quant-ph0104119/let0.tex
1: %\documentstyle[preprint,epsf,bm,prl,aps]{revtex}
2: \documentstyle[multicol,epsf,bm,prl,aps]{revtex}
3: %\documentstyle[aps]{revtex}
4: 
5: 
6: \begin{document}
7: 
8: %---------- FRONT MATTERS ---------------------------------------
9: \draft
10: \preprint{Preprint No.}
11: \title{Stimulated emission with a non-equilibrium state of radiation}
12: \author{L.Accardi\thanks{accardi@volterra.mat.uniroma2.it}, ~K.Imafuku\thanks{imafuku@volterra.mat.uniroma2.it}, ~S.V.Kozyrev$^1$\thanks{kozyrev@mi.ras.ru}}
13: \address{Centro Vito Volterra, Universita' di Roma Tor Vergata, 
14: 00133 Rome, Italy\\
15: N.N. Semenov Institute of Chemical Physics, Russian Academy of Science, \\
16: 117334 Moscow, Russia$^1$}
17: \date{April 24, 2001}
18: \maketitle
19: \begin{abstract}
20: The stimulated emission from an atom interacting with radiation 
21: in non-equilibrium state is considered. 
22: The stochastic limit,  applied to the  
23: non-relativistic Hamiltonian describing the interaction, shows 
24: that the state of atoms,  driven by some non-equilibrium 
25: state of the field approaches a stationary state which can 
26: continuously emit photon, unlike the case 
27: with an equilibrium state.
28: \end{abstract}
29: \pacs{}
30: 
31: %------------------------------------------------------------------
32: %\begin{multicols}{2}
33: %\narrowtext
34: Einstein applied Planck's radiation theory to describe the equilibrium 
35: state between a atom and field\cite{ref:Einstein}.  He perceived that such an 
36: equilibrium state can be realized through the detailed balance 
37: condition, i.e. the balance in each mode between spontaneous and 
38: ``stimulated emission", i.e. the emission from the atom 
39: stimulated by the field. Einstein's investigation can be said to give 
40: the clearest insight into the origin of Planck's radiation because, with this 
41: notion, we can understand the Planck's law on the density of states of 
42: the photons from an equilibrium field, i.e.
43: \begin{equation}
44: \rho(\omega)d\omega=\frac{\hbar \omega^3}{\pi^2 c^3}\frac{1}
45: {\exp\left(\frac{\hbar\omega}{kT}\right)-1}d\omega .
46: \end{equation}
47: 
48: On the other hand, with the development of technology,
49: a controlled  emission with a controlled stimulating 
50: field has been realized in experimental situation,
51: for example, a laser system.
52: A laser is often described in terms of stimulated emission 
53: from an equilibrium state at {\it negative temperature} \cite{ref:laser}. 
54: However  the concept of the negative temperature can't be accepted literally, 
55: but rather as an expediency to understand an inverse 
56: population state of the atoms which can emit photons.
57: More precisely, a laser should be considered as a stimulated emission 
58: due to a non-equilibrium state of a field.
59: In this letter we propose a way to understand such kind of
60: stimulated emissions from non-equilibrium states, 
61: without introducing any phenomenological expediency like 
62: a negative temperature. 
63: 
64: We apply the stochastic limit \cite{ref:Accardi_text} 
65: to the non-relativistic QED Hamiltonian and derive two types of 
66: equations: one is the so-called rate equation for the atom and the other 
67: is  an equation describing the time evolution of the number of photons.  
68: In addition, based on these equations, 
69: we discuss the stimulated emission from the atom interacting 
70: with some non-equilibrium field.  The state of an atom driven by the 
71: non-equilibrium state of the field approaches 
72: a stationary state which can continuously emit photon, unlike the 
73: case of the equilibrium field.  The conditions of  
74: the non-equilibrium state  of the field for such phenomena 
75: is also made clear.
76: 
77: We shall consider an atom interacting with the EM-filed 
78: described with the standard non-relativistic QED Hamiltonian
79: \begin{equation}
80: H=H_0+\lambda ~V,\quad H_0=H_A+H_F
81: \end{equation}
82: where
83: \begin{equation}
84: H_A=\frac{p^2}{2m}+V(q),\quad
85: \left([q,p]=i\right)
86: \end{equation}
87: \begin{equation}
88: H_F=\int \omega(k) a^{\dagger}_{k,\sigma}a_{k,\sigma},~\quad 
89: \left([a_{k,\sigma},a^{\dagger}_{k',\sigma'}]=\delta_{\sigma\sigma'}\delta(k-k')\right)
90: \end{equation}
91: and
92: \begin{equation}
93: V= \sum_{\sigma} \int dk \frac{1}{|k|^{1/2}}
94: \varepsilon_{k,\sigma}\left(a_{\sigma,k}e^{ik\cdot q}
95: +a_{\sigma,k}^\dagger e^{-ik\cdot q}\right).
96: \end{equation}
97: $\sigma$ is polarization index ($\sigma=\leftrightarrow,\updownarrow$).
98: In the following discussion we assume that $H_A$ has discrete 
99: spectrum or 
100: $
101: H_A|\epsilon_a\rangle=\epsilon_a|\epsilon_a\rangle
102: $. 
103: The interaction Hamiltonian in interaction picture 
104: $V(t)$ can be witten as \cite{ref:Accardi_Kozyrev_note}
105: \begin{equation}
106: V(t)=\sum_{\omega \in F} \sum_{a,b}\int dk \left(
107: \overline{g_{ab}(k,\sigma)}
108: E_\omega^{\dagger}\Big(|\epsilon_a \rangle\langle \epsilon_b|\Big)
109: a_{k,\sigma} e^{-i(\omega(k)-\omega)t}
110: +h.c \right)
111: \end{equation}
112: where
113: \begin{eqnarray}
114: \overline{g_{ab}(k,\sigma)}
115: &=&\frac{1}{|k|^{1/2}} 
116: ~\langle \epsilon_a|e^{ik\cdot q}p \cdot \varepsilon_\sigma|\epsilon_b\rangle,
117: \\
118: E^\dagger_\omega(X)&=&
119: \sum_{\epsilon_r \in F_\omega}P_{\epsilon_r}
120: XP_{\epsilon_r-\omega},\quad 
121: P_{\epsilon_r}:=|\epsilon_r\rangle\langle \epsilon_r|\\
122: F&=&\{\omega=\epsilon_r-\epsilon_r';~\epsilon_r,\epsilon_r'\in Spec ~H_A\},\quad\mbox{(Bohr frequency)}\\
123: F_\omega&=&\{\epsilon_r\in Spec ~H_s ;~\epsilon_r-\omega\in Spec ~H_A\}.
124: \end{eqnarray}
125: We investigate the dynamics of the system described by the above 
126: Hamiltonian using the stochastic limit, which describes the quantum 
127: dynamics in the regime of weak coupling ($\lambda\rightarrow 0$) 
128: and large times ($t \rightarrow t/\lambda^2$) \cite{ref:Accardi_text}.  
129: The main result of this theory (stochastic golden rule) 
130: \cite{ref:Accardi_text} is that the time rescaling $t\rightarrow t/\lambda^2$ 
131: induces a rescaling of the quantum field
132: \begin{equation}\label{rescaled_fielsd}
133: a_{k,\sigma}\rightarrow\frac{1}{\lambda}e^{-i\frac{t}{\lambda^2}(\omega(k)-\omega)}a_{k,\sigma}
134: \end{equation}
135: and, in the limit $\lambda\rightarrow 0$ the rescaled field 
136: (\ref{rescaled_fielsd}) becomes a quantum white noise (or master field)
137: $b_{\omega,\sigma}(t,k)$ satisfying the commutation 
138: relations
139: \begin{equation}
140: [b_{\omega,\sigma}(t,k),b_{\omega',\sigma'}^\dagger(t',k')]
141: =\delta_{\sigma\sigma'}\delta_{\omega\omega'}2\pi \delta(t-t')\delta(\omega(k)-\omega)\delta(k-k').
142: \end{equation}
143: Moreover, if the initial state of the field is 
144: the mean zero gauge invariant Gaussian state with correlations 
145: $$\langle a^\dagger_k a_k'\rangle=N(k)\delta(k-k')$$, 
146: then the state of the limit white noise will be of the some 
147: type with correlations
148: \begin{eqnarray}\label{correlation1}
149: \langle b_{\omega,\sigma}(t,k)
150: b_{\omega',\sigma'}^\dagger (t',k') \rangle &=&
151: \delta_{\sigma\sigma'}\delta_{\omega\omega'}2\pi \delta(t-t')\delta(\omega(k)-\omega)\delta(k-k')N_{\sigma}(k)\\
152: \langle b^\dagger_{\omega,\sigma}(t,k)
153: b_{\omega',\sigma'}(t',k') \rangle &=&
154: \delta_{\sigma\sigma'}\delta_{\omega\omega'}2\pi \delta(t-t')\delta(\omega(k)-\omega)\delta(k-k')(N_{\sigma}(k)+1).\label{correlation2}
155: \end{eqnarray}
156: The Schr\"{o}dinger equation becomes a quantum white noise equation 
157: which, after having been normally ordered, takes the form 
158: \cite{ref:Accardi_text,ref:Accardi_Kozyrev_note}
159: \begin{equation}
160: dU_t=(-idH(t)-Gdt)U_t\quad;\quad t>0
161: \end{equation}
162: with the initial condition $U_0=1$ and where
163: $dH(t)$, {\it called the martingale term}, is the stochastic differential:
164: \begin{equation}
165: dH(t)=\sum_{\omega \in F} \sum_{a,b}\int dk \left(
166: \overline{g_{ab}(k,\sigma)}
167: E_\omega^{\dagger}\Big(|\epsilon_a \rangle\langle \epsilon_b|\Big)
168: dB_{\omega,\sigma}(t)+
169: E_\omega\Big(|\epsilon_a \rangle\langle \epsilon_b|\Big)
170: dB_{\omega,\sigma}^\dagger(t)
171: \right)
172: \end{equation}
173: driven by the quantum Brownian motions
174: \begin{equation}
175: dB_{\omega,\sigma}=\sum_{\epsilon_a-\epsilon_b=\omega}\int^{t+dt}_{t}d\tau
176: ~\int dk \overline{g_{ab}(k,\sigma)}b_{\omega,\sigma}(\tau,k),
177: \end{equation}
178: and the operator $G$, {\it called the drift term}, is given by
179: \begin{equation}
180: G=\sum_{\sigma}\sum_{\omega \in F} \left(
181: (g|g)^{-}_{\omega,\sigma}
182: E^\dagger_{\omega}\Big(|\epsilon_a \rangle\langle \epsilon_b|\Big)
183: E_{\omega}\Big(|\epsilon_a \rangle\langle \epsilon_b|\Big)
184: +\overline{(g|g)}^{+}_{\omega,\sigma}
185: E_{\omega}\Big(|\epsilon_a \rangle\langle \epsilon_b|\Big)
186: E^\dagger_{\omega}\Big(|\epsilon_a \rangle\langle \epsilon_b|\Big)
187: \right)
188: \end{equation}
189: \begin{eqnarray}
190: (g|g)^{-}_{\omega,\sigma}&=&\sum_{\epsilon_a-\epsilon_b=\omega}\int dk |g_{ab}(k,\sigma)|^2 \frac{-i(N_{\sigma}(k)+1)}{\omega(k)-\omega-i0},\label{19}\\
191: (g|g)^{+}_{\omega,\sigma}&=&\sum_{\epsilon_a-\epsilon_b=\omega}\int dk |g_{ab}(k,\sigma)|^2 \frac{-iN_{\sigma}(k)}{\omega(k)-\omega-i0}.\label{20}
192: \end{eqnarray}
193: With this quantum stochastic differential equation, we can always 
194: derive equations for the atom and field, that is, 
195: (i)master equation for reduced density operator for the atom,
196: and 
197: (ii)equation for the field\cite{ref:Accardi_Kozyrev_note,ref:Accardi_KI_Kozyrev}. 
198: 
199: 
200: One should notice that we do not assume the state of the field to be 
201: equilibrium. Therefore $N_{\sigma}(k)$ in (\ref{correlation1}) 
202: and (\ref{correlation2}) is not necessary to be 
203: $$
204: N_{\sigma}(k)=\frac{1}{e^{\beta \omega(k)}-1}, 
205: ~\left(\beta=\frac{1}{kT}\right),
206: $$
207: but can be, for some general nonlinear functions 
208: $\beta_{\sigma}(\omega)$,
209: \begin{equation}\label{nesf}
210: N_{\sigma}(k)=\frac{1}{e^{\beta_{\sigma}\left(\omega(k)\right)}-1}.
211: \end{equation}
212: 
213: In this letter, let us consider a 3-level atom 
214: ($\epsilon_1 < \epsilon_2 < \epsilon_3$) whose matrix 
215: elements satisfy the conditions (Fig.~1)
216: \begin{equation}\label{assumption0}
217: \langle \epsilon_1| e^{ikq}p \cdot\varepsilon_{\updownarrow}
218: |\epsilon_2\rangle=0,
219: \quad
220: \langle \epsilon_2| e^{ikq}p \cdot
221: \varepsilon_{\leftrightarrow}|\epsilon_3\rangle
222: =\langle \epsilon_1| e^{ikq}p \cdot
223: \varepsilon_{\leftrightarrow}|\epsilon_3\rangle=0,\quad
224: others\neq 0. 
225: \end{equation}
226: There are rather standard conditions in laser theory~\cite{ref:laser}.
227: Notice however that condition (\ref{assumption0}) requires the 
228: preparation of the atom in a situation in which the longitudinal
229: ($\updownarrow$) and transverse ($\leftrightarrow$) polarization 
230: do not enter symmmetrically. We believe that atoms, satisfying the condition 
231: (\ref{assumption0}), can be experimentally prepared.
232: In addition, for simplicity, we restrict ourselves to a generic 
233: system \cite{ref:Accardi_Kozyrev_note}: this means in our case that the 
234: 3 Bohr frequencies  $\omega_{21}$, $\omega_{31}$, $\omega_{32}$  
235: ($\omega_{jk}=\epsilon_{j}-\epsilon_{k}$) are all different among 
236: themselves.
237: 
238: With these assumptions, we can derive the so-called rate equation 
239: for the atom 
240: \begin{eqnarray}\label{rate_3_eq2_1}
241: \frac{d}{dt}P_1(t)&=& -\left\{
242: 2(\gamma_{21,\leftrightarrow}^{(+)}+\gamma_{31,\updownarrow}^{(+)})P_{1}(t)
243: -2(\gamma^{(-)}_{21,\leftrightarrow}P_{2}(t)+\gamma_{31,\updownarrow}^{(-)}
244: P_{3}(t))
245: \right\}\nonumber\\
246: \frac{d}{dt}P_2(t)&=& -\left\{
247: 2(\gamma_{21,\leftrightarrow}^{(-)}+\gamma_{32,\updownarrow}^{(+)})P_{2}(t)
248: -2(\gamma^{(-)}_{32,\updownarrow}P_{3}(t)+\gamma_{21,\leftrightarrow}^{(+)}P_{1}(t))
249: \right\}\nonumber\\
250: \frac{d}{dt}P_3(t)&=& -\left\{
251: 2(\gamma_{32,\updownarrow}^{(-)}+\gamma_{31,\updownarrow}^{(-)})P_{3}(t)
252: -2(\gamma^{(+)}_{31,\updownarrow}
253: P_{1}(t)+\gamma_{32,\updownarrow}^{(+)}P_{2}(t))
254: \right\}\nonumber\\
255: \label{rate_equation}
256: \end{eqnarray}
257: and the corresponding equations for  polarized photons
258: \begin{eqnarray}\label{n_tr}
259: \frac{d}{dt}n_{\leftrightarrow}(t)&=&
260: 2~\left(\gamma^{(-)}_{21,\leftrightarrow}P_2(t)
261: -\gamma^{(+)}_{21,\leftrightarrow}P_1(t)\right)\label{n_lr(t)}\\
262: \frac{d}{dt}n_{\updownarrow}(t)&=&2~\left(\gamma^{(-)}_{31,\updownarrow}P_3(t)
263: -\gamma^{(+)}_{31,\updownarrow}P_1(t)+
264: \gamma^{(-)}_{32,\updownarrow}P_3(t)
265: -\gamma^{(+)}_{32,\updownarrow}P_2(t)\right)\label{n_ud(t)}
266: \end{eqnarray}
267: where
268: \begin{equation}
269: P_j(t)=Tr\Big(\rho_{tot}(0)U^\dagger_t|\epsilon_j\rangle\langle\epsilon_j|U_t \Big),\quad
270: n_{\sigma}(t)=Tr\Big(\rho_{tot}(0)U^\dagger_t \int dk 
271: a^\dagger_{k,\sigma}a_{k,\sigma}~U_t \Big),
272: \end{equation}
273: \begin{equation}
274: \gamma_{ij,\sigma}^{(\pm)}={\rm Re}(g|g)^{\pm}_{\omega_{ij},\sigma},
275: \end{equation}
276: and the $(g|g)^{\pm}_{\omega_{ij},\sigma}$ 
277: are given by (\ref{19}) and (\ref{20}).
278: Equations (\ref{rate_equation}) have a stationary solution 
279: which satisfies 
280: \begin{eqnarray}\label{st_sol_2}
281: \frac{P_2(\infty)}{P_1(\infty)}&=&
282: \frac{\gamma_{31,\updownarrow}^{(-)}\gamma_{21,\leftrightarrow}^{(+)}+
283: \gamma_{32,\updownarrow}^{(-)}\gamma_{21,\leftrightarrow}^{(+)}+
284: \gamma_{31,\updownarrow}^{(+)}\gamma_{32,\updownarrow}^{(-)}}
285: {\gamma_{31,\updownarrow}^{(-)}\gamma_{21,\leftrightarrow}^{(-)}+
286: \gamma_{31,\updownarrow}^{(-)}\gamma_{32,\updownarrow}^{(+)}+
287: \gamma_{32,\updownarrow}^{(-)}\gamma_{21,\leftrightarrow}^{(-)}}
288: \nonumber\\
289: \frac{P_3(\infty)}{P_1(\infty)}&=&
290: \frac{
291: \gamma_{31,\updownarrow}^{(+)}\gamma_{21,\leftrightarrow}^{(-)}+
292: \gamma_{32,\updownarrow}^{(+)}\gamma_{21,\leftrightarrow}^{(+)}+
293: \gamma_{31,\updownarrow}^{(+)}\gamma_{32,\updownarrow}^{(+)}}
294: {\gamma_{32,\updownarrow}^{(-)}\gamma_{21,\leftrightarrow}^{(-)}+
295: \gamma_{31,\updownarrow}^{(-)}\gamma_{21,\leftrightarrow}^{(-)}+
296: \gamma_{31,\updownarrow}^{(-)}\gamma_{32,\updownarrow}^{(+)}}
297: \nonumber\\
298: \frac{P_3(\infty)}{P_2(\infty)}&=&
299: \frac
300: {\gamma_{32,\updownarrow}^{(+)}\gamma_{21,\leftrightarrow}^{(+)}+
301: \gamma_{31,\updownarrow}^{(+)}\gamma_{21,\leftrightarrow}^{(-)}+
302: \gamma_{32,\updownarrow}^{(+)}\gamma_{31,\updownarrow}^{(+)}}
303: {\gamma_{31,\updownarrow}^{(-)}\gamma_{21,\leftrightarrow}^{(+)}+
304: \gamma_{32,\updownarrow}^{(-)}\gamma_{21,\leftrightarrow}^{(+)}+
305: \gamma_{31,\updownarrow}^{(+)}\gamma_{32,\updownarrow}^{(-)}}.
306: \nonumber\\
307: \end{eqnarray}
308: 
309: The right-hand side of equation (\ref{n_tr}) shows that, 
310: if
311: $
312: \gamma^{(-)}_{21,\leftrightarrow}P_2(t)
313: -\gamma^{(+)}_{21,\leftrightarrow}P_1(t) > 0
314: $,
315: i.e. if
316: \begin{equation}
317: \frac{P_2(t)}{P_1(t)} > \frac{\gamma_{21,\leftrightarrow}^{(+)}}{\gamma_{21,\leftrightarrow}^{(-)}}
318: \end{equation}
319: then, $n_{\leftrightarrow}(t)$ must increase.  
320: {\it Also  with the  stationary state of the atom}, 
321: $n_{\leftrightarrow}(t)$ must keep increasing if
322: \begin{equation}\label{emission_condition}
323: \frac{P_2(\infty)}{P_1(\infty)} > \frac{\gamma_{21,\leftrightarrow}^{(+)}}{\gamma_{21,\leftrightarrow}^{(-)}}
324: \end{equation}
325: is satisfied. This means that, if we can experimentally realize 
326: condition (\ref{emission_condition}), then we can also realize a 
327: continuous emission of 
328: $\leftrightarrow$-photon from the atom, stimulated by 
329: a non-equilibrium state of the field. 
330: Using (\ref{st_sol_2}), the condition (\ref{emission_condition}) 
331: can be written as 
332: \begin{equation}\label{condition_0}
333: \frac{\gamma_{21,\leftrightarrow}^{(+)}}{\gamma_{21,\leftrightarrow}^{(-)}}
334: < \frac{\gamma_{31,\updownarrow}^{(-)}\gamma_{21,\leftrightarrow}^{(+)}+
335: \gamma_{32,\updownarrow}^{(-)}\gamma_{21,\leftrightarrow}^{(+)}+
336: \gamma_{31,\updownarrow}^{(+)}\gamma_{32,\updownarrow}^{(-)}}
337: {\gamma_{31,\updownarrow}^{(-)}\gamma_{21,\leftrightarrow}^{(-)}+
338: \gamma_{31,\updownarrow}^{(-)}\gamma_{32,\updownarrow}^{(+)}+
339: \gamma_{32,\updownarrow}^{(-)}\gamma_{21,\leftrightarrow}^{(-)}}
340: ~~\Leftrightarrow
341: ~~\frac{\gamma_{21,\leftrightarrow}^{(+)}}{\gamma_{21,\leftrightarrow}^{(-)}}
342: <
343: \frac{\gamma_{31,\updownarrow}^{(+)}}{\gamma_{31,\updownarrow}^{(-)}}
344: \frac{\gamma_{32,\updownarrow}^{(-)}}{\gamma_{32,\updownarrow}^{(+)}}.
345: \end{equation}
346: If we suppose that the photon densities in the state (\ref{nesf}) satisfy
347: $$
348: N_{\leftrightarrow}(\omega)=N_{\updownarrow}(\omega)=\frac{1}{e^{\beta(\omega)}-1},
349: $$
350: the condition (\ref{condition_0}) becomes equivalent to
351: \begin{equation}\label{condition_1}
352: \beta(\omega_{31}) < \beta(\omega_{32})+\beta(\omega_{21}).
353: \end{equation}
354: This process does not break the energy-conservation law. In fact 
355: one can easily check that 
356: \begin{equation}
357: \frac{d}{dt} \left(\sum_{j=1}^3 \epsilon_jP_j(t)+\sum_{\sigma}
358: \int dk~\omega(k) Tr\Big(\rho_{tot}(0)U^\dagger_t 
359: a^\dagger_{k,\sigma}a_{k,\sigma}~U_t \Big) \right)=0,
360: \end{equation}
361: and one should notice that 
362: the energy of the $\updownarrow$-field would be 
363: converted to the energy of the $\leftrightarrow$-field through the 
364: stationary state of the atom (Fig.~2).
365: This is  a peculiar property of the 
366: stimulated emission from a  non-equilibrium state of the field. 
367: In fact the usual Gibbs states are characterized by the fact 
368: that the function $\beta$ in (\ref{nesf}) is linear, i.e. 
369: $\beta(\omega)=\beta \omega$, and in this case condition (\ref{condition_1}) is never satisfied.
370: The non-equilibrium field can therefore drive the atom to a stationary 
371: state which can emit photon, and this means that this stationary state 
372: gives an example of dissipative structure in the Prigogine 
373: sense~\cite{ref:Prigogine}.
374: 
375: In order to understand better the physical meanings of our condition 
376: (\ref{assumption0}), 
377: let us consider an extreme case such as  
378: \begin{equation}
379: N_{\leftrightarrow}(\omega) \ll N_{\updownarrow}(\omega).
380: \end{equation}
381: In this situation, 
382: the relation (\ref{st_sol_2}), defining the stationary solution, 
383: becomes approximately 
384: \begin{eqnarray}
385: \frac{P_2}{P_1}&\sim& \frac{\gamma_{31,\updownarrow}^{(+)}
386: \gamma_{32,\updownarrow}^{(-)}}
387: {\gamma_{32,\updownarrow}^{(-)}\gamma_{31,\updownarrow}^{(+)}}
388: =\frac{N_{\updownarrow}(\omega_{32})}{N_{\updownarrow}
389: (\omega_{32})+1}\frac{N_{\updownarrow}(\omega_{31})+1}
390: {N_{\updownarrow}(\omega_{31})}
391: \left(\neq \frac{N_{\updownarrow}(\omega_{21})}
392: {N_{\updownarrow}(\omega_{21})+1},~\mbox{in general but Gibbs}\right)
393: \label{double_einstein}\\
394: \frac{P_3}{P_1}&\sim& \frac{\gamma_{31,\updownarrow}^{(+)}}
395: {\gamma_{31,\updownarrow}^{(-)}}\qquad=
396: \frac{N_{\updownarrow}(\omega_{31})}{N_{\updownarrow}
397: (\omega_{31})+1}\\
398: \frac{P_3}{P_2}&\sim& \frac{\gamma_{32,\updownarrow}^{(+)}}
399: {\gamma_{32,\updownarrow}^{(-)}}\qquad=\frac{N_{\updownarrow}(\omega_{32})}
400: {N_{\updownarrow}(\omega_{32})+1}.
401: \end{eqnarray}
402: Notice that (\ref{double_einstein}) is the {\it Double Einstein formula}
403: discussed in \cite{ref:Accardi_KI_Kozyrev} and that 
404: this stationary state is not a detailed balance but a {\it distorted balance} 
405: state. 
406: It is obvious that we need at least a 3-level atom to get such 
407: a distorted balance state because, since 
408: detailed balance means balance at each transition frequency, 
409: it follows that, if there is only one such frequency 
410: as in a 2-level atom, then every stationary 
411: state is a detailed balance state.
412: This distorted balance state should play the role of the {\it negative temperature} state introduced in the phenomenology of laser systems\cite{ref:laser}.
413: Under the assumption (\ref{assumption0}) $\updownarrow$-field can be 
414: interpreted as the {\it pumping field} which realizes the so-called 
415: inverse population state of atom and the  
416: $\leftrightarrow$-field  as a {\it stimulating field} which stimulates 
417: the emission from 2 to 1.
418: 
419: 
420: 
421: Finally, let us discuss to the stability of the non-equilibrium 
422: state of the field.  
423: With the above analysis, the non-equilibrium state seems
424: extremely fragile against the interaction with atoms, while the 
425: equilibrium state seems robust. 
426: This difference should play an essential 
427: role to understand the dynamical properties of the 
428: nonequilibrium state of the field itself.
429: It is an interesting problem to consider whether the non-equilibrium 
430: state after some interaction can become stable or not, and 
431: this problem should be related to the discussion on the stability 
432: of the macroscopic state. 
433: In order to approach these problems, we would have to take account of some kind of self-interaction of the field, that is the re-stimulation by 
434: the emitted photon, which we neglected in our discussion.
435: Taking into account these effects, the equations of this system 
436: would be non-linear equations between $P_i(t)$, $n_{\sigma}(t)$ 
437: and other quantities, unlike in this letter.  
438: Our treatment in this letter is not sufficient to deal with 
439: the stability problem of the field.
440: But in another application of the stochastic limit to 
441: superfluidity\cite{ref:Accardi_Kozyrev:superfluid}, 
442: a procedure to deal with this kind of self interaction 
443: has been already 
444: discussed and a non-linear equation has been derived. 
445: We think that we will be able to approach  the stability problems of 
446: the non-equilibrium field also with a similar procedure.
447: 
448: 
449: %---------- ACKNOWLEDGEMENT ------------------------------------
450: \bigskip
451: The authors are grateful to I.V.Volovich for discussion.
452: Kentaro Imafuku and Sergei Kozyrev are grateful to Centro Vito 
453: Volterra and Luigi Accardi for kind hospitality.
454: This work was partially supported by INTAS 9900545 grant. 
455: Kentaro Imafuku is supported by a overseas research fellowship 
456: of Japan Science and Technology Corporation.
457: Sergei Kozyrev was partially supported by RFFI 990100866 grant.
458: 
459: 
460: %---------- REFERENCES ------------------------------------------
461: 
462: \begin{references}
463: \bibitem{ref:Einstein}
464: Einstein, A. Phys. Z. {\bf 18}, 121 (1917)
465: 
466: \bibitem{ref:laser}
467: See, e.g. H. Haken,  In 
468: {\it Encyclopedia of Physics, Vol. XXV/2c, Laser theory}
469: (Springer, Berlin-Heidelberg-New York 1970)
470: \bibitem{ref:Accardi_text}
471: L.Accardi, ~Y.G.Lu, ~I.V.Volovich,  
472: {\it ~Quantum theory and its stochastic limit},  Springer-Verlag (in press)
473: 
474: \bibitem{ref:Accardi_Kozyrev_note}
475: L.Accardi,~S.V.Kozyrev,
476: ~{\it Quantum interacting particle systems}, 
477: Lecture Note of Levico school, September 2000, Preprint Centro Vito Volterra 
478: N.431
479: \bibitem{ref:Accardi_KI_Kozyrev}
480: L.Accardi, ~K.Imafuku, ~S.V.Kozyrev, 
481: {\it Interaction of 3-level atom with radiation}, Preprint Centro Vito Volterra
482: 
483: \bibitem{ref:Prigogine}
484: P. Glansdorff, I.Prigogine, {\it Thermodynamic theory of structure, stability 
485: and fluctuations} 
486: (Wiley-Interscience, London, 1971) 
487: 
488: \bibitem{ref:Accardi_Kozyrev:superfluid}
489: L.Accardi, ~S.V.Kozyrev, Superfluidity in the stochastic limit
490: , Preprint Centro Vito Volterra.
491: \end{references}
492: %---------- FIGURE -----------------------------------------
493: 
494: \newpage
495: \vspace*{2cm}
496: 
497: \begin{figure}
498: \epsfxsize=0.9\textwidth
499: \begin{center}
500: \epsfbox{pfig1.eps}
501: \end{center}
502: \caption{Schematical illustration of our assumption (\ref{assumption0}).}
503: \label{fig1}
504: 
505: \vspace{2cm}
506: 
507: \epsfxsize=0.9\textwidth
508: \begin{center}
509: \epsfbox{pfig2.eps}
510: \end{center}
511: \caption{The energy flow from $\updownarrow$-photon to 
512: $\leftrightarrow$-photon.}
513: \label{fig2}
514: \end{figure}
515: %---------- END OF MAIN -----------------------------------------
516: %\end{multicols}
517: 
518: \end{document}
519: 
520: 
521: 
522: