(NDWP1D) Schlagheck Epjd03
(NDWP1D) Schlagheck Epjd03
(NDWP1D) Schlagheck Epjd03
Abstract. We provide a detailed quantum treatment of the spectral characteristics and of the dynamics of
nondispersive two-electron wave packets along the periodically driven, collinear frozen planet configuration
of helium. These highly correlated, long-lived wave packets arise as a quantum manifestation of regular
islands in a mixed classical phase space, which are induced by nonlinear resonances between the external
driving and the unperturbed dynamics of the frozen-planet configuration. Particular emphasis is given to
the dependence of the ionization rates of the wave packet states on the driving field parameters and on
the quantum mechanical phase space resolution, preceded by a comparison of 1D and 3D life times of
the unperturbed frozen planet. Furthermore, we study the effect of a superimposed static electric field
component, which, on the grounds of classical considerations, is expected to stabilize the real 3D dynamics
against large (and possibly ionizing) deviations from collinearity.
PACS. 32.80.Rm Multiphoton ionization and excitation to highly excited states (e.g., Rydberg states) –
05.45.Mt Semiclassical chaos (“quantum chaos”) – 31.25.Jf Electron correlation calculations for atoms
and ions: excited states
The above interpretation in terms of nonlinear reso- the quantum dynamics along this axis. Indeed, nondisper-
nances in a mixed phase space suggests that the concept of sive two-electron wave packets associated with the classi-
nondispersive wave packets can be generalized to atomic cal resonances are found in quantum calculations on the
(or molecular) systems that are more complex than hy- collinear model atom, as we have already reported in a
drogen or hydrogen-like Rydberg atoms – such as, e.g., previous letter [24]. Here we shall give detailed account of
the correlated dynamics in doubly excited helium. How- our quantum treatment of these objects, identify them as
ever, such a generalization is not at all obvious, since most special eigenstates in the Floquet spectrum of the driven
doubly excited two-electron orbits of the bare three-body atomic system (alike their counterparts in driven hydrogen
Coulomb problem (as the classical version of the doubly atoms), and investigate their characteristic properties.
excited helium atom) are unstable and promote rapid au- To make the present contribution self-contained, we
toionization of the configuration. This prevents the con- start out in Section 2 with a review of the classical dy-
finement of an initially localized wave packet on longer namics of the driven frozen-planet configuration, which
time scales. was already described elsewhere [20,21]. Section 3 is de-
Nonetheless, extensive studies on the classical dynam- voted to the formulation of the quantum eigenvalue prob-
ics of helium [17] have identified local regions of regu- lem for the collinear model. As a first application thereof,
lar motion which are associated with configurations of we study the unperturbed collinear atom in Section 4,
particular symmetry of the two-electron atom. The most what provides a remarkable example of regular dynamics
prominent and surprising example is the “frozen planet” in an atomic two-electron system. In Section 5, we show
orbit [18,19], a highly polarized, nearly collinear con- how the nondispersive wave packets emerge in the spec-
figuration with asymmetric excitations of the electrons. trum of driven helium (Sect. 5.1) and discuss their life-
The frozen planet’s stability essentially relies on the non- times (Sect. 5.2), as well as their properties in presence of
negligible electron-electron interaction (the electrons are, an additional static electric field (Sect. 5.3).
contrary to intuition, located on the same side of the
nucleus). Stable eigenmodes sustain radial, vibration-like
oscillations of the outer electron, as well as coupled trans- 2 Classical dynamics of the driven
verse excursions of the outer and inner electron. Well- frozen-planet configuration
defined eigenfrequencies are associated with these eigen-
modes, what makes the configuration a good candidate 2.1 The frozen-planet configuration of helium
for inducing nonspreading wave packets in the correlated
two-electron dynamics, by near-resonant external driving. In atomic units, which are used throughout this paper,
In fact, our earlier analysis of the classical dynamics of the classical Hamiltonian of the electromagnetically driven
helium in a linearly polarized electromagnetic field already helium atom reads
provided strong support for nondispersive wave packets p21 p2 Z Z 1
along the frozen-planet orbit [20,21]. The external per- H= + 2 − − +
2 2 |r1 | |r2 | |r1 − r2 |
turbation induces regular islands within the phase space
of collinear motion (where electrons and nucleus are lined + F(t) (r1 + r2 ), (1)
up parallel to the driving field polarization axis), which
with ri = (xi , yi , zi ) and pi = (pix , piy , piz ) the position
are due to nonlinear resonances between the drive and
and the momentum of electron i = 1, 2, respectively, Z = 2
the oscillation of the outer electron. The dynamics within
the nuclear charge, and
these resonances is in general unstable with respect to
deviations from collinearity (in contrast to the analogous F = (F cos ωt + Fst ) ez (2)
scenario in atomic hydrogen, the resonantly driven one
dimensional Kepler orbit [22]). However, a static electric the external, linearly polarized driving at frequency ω
field applied parallel to the microwave polarization axis and amplitude F , possibly superimposed by an additional,
can be used to enforce stability of the driven configura- static component with field strength Fst (ez represents the
tion in all three spatial dimensions [20]. This suggests that unit vector along the z-axis). In analogy to driven hydro-
nondispersive two-electron wave packets can be built on gen [25] as well as to the unperturbed helium atom [26],
these stabilized resonances, and, on the basis of semiclas- the Hamiltonian (1) exhibits general scaling laws: the clas-
sical considerations, at experimentally accessible quantum sical dynamics generated by (1) remain invariant if all
numbers of doubly excited helium [21,23]. Yet, this con- variables and parameters of the system are transformed
jecture remains to be corroborated by an exact quantum according to
treatment.
With the presently available computational facilities, ri −→ N 2 ri (i = 1, 2), (3)
full blown quantum ab initio calculations of highly cor- pi −→ N −1 pi (i = 1, 2),
related two-electron wave packet states under electromag- t −→ N 3 t,
netic driving are prohibitive. However, since the stabilizing
F −→ N −4 F,
static field confines the wave packet to the near vicinity
of the field polarization axis, essential properties of these ω −→ N −3 ω,
wave packets appear accessible through the restriction of H −→ N −2 H,
P. Schlagheck and A. Buchleitner: Nondispersive two-electron wave packets in driven helium 403
0.6
the outer electron’s slow oscillation, i.e., a field with am- a b c
plitude F < FI and frequency ω ωI .
−1
pz1 / N
0
−0.6
0 5 0 5 0 5
2.2 Phase space of the collinear, driven configuration 2
z1 / N
In the presence of the time-periodic perturbation, the dy- Fig. 4. Phase space of the collinear, driven frozen-planet
namics of the collinear, driven frozen-planet evolves in a configuration, for variable field amplitudes, (a) F = 0,
five-dimensional phase space spanned by the positions and (b) 0.001 N −4 , (c) 0.005 N −4 , at fixed driving frequency ω =
momenta of the electrons, and by the phase ωt of the driv- 0.2 N −3 . The momentum pz1 of the outer electron is plotted
ing field. A complete visualization of the driven dynamics versus its position z1 at fixed action N = 1 of the Kepler
mode, fixed position z2 = 0 of the inner electron, and fixed
within a simple two-dimensional Poincaré surface of sec-
phase ωt = 0 (mod 2π) of the driving field [21, 23]. The regular
tion is therefore not possible. However, for ω ωI , and for
phase space of the unperturbed atom (a) turns mixed regular-
field amplitudes F < FI which do not appreciably affect chaotic in presence of the external perturbation (b, c). The
the Kepler motion of the inner electron, the separation of nonlinear 1:1 resonance between the driving field and the un-
time scales allows to map the phase space structure onto a perturbed oscillation of the electron gives rise to a substructure
two-dimensional surface by means of a two-step Poincaré at F = 0.001 N −4 (b) (centered around z1 4.8 N 2 ), which
section method [21,23]. We begin with a simple Poincaré transforms into a separate regular island within the chaotic sea
section by plotting the outer electron’s phase space vari- at larger field amplitude F = 0.005 N −4 (c).
ables z1 , pz1 whenever the inner electron reaches the nu-
cleus (z2 = 0) – i.e., whenever the phase of the Kepler
oscillation assumes a fixed value. In the time domain, the
resonances between the external driving and the unper-
points obtained by this section are separated by the Kepler
turbed oscillation of the outer electron, emerge within
period which is much shorter than any other time scale in
the regular region. The prominent substructure located
the system. The dynamics generated by the slow modes
around z1 4.8 N 2 for F = 0.001 N −4 (Fig. 4b) corre-
– i.e., the outer electron’s oscillation within the effective
sponds to the 1:1 resonance where one oscillation cycle of
potential, as well as the external driving – is therefore very
the outer electron is completed after precisely one period
well monitored by this sequence of points.
of the driving field.
Neglecting the effect of resonances between fast and
slow dynamics, the intersection of the Lagrangian mani- At larger field amplitude (F 0.003 N −4 for the fre-
fold with the surface defined by z2 = 0 can now be very quency used in Fig. 4), the field-induced resonance even-
well approximated by a cubic interpolation between sub- tually disconnects from the rest of the regular domain.
sequent points in the above sequence. This yields a con- The phase space then exhibits two separate, large reg-
tinuous trajectory which can be used to perform a second ular islands embedded into the chaotic sea (see Fig. 4c
Poincaré section, by fixing the phase of the driving field at F = 0.005 N −4): the intrinsic island (centered around
ωt = φ0 (mod 2π). As a consequence, we effectively plot z1 2.5 N 2 in Fig. 4c) which is essentially due to the
the outer electron’s position and momentum for z2 = 0 intrinsic nonlinear frozen-planet dynamics with the ex-
and for fixed ωt. The reduction of the phase space to a ternal field acting only as a small perturbation, and the
two-dimensional surface is completed by restricting the field-induced 1:1 resonance island (centered around z1
initial conditions to those with fixed value N = 1 of the 4.9 N 2 in Fig. 4c) which arises from the combination of two
inner electron’s action. This latter quantity represents an equally important nonlinear components: the coupling to
adiabatic invariant of the system and therefore remains the external driving field and the internal interactions be-
nearly constant as time evolves. tween the charged particles of the atom.
Figure 4 shows the Poincaré section that is obtained Figure 5 illustrates the effect of tuning the frequency
by the two-step method described above, for fixed fre- ω at fixed field amplitude F = 0.005 N −4. With increas-
quency ω = 0.2 N −3 < ωI , fixed driving phase ωt = 0, ing ω, the field-induced resonance shifts towards lower ex-
and for variable field amplitude F . In the unperturbed citations within the effective potential, i.e., towards un-
case (F = 0, Fig. 4a), we recognize a regular phase space perturbed orbits with higher frequencies. Consequently,
structure, with closed curves corresponding to the regu- the resonance island is located closer (as compared to
lar oscillation of the outer electron within the effective Fig. 4c) to the equilibrium distance for ω = 0.25 N −3
potential. If we switch on the driving field, phase space (Fig. 5a) and finally undergoes a smooth transition into
turns mixed regular-chaotic. The bounded frozen-planet the intrinsic island at ω = 0.3 N −3 ωI (Fig. 5b). Above
dynamics is now represented by a local regular region cen- the intrinsic frequency ωI – which corresponds to the max-
tered around the equilibrium point, outside which the dy- imum frequency of free oscillations within the effective po-
namics is chaotic and leads to ionization. This remaining tential – the electromagnetic field can no longer induce a
regular region is, due to the Kolmogorov-Arnold-Moser 1:1 resonance. Resonances of higher order, however, are
(KAM) theorem [27], still large for weak field amplitudes still possible (e.g., the 2:1 resonance, where one period of
(F = 0.001 N −4, Fig. 4b), but decreases in size with in- the outer electron is completed precisely after two field
creasing F . As a consequence of the Poincaré-Birkhoff cycles; the elliptic islands created by this resonance are
theorem [27], elliptic substructures, induced by nonlinear located at z1 4 N 2 and pz1 ±0.2 N −1 in Fig. 5c).
P. Schlagheck and A. Buchleitner: Nondispersive two-electron wave packets in driven helium 405
0.6
a b c
−1
pz1 / N
−0.6
0 5 0 5 0 5
2
z1 / N
Hamiltonian, the external electric field consists of an oscil- in triply excited lithium). In order to separate these res-
lating part F – which, in contrast to (1), is incorporated in onances from the flat background of the continuous spec-
the velocity gauge in order to ensure better convergence of trum, we use the method of complex scaling [45–47]. It
the numerical calculation [39] – and of a static component consists in the complexification of coordinates and mo-
Fst > 0 which forces the electrons away from the nucleus. menta (not of time!) according to1
In a last step, we introduce the coordinates
∂ ∂
ξ −→ ξeiθ −i −→ −i e−iθ
ξ = z1 − z2 , η = z2 , (5) ∂ξ ∂ξ
∂ ∂
which independently lie in the range 0 ≤ ξ, η ≤ ∞. η −→ ηeiθ −i −→ −i e−iθ , (10)
∂η ∂η
In these perimetric coordinates [40], the Hamiltonian is
rewritten as through application of the nonunitary complex scaling op-
(E)
erator R(θ) on ψt ,
∂2 1 ∂ ∂2 Z Z 1
H =− 2
− 2
+ − − + (E) (E) (E)
∂ξ 2 ∂η ∂ξ∂η ξ+η η ξ R(θ)ψt (ξ, η) = eiθ ψt (ξeiθ , ηeiθ ) =: ψθ,t (ξ, η) . (11)
F 1 ∂
− sin ωt − Fst (ξ + 2η), (6) Accordingly, the Hamiltonian is transformed as
ω i ∂η
what will turn out as a convenient form for our further H −→ Hθ = R(θ) H R(−θ), (12)
treatment.
what turns (8) into a complex symmetric rather than
hermitian eigenvalue problem, with eigenvalues in the
lower half of the complex plane. In particular, resonances
3.2 Floquet theory and complex scaling
of the “real” eigenvalue problem (i.e., at θ = 0) ap-
pear, for finite θ, as isolated, discrete complex eigenvalues
Due to the temporal periodicity of the Hamiltonian, the E = E − iΓ/2: their real and imaginary parts correspond
Schrödinger problem represented by (6) is conveniently to the energies E and to the half widths at half maximum
treated in the framework of Floquet theory [41,42]: any Γ/2 (HWHM) of the resonances, respectively.
solution of the Schrödinger equation can be expanded in
(E)
a set of 2π/ω-periodic quasienergy-eigenfunctions ψt ,
3.3 Expansion in Sturmian basis functions
(E)
ψt = dECE ψt e−iE t , The complex-scaled Floquet Hamiltonian is now expanded
(E) (E) in the product basis
with ψt+2π/ω = ψt , CE ∈ C (7)
{Sn(α) (ξ)Sm
(β)
(η): n, m ≥ 1} (13)
which satisfy the Floquet eigenvalue equation
composed of the real-valued Sturmian functions [48]
∂ (E) (E)
H −i ψt = E ψt . (8) (−1)n 2ξ ξ 2ξ
∂t (α)
Sn (ξ) = √ exp −
(1)
Ln−1 , (14)
n α α α
(E)
A Fourier series expansion of ψt , (1)
where the Ln−1 denote the associated Laguerre polyno-
∞
mials [49]. The real parameter α > 0 fixes the length
(E) (E)
ψt = ψ̂k ei k ω t , (9) scale of the Sturmians. As a consequence, the product ba-
k=−∞ sis (13) exhibits two independent scaling parameters, α
and β, which permit the optimal tuning of the basis with
yields a time-independent eigenvalue problem for the respect to the relative excitation of the electrons. Also
(E) (α)
Fourier components ψ̂k , where the effective Hamiltonian note that the Sturmians Sn scale at least linearly with
matrix contains the stationary part of the Hamiltonian ξ, for ξ → 0. This implies that the Coulomb singularities
shifted by kω as diagonal, and the Fourier components of of the Hamiltonian (6) do not lead to divergent matrix
the periodic driving as off-diagonal (block) elements. This elements in the product basis (13).
results in an ω-periodic spectrum of quasienergies E. Multiplication of (8) by ξη leaves 1/(ξ + η) as the only
For atomic systems, the Floquet spectrum is absolutely non-polynomial contribution in ξ and η, after inserting (6).
continuous: each bound state of the unperturbed atom is We factor out (ξ + η) from the eigenvectors of (8),
coupled to the atomic continuum via multiphoton transi-
tions, and therefore appears as a resonance structure in (E) (E)
ψt (ξ, η) =: (ξ + η) φt (ξ, η) , (15)
the spectrum – in analogy to autoionizing states in mul-
1
tiply excited atoms, which are coupled to the continuum Note that, due to (5), complexification of ξ, η and of z1 , z2
via configuration interaction [43] (see, e.g., [44] for a case is equivalent.
408 The European Physical Journal D
All together, the generalized Floquet eigenvalue equations 3.4 Visualization of the wavefunction
finally reads
The energies E and widths Γ extracted from the complex
(E)
(ξ + η)ξη(H0θ + kω − E)(ξ + η)φ̂θ,k (ξ, η) + (ξ + η) eigenvalues E allow to classify the Floquet states in se-
(E) (E) ries and to determine their stability; in general, however,
× ξηVθ (ξ + η)(φ̂θ,k−1 (ξ, η) − φ̂θ,k+1 (ξ, η)) = 0 , (17) they do not provide any direct information about the lo-
calization properties of these states in configuration or
with the complex scaled operators H0θ , Vθ given by in phase space, which in themselves can carry important
physical information. To extract the latter from the eigen-
∂ 1 ∂ ∂2
H0θ = − 2 − + e−2iθ states of the complex symmetric eigenvalue problem (17),
∂ξ 2 ∂η 2 ∂ξ∂η some care has to be taken, due to the non-unitarity of the
Z Z 1 −iθ dilation operator R(θ). The technical and computational
+ − − + e
ξ+η η ξ details of this procedure are described in detail in [23,51],
and we only import here those results of immediate rele-
−Fst (ξ + 2η) eiθ , (18)
vance for our specific purpose.
F ∂ −iθ We first need the density of the physical wave function
Vθ = e , (19)
2ω ∂η (E)
(at real energy E) ψt (z1 , z2 ), which is obtained (up to
(E) a normalization constant), for a well-isolated resonance
and the φ̂θ,k (ξ, η) the Fourier components (see (9)) of the energy E = E−iΓ/2 (i.e., |Ej −E| Γ , ∀Ej = E), from the
(E) (E)
wavefunction φθ,t (ξ, η). Introducing the scalar product image of ψθ,t under the inverse dilation operator R(−θ),
∞ ∞ ∞
1
f |g
≡ f (ξ, η) g(ξ, η) dξ dη , (20) (E)
R(−θ)ψθ,t = R(−θ)
(E)
z1 φ̂θ,k (z1 − z2 , z2 ) eikωt ,
0 0 ξ η
k=−∞
with respect to which the Sturmians are orthogonal, we (22)
obtain a complex symmetric eigenvalue problem.
(E) together with the image of the corresponding left eigen-
All the opertor-valued terms acting on the φ̂θ,k and (E)
(E) vector ψθ,−t [23,51]3 :
φ̂θ,k±1 in (17) can now be expressed as polynomials of the
elementary ladder operators of the Sturmian basis (14) (E) (E)
|ψt (z1 , z2 )|2 Re z1 , z2 |R(−θ)|ψθ,t
- -
- -
- -
- -
- -
- -
- -
simulations are feasible), and that similar stabilization 23. P. Schlagheck, Das Drei-Körper-Coulombproblem unter pe-
phenomena are possible also in other, high dimensional riodischem Antrieb, Ph.D. thesis, Technische Universität
atomic or molecular systems under external forcing. How- München, Herbert Utz Verlag, München, 1999
ever, our comparison of the life times of 1D and 3D 24. P. Schlagheck, A. Buchleitner, Europhys. Lett. 46, 24
collinear frozen planet states in the absence of any external (1999)
perturbation also suggest that the life times of nondisper- 25. J.G. Leopold, I.C. Percival, Phys. Rev. Lett. 41, 944 (1978)
26. I.C. Percival, Adv. Chem. Phys. 36, 1 (1977)
sive wave packets dramatically depend on the accessible 27. A.J. Lichtenberg, M.A. Lieberman, Regular and Stochastic
phase space dimensions, as a direct consequence of the Motion (Springer-Verlag, New York, 1983)
interelectronic repulsion – absent in single electron Ryd- 28. V.N. Ostrovsky, N.V. Prudov, J. Phys. B 28, 4435 (1995)
berg dynamics. It remains to be seen whether opening one 29. A. Buchleitner, D. Delande, Phys. Rev. A 55, R1585
single transverse degree of freedom – by extending config- (1997)
uration space from the z-axis to the plane, in a 2D model 30. K. Sacha, J. Zakrzewski, D. Delande, Eur. Phys. J. D 1,
(now within reach for most advanced supercomputing fa- 231 (1998)
cilities) – suffices to restablish the order of magnitude of 31. J.B. Keller, Ann. Phys. (N.Y.) 4, 180 (1958)
the (field free) 3D decay rates, and whether the confine- 32. J. Müller, J. Burgdörfer, D. Noid, Phys. Rev. A 45, 1471
(1992); J. Müller, J. Burgdörfer, Phys. Rev. Lett. 70, 2375
ment by the static field has a systematic effect on the rates
(1993)
in 2D or 3D (with or without an external drive). 33. M. Domke et al., Phys. Rev. A 53, 1424 (1996); B. Püttner
et al., J. Electron. Spectrosc. Relat. Phenom. 101–103, 27
It is a pleasure to acknowledge fruitful and inspiring discussions (1999); B. Püttner et al., Phys. Rev. Lett. 86, 3747 (2001)
with Dominique Delande, Benoı̂t Grémaud, Laurent Hilico, 34. W.E. Cooke, T.F. Gallagher, Phys. Rev. Lett. 41, 1648
Andreas Krug, and Javier Madroñero, as well as financial sup- (1978)
port by the DAAD through the PROCOPE program. 35. U. Eichmann, V. Lange, W. Sandner, Phys. Rev. Lett. 64,
274 (1990)
36. P.M. Koch, Physica D 83, 178 (1995)
37. B. Grémaud, D. Delande, Europhys. Lett. 40, 363 (1997)
References 38. R. Grobe, J.H. Eberly, Phys. Rev. Lett. 68, 2905 (1992)
39. R. Shakeshaft, Z. Phys. D 8, 47 (1988)
1. T.C. Weinacht, J. Ahn, P.H. Bucksbaum, Nature 397, 233 40. C.L. Pekeris, Phys. Rev. 112, 1649 (1958)
(1999) 41. J.H. Shirley, Phys. Rev. 138, B979 (1965)
2. X. Chen, J.A. Yeazell, Phys. Rev. Lett. 81, 5772 (1998) 42. Ya.B. Zel’dovich, Sov. Phys. JETP 24, 1006 (1967)
3. E. Schrödinger, Naturwissenschaften 14, 664 (1926) 43. U. Fano, Phys. Rev. 124, 1866 (1961)
44. L.B. Madsen, P. Schlagheck, P. Lambropoulos, Phys. Rev.
4. B. Kohler et al., Phys. Rev. Lett. 74, 3360 (1995)
Lett. 85, 42 (2000)
5. I.S. Averbukh, M.J.J. Vrakking, D.M. Villeneuve, A.
45. E. Balslev, J.M. Combes, Commun. Math. Phys. 22, 280
Stolow, Phys. Rev. Lett. 77, 3518 (1996)
(1971)
6. A. Buchleitner, Ph.D. thesis, Université Pierre et Marie 46. Y.K. Ho, Phys. Rep. 99, 1 (1983)
Curie, Paris, 1993 47. W.P. Reinhardt, Ann. Rev. Phys. Chem. 33, 223 (1982)
7. D. Delande, A. Buchleitner, Adv. At. Mol. Opt. Phys. 35, 48. R. Blümel, U. Smilansky, Z. Phys. D 6, 83 (1987)
85 (1994) 49. M. Abramowitz, I.A. Stegun, Handbook of Mathemati-
8. A. Buchleitner, D. Delande, Phys. Rev. Lett. 75, 1487 cal Functions, 10th edn. (National Bureau of Standards,
(1995) Washington, 1972), Chap. 22
9. I. Bialynicki-Birula, M. Kaliński, J.H. Eberly, Phys. Rev. 50. D. Delande, A. Bommier, J.C. Gay, Phys. Rev. Lett. 66,
Lett. 73, 1777 (1994) 141 (1991)
10. J. Zakrzewski, D. Delande, A. Buchleitner, Phys. Rev. 51. A. Buchleitner, B. Grémaud, D. Delande, J. Phys. B 27,
Lett. 75, 4015 (1995) 2663 (1994)
11. K. Sacha, J. Zakrzewski, Phys. Rev. A 59, 1707 (1999) 52. R. Blümel, W.P. Reinhardt, in Directions in Chaos, edited
12. J. Zakrzewski, D. Delande, A. Buchleitner, Phys. Rev. E by B.L. Hao et al. (World Scientific, Hong Kong, 1991),
57, 1458 (1998) Vol. 4
13. B.E. Sauer, M.R.W. Bellermann, P.M. Koch, Phys. Rev. 53. H. Friedrich, Theroretical Atomic Physics, 2nd edn.
Lett. 68, 1633 (1992) (Springer, Berlin, 1998)
14. G.P. Berman, G.M. Zaslavsky, Phys. Lett. 61A, 295 (1977) 54. D. Lappas et al., J. Phys. B 29, L619 (1996)
15. J. Henkel, M. Holthaus, Phys. Rev. A 45, 1978 (1992) 55. K. Rza̧żewski, M. Lewenstein, P. Salières, Phys. Rev. A
16. A. Buchleitner, D. Delande, J. Zakrzewski, Phys. Rep. 49, 1196 (1994)
56. F. Grossmann, T. Dittrich, P. Jung, P. Hänggi, Phys. Rev.
368, 409 (2002)
Lett. 67, 516 (1991)
17. K. Richter, G. Tanner, D. Wintgen, Phys. Rev. A 48, 4182
57. H.P. Breuer, K. Dietz, M. Holthaus, Z. Phys. D 8, 349
(1993); D. Wintgen, K. Richter, G. Tanner, Chaos 2, 19
(1988)
(1992) 58. O. Bohigas, S. Tomsovic, D. Ullmo, Phys. Rep. 223, 43
18. K. Richter, D. Wintgen, Phys. Rev. Lett. 65, 1965 (1990) (1993); S. Tomsovic, D. Ullmo, Phys. Rev. E 50, 145 (1994)
19. K. Richter, J.S. Briggs, D. Wintgen, E.A. Solov’ev, J. 59. K. Hornberger, A. Buchleitner, Europhys. Lett. 41, 383
Phys. B 25, 3929 (1992) (1998)
20. P. Schlagheck, A. Buchleitner, J. Phys. B 31, L489 (1998) 60. O. Brodier, P. Schlagheck, D. Ullmo, Phys. Rev. Lett. 87,
21. P. Schlagheck, A. Buchleitner, Physica D 131, 110 (1999) 064101 (2001); Ann. Phys. (N.Y.) 300, 88 (2002)
22. J.G. Leopold, D. Richards, J. Phys. B 20, 2369 (1987)