bf kv})]+|G|^2}d{\bf v}.
474: \label{gn}
475: \end{equation}
476: The ``coherent'' term, $\gamma_{coh}$, originates from
477: $\overline\rho^{(0)}_o(n|m;{\bf v})$, in Eq.~(\ref{ro3}),
478: \begin{equation}
479: \gamma_{coh} = |V|^2\Gamma^{-1}Re\int
480: \frac{p(\Omega,{\bf v})[\Gamma-i(\Omega-{\bf kv})]}
481: {(\Gamma+i\omega)[\Gamma+i(\Omega+\omega-{\bf kv})]+|G|^2}d{\bf v}.
482: \label{gcoh}
483: \end{equation}
484:
485: General expressions for the conversion rates are rather difficult to
486: analyse. It will be done elsewhere. Below we compare the results
487: \cite{Chap01JPB} with the present calculations. Solution to
488: Eq.~(\ref{final}) can be presented as,
489: \begin{equation}
490: \rho_o = \overline\rho_o + (\rho_o(0)-\overline\rho_o)\exp(-\gamma t);\ \ \
491: \overline\rho_o = N\gamma'_{op}/\gamma.
492: \label{overro}
493: \end{equation}
494: Here $\overline\rho_o$ is the stationary concentration of ortho molecules.
495: Without an external radiation (at the instant $t=0$), the equilibrium
496: concentration of para molecules reads,
497: \begin{equation}
498: \rho_p(0)=N-\rho_o(0)=N\gamma'_{po}/\gamma_{free}= N/2,
499: \label{rop}
500: \end{equation}
501: if the Boltzmann factors are assumed to be equal, $w(k')=w(m')$.
502: This implies $\gamma'_{op}=\gamma'_{po}$ (see Eq.~(\ref{g'})).
503: A stationary enrichment of para molecules is defined as,
504: \begin{equation}
505: \beta \equiv \frac{\overline\rho_p}{\rho_p(0)}-1 =
506: 1-2\frac{\gamma'_{op}}{\gamma}.
507: \label{bp}
508: \end{equation}
509: One can see from this equation that the external field produces
510: an enrichment of para isomers if $\gamma \neq \gamma_{free}$.
511:
512: We assume in further analysis the following parameters,
513: the Doppler width of the absorbing transition, $kv_0=30$~MHz;
514: $\omega=100$~MHz, $\omega'=130$~MHz, $V'_{m'k'}=5$~kHz,
515: $\Gamma/P=2\cdot10^8$~s$^{-1}$/Torr and the
516: Boltzmann factors of the states $m'$, $k'$, $m$, and $k$ all equal
517: $10^{-2}$. This set of parameters gives the field free conversion rate,
518: $\gamma_{free}=10^{-2}$~s$^{-1}$/Torr, which coincides with the
519: conversion rate in $^{13}$CH$_3$F. Nuclear spin conversion in these
520: molecules is governed by quantum relaxation (see the review \cite{Chap99ARPC}).
521: The translational, rotational and vibrational relaxation rates will be taken equal,
522: $\nu_t=0.1\Gamma$, $\nu_r=0.1\Gamma$, and $\nu_v=0.01\Gamma$, respectively.
523:
524: Close similarity between noncoherent, $\gamma_n$, and coherent, $\gamma_{coh}$,
525: terms (see Eqs.~(\ref{gn}) and (\ref{gcoh})) suggests that their
526: contributions to the total rate, $\gamma$, are rather difficult
527: to separate. It is indeed so for a wide range of parameters. On the
528: other hand, at some conditions it is possible. For example, it can
529: be done for weak external fields, thus small $G$. In this case one has,
530: similar to \cite{Chap01JPB}, two peaks in enrichment at frequencies
531: $\Omega\simeq-\omega$ and $\Omega=0$, see Fig.~\ref{fig2}. An obvious
532: difference is that these two peaks are much wider now having the
533: Doppler width, $kv_0$. The data shown in this figure
534: correspond to $V_{mk}=3$~kHz, and $\Gamma=2$~MHz. The peak at $\Omega=0$
535: appears because the excitation probability, $p(\Omega)$, has maximum at
536: $\Omega=0$. This peak has positive contribution from the rate $\gamma_n$
537: and negative contributions from the rates $\gamma'_n$ and $\gamma_{coh}$.
538: At high $G$ the main contribution originated from $\gamma_n$ reaches the value,
539: \[
540: \frac{\gamma_n}{\gamma_{free}}\sim \frac{1}{2}\left(\frac{V\omega'}
541: {V'\omega}\right)^2\frac{\tau_{ev}w(m)+\tau_{et}w(m)+\tau_e}{\tau_1+\tau_2},
542: \]
543: which constitutes $\simeq 15\%$ in our numerical example.
544: The homogeneous width of the peak at $\Omega=0$ is equal to $\Gamma_B$.
545: Consequently, this peak has rather large power broadening.
546: The peak at $\Omega=0$ in isomer enrichment was predicted in
547: \cite{Ilichov98CPL} by considering only the level population effects.
548:
549: The ``off-resonant'' peak at $\Omega\simeq -\omega$ appears
550: because the external field splits the ortho state $m$ into
551: two components and one sublevel crosses the para state $k$ at
552: this frequency of the external field (see Ref.~\cite{Chap01JPB}
553: for more details). At low $G$ this peak is determined by
554: $\gamma_{coh}$. When $G$ increases the peak at $\Omega\simeq -\omega$
555: grows to much bigger values than the peak at $\Omega=0$.
556: At high $G$ the main contribution to the $\Omega\simeq -\omega$
557: peak comes from the rate $\gamma_n$. Position of this peak is given by
558: $-\omega[1-|G|^2/(\Gamma^2+\omega^2)]$,
559: and the homogeneous width is equal to,
560: $\Gamma[1+|G|^2/(\Gamma^2+\omega^2)]$,
561: thus this peak has much smaller homogeneous width than the peak
562: at $\Omega=0$ if $\omega\gg\Gamma$.
563:
564: The data shown in Fig.~\ref{fig3} correspond to strong optical field,
565: $G=50$~MHz, and three values of $V_{mk}$. $\Gamma$ was taken equal 2~MHz.
566: One can see, that strong optical field is able to convert
567: almost all molecules to the para state if $V_{mk}\simeq V_{m'k'}$.
568: Thus relatively weak (3~kHz) coupling in upper state produces
569: macroscopic effect, viz., almost complete enrichment of spin isomers.
570: In the case of much weaker coupling
571: in upper state, enrichment is still significant. For example, if the
572: perturbation in upper state, $V_{mk}=100$~Hz, one has the
573: enrichment, $\beta\simeq1$\%. Enrichment at this level can
574: easily be measured. The enrichment peak at
575: $\Omega\simeq -\omega$ is narrow (having almost the Doppler width)
576: and thus can be distinguished from much wider structures
577: (the width $\simeq \Gamma_B$) caused by the $\Omega=0$ peaks.
578:
579: Conversion rate in the system is equal to $\gamma$, see Eq.~(\ref{overro}).
580: It is convenient to characterize it in relative units,
581: \begin{equation}
582: \gamma_{rel} = \gamma/\gamma_{free} - 1.
583: \label{grel}
584: \end{equation}
585: Thus, $\gamma_{rel}=0$ without an external field.
586: At low $G$, conversion rate, like enrichment, has two peaks in its frequency
587: dependence. If Rabi frequency, $|G|$, is large and the ortho-para couplings
588: in upper and low states have the same order of magnitude, conversion can be
589: significantly enhanced (Fig.~\ref{fig3}, upper panel),
590: although this enhancement is smaller than in the case of the molecules at rest
591: \cite{Chap01JPB}. This
592: enhancement appears because of the crossing of ortho and para states in
593: upper vibrational state by external field. The main difference between the
594: present case and Ref.~\cite{Chap01JPB} is that the peaks in conversion rate
595: are much wider now having the Doppler width, $kv_0$.
596:
597: \section{Conclusions}
598:
599: We have developed the model of coherent control of nuclear
600: spin isomers taking into account the molecular center-of-mass motion. This
601: model is more realistic than the model \cite{Chap01JPB} in which the molecules
602: were assumed to be at rest. It was shown that the main conclusions of \cite{Chap01JPB}
603: still hold. One has two peaks in the enrichment and conversion rate dependences
604: on the laser radiation frequency. One can achieve a substantial enrichment and
605: speed up the conversion rate. It is also valid that this
606: enrichment scheme can be used to access very weak perturbations in
607: molecules. On the other hand, molecular motion causes some decrease in sensitivity
608: of the method. For example, 1\% enrichment can be achieved for the
609: intramolecular perturbation equal $V\simeq100$~Hz in the present model,
610: in comparison with 1\% enrichment at $V\simeq30$~Hz in Ref.~\cite{Chap01JPB}.
611: This is not a significant difference. But even this loss in sensitivity can
612: probably be avoided in special excitation scheme, e.g., by using a multifrequency
613: excitation instead of a monochromatic excitation. This needs further investigations.
614:
615: Search of proper resonances will be one of the main difficulty in
616: practical implementation of the proposed enrichment scheme. An existence
617: of the off-resonance peak in the isomer enrichment has particular
618: importance. First, one does not need to search for an exact
619: coincidence of the ortho and para states in upper
620: vibrational state. The energy gap can be compensated by detuning
621: the laser frequency. Second, the excitation in the off-resonant peak
622: does produce even higher enrichment than in the resonant peak but does not
623: produce significant population in upper vibrational state. The latter
624: helps not to populate the excited vibrational state of other spin isomer and
625: consequently suppress the spurious back conversion through the upper channel.
626:
627: \section*{Acknowledgments}
628:
629: This work was made possible by financial support from the
630: Russian Foundation for Basic Research (RFBR), grant No.~01-03-32905
631: and from The Institute of Physical and Chemical Research (RIKEN),
632: Japan.
633:
634: \begin{thebibliography}{10}
635:
636: \bibitem{Landau81}
637: L.~D. Landau and E.~M. Lifshitz, {\em Quantum Mechanics, 3rd ed.} (Pergamon
638: Press, Oxford, 1981).
639:
640: \bibitem{Farkas35}
641: A. Farkas, {\em Orthohydrogen, Parahydrogen and Heavy Hydrogen} (Cambridge
642: University Press, London, 1935), p.\ 215.
643:
644: \bibitem{Chap99ARPC}
645: P.~L. Chapovsky and L.~J.~F. Hermans, Annu. Rev. Phys. Chem. {\bf 50}, 315
646: (1999).
647:
648: \bibitem{Ilichov98CPL}
649: L.~V. Il'ichov, L.~J.~F. Hermans, A.~M. Shalagin, and P.~L. Chapovsky, Chem.
650: Phys. Lett. {\bf 297}, 439 (1998).
651:
652: \bibitem{Shalagin99JETPL}
653: A.~M. Shalagin and L.~V. Il'ichov, Pis'ma Zh. Eksp. Teor. Fiz. {\bf 70}, 498
654: (1999).
655:
656: \bibitem{Chap01PRA2}
657: P.~L. Chapovsky, Phys. Rev. A {\bf 63}, 063402 (2001), \\
658: http://arXiv.org/abs/physics/0011012.
659:
660: \bibitem{Chap01JPB}
661: P.~L. Chapovsky, J. Phys. B: At. Mol. Opt. Phys. {\bf 34}, 1123 (2001), \\
662: http://arXiv.org/abs/physics/0011043.
663:
664: \bibitem{Curl67JCP}
665: R.~F. {Curl, Jr.}, J.~V.~V. Kasper, and K.~S. Pitzer, J. Chem. Phys. {\bf 46},
666: 3220 (1967).
667:
668: \bibitem{Chap91PRA}
669: P.~L. Chapovsky, Phys. Rev. A {\bf 43}, 3624 (1991).
670:
671: \bibitem{Guskov95JETP}
672: K.~I. Gus'kov, Zh. Eksp. Teor. Fiz. {\bf 107}, 704 (1995), [JETP. {\bf 80},
673: 400-414 (1995)].
674:
675: \bibitem{Ilisca98PRA}
676: E. Ilisca and K. Bahloul, Phys. Rev. A {\bf 57}, 4296 (1998).
677:
678: \bibitem{Rautian79}
679: S.~G. Rautian, G.~I. Smirnov, and A.~M. Shalagin, {\em Nonlinear resonances in
680: atom and molecular spectra} (Nauka, Siberian Branch, Novosibirsk, Russia,
681: 1979), p.\ 310.
682:
683: \bibitem{Trappeniers79JCP}
684: N.~J. Trappeniers and E.~W.~A. {Elenbaas-Bunschoten}, J. Chem. Phys. {\bf 64},
685: 205 (1979).
686:
687: \bibitem{Dykhne80JETP}
688: A.~M. Dykhne and A.~N. Starostin, Zh. Eksp. Teor. Fiz. {\bf 79}, 1211 (1980).
689:
690: \bibitem{Mironenko81IANS}
691: V.~R. Mironenko and A.~M. Shalagin, Izv. Akad. Nauk SSSR, Seriya Fiz. {\bf 45},
692: 995 (1981), bull. Acad. Sci. USSR, Phys. Ser. {\bf 45}, 87 (1981).
693:
694: \end{thebibliography}
695:
696:
697: \newpage
698: %Fig.1
699: \begin{figure}[hb]
700: \centerline{\psfig
701: {figure=fig1.eps,width=12cm}}
702: \vspace{0.5cm}
703: \caption{\sl Level scheme. Vertical solid line indicates infrared excitation.
704: Vibrational relaxation is shown by the dashed vertical lines. $V$ and $V'$
705: indicate the intramolecular ortho-para state mixing.}
706: \label{fig1}
707: \end{figure}
708:
709: \newpage
710: %Fig.2
711: \begin{figure}[htb]
712: \centerline{\psfig
713: {figure=fig2.eps,height=14cm}}
714: \vspace{2cm}
715: \caption{\sl Frequency dependence of the enrichment of para molecules,
716: $\beta$, at the Rabi frequencies $|G|=1$, 2, and 4~MHz.}
717: \label{fig2}
718: \end{figure}
719:
720: \newpage
721: %Fig.3
722: \begin{figure}[htb]
723: \centerline{\psfig
724: {figure=fig3.eps,height=20cm}}
725: \vspace{2cm}
726: \caption{\sl Enrichment of para molecules, $\beta$, and conversion rate,
727: $\gamma_{rel}$, for the three values of $V_{mk}$. The wide negative structures
728: in the middle and low panels (the width equal $\Gamma_B$) are mainly due to
729: $\gamma'_n$.}
730: \label{fig3}
731: \end{figure}
732:
733: \end{document}
734:
735:
736: