1: \documentstyle[12pt]{article}
2: \topmargin -25mm
3: \oddsidemargin 0.25cm
4: \textwidth 165mm
5: \textheight 24cm
6: \def\baselinestretch{1.3}
7: \begin{document}
8:
9: \title{Test of the Quantum Chaoticity Criterion for
10: Diamagnetic Kepler Problem}
11: \author{V.E. Bunakov, I.B. Ivanov, R.B. Panin
12: \footnote{e-mails: Vadim.Bunakov@pobox.spbu.ru, Ivan.Ivanov@thd.pnpi.spb.ru}
13: \\Petersburg Nuclear Physics Institut, 188350, Gatchina, Russia}
14:
15:
16: \date{}
17: \maketitle
18: \begin{center}
19: {\bf Abstract:}
20: \end{center}
21: {\small
22: The earlier suggested criterion of quantum chaoticity, borrowed from
23: the nuclear compound resonance theory, is used in the analysis of the
24: quantum diamagnetic Kepler problem (the spinless charged particle motion
25: in the Coulomb and homogenious magnetic fields).
26: }
27:
28: \section{Introduction}
29:
30: It is believed for decades that the main feature of the classically
31: chaotic system is the instability of its trajectories to minor variations
32: of the initial conditions. Since the concept of the trajectory in phase space
33: does not apply in quantum mechanics, the possibility of quantum chaos is still
34: opened to discussion. It was suggested [1] to look for the "quantum signatures
35: of classical chaos". The only more or less generally accepted "signature"
36: found up to now is the Wigner level repulsion. In this paper we proceed with
37: the illustration of our alternative suggestions [2-5]
38: concerning the definition of chaos for the Hamiltonian quantum (and classical)
39: systems.
40:
41: According to Liouville-Arnold theorem in classical mechanics,
42: the Hamiltonian system with $N$ degrees of freedom is regular if it has
43: $M=N$ independent global integrals of motion. If the number $M$ of global
44: integrals becomes less than $N$, the system becomes chaotic. The well
45: known Noether's theorem connects the existence of global integrals of
46: the system with the symmetries of its Hamiltonian. According to this theorem,
47: breaking the symmetry of the initially regular system decreases the number
48: of its independent global integrals of motion. Thus the system becomes
49: chaotic {\bf only} in the case of such a symmetry-breaking which makes the
50: number $M$ of global integrals {\bf less} than $N$.
51:
52: Our first (and major) suggestion is to generalize this definition of
53: chaoticity for the case of quantum systems. Since the concept of
54: symmetry (unlike the trajectory) is universal for both classical and quantum
55: mechanics, this
56: generalization seems to be quite straightforward - one should simply
57: substitute
58: the integrals of motion by the corresponding 'good' quantum numbers, resulting
59: from the symmetries of quantum Hamiltonian. This approach immediately
60: allows to treat the only generally accepted signature of quantum chaos -
61: Wigner's level repulsion - as a signature of symmetry-breaking leading to
62: chaos. Indeed, the general property of the highly symmetrical regular quantum
63: system is the high degeneracy of its eigenstates. The immediate consequence
64: of perturbation breaking the original symmetry is the removal of this
65: degeneracy (in other words, Wigner's level repulsion). It is worthwhile to
66: remind that Wigner's level repulsion was first observed for the resonance
67: states of compound nucleus, whose only good quantum numbers are their energy
68: and spin. This property comes from the fact that the symmetries of the
69: nuclear mean field are destroyed by the pair-wise "residual" interactions
70: [2--5].
71:
72: Our second (rather technical) suggestion is to use the concept of spreading
73: width $\Gamma_{spr}$ (and the related criterion $\ae$) as a sensitive measure
74: of symmetry-breaking of the Hamiltonian $H_0$ caused by the perturbation
75: $ V$. Indeed, consider a Hamiltonian $H$ of the non-integrable
76: system as a sum:%\1
77: \begin{equation}
78: H\ =\ H_0+V
79: \end{equation}
80: of the highly symmetrical regular Hamiltonian $H_0$ (say, of non-interacting
81: particles or quasi-particles in the spherically-symmetrical mean field):%\2
82: \begin{equation}
83: H_0\phi_k=\epsilon_k\phi_k
84: \end{equation}
85: and of the perturbation $V$ which destroys the symmetries of $H_0$ (the
86: pair-wise particle-particle forces in nuclear case). Expand now the eigenstates
87: $\psi_i$ of $H$ over the "regular" basis $\phi$:%\3
88: \begin{equation}
89: \psi_i=\sum_k c^k_i\phi_k
90: \end{equation}
91: and look for the probability $P_k(E_i)=|c^k_i|^2$ to find the original
92: "regular" component $\phi_k$ in the different eigenstates $\psi_i$ (with
93: eigenenergies $E_i$) of our nonintegrable system. It is obvious, that for
94: sufficiently small perturbations $V$ the probability $P_k(E_i)$ is centered
95: around the "original" energy $\epsilon_k$ and tends to saturate to unity
96: over some characteristic energy interval $\Gamma_{spr}$ which is called
97: "the spreading width" of the initially unperturbed state $\phi_k$. Various
98: realistic models (see e.g. chapter 2 of ref.[6]) give the Lorentzian shape
99: of the strength function energy dependence:%\4
100: \begin{equation}
101: S_k(E_i)=\frac{|c^k_i|^2}{D}\approx\frac{}{2\pi}
102: \frac{\Gamma_{spr}}{(E_i-\epsilon_k)^2+\Gamma_{spr}^2/4}
103: \end{equation}
104: where D is the average level spacing of the nonintegrable system. A slight
105: generalization of the derivation given in ref. [6] allows [3,4] to express the
106: spreading width in terms of the "mean square root" matrix element
107: $\tilde{v}=\sqrt{<v^2>}$ of the interaction $V$ mixing the basic states $\phi$
108: (angular brackets imply averaging over all the basic components admixed by
109: $V$ to a given one):%\5
110: \begin{equation}
111: \Gamma_{spr}\sim \tilde{v}\sqrt{N_d}
112: \end{equation}
113: Here $N_d$ stands for the degeneracy rank of the initial level $\epsilon_k$.
114:
115: Thus the system formally becomes nonintegrable as soon as $\Gamma_{spr}$
116: deviates from zero. However, while the ratio %\6
117: \begin{equation}
118: \ae=\frac{\Gamma_{spr}}{D_0}
119: \end{equation}
120: (where $D_0$ is the level spacing of the initial regular system) is smaller
121: than unity - the traces of the initial good quantum numbers are quite obvious
122: as isolated maxima of the strength function. We can easily distinguish
123: between the maxima corresponding to the different values of the originally
124: good quantum numbers.
125: This is the analogue of the classical "weak chaos" governed by the KAM
126: theorem.
127: When $\ae$ exceeds unity these traces of regularity disappear since it
128: becomes impossible to distinguish between the successive maxima of the
129: strength function corresponding to the different values of the original
130: quantum numbers $k$. This situation is the quantum analogue of the smearing
131: out and disappearance of the invariant tori. It means that we
132: approach the domain of "global" or "hard" chaos.
133:
134: Furrier transforming Eq. (4) one can show (see e.g. [6]) that
135: $\Gamma_{spr}/\hbar$ defines the rate of decay of the "regular" states
136: $\phi$ resulting from the instability caused by the perturbation $V$. One
137: can even
138: form the wave packets $|A>$ of the states $\phi_k$ and analyze the recurrence
139: probabilities $P(t)=|<A(t)|A(0)>|^2$. This analysis shows [3] the periodic
140: recurrences with classical period $T$ modulated exponentially by the factor
141: $exp(-\Gamma_{spr}t/\hbar)$ arising from the above instability. Combining
142: these results with the results of Heller's wave-packet experience
143: (see e.g. [7] or paragraph 15.6 of ref. [8]),
144: one can show that the quantity $\Gamma_{spr}/\hbar$ transforms in the
145: classical limit into the Lyapunov exponent $\Lambda$:%\7
146: \begin{equation}
147: \frac{\Gamma_{spr}}{\hbar}\rightarrow\Lambda
148: \end{equation}
149: The corresponding classical limit for the dimensionless chaoticity criterion
150: is:%\8
151: \begin{equation}
152: \ae\rightarrow\frac{\Lambda T}{2\pi}=\frac{\chi}{2\pi}
153: \end{equation}
154: where $T$ is the classical period and $\chi$ is the stability parameter of
155: the classical monodromy matrix (see, e.g. [8]).
156:
157: Thus the particular quantity $\Gamma_{spr}$ and the parameter $\ae$ seem to
158: be more accurate numerical
159: measures of quantum chaoticity than the level distribution law - this is
160: proved by the nuclear physics experience [2-4] and by its application
161: to one of the most popular in classical mechanics cases of transition
162: from regularity to chaos - Hennon-Heiles problem [5].
163:
164:
165: \section{Diamagnetic Kepler Problem}
166:
167: Another very popular model for studies of transition from regularity
168: to chaos in classical mechanics is the non-relativistic hydrogen atom in
169: the uniform magnetic field (see e.g. [8,10,14]) with the Hamiltonian:%9
170: \begin{equation}
171: H=p^2/2m -e^2/r +\omega l_z +\frac{1}{2}m\omega ^2(x^2+y^2)
172: \end{equation}
173: Here the frequency $\omega=eB/2mc$ is a half of the cyclotron frequency
174: and $B$ is the strength of
175: the magnetic field acting along z-axis. The dimensionless field strength
176: parameter $\gamma=\hbar\omega/{\cal R}$ (here ${\cal R}$ is the Rydberg energy) is
177: usually combined with the electron energy $E$ to produce the scaled energy
178: $\epsilon=E\gamma^{-2/3}$. When the scaled energy varies from $-\infty$
179: (for $B=0$) to 0 (for $B=\infty$) the regular motion of the system becomes
180: more and more chaotic. The fraction $R$ of available phase space covered by
181: regular trajectories was calculated in ref. [9 - 10] as a function of scaled
182: energy for the case of $l_z=0$ (see Fig.1), showing the rapid chaotisation of
183: the system in the range $-0.48\le\epsilon\le -0.125$.
184:
185: We analyzed the quantum analogue of this system (with the Hamiltonian
186: (9))
187: on the same lines as it was done in [5] for the quantum Henon-Heiles problem,
188: namely we traced the gradual destruction of the $O(4)$ symmetry characteristic
189: of the unperturbed motion in Coulomb potential by the external magnetic field
190: B. In other words, we traced the disappearance of the "good" quantum numbers
191: (integrals of motion) which characterize the regular motion in this potential.
192: In order to do this, we diagonalized the Hamiltonian matrix (9) in parabolic
193: coordinates on the basis of purely Coulomb wave functions $\phi_{n_1 n_2 m}$,
194: whose eigenvalues in the unperturbed case are defined by the principal
195: quantum number n:
196: $$ n=n_1+n_2+|m|+1 $$
197: and are highly ($n^2$ times) degenerate. Diagonalizing the Hamiltonian matrix,
198: we defined the new
199: eigenvalues $E_i$ and the eigenstates $\psi_i$ in terms of the expansion
200: coefficients $c^k_i$ (see Eq. (3)).
201: As a next step, we plotted the energy distribution
202: of Eq. (4) for the coefficients' squares of the $n$-th shell
203: over the "new" eigenstates. Fig. 2 shows the examples of these
204: distributions for $n=10$, $m=0$ and the magnetic field $\gamma$ equal to
205: $4\cdot 10^{-4}$, $6\cdot 10^{-4}$, $8\cdot 10^{-4}$ and $12\cdot 10^{-4}$,
206: respectively. In order to increase the statistical accuracy, we performed
207: the averaging over all the components of the basis with the same $n$ value,
208: as it is usually done in nuclear physics and as it was done in the case
209: of quantum Henon-Heiles problem [5].
210: Assuming now that the shape of these distributions is approximately
211: Lorentzian, like in the case of the neutron strength function in nuclear
212: physics, we define $\Gamma_{spr}$ as the energy range around the maximum
213: where the sum of the squares of the coefficients $\sum_i|c^k_i|^2$
214: saturates to 0.5. Thus obtained values of $\Gamma_{spr}$ were divided
215: then by the level spacing $D_0$ between the adjacent maxima of the strength
216: function to give the desired parameter
217: $\ae$. The plot of this parameter versus the scaled energy $\epsilon$ is
218: given in Fig. 1.
219:
220: We see that our parameter reaches the critical value of $\ae=1$ at the
221: critical scaled energy $\epsilon\approx -0.45$ in fairly good agreement
222: with the classical critical value $\epsilon\approx -0.48$ of refs. [9,10].
223: It is worthwhile to remind here that in the previous studies of the quantum
224: diamagnetic Kepler problem [11 - 14] the existence was pointed of the
225: approximately good quantum number $K$, corresponding to the eigenvalues
226: of the operator $\Sigma$ built as a combination of the Runge-Lenz vector
227: $A$:%10
228: \begin{equation}
229: \Sigma=4A^2 - 5A_z^2
230: \end{equation}
231: The eigenstates of this operator are obtained [12, 13] by prediagonalization
232: of the unperturbed Coulomb basis within a single manifold $n$ (which
233: physically corresponds to the values of our $\ae\leq 1$). The appreciable
234: $K$-mixing (disappearance of the integral of motion $\Sigma$) starts [13, 14]
235: when $\gamma^2 n^7\approx 16$. In our case of $n=10$ this corresponds to
236: the scaled energy $\epsilon\approx -0.45$.
237:
238: \section{Conclusion}
239: Thus we confirmed once more the plausibility of the suggested approach to
240: quantum chaoticity, based on its connection with symmetry-breaking of the
241: regular motion which makes the number $M$ of the system's global integrals of
242: motion less than the number $N$ of its degrees of freedom. We had also
243: demonstrated that the spreading width $\Gamma_{spr}$ and the dimensionless
244: parameter $\ae$ might serve a good quantitative criterion of quantum
245: chaoticity. Likewise in the case of Henon-Heiles problem [4], the critical
246: scaled energy value $\epsilon_c$ when the parameter $\ae$ reaches unity
247: corresponds to the onset of "global" chaos on the classical phase portrait
248: for the diamagnetic Kepler problem. Here, however, the origin of the
249: approximate regularity of the perturbed system for $\epsilon\leq\epsilon_c$
250: is more evident. Although formally the external magnetic field makes the
251: system nonintegrable by reducing the number of global integrals of motion
252: to $M=2$ (energy and $l_z$), the third approximate integral of motion
253: ($\Sigma$)
254: survives much longer making the system practically regular.
255:
256: We should add in conclusion that the importance of studying the particular
257: example of hydrogen atom in the uniformed magnetic field was stressed (see,
258: e.g. [14]) because it "is not an abstract model system but a real physical
259: system that can be and has been studied in the laboratory". These studies
260: were indeed started in 1986 (see [15]). One should point, however, that
261: atomic nucleus is also "not an abstract model", whose experimental and
262: theoretical studies are going on for more than half a century. As we had
263: already mentioned, Wigner developed his random matrix approach in order
264: to describe the experimentally observed properties of compound nuclear
265: resonances. Since those times nuclear physics accumulated a vast arsenal
266: of theoretical methods which allow Schr{\"o}dinger's equation to be
267: solved in some effective manner, even when the system is not integrable and
268: its behavior is chaotic by the criteria of level repulsion. As shown in [2-4],
269: most of
270: them are based on the smallness of the chaoticity parameter $\ae$, which seems
271: to be the most important small parameter of nuclear physics.
272:
273:
274: \begin{thebibliography}{99}
275: \bibitem{1} M.Berry,in
276: "Quantum Chaos. Adriatico Research Conf., Trieste,1990",
277: H.A.Cerdeira, R.Ramaswamy, M.Gutzwiller, G.Casati
278: (editors), World Scientific, 1991, p.VII
279: \bibitem{2} V.E.Bunakov, in: Proc.Int.Conf. on Selected Topics in Nuclear
280: Structure, Dubna, 1994, (JINR publications E4-94-370, Dubna, 1994) 310
281: \bibitem{3} V.E.Bunakov, in: Lecture Notes of XXX PNPI Winter School,
282: (St.Petersburg, 1996) 135
283: \bibitem{4} V.E. Bunakov, Physics of Atomic Nuclei (to be published in
284: No 1, 1999)
285: \bibitem{5} V.E.Bunakov , F.F.Valiev , Yu.M.Tchuvilsky, Bull.Russ.Acad.Sci.
286: Physics. 62 (1998) 41; Phys. Lett. A (in press)
287: \bibitem{6} O.Bohr, B.Mottelson, Nuclear Structure, v.I, (Benjamin, N-Y,
288: 1969)
289: \bibitem {7} E.Heller, in: Chaos and Quantum Physics, Les Houches,
290: Ses.LII, 1989, (Elsevier, 1991) 548
291: \bibitem{8} M.Gutzwiller, Chaos in Classical and Quantum Mechanics,
292: (Springer, 1990)
293: \bibitem{9} Harada A., Hasegawa J., J.Phys. A16 (1983) L259.
294: \bibitem{10} H.Hasegawa, M.Robnik, G.Wunner, Prog. Theor. Phys. Suppl.
295: 98 (1989) 198
296: \bibitem{11} E.A.Soloviev, JETP Lett. 34 (1981) 265
297: \bibitem{12} D.Delande, J.C.Gay, J.Phys. B 17 (1984) L335
298: \bibitem{13} D.Wintden, H.Friedrich, J.Phys. B 19 (1986) 1261
299: \bibitem{14} H.Friedrich, D.Wintgen, Phys.Rep. 183 (1989) 38.
300: \bibitem{15} J.Main, G.Wiebusch, A.Holle, K.H.Welge, Phys.Rev.Lett. 57 (1986)
301: 2789
302: \end{thebibliography}
303: \end{document}