1: \documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
2:
3:
4: \usepackage{graphicx}% Include figure files
5: \usepackage{bm}% bold math
6:
7: \begin{document}
8: \title{
9: Scaling laws for the photo-ionisation cross section of two-electron atoms
10: }
11:
12: \author{Chang Woo Byun}
13: \author{Nark Nyul Choi}
14: \author{Min-Ho Lee}
15: \affiliation{School of Natural Science, Kumoh National Institute of Technology,
16: Kumi, Kyungbook 730-701, Korea.}
17: \author{Gregor Tanner}
18: \affiliation{School of Mathematical Sciences,
19: University of Nottingham, University Park, Nottingham NG7 2RD, UK.}
20:
21: \date{\today}
22:
23:
24:
25:
26: \begin{abstract}
27: The cross sections for single-electron photo-ionisation in two-electron atoms
28: show fluctuations which
29: decrease in amplitude when approaching the double-ionisation threshold. Based
30: on semiclassical closed orbit theory, we show that the algebraic decay of the
31: fluctuations can be characterised in terms of a threshold law
32: $\sigma \propto |E|^{\mu}$ as $E \to 0_-$ with exponent $\mu$
33: obtained
34: as a combination of stability exponents of the triple-collision
35: singularity. It differs from Wannier's exponent dominating double
36: ionisation processes. The details of the fluctuations are linked to a set of
37: infinitely unstable classical orbits starting and ending in the
38: non-regularisable triple collision. The findings are compared with quantum
39: calculations for a model system, namely collinear helium.
40: \end{abstract}
41:
42: \pacs{32.80.Fb,03.65.Sq,05.45.Mt,05.45.-a}
43: \maketitle
44:
45: Recent experimental progress has significantly improved the energy
46: resolutions of highly excited two electron states
47: below \cite{Pue01,Cza04} and above \cite{double} the
48: double ionisation threshold $E=0$. Near the threshold, electron-electron
49: correlation effects become dominant which are directly observable in
50: total and partial cross sections, see \cite{KA00,Sad00,TRR00} for
51: recent reviews; an example is Wannier's celebrated threshold law
52: for double ionisation \cite{Wa53} confirmed experimentally in \cite{KSA88}.
53: For $E<0$, two-electron atoms exhibit a rich resonance
54: spectrum while the classical dynamics of this three-body Coulomb problem
55: becomes chaotic. Approaching the double ionisation threshold from
56: below has thus proved challenging \cite{TRR00} and recent experimental
57: and theoretical efforts still only reach single-ionisation
58: thresholds $I_N$ with $N \le 15$ \cite{Pue01,Cza04,Bue95}.
59: Semiclassical methods need to address the chaotic
60: nature of the classical dynamics which is dominated by the complex folding
61: patterns of the stable/unstable manifolds of the triple collision \cite{CLT04}.
62: Due to the high dimensionality of the system semiclassical applications have
63: been restricted to subsets of the full spectrum and again small
64: $N$ values \cite{TRR00}.
65:
66: In this letter, we show that the fluctuations in the total
67: cross section for single electron photo-ionisation below the
68: three-particle breakup energy
69: decays algebraically with an exponent determined by the triple collision
70: singularity {\em different} from Wannier's exponent. Writing the
71: cross section in dipole approximation \cite{Bog88, DD88}, we obtain
72: \begin{equation} \label{cs-G}
73: \sigma(E) = -4 \pi \, \alpha \, \hbar \omega \,
74: \Im \langle D \phi_i |G(E) |D \phi_i\rangle
75: \end{equation}
76: where $\phi_i$ is the wave function of the initial state
77: and $D = {\bm \pi} \cdot ({\bf r}_1 + {\bf r}_2)$ is the dipole operator with
78: $\bf \pi$, the polarisation of the incoming photon and ${\bf r}_j$, the
79: position of electron $j$. Furthermore, $G(E)$ is the Green function of the
80: system at energy $E = E_i + \hbar \omega$ and $\alpha = e^2/\hbar c$.
81: Note that we work in the infinite nucleus mass
82: approximation, that is, the position of the nucleus is fixed at
83: the origin.
84:
85: By expressing the Green function semiclassically in terms of classical
86: trajectories \cite{Bog88}, fluctuations in the cross section
87: of hydrogen-like atoms in external fields have been analysed
88: successfully using {\em closed orbit theory} (COT) \cite{Bog88,DD88}.
89: In the semiclassical limit, the support of the wave function $\phi_i$
90: shrinks to zero relative to the size of the system reducing the
91: integration in (\ref{cs-G}) to an evaluation of the Green function at the
92: origin. This is strictly valid only for potentials sufficiently smooth
93: at the origin, corrections due to diffractive scattering at the
94: central singularity give additional contributions often
95: treated in quantum defect approximation \cite{Del94,QD}.
96: The situation is different for two-electron atoms where the dynamics
97: near the origin is dominated by the non-regularisable triple collision
98: singularity. Closed orbit theory has been used to analyse experimental
99: photoabsorbtion spectra of helium with and without
100: external fields \cite{Del94,He} for highly asymmetric states;
101: accompanying theoretical considerations treat the system in a single
102: electron approximation thus not considering the triple collision dynamics
103: which becomes important for doubly-excited resonances.
104:
105: In the following, we will discuss a COT treatment of two-electron atoms
106: explicitly including the triple collision dynamics when approaching
107: the limit $E\to 0_-$. Introducing the hyperradius $R = ({\bf r}^2_1 +
108: {\bf r}^2_2)^{1/2}$, we fix $R_0$
109: such that a surface $\Sigma$
110: defined as $R = R_0$ encloses the support of initial state $\phi_i$.
111: The surface $\Sigma$ naturally leads to a partition
112: of the configuration space into physically distinct regions.
113: In particular, quantum contributions to (\ref{cs-G}) from the
114: inner region are insensitive to the total energy. Contributions
115: from the outer region test the full scale of the classically allowed region
116: of size $|E|^{-1}$ and will be responsible for the resonance structures
117: near the double ionisation threshold $E=0$.
118: Following Granger and Greene \cite{GG00}, we write the photo-ionisation
119: cross section (\ref{cs-G}) in terms of local scattering matrices, that is,
120: \begin{eqnarray} \label{cs-S1}
121: \sigma &=& 4 \pi^2 \alpha \hbar \omega \, \Re \left[\, d^\dagger
122: \left( 1 - S^{\uparrow}S^{\downarrow}\right)^{-1}
123: \left( 1+S^{\uparrow}S^{\downarrow}\right) d\right]\\
124: \label{cs-S2} &=&4 \pi^2 \alpha \hbar \omega \, \Re \left[\,
125: d^\dagger\left( 1 + 2 S^{\uparrow}S^{\downarrow} +
126: 2 (S^{\uparrow}S^{\downarrow})^2 + \ldots \right) d \right] \, .
127: \end{eqnarray}
128: Here, $S^{\downarrow}$ is a core-region scattering matrix which maps
129: amplitudes of waves coming in at $\Sigma$
130: onto amplitudes of the wave components
131: going out at $\Sigma$; it thus contains
132: all the information about the correlated two-electron dynamics
133: near the nucleus. Likewise, $S^{\uparrow}$ describes the wave dynamics
134: of the two-electron wave function emanating from and returning to $\Sigma$
135: and thus picks up long-range correlation in the exterior of $\Sigma$.
136: Furthermore, $d$ is the atomic dipole vector,
137: $ d_n(E) = <\Psi_n^\downarrow(E) | D \phi_i > $
138: with $\Psi_n^\downarrow(E)$ being the $n$th linearly-independent
139: energy-normalised solution of the Schr\"odinger equation inside $\Sigma$
140: with incoming wave boundary conditions at $\Sigma$ \cite{GG00}.
141: This type of scattering formulation was independently developed
142: in \cite{Pro95} for general surfaces $\Sigma$,
143: for a semiclassical formulation, see \cite{Bog92}. The series
144: expansion (\ref{cs-S2}) was exploited in \cite{Del94,QD,series} in order
145: to include core-region scattering or quantum defect effects in COT.
146:
147: For the wave dynamics inside $\Sigma$, the double ionisation threshold
148: $E=0$ is irrelevant and the core-region scattering matrix $S^{\downarrow}$
149: as well as the dipole vector $d$ vary smoothly across $E=0$; they
150: can be regarded as constant for energies sufficiently close to the threshold.
151: The information about the increasing number of overlapping resonances near
152: the threshold is thus largely contained in $S^{\uparrow}$.
153:
154: Semiclassical approximations for the quantities introduced above
155: are valid in the outer-region $R > R_0$. The long-range scattering matrix,
156: $S^\uparrow$, can thus be treated semiclassically while $d$
157: and $S^\downarrow$ demand a full quantum treatment.
158: The semiclassical representation of $S^\uparrow$ in position space reads
159: \cite{GG00,Pro95,Bog92}
160: \begin{equation} \label{G-semi}
161: S^{\uparrow}(x,x',E)\approx (2\pi i\hbar)^{-\frac{f-1}{2}} \sum_j
162: |M_{12}|^{-1/2}_j e^{i\frac{S_j}{\hbar}-i\frac{\pi\nu_j}{2}}\, ,
163: \end{equation}
164: where the sum is taken over all classical paths $j$ connecting
165: points $x$ and $x'$ on $\Sigma$ without crossing $\Sigma$;
166: $S_j(E)$ denotes the action of that path, $\nu_j$ is
167: the Maslov index and $f=4$ is the dimension of the system for
168: fixed angular momentum. Furthermore,
169: $|M_{12}|^{-1/2}_j = |\partial^2 S_j(x, x')/
170: \partial x \partial x'|^{1/2}$, where
171: $M_{12}$ is a $(3 \times 3)$ sub-matrix of the $6$-dim.\
172: Monodromy matrix describing the linearised flow near a trajectory.
173: Note that due to the strong instability of the classical dynamics
174: near the triple collision, these matrix elements become singular
175: for triple collision orbits (TCO)
176: starting from or falling into the triple collision $R = 0$.
177: It is thus important here that trajectories
178: contributing to (\ref{G-semi}) start at a fixed hyper-radius $R_0 >0$
179: away from the triple collision \cite{note1}.
180:
181: Making use of the scaling properties of the classical dynamics, we introduce
182: the transformation \cite{RTW93}
183: \[ {\bf r} = \tilde{\bf r}/|E|; \quad {\bf p} = \sqrt{|E|}\tilde{\bf p};
184: \quad S = \tilde{S}/\sqrt{|E|}, \quad {\bf L} = \tilde{\bf L}/\sqrt{|E|}\]
185: where $\tilde{\bf r}, \tilde{\bf p}$ corresponds to coordinates and
186: momenta at fixed energy $E = -1$ and $\bf L$ is the total angular momentum.
187: Expressing $\bf L$ in scaled coordinates,
188: we have $\tilde{\bf L} \to 0$ as $E\to 0$
189: and can thus restrict the analysis to the three degrees of freedom
190: subspace $\tilde{L} = 0$ (for fixed $L$) \cite{CLT04}.
191:
192: In scaled coordinates, the inner-region shrinks according to
193: $\tilde{R}_0 = |E| R_0 \to 0$ for $E\to 0_-$, and
194: the part of the dynamics contributing to $S^\uparrow$ in
195: (\ref{G-semi}) is formed by trajectories starting and ending closer and
196: closer to the triple collision $R =0$ as $|E| \propto \tilde{R}_0 \to 0$.
197: TCOs only occur in the so-called eZe space
198: \cite{CLT04}, a collinear subspace of the full three body
199: dynamics where the two electrons are on opposite sides of the nucleus
200: \cite{RTW93}. As $\tilde{R}_0 \to 0$, only orbits coming close to the
201: eZe space can start and return to $\tilde{\Sigma}$ and they will do so in
202: the vicinity of a {\em closed triple collision orbit} (CTCO) starting
203: and ending exactly in the triple collision. The dynamics in the eZe space
204: is relatively simple as it is conjectured to be fully chaotic
205: with a complete binary symbolic dynamics. In particular, for every finite
206: binary symbols string there is a CTCO, the shortest being the so-called
207: Wannier orbit (WO) of symmetric collinear electron dynamics.
208: Furthermore TCOs escape from or approach the triple collision at
209: $R=0$ always symmetrically along the $r_1 = r_2$ axis in the
210: eZe space, that is, along the WO \cite{McG74}. This universality will
211: be exploited below when treating the energy-dependence of $M_{12}$
212: in (\ref{G-semi}).
213:
214: Returning to the cross-section (\ref{cs-S1}), we
215: write $\sigma = \sigma_0 + \sigma_{fl}$, where we identify the smooth
216: contribution $\sigma_0$ with the leading term in the series expansion
217: (\ref{cs-S2}). The main contribution to the fluctuating part of the cross section
218: $\sigma_{fl}$ is contained in $S^\uparrow$ which in semiclassical
219: approximation (\ref{G-semi}) can be expressed in terms of orbits returning
220: to $\tilde{\Sigma}$ once; multiple traversals of $\tilde{\Sigma}$ represented
221: by $\left(S^\uparrow S^\downarrow \right)^k$ with $k \ge 2$ will give
222: sub-leading contributions in the semiclassical limit $E \to 0_-$ due to
223: the unstable dynamics near the triple collision. Furthermore, swarms of
224: trajectories starting on $\tilde\Sigma$ and returning to $\tilde\Sigma$
225: will do so close to the eZe subspace and thus in the neighbourhood of a
226: CTCO with actions and amplitudes approaching those of the CTCO trajectory
227: as $\tilde{R}_0 \to 0$. The fluctuating part can thus in leading order be
228: written in the form
229: \begin{equation} \label{s_fl_2}
230: \sigma_{fl}(E) \approx
231: \Re \sum_{{\rm CTCO}_j} A_j(E)
232: e^{i z \tilde{S}_j }
233: \end{equation}
234: with
235: \begin{equation} \label{M_j}
236: A_j(E) \propto |M_{12}(E)|_j^{-1/2}
237: = |E|^{9/4} |\tilde{M}_{12}(\tilde{R}_0)|_j^{-1/2}
238: \end{equation}
239: and $z = 1/\hbar\sqrt{|E|}$.
240: In (\ref{s_fl_2}),
241: the sum is taken over all CTCO's $j$ starting and ending at
242: $\tilde{\Sigma}$. Note that the stability $\tilde{M}_{12}(\tilde{R}_0)$ in
243: scaled coordinates depends on energy implicitly through the scaled radius
244: $\tilde{R}_0 (E) = |E| R_0$. As $E\to 0_-$, $\tilde{M}_{12}$ picks up
245: additional contributions of parts of the CTCO closer and closer to
246: the triple collision. Asymptotically, all CTCOs approach the triple collision
247: along the WO and the contributions to $M_{12}$ become
248: orbit-independent.
249: The $R$-dependence for the Monodromy matrix of the WO can for small $\tilde{R}$
250: be obtained {\em analytically} \cite{NNC06} leading to
251: \begin{equation} \label{tildeM_j}
252: |\tilde{M}_{12} (\tilde{R}_0)| \propto |\tilde{R_0}|^{-2\mu + 9/2}
253: \; \mbox{for} \; \tilde{R}_0 \to 0
254: \end{equation}
255: with exponent
256: \begin{equation} \label{mu_t}
257: \mu= \mu_{eZe} + 2 \mu_{wr} =
258: \frac{1}{4}\left[\sqrt{\frac{100 Z-9}{4Z-1}} +
259: 2\sqrt{\frac{4 Z - 9}{4Z -1}}\right].
260: \end{equation}
261: Thus, in unscaled coordinates, $M_{12}$ in (\ref{M_j})
262: diverges which is a direct consequence
263: of the non-regularisability of the triple collision acting as an infinitely
264: unstable point in phase space; details will be presented in \cite{NNC06}.
265: The exponent $\mu$ in (\ref{mu_t}) is universal for all CTCOs and
266: consists of two components: $\mu_{eZe}$ is related to the linearised dynamics
267: in the eZe space and $\mu_{wr}$ picks up
268: contributions from two equivalent expanding degrees of freedom orthogonal
269: to the eZe space in the so-called Wannier ridge (WR). The latter is
270: the invariant subspace of symmetric electron dynamics with $|r_1| = |r_2|$ at
271: all times \cite{RTW93}. The fluctuations in the photo-ionisation cross
272: section thus vanish in amplitude as $E \to 0_-$ according to
273: \begin{equation} \label{CTCO1}
274: \sigma_{fl}(E) \propto |E|^{\mu} \, \Re \sum_{{\rm CTCO}_j} a_j
275: e^{i z \tilde{S}_j}
276: \end{equation}
277: where the rescaled amplitudes $a_j = |E|^{-\mu} A_j$ depend only weakly on $E$.
278: These
279: amplitudes contain contributions from the linearised dynamics along the
280: orbit far from $\Sigma$ as well as information about the inner quantum region
281: $R<R_0$ via the dipole vector $d$ and the core-region scattering matrix
282: $S^\downarrow$.
283: Furthermore, multiple traversals of $\Sigma$ contained in
284: $\left(S^\uparrow S^\downarrow\right)^k$ with $k \ge 2$ in (\ref{cs-S2})
285: approach the triple collision $k$ times from the semiclassical
286: side and will thus contribute at lower order
287: with weights scaling at least as $A_{kj} \sim |E|^{k \mu}$.
288: The exponent $\mu$
289: in (\ref{mu_t}) is different from Wannier's exponent $\mu_w$ with
290: \begin{equation}\label{mu_W}
291: \mu_w = \frac{1}{4}\sqrt{\frac{100 Z - 9}{4Z -1}} - \frac{1}{4}\, .
292: \end{equation}
293: One obtains, for example, $\mu = 1.30589...$ compared to
294: $\mu_w = 1.05589...$ for helium; the WR contributes to the decay
295: for $Z> 9/4$ when $\mu_{wr}$ is real.
296:
297: The exponent $\mu$ can be interpreted in terms of stability
298: exponents of the triple collision singularity.
299: Using an appropriate scaling of space and time by, for example, employing McGehee's
300: technique \cite{McG74},
301: the dynamics near the singularity is dominated by
302: two unstable fixed points in scaled phase space,
303: the {\em double escape point} (DEP) and the {\em triple collision point} (TCP).
304: In unscaled coordinates, these fixed
305: points correspond to the WO at energy $E=0$, that is, the DEP is the
306: trajectory of symmetric double escape while
307: the TCP corresponds to the symmetric triple collision and is the time reversed
308: of the DEP. The triple collision itself can be mapped onto the
309: classical dynamics at $E=0$; likewise, $\tilde{R}_0 = |E| R_0$ acts as
310: a parameter measuring the closeness to the $E=0$ manifold which contains the
311: fixed points, see \cite{CLT04,McG74}. Most classical trajectories emerging from
312: $\tilde{\Sigma}$ in the vicinity of the triple collision $\tilde{R} =0$ will lead to
313: immediate ionisation of one electron carrying away a larger
314: amount of kinetic energy. Only a fraction of orbits starting on $\tilde{\Sigma}$
315: near the WO will enter a chaotic scattering region; the WO at $E<0$ thus acts as
316: an unstable direction of the DEP, $U_D^{wo}$, with a stability exponent
317: $\lambda_{U_D}^{wo}$. From there, they can return to
318: the triple collision region and thus approach the surface $\tilde{\Sigma}$ again
319: along the WO, that is, along the stable direction $S_T^{wo}$.
320: The transition from the DEP into the chaotic scattering region
321: and from this scattering region to the TCP is limited by the least stable
322: eigendirections of the fixed points in each of
323: the invariant subspaces (eZe or WR) perpendicular to the WO.
324:
325: Trajectories leaving $\tilde{\Sigma}$ along the WO in the eZe space diverge
326: from the WO along an unstable direction $U_D^{eZe}$ with a stability exponent
327: $\lambda_{U_D}^{eZe}$.
328: Competition of the instability in $U_D^{wo}$ with that in $U_D^{eZe}$
329: determines the fraction $\Delta_{eZe}^D$ of orbits entering the chaotic
330: scattering region. By using methods as in \cite{CLT04,Sie41}, one finds that
331: \begin{equation} \label{DeltaD}
332: \Delta_{eZe}^D \propto \tilde{R}_0^{\lambda_{U_D}^{eZe}/\lambda_{U_D}^{wo}}
333: = \tilde{R}_0^{\mu_w}\,
334: \end{equation}
335: with exponent equal to Wannier's exponent (\ref{mu_W}). For
336: the cross section (\ref{cs-G}), information about the phase
337: space region returning from the chaotic scattering region
338: to the surface $\tilde{\Sigma}$ along the WO is also needed.
339: While approaching $\tilde{\Sigma}$, these orbits are
340: deflected away from the triple collision
341: along an unstable direction $U_T^{eZe}$ of the TCP fixed point.
342: The fraction of orbits reaching the surface $\tilde{\Sigma}$
343: among those leaving the chaotic scattering region
344: scales thus as in (\ref{DeltaD}) now with exponent
345: $\lambda_{U_T}^{eZe}/|\lambda_{S_T}^{wo}|$ .
346: Similar mechanisms apply for the WR dynamics.
347:
348: The fraction $\Delta$ of two-electron trajectories making the transition
349: from $\tilde{\Sigma}$ back to $\tilde{\Sigma}$ can thus in the limit
350: $E \to 0_-$ be estimated in terms of the stability exponents of the fixed
351: points, also referred to as Siegel exponents \cite{Sie41} in celestial
352: mechanics. One obtains $\Delta \propto \tilde{R}_0^{2\mu}$ with
353: $\mu$ as in (\ref{mu_t}) which can be written as
354: \[
355: \mu_{eZe} = \frac{1}{2}
356: \left(\frac{\lambda_{U_D}^{eZe}}{\lambda_{U_D}^{wo}} +
357: \frac{\lambda_{U_T}^{eZe}}{\left|\lambda_{S_T}^{wo}\right|}\right); \;\;
358: \mu_{wr} = \frac{1}{2} \left(\frac{\lambda_{S_D}^{wr}}{\lambda_{U_D}^{wo}} +
359: \frac{\lambda_{U_T}^{wr}}{\left|\lambda_{S_T}^{wo}\right|}\right);
360: \]
361: for the actual values of the stability exponents $\lambda$, see
362: \cite{Wa53, CLT04}. The mean amplitude of
363: the fluctuations in the quantum signal is thus asymptotically related to the
364: fraction of phase space volume starting and ending at $\tilde{R}_0$.\\
365: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
366: \begin{figure}
367: \includegraphics[scale=.32]{fig1.eps}
368: \caption[]{\small The Fourier spectrum of the fluctuating part of the
369: eZe cross section rescaled according to (\ref{Fz}); the circles denote
370: the position $\tilde{S}_j$ and (relative) size of $|M_{12}|^{-1/2}_j$
371: for CTCO's with $\tilde{S}/2 \pi < 13$. Corresponding trajectories
372: in configuration space are shown for the first 4 peaks.
373: Inset: $\sigma_{fl}$ for $N \le 52$.
374: }
375: \label{fig:fig1}
376: \end{figure}
377: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
378:
379: A numerical study of the full three-body quantum problem is still out
380: of reach for energies $E < I_N$ with $N \sim 15$ \cite{Pue01,Bue95};
381: we therefore chose a model system, namely eZe collinear helium
382: first studied quantum mechanically in \cite{TH74}.
383: We calculated the cross
384: section (\ref{cs-G}) directly in a large set of basis functions using the
385: method of complex rotation and obtained a converged signal for $N \sim 55$.
386: Semiclassically, we consider now the dynamics in the eZe space alone,
387: which contains all the important parts
388: regarding the algebraic decay in the fluctuations.
389: The number of basis functions used are scaled with energy
390: to cover a fixed scaled region in $\tilde{R}$ containing CTCOs
391: with $\tilde{S}/2\pi \le 20$.
392: Adopting the basis functions used by P\"uttner {\em et al} \cite{Pue01}
393: leading to a strongly banded Hamiltonian matrix, it is possible to increase the
394: basis size to $10^6$.
395: Starting with an odd initial state
396: $\phi_i$, we obtain the cross section for the even parity eZe spectrum;
397: its fluctuating part after numerically subtracting a
398: smooth background is shown in Fig.\ \ref{fig:fig1}. The numerical value
399: of the exponent $\mu$ is determined by rescaling the signal according to
400: \begin{equation} \label{Fz}
401: F(z) = |E|^{-\mu} \sigma_{fl}(z) / \hbar \omega
402: \end{equation}
403: and testing the stationarity of the Fourier transform of $F(z)$ in
404: different energy windows \cite{NNC06}. The best value thus obtained
405: is $\mu = 1.306 \pm 0.035$ in good agreement with the
406: theoretical prediction (\ref{mu_t}). (Note that the real parts of
407: the exponents for 3-dim. helium and for eZe helium coincide
408: as $\Re \mu_{wr} = 0$). Furthermore,
409: the peaks in the Fourier transform can be associated one-by-one with
410: CTCO's in the eZe system, see Fig.\ \ref{fig:fig1}. We do not observe
411: peaks associated with the concatenation of different CTCO's or repetitions
412: of single CTCO's. This is consistent with the expected suppression of orbit
413: contributions traversing $\Sigma$ more than once as discussed earlier.
414: Furthermore, we calculated the geometrical contribution to the coefficients
415: $a_j$ in (\ref{CTCO1}) directly from the matrix-elements $M_{12}$ by
416: scaling out the leading order divergence; a clear correlation with the peak
417: heights can be seen in Fig.\ \ref {fig:fig1}. Quantum
418: contributions from the core region are thus indeed roughly the same for all
419: CTCOs.
420:
421: In conclusion, we show that the fluctuations in the total photo-ionisation
422: cross section below the double ionisation threshold follow an algebraic law
423: with a novel exponent which can be
424: written in terms of stability exponents of the triple collision.
425: Our findings are verified numerically for a collinear model systems; we
426: furthermore predict that the algebraic decay law is valid for the physically
427: relevant 3 dimensional cases with an additional
428: contribution from the WR dynamics for $Z>9/4$. Our findings will provide new
429: impetus for experimentalists and theoreticians alike to study highly
430: doubly excited states in two-electron atoms.
431:
432: This work was supported in part by the Korea Research
433: Foundation (KRF-2006-521-C00019). We thank B.\ Gremaud, J.\ Madro\~{n}ero
434: and P.\ Schlagheck for useful discussions regarding the
435: numerical calculations, and the KISTI supercomputing centre for numerical
436: support.
437:
438:
439:
440: \begin{thebibliography}{999999}
441: \bibitem{Pue01} R.\ P\"uttner {\em et al}, Phys.\ Rev.\ Lett.\
442: {\bf 86}, 3747, (2001); Y.\ H.\ Jiang {\em et al}, Phys.\ Rev.\ A
443: {\bf 69}, 042706 (2004); Y.\ H.\ Jiang, {\em Doctoral Dissertation}
444: (Freien Universit\"at Berlin, Berlin, Germany, 2006).
445: \bibitem{Cza04}
446: A.\ Czasch {\em et al}, Phys.\ Rev.\ Lett.\ {\bf 95}, 243003, (2005).
447: \bibitem{double} R.\ D\"orner {\em et al}, Phys.\ Rep.\ {\bf 330}, 95 (2000);
448: Th.\ Weber {\em et al}, Nature {\bf 405}, 658 (2000); A.\ Knapp {\em et al}
449: Phys.\ Rev.\ Lett.\ {\bf 89}, 033004, (2002).
450: \bibitem{KA00} G.\ C.\ King and L.\ Avaldi, J Phys B {\bf 33}, R215 (2000).
451: \bibitem{Sad00} H,\ R.\ Sadeghpour {\em et al}, J Phys B {\bf 33}, R93 (2000).
452: \bibitem{TRR00} G.\ Tanner {\em et al}, Rev.\ Mod.\ Phys.\
453: {\bf 72}, 497 (2000).
454: \bibitem{Wa53} G.\ H.\ Wannier, Phys.\ Rev.\ {\bf 90}, 817 (1953).
455: \bibitem{KSA88} H.\ Kossmann {\em et al}, Phys.\ Rev.\ Lett.\ {\bf 60},
456: 1266 (1988).
457: \bibitem{Bue95} A.\ B\"urgers {\em et al}, J.\ Phys.\ B {\bf 28}, 3163 (1995);
458: H.\ W.\ van der Hart and C.\ H.\ Greene, Phys.\ Rev.\ A, {\bf 66},
459: 022710 (2002).
460: \bibitem{CLT04} N.\ N.\ Choi {\em et al} Phys.\ Rev.\ Lett.
461: {\bf 93}, 054302 (2004); M.-H.\ Lee {\em et al}
462: Phys.\ Rev.\ E {\bf 71}, 056208 (2005).
463: \bibitem{Bog88} E.\ B.\ Bogomolny, Sov.\ Phys.\ JETP {\bf 69}, 275 (1988).
464: \bibitem{DD88} M.\ L.\ Du and J.\ B.\ Delos, Phys.\ Rev.\ A {\bf 38},
465: 1896 and 1913 (1988).
466: \bibitem{Del94} D.\ Delande {\em et al}, J.\ Phys.\ B {\bf 27} 2771 (1994).
467: \bibitem{QD} P.\ A.\ Dando {\em et al} Phys.\ Rev.\ Lett. {\bf 74},
468: 1099 (1995); B.\ H\"upper {\em et al} Phys.\ Rev.\ Lett.\ {\bf 74},
469: 2650 (1995).
470: \bibitem{He} M.\ Keeler and T.\ J.\ Morgan, Phys.\ Rev.\ Lett.\ {\bf 80},
471: 5726 (1998); W.\ Huang and T.\ J.\ Morgan, Phys.\ Rev.\ A {\bf 65},
472: 033409 (2002).
473: \bibitem{GG00} B.\ E.\ Granger and C.\ H.\ Greene,
474: Phys.\ Rev.\ A {\bf 62}, 012511 (2000); T. Bartsch {\em et al},
475: Phys.\ Rev.\ A {\bf 67}, 063411 (2003).
476: \bibitem{Pro95} T.\ Prosen, J.\ Phys.\ A {\bf 28}, 4133 (1995);
477: C.\ Rouvinez and U.\ Smilansky, J.\ Phys.\ A {\bf 28}, 77 (1995).
478: \bibitem{Bog92} E.\ B.\ Bogomolny, Nonlinerity {\bf 5}, 805 (1992).
479: \bibitem{series} P.\ A.\ Dando {\em et al}, Phys.\ Rev.\ A {\bf 54}, 127 (1996);
480: J.\ A.\ Shaw and F.\ Robicheaux, Phys.\ Rev.\ A {\bf 58}, 3561 (1998).
481: \bibitem{note1}
482: For regular potentials, Eqn.\ (\ref{G-semi}) can be written in terms of
483: closed orbits immediatly as contributions due to the finite radius become
484: negligible in the semiclassical limit.
485: \bibitem{RTW93} K.\ Richter {\em et al}, Phys. Rev. A {\bf 48}, 4182 (1993).
486: \bibitem{McG74} R.\ McGehee, Inventiones Math. {\bf 27}, 191 (1974).
487: \bibitem{NNC06} N.\ N.\ Choi {\em et al}, to be submitted to Phys.\ Rev.\ A.
488: \bibitem{Sie41} C.\ S.\ Siegel, Ann.\ of Math.\ {\bf 42}, 127 (1941);
489: J.\ Waldvogel, Lecture Notes in Physics {\bf 590}, 81 (2002).
490: \bibitem{TH74} A.\ Temkin and Y.\ Hahn, Phys.\ Rev.\ A {\bf 9}, 708 (1974).
491: \end{thebibliography}
492:
493: \end{document}
494: