1: \documentclass[12pt]{iopart}
2: \usepackage{iopams}
3: \usepackage{graphicx}
4: \usepackage{epsf}
5:
6: \begin{document}
7:
8: \sloppy
9:
10: \jl{2}
11:
12: \title[Recombination of low energy electrons with U$^{28+}$]
13: {\bf Recombination of low energy electrons with U$^{28+}$}
14:
15: \author{G. F. Gribakin and S. Sahoo}
16:
17: \address{Department of Applied Mathematics and Theoretical Physics, Queens University, Belfast BT7 1NN, NI UK}
18: \author{V. Dzuba}
19: \address{School of Physics, University of New South Wales,Sydny 2052, Australia }
20:
21:
22: %\date{\today }
23:
24: %\maketitle
25: %\vspace{1cm}
26:
27: \begin{abstract}
28: We performed an extensive study of configuration mixing between the
29: doubly excited (doorway) states and complex multiply excited states of
30: U$^{28+}$ near its ionization threshold. The detailed investigation of
31: complex spectrum and analysis of the statistics of eigenstate
32: components show that the dielectronic (doorway) states weakly mixed with
33: each other. However, they show substantial mixing with the complex
34: multiply excited states. This situation explains the mechanism of low
35: energy electron recombination with U$^{28+}$. We calculated the energy
36: averaged capture cross sections as a sum over dielectronic doorway states and found our present calculation interprets well the experimental recombination rates in the energy range of 1 to 100 eV.
37:
38: \end{abstract}
39: \vspace{1cm}
40:
41: \pacs{PACS: 34.80.Lx, 31.10.+z, 34.10.+x, 32.80.Dz}
42:
43: \section{Introduction}
44: Atomic processes are of great importance from astrophysical and other various point of view. Especially the low energy recombination cross sections and rate coefficients are required for understanding the fusion and astrophysical plasmas. In particular dielectronic recombination (DR) plays a significant role in astrophysics because it is the dominant electron-ion recombination process for most ions in low density, photo-ionized and electron-ionized cosmic plasmas \cite {Arnaud:85}. Accurate rate coefficients are needed to calculate reliably the ionization balance, thermal structure and line emission of cosmic plasmas. Most importantly the low energy positron antiproton recombination provides a challenging scheme for the production of antihydrogen\cite{Holzscheiter:99}.
45:
46:
47: It is well-known that due to the presence of additional recombination
48: channels such as DR, the rate coefficients are found to be larger
49: than radiative recombination (RR) rates for many electron complex ions. In DR process the incident electron is captured in a doubly-excited state of the compound ion, which is then
50: stabilized by photoemission. This process originally suggested by J.~Sayers and was first considered by Massey and Bates\cite{Massey:43} in the study of ionospheric oxygen. Electron-ion recombination has been measured directly in the laboratory
51: since early 1980's \cite{Recomb:92}. More recently the use of heavy-ion
52: accelerators and electron coolers of ion storage rings has greatly advanced
53: the experiment \cite{Andersen:89,Kilgus:90}. Recombination rates for various
54: ions have been measured at electron energies from threshold to hundreds of
55: electron volts (eV) with a fraction-of-an-eV resolution
56: \cite{Schennach:94,Gao:95,Schuch:96,Uwira:96,Uwira:97,Zong:97,Mannervik:98}.
57: For few-electron ions the measured rates were found to be in good agreement with theoretical predictions which included the contribution of DR resonances on top of the RR background, e.g., in He$^+$ \cite{DeWitt:94}, Li-like C$^{4+}$
58: \cite{Mannervik:98} and Ar$^{15+}$ \cite{Schennach:94,Zong:97}, and B-like
59: Ar$^{13+}$ \cite{DeWitt:96}. However, more complicated ions, e.g.,
60: Au$^{50+}$ \cite{Uwira:97}, U$^{28+}$\cite{Uwira:96} and
61: Au$^{25+}$\cite{Hoffknecht:98}, showed complicated resonance spectra and
62: strongly-enhanced recombination rates at low electron energies. The
63: Au$^{25+}$ ion has been studied in detail by Gribakin et al.\cite{Au}
64: and Flambaum et al.\cite{Fl:02} using statistical methods. They
65: suggested that the strongly enhanced low energy electron recombination
66: observed in this ion is mediated by complex multiply-excited states
67: rather than simple dielectronic resonances and the dielectronic
68: resonances play the role of doorways to the electron capture
69: process. The statistical method developed by Flambaum et al.\cite{Fl:02}
70: is based on the assumption of strong (chaotic) configuration
71: mixing. This assumption has been verified by Gribakin and Sahoo \cite{Sahoo:JPB} in a recent study. However, for U$^{28+}$, no theory so far has described the low energy DR process successfully.
72:
73: Historically the recombination rate enhancement was observed first in
74: U$^{28+}$\cite{Uwira:96}. This measurement has been performed in a
75: merged beam experiments at UNILAC accelerator in Darmstadt and at
76: heavy-ion storage ring TSR in Heidelberg. The experiment found rate
77: enhancement in the U$^{28+}$ spectrum exceeds the theoretical
78: calculation by at least a factor 20 in the energy range below 10
79: eV. Later in 1998, they extended the experiment to high energies up to
80: 420 eV \cite{Mitnik:98} and a comparison has been made with the
81: theoretical calculations which is based on distorted wave
82: approximations. The DR cross sections are calculated in this method are
83: able to explain the main resonant features in the range 80--180 eV, but
84: failed to identify the resonances and reproduce the rate at smaller
85: energies. The cross sections involving the excitations from the
86: 5s$^{2}$5p$^{2}$ ground state configuration calculated in this method are well described by using either semirelativistic wave functions as found in AUTOSTRUCTURE codes or fully relativistic wave functions in HULLAC codes. However, the resonance structure observed at low energies i.e. below 80 eV largely remains unexplained. They finally concluded that in complex ions, particularly in U$^{28+}$ ion, what are the resonances just above the threshold and how they contribute to the low energy recombination remain a 'mystery'.
87: In this paper we performed an extensive study of the excited spectra and eigenstates of U$^{28+}$ near its ionization threshold and calculated recombination rate coefficients for electron recombination with U$^{28+}$. We identified some of the resonances near the threshold those contribute significantly to the low energy recombination. The present results are found to be in good agreement with the experiment. This work develops further a statistical theory towards the full understanding of the mechanism of low energy electron recombination with U$^{28+}$ and other similar complex ions.
88:
89:
90: \section{Many-electron excitations}\label{sec:mix}
91: We consider the recombination of an electron with U$^{28+}$. Due to
92: electron correlation the slow electron can be captured in one of the
93: excited states of the compound ion U$^{27+}$. It has 65 electrons and
94: its ground state configuration belongs to
95: 1$s^{2}$........5$s^{2}$5$p^{3}$. Figure 1 shows the spectrum of
96: relativistic $nlj$ orbitals obtained by relativistic Hartree-Fock
97: calculations. Atomic units (a.u.) are used unless otherwise stated. All
98: the occupied orbitals below fermi level ($\sim 31.19~ a.u.$) are obtained by a
99: self consistent calculation of U$^{28+}$ground state. Each of excited
100: orbitals above the fermi level are calculated by placing one electron on
101: to it in the frozen 1$s^{2}$........5$s^{2}$5$p^{2}$ (U$^{28+}$)
102: core. Our configuration interaction (CI) calculation shows that the
103: ground state of U$^{28+}$ and U$^{27+}$ ions are characterized by their
104: total angular momentum J $=$ 0.0 and 1.5 respectively. The difference
105: between their total energies $-$27741.40 a.u. and $-$27771.37 a.u.,
106: gives us an estimation of ionization threshold (I.T.) $=$ 29.97 a.u. (815.184 eV).
107:
108: The excited states of the ion are generated by transferring one, two, three, etc., electrons from the ground state into the empty orbitals above fermi level. Since we are interested in the excited spectrum of the ion near its ionization threshold (29.97 a.u.), we consider it as a system of having only 29 electrons above the frozen Kr-like 1$s^{2}$....4$p^{6}$ core. The number of many-electron states obtained by distributing 29 electrons over 40 relativistic orbitals, 4$d_{3/2}$ through to 8$g_{9/2}$ are huge in number. It is practically impossible to perform a CI calculation for all of them. We construct the excite spectrum by using mean field approach by calculating their mean energies $E_i$, and number of many electron states $N_i$ associated with each of them:
109: \begin{equation}\label{eq:E}
110: E_{i}= E_{core}+\sum_a{\epsilon_a n_a}+\sum_{a<b}\frac{n_a(N_b-\delta_{ab})}{1+\delta_{ab}}U_{ab} ,
111: \end{equation}
112: \begin {equation}\label{eq:N}
113: N_i=\prod_a\frac{g_a!}{n_a! (g_a-n_a)!},
114: \end {equation}
115: where $n_a$ are the orbital occupation numbers of the relativistic
116: orbitals in a given configuration and
117: $\sum_a{n_a}=n$. $\epsilon_a=<a|H_{core}|a>$ is the s
118:
119: %\begin{figure}[h]
120: %\epsfxsize=10cm
121: %\includegraphics[width=\textwidth]{en_U28+.eps}
122: %\centering\leavevmode\epsfbox{en_U28+.eps}
123: %\vspace{8pt}
124: %\caption{Energies of occupied and vacant single-particle
125: %orbitals of Au$^{28+}$ obtained in a Dirac-Fock calculation.}
126: %\label{fig:orb}
127: %\end{figure}
128:
129: ingle-particle energy of the orbital $a$ in the field of the core, $g_a=2j_a+1$, and $U_{ab}$ are the average Coulomb matrix elements for the electrons in orbitals $a$ and $b$ ( direct minus exchange):
130:
131: \begin{equation}\label{eq:U}
132: U_{ab}=\frac{g_a}{g_a-\delta_{ab}}\left[R_{abab}^{(0)}-\sum_{\lambda}\delta_p R_{abba}^{(\lambda)}\left\{ {j_a \atop \frac{1}{2}}{\j_b \atop-\frac{1}{2}}{\lambda \atop \ 0}\right\}^{2}\right]
133: \end{equation}
134:
135:
136: $R_{abba}^{(\lambda)}$ is the two-body radial Coulomb integral of $\lambda$ multipole, and $\delta_p=1$ when $l_a +l_b +\lambda$ is even and 0 otherwise. Using Eqs. (1)-(3) we obtained about 353 configurations within $\pm$ 1 a.u. of ionization threshold. They comprise a total of $1.9\times 10^5$ many electron states.
137: The single-particle spectrum of U$^{28+}$ does not show large gaps. Owing to the ``gapless''single-particle spectrum, the density increases rapidly as a function of energy, as described by the Fermi-gas-model ansatz
138: \cite{Bohr:69}
139: \begin{equation}\label{eq:rho}
140: \rho (E)=AE^{-\nu }\exp (a \sqrt{E}),
141: \end{equation}
142: with $A$ = 0.0885, $\nu$ = 2.33, and $a$ = 3.99 a.u.\cite{comment3},
143: where $E$ is the energy above the ground state in atomic units.
144: Figure 2 shows the level density calculated by averaging with a Gaussian with 1
145: a.u. variance. This figure depicts the level densities for both odd and
146: even parity configurations. Also included in the figure are only the doubly
147: excited configurations of even and odd parity found within $\pm$ 1
148: a.u. of ionization threshold. It is found that the level density of the
149: odd configurations near its ionization threshold is of the order is of
150: $1\times 10^5$ and
151:
152: \begin{figure}[h]
153: \epsfxsize=10cm
154: \centering\leavevmode\epsfbox{g-den_new.eps}
155: \vspace{8pt}
156: \caption{Level densities in U$^{28+}$ within $\pm 1$ a.u. of Ionization threshold. Thick solid line: odd configurations, Thin solid line: odd dielectronic configurations, Dash-dot line: even configurations and dashed line: even dieclectronic configurations.}
157: \end{figure}
158: those of
159: even configurations it is of the order of $1.3\times 10^4$. This
160: provides an evidence that this system is characterized by huge level
161: density. This system may be compared with Au$^{25+}$, in which the level
162: density near its ionization threshold is about $1\times
163: 10^7$\cite{Au}. This shows that U$^{28+}$ ion is less
164: complicated than Au$^{25+}$ from the view point of dense excited spectrum. Apart from this one can estimate the mean level spacing from the distribution of total angular momentum J as shown in Figure 3.
165:
166: Ionic eigenstates are characterized by their total angular momentum and
167: parity $J^\pi $, and are $2J+1$ times degenerate. Therefore the total level
168: density can be broken into a sum of partial level densities:
169: $\rho (E)=\sum _{J^\pi }(2J+1)\rho _{J^\pi }(E)$. The excitation spectrum of
170: U$^{28+}$ near the ionization threshold, $E=I\approx 29.97$ a.u.,
171: contains many $J$ ranging from $\frac{1}{2}$ to $\frac{25}{2}$.
172: Their distribution is in agreement with statistical theory
173: \cite{Bohr:69,Bauche:87}, which predicts that at a given energy $\rho _{J^\pi }$
174: are proportional to the function
175: \begin{equation}\label{eq:f_J}
176: f(J)= \frac{2(2J+1)}{(2J_m+1)^2}\exp \left[ -\frac{(2J+1)^2}
177: {2(2J_m+1)^2}\right] ,
178: \end{equation}
179: where $J_m$ is the most abundant $J$ value. Numerically for U$^{27+}$ we find
180: $J_m \approx \frac{7}{2}$.
181: \begin{figure}[h]
182: \epsfxsize=10cm
183: \centering\leavevmode\epsfbox{J-graph.eps}
184: %\vspace{8pt}
185: \caption{Densities of states with different J near ionization threshold
186: of U$^{28+}$. Histograms show their distribution and the solid line is
187: non linear fitting obtained from Eq.(5) which predicts the most abundant
188: value of $J\sim3.2$ and the proportionality constant $\pm$ 11075.9.}
189: \end{figure}
190:
191: Using Eq. (\ref{eq:f_J}) one can estimate the partial densities by
192: $\rho _{J}=\rho _{J^+}+\rho _{J^-}\simeq f(J)\rho / \langle 2J+1\rangle $,
193: where $\langle 2J+1\rangle $ is an average over $f(J)$. For the most abundant
194: angular momenta $J\sim J_m$, and assuming $\rho _{J^+}\approx \rho _{J^-}$,
195: we have $\rho _{J^\pi }(E)=A_{J^\pi }E^{-\nu }\exp (a \sqrt{E})$.
196: Near the ionization threshold this gives
197: $\rho _{J^\pi }\approx 1.8\times 10^3$ a.u. (Fig. 3), which means that the
198: spacing between the multiply-excited states with a given $J^\pi $ is
199: small: $D=1/\rho _{J^\pi }\sim 15$ meV. Whereas in Au$^{25+}$ it is about 1 meV, i.e. 15 times larger than that found in Au$^{25+}$. This situation explains why individual resonances appear in recombination rates of U$^{28+}$ but the same is not observed in Au$^{25+}$ experimentally even at an energy resolution of 0.1 eV. However, the large density of
200: multiply-excited states is only a ``kinematic'' reason behind the
201: experimental observation. To explain the fact that the electron can actually be captured into these states, we need to analyze the dynamics
202: of electron capture and show that the residual Coulomb interaction between the
203: electrons (i.e. that beyond the mean field) makes for an efficient capture and
204: accounts for the observed enhanced recombination rate.
205: \section{Configuration mixing}
206: Taking into account the fact that the rasidual Coulomb interaction is the key problem in many-electron
207: processes, we construct the basis of many-electron
208: states $\Phi _k $ from single-particle (e.g., Dirac-Fock) orbitals, and
209: solve the eigenvalue problem for the Hamiltonian matrix
210: $H_{ik}=\langle \Phi _i|\hat H|\Phi _k\rangle $, which yields the
211: eigenvalues $E_\nu $ and eigenstates $|\Psi _\nu \rangle =\sum _kC^{(\nu )}_k
212: |\Phi _k \rangle $ of the system (configuration interaction method).
213: We performed two sets of model calculations. One includs all the
214: configurations within $\pm 1$ a.u. of ionization threshold of U$^{28+}$,
215: which produces 2516 states for $J^{\pi}$=$\frac{7}{2}^{-}$
216: sequence. Similarly the second calculation includes all the dielectronic
217: (doorway) states within $\pm 1$ a.u. of ionization threshold and produces 108
218: states for $J^{\pi}$=$\frac{7}{2}^{-}$ sequence. This shows that the
219: number of states associated with the dielectronic configurations for a
220: given $J^{\pi}$ are not large in number. As a result we performed the
221: full CI calculations for Hamiltonian matrix of size 108 and 2516
222: respectively and obtained the eigen values and eigenstate components. To
223: study the mixing between the doubly excited states we analyze the
224: eigenstate components by calculating the weight of a given doorway
225: configuration shown in Figure 4(a)-(e). The weight ($w$) of a doorway
226: state can be calculated as $\Sigma_{k=1,Nc}|C^{\nu}_{k}|^{2}$, where
227: $N_{c}$ is the number of states in each configuration. It has been
228: shown in Ref\cite{Sahoo:JPB} that when there is strong and uniform
229: configuration mixing the
230: weights significantly reduce from 1, but in the present case the weights
231: go down from 1 but not significantly. So one can say that these doubly
232: excited (doorways) states weakly mix with each other. When they are
233: included in the large calculation (2516 states), their weights
234: significantly go down from 1 as shown in Figure 4(f)-(j). This gives a
235: signature of strong mixing i.e., the doorways show a significant mixing with multiply excited states. It has also been
236: observed that the multiply excited configurations mix with each other
237: quite comfortably. This mixing is mainly responsible for an enhancement of recombination rates
238: over RR. In the recombination cross sections, the doorways which do not mix completely with either multiply excited states or with other doorways (in a sense they remain 'isolated') appear in the form of narrow peaks (resonances) which explains the experiment as well as the theory\cite{Mitnik:98}.
239:
240: On the other hand, when the level density is high and the two-body interaction
241: is sufficiently strong the system is driven into a regime of {\em many-body
242: quantum chaos}, where the effect of configuration mixing can be described
243: statistically in the case of Au$^{25+}$\cite{Au}. This regime is characterized by the following
244: \cite{Zel,Ce}. (i) Each eigenstate contains a large number $N$ of
245: {\em principal} components $C^{(\nu )}_k\sim 1/\sqrt{N}$, corresponding to
246: the basis states $\Phi _k $ which are strongly mixed together. (ii) Owing to
247: the strong mixing, the only good quantum numbers that can be used to classify
248: the eigenstates, are the exactly conserved total angular momentum and parity
249: $J^\pi $ and the energy. (iii) The degree of mixing in this regime is in some
250: sense complete, i.e. all basis states that can be mixed (within a
251: certain energy rage, see below) are mixed. The notion of configurations based
252: on the single-particle orbitals becomes largely irrelevant for the purpose of
253: classifying the eigenstates. Each eigenstate contains substantial
254: contributions of a few nearby configurations. As mentioned above, in the present case we
255: found that there is a weak configuration mixing between the
256: dielectronic doorway configurations and complicated multiply excited
257: states, and the multiply excited states show a substantially strong
258: configuration mixing with each other. But the degree of mixing is not
259: sufficiently strong to drive the system into chaotic regime.
260:
261:
262: \begin{figure}[h]
263: \epsfxsize=10cm
264: \centering\leavevmode\epsfbox{Wt_door+2516.eps}
265: %\vspace{8pt}
266: \caption{Weights of odd dielectronic doorway states obtained from two different calculations. Figure (a)-(e) are the weights of doorways obtained from a calculation involving 108 states. Figure (f)-(j) are the weights of the same doorways obtained from a calculation involving 2516 states}.
267:
268: {(a):$4d_{3/2}^{4}$$4d_{5/2}^{6}$$4f_{5/2}^{6}$$4f_{7/2}^{\bf7}$$5s_{1/2}^{2}$$5p_{1/2}^{2}$$5p_{3/2}^{1}$$7f_{5/2}^{1}$($N_{c}$=$4$)
269: (b):$4d_{3/2}^{4}$$4d_{5/2}^{6}$$4f_{5/2}^{6}$$4f_{7/2}^{\bf7}$$5s_{1/2}^{2}$$5p_{1/2}^{2}$$5p_{3/2}^{1}$$7f_{7/2}^{1}$($N_{c}$=$4$)
270: (c):$4d_{3/2}^{4}$$4d_{5/2}^{6}$$4f_{5/2}^{6}$$4f_{7/2}^{8}$$5s_{1/2}^{2}$$5p_{1/2}^{\bf1}$$5g_{9/2}^{1}$$6p_{3/2}^{1}$($N_{c}$=$2$)
271: (d):$4d_{3/2}^{4}$$4d_{5/2}^{6}$$4f_{5/2}^{6}$$4f_{7/2}^{8}$$5s_{1/2}^{\bf1}$$5p_{1/2}^{2}$$5d_{5/2}^{1}$$7f_{5/2}^{1}$($N_{c}$=$2$)
272: (e):$4d_{3/2}^{4}$$4d_{5/2}^{6}$$4f_{5/2}^{\bf5}$$4f_{7/2}^{8}$$5s_{1/2}^{2}$$5p_{1/2}^{2}$$5f_{5/2}^{1}$$6p_{1/2}^{1}$($N_{c}$=$2$)}
273:
274: {$N_{c}$ is the number of states associated with each of the configurations}.
275: \end{figure}
276:
277: This feature can be studied from the inverse
278: participation ratio (IPR): $\Sigma_{j=1,Nc}|C^{\nu}_{k}|^{-4}$.
279: \begin{figure}[h]
280: \epsfxsize=10cm
281: \centering\leavevmode\epsfbox{IPR_FINAL.eps}
282: %\vspace{8pt}
283: \caption{Inverse participation ratio (IPR). Top figure: for doorways
284: having 108 states, the middle figure: for all configurations having
285: 2516 states and the bottom one is same as middle figure but in
286: an extended scale.}
287: \end{figure}
288: Figure 5
289: shows the inverse participation ratio for the two sets of model
290: calculations. The top figure depicts the inverse participation ratio of
291: the doorways only and the middle one for the large (2516 states)
292: calculation that includes all the configurations (both dielectronic and
293: multiply excited states). We found most of the doorways are associated
294: with number of states either 4 or 2 or even less. The top figure shows
295: that the inverse participation ratio lies flat between 2 and 4 except
296: at the energy range where the eigen energies are close to the ionization
297: threshold. This indicates that a few number of the doorways having
298: energies close to
299: the threshold participate in mixing. However the middle figure shows a
300: picture of strong but non uniform mixing which involves a lots of multiply excited
301: states. When we included the IPR of doorways with the IPR obtained from
302: large calculation as shown in the bottom figure, it
303: lies well below as indicated by solid circles. It also provides us with
304: an information that some of the doorways take part in mixing with multiply excited
305: configurations and other doorways are weakly mixed or even remain isolated.
306:
307: This strong mixing takes place in a certain energy range $|E_k-E_\nu |\lesssim
308: \Gamma _{\rm spr}$, where $E_k\equiv H_{kk}$ is the mean energy of the basis
309: state and $\Gamma _{\rm spr}$ is the so-called {\em spreading width}.
310: More precisely, the mean-squared value of $C^{(\nu )}_k$
311: as a function of $E_k-E_\nu $, can be described by a Breit-Wigner (BW) formula
312: \begin{equation}\label{eq:C^2}
313: \overline {\left| C_k^{(\nu)}\right|^2}=N^{-1}\frac{\Gamma _{\rm spr}^2/4}
314: {(E_k-E_\nu )^2+\Gamma _{\rm spr}^2/4},
315: \end{equation}
316: with $N=\pi \Gamma _{\rm spr}/2D$ fixed by normalization
317: $\sum _k\left| C_k^{(\nu)}\right|^2\simeq
318: \int \overline {\left| C_k^{(\nu)}\right|^2} dE_k/D=1$.
319:
320: \begin{figure}[h]
321: \epsfxsize=10cm
322: \centering\leavevmode\epsfbox{bas_spread_door_2516.eps}
323: %\vspace{8pt}
324: \caption{Spreading of doorway configurations in complicated multiply
325: excited states obtained from 2516 x 2516 calculations for
326: $J^{\pi}=3.5^{-}$ sequence. Each of the doorway configuration spreads as (a) 0.1433 a.u., (b) 0.1310 a.u., (c) 0.1661 a.u., (d) 0.2321 a.u. and (e) 0.2310 a.u.. These doorways are same as indicated in Fig4.}
327:
328: \end{figure}
329:
330:
331: The mean-squared components are obtained by averaging over the
332: basis states associated with each of the doorway configurations and are
333: plotted as a function of eigen energies. We obtained $\Gamma _{\rm
334: spr}$ of doorways (close to the
335: ionization threshold) from the BW fit which is shown in Figure
336: 6(a)-(e). From the BW fit we observed
337: that the $\Gamma _{\rm spr}$ is not constant and it varies from 0.1
338: a.u. to 0.2 a.u. in these doorways. There may be doorways which even
339: show much less spreading. Roughly one can say the mixing takes place
340: within 0.1 a.u.. So we calculated the recombination rates with two
341: different values of spreading width which can be found in section IV. It
342: may be pointed out that though the single BW fit is not as accurate as it should be, still it gives an estimation of the important quantity i.e. $\Gamma_{\rm spr}$. It is worth mentioning that the spreading width found in Au$^{25+}$ is about 0.5 a.u.\cite{Au} and these value does not change if one performs calculations by adding more configurations\cite{Sahoo:JPB}.
343:
344:
345:
346: \section{Recombination}\label{sec:rec}
347: \subsection{Theory}
348: For low-energy electrons the contribution of the autoionising
349: states (resonances) to the recombination cross section is given by
350: (see, e.g., Ref. \cite{Landau})
351: \begin{equation}\label{eq:sigma_res}
352: \sigma _r=\frac{\pi }{k^2}\sum _\nu \frac{2J+1}{2(2J_i+1)}\,
353: \frac{\Gamma _\nu ^{(r)}\Gamma _\nu ^{(a)}}{(\varepsilon -\varepsilon _\nu )^2
354: +\Gamma _\nu ^2/4},
355: \end{equation}
356: where $\varepsilon =k^2/2$ is the electron energy, $J_i$ is the angular
357: momentum of the initial (ground) target state, $J$ are the angular momenta
358: of the resonances, $\varepsilon _\nu =E_\nu -I$ is the position of the
359: $\nu $th resonance relative to the ionization threshold of the compound
360: (final-state) ion, and $\Gamma _\nu ^{(a)}$, $\Gamma _\nu ^{(r)}$,
361: and $\Gamma _\nu =\Gamma _\nu ^{(r)}+\Gamma _\nu ^{(a)}$ are its
362: autoionisation, radiative, and total widths, respectively
363: \cite{comment4}. When the resonance spectrum is dense, $\sigma _r$
364: can be averaged over an energy interval $\Delta \varepsilon$ which
365: contains many resonances,
366: $D\ll \Delta \varepsilon \ll \varepsilon $, yielding
367: \begin{equation}\label{eq:sigres_av}
368: \bar \sigma _r=\frac{2\pi ^2}{k^2}\sum _{J^\pi }
369: \frac{2J +1}{2(2J_i+1)D}\left\langle \frac{\Gamma _\nu ^{(r)}
370: \Gamma _\nu ^{(a)}}{\Gamma _\nu ^{(r)}+\Gamma _\nu ^{(a)}} \right\rangle ,
371: \end{equation}
372: where $\langle \dots \rangle $ means averaging. If the fluorescence yield,
373: $\omega _f\equiv \Gamma _\nu ^{(r)}/(\Gamma _\nu ^{(r)}+\Gamma _\nu ^{(a)})$,
374: fluctuates weakly from resonance to resonance (see below), one can
375: write $\bar \sigma _r=\bar \sigma _c\omega _f$, where
376: \begin{equation}\label{eq:sig_cap}
377: \bar \sigma _c=\frac{\pi ^2}{k^2}\sum _{J^\pi } \frac{(2J +1)
378: \Gamma ^{(a)}}{(2J_i+1)D}
379: \end{equation}
380: is the energy-averaged capture cross section, and $\Gamma ^{(a)}$ is the
381: average autoionisation width.
382:
383: In a situation when there is a strong configuration mixing between the
384: dielectronic doorway states and multiply excited states, the capture
385: cross sections can be obtained as a sum over the single-electron excited states
386: $\alpha ,~\beta $ and hole states $\gamma $, as well as the partial waves
387: $lj$ of the continuous-spectrum electron $\varepsilon $. As a result, we have
388: \begin{eqnarray}\label{eq:sig_cap2}
389: \bar \sigma _c&=&\frac{\pi ^2}{k^2}\sum _{\alpha \beta \gamma ,lj}
390: \frac{\Gamma _{\rm spr}}{(\varepsilon -\varepsilon _\alpha -
391: \varepsilon _\beta +\varepsilon _\gamma )^2 +\Gamma _{\rm spr}^2/4}
392: \sum _\lambda \frac{\langle \alpha ,\beta \| V_\lambda \| \gamma ,
393: \varepsilon lj \rangle }{2\lambda +1}\nonumber \\
394: &\times &\Biggl[ \langle \alpha ,\beta \| V _\lambda \| \gamma ,
395: \varepsilon lj\rangle -(2\lambda +1) \sum _{\lambda '}
396: (-1)^{\lambda +\lambda '+1}\left\{ {\lambda \atop \lambda '}{j_\alpha
397: \atop j_\beta }{j \atop j_\gamma }\right\} \langle \alpha ,\beta \| V
398: _{\lambda '}\| \varepsilon lj,\gamma \rangle \Biggr] ,
399: \end{eqnarray}
400: where $\varepsilon _\alpha $, $\varepsilon _\beta $ and $\varepsilon _\gamma $
401: are the orbital energies, the two terms in square brackets represent the
402: direct and exchange contributions, and
403: $\langle \alpha ,\beta \| V_\lambda \| \gamma ,
404: \varepsilon lj \rangle $ is the reduced Coulomb matrix element (see Ref.\cite{Fl:02}).
405:
406: It is assumed that the energies of dielectronic doorway states relative
407: to the threshold is given by $\varepsilon _\alpha $ + $\varepsilon
408: _\beta $ - $\varepsilon _\gamma$.
409: A more accurate value can be obtained by using mean field energies
410: (configuration energies) of
411: doorway configurations in Eq.(10). The effect of using these two
412: different energies can be found in next subsection. We have also shown
413: the the qualitative difference between the results obtained by using two
414: different values of $\Gamma_{spr}$. Because $\Gamma_{spr}$ is well defined in
415: the case of a strong and chaotic configuration mixing. However, it can
416: not be
417: properly defined if the mixing of the configurations is weak and non-uniform.
418:
419: Equation (\ref{eq:sig_cap2}) is directly applicable to targets with
420: closed-shell ground states. If the target ground state contains partially
421: occupied orbitals, a factor
422: \begin{equation}\label{eq:occup}
423: \frac{n_\gamma }{2j_\gamma +1}\left(1-\frac{n_\alpha }{2j_\alpha
424: +1}\right)\left(1-\frac{n_\beta }{2j_\beta +1}\right),
425: \end{equation}
426: where $n_\alpha $, $n_\beta $, and $n_\gamma $ are the orbital occupation
427: numbers in the ground state $\Phi _i$, must be introduced on the right-hand
428: side of Eq. (\ref{eq:sig_cap2}). Steps similar to those that lead to
429: Eq. (\ref{eq:sig_cap2}) were used to obtain mean-squared matrix elements of
430: operators between chaotic many-body states \cite{Ce,Flambaum:93}.
431:
432: The chaotic nature of the multiply-excited states $\Psi _\nu $ can also be
433: employed to estimate their radiative widths $\Gamma _\nu ^{(r)}$.
434: Electron-photon interaction is described by a single-particle dipole operator
435: $\hat d$. Any excited electron in $\Psi _\nu $ may emit a photon, thus leading
436: to radiative stabilization of this state. The total photo-emission rate
437: $\Gamma _\nu ^{(r)}$ can be estimated as a weighted sum of the single-particle
438: rates,
439: \begin{equation}\label{eq:Gamma_r}
440: \Gamma _\nu ^{(r)}\simeq \sum _{\alpha ,\beta }
441: \frac{4\omega _{\beta \alpha }^3} {3c^3}
442: |\langle \alpha \|\hat d\|\beta \rangle |^2
443: \left\langle \frac{n_\beta }{2j_\beta +1}\left( 1-\frac{n_\alpha }
444: {2j_\alpha +1}\right) \right\rangle _\nu ,
445: \end{equation}
446: where $\omega _{\beta \alpha }=\varepsilon _\beta -\varepsilon _\alpha >0$,
447: $\langle \alpha \|\hat d\|\beta \rangle $ is the reduced dipole operator
448: between the orbitals $\alpha $ and $\beta $, and
449: $\langle \dots \rangle _\nu $ is the
450: mean occupation number factor. Since $\Psi _\nu $ have large numbers
451: of principal components $N$, their radiative widths display small $1/\sqrt{N}$
452: fluctuations. This can also be seen if one recalls that a chaotic
453: multiply-excited state is coupled by photo-emission to many lower-lying states,
454: and the total radiative width is the sum of a large number of (strongly
455: fluctuating) partial widths. A similar effect is known in compound nucleus
456: resonances in low-energy neutron scattering \cite{Bohr:69}.
457:
458: There is a certain similarity between Eqs. (\ref{eq:sig_cap2}) and
459: (\ref{eq:Gamma_r}) and those for autoionisation and radiative rates obtained
460: in a so-called configuration-average approximation \cite{Griffin:85}.
461: In both cases the answers involve squares or products of two-body Coulomb
462: matrix elements [see the direct and exchange terms in Eq. (\ref{eq:sig_cap2})],
463: or single-particle dipole amplitudes [Eq. (\ref{eq:Gamma_r})]. However, there
464: are a number of important differences between the present results and
465: the configuration-average approximation. The latter considers dielectronic
466: recombination and introduces averaging over configurations as a means of
467: simplifying the calculation. The DR cross section is averaged over an
468: arbitrary energy interval $\Delta \varepsilon $, and only the configurations
469: within this energy range contribute to the average. Effects of configuration
470: mixing as well as level mixing within a configuration are neglected.
471:
472: It is important to compare the radiative and autoionisation widths of
473: chaotic multiply-excited states. Equation (\ref{eq:Gamma_r}) shows that
474: $\Gamma ^{(r)}$ is comparable to the single-particle radiative widths.
475: On the other hand, the autoionisation width $\Gamma ^{(a)}$, is suppressed by a factor
476: $\left| C_k^{(\nu )}\right| ^2\sim N^{-1}$ relative to that of a typical
477: dielectronic resonance. Therefore, in systems
478: with dense spectra of chaotic multiply-excited states the autoionisation
479: widths are small. Physically this happens because the coupling strength
480: of a two-electron doorway state to the continuum is shared between many
481: complicated multiply-excited eigenstates.
482: As a result, the radiative width may dominate the total width of the
483: resonances, $\Gamma ^{(r)}\gg \Gamma ^{(a)}$, making their
484: fluorescence yield close to unity. However, in the case of U$^{28+}$, $\Gamma ^{(r)}\sim \Gamma ^{(a)}$. Our numerical results for the
485: recombination of U$^{28+}$ presented in Sec. \ref{sec:num}, confirm this
486: scenario.
487:
488: The resonance recombination cross section should be compared with the
489: direct radiative recombination cross section
490: \begin{equation}\label{eq:sigmad}
491: \sigma _d= \frac{32\pi }{3\sqrt{3}c^3}\,\frac{Z_i^2}
492: {k^2} \ln \left( \frac{Z_i}{n_0k}\right) ,
493: \end{equation}
494: obtained from the Kramers formula by summing over the principal
495: quantum number of the final state \cite{Au}. Here $Z_i$ is the ionic charge
496: ($Z_i=28$ for U$^{28+}$), and $n_0$ is the principal quantum number of
497: the lowest unoccupied ionic orbital ($n_0=5$). Note that the direct
498: and energy-averaged resonance recombination cross sections of
499: Eqs. (\ref{eq:sigmad})
500: and (\ref{eq:sigres_av}) have similar energy dependences.
501:
502: \subsection{Numerical results}\label{sec:num}
503:
504: Numerical calculations of the cross section from Eqs. (\ref{eq:sig_cap2})
505: and (\ref{eq:occup}) involve summation over the orbitals shown in
506: Fig. 1 with electron partial waves up to $i_{11/2}$. The results of the calculations for the
507: recombination rates are displayed in Figure 7. We calculated the
508: recombination rates using two different energies, i. e., configuration
509: energies and relativistic energies. It is found that both the
510: calculations are in good agreement. However, the magnitude of
511: recombination rates obtained from both the present calculations lie
512: above the experimental rates. This is due to the fact that we used
513: $\omega_{f}$ = 1 in the present calculations similar to that in
514: Au$^{25+}$. It is clear that in the present case $\omega_{f}$ is less
515: than 1 and remains constant throughout the energy range considered. The
516: exchange contributions are found to be about 200 times smaller than the
517: direct one so we did not include these when calculated the final rates. The
518: radiative rates are found to be smaller in magnitude in comparison to DR
519: rates and have almost negligible interference with the resonances. It may
520: be pointed out that the present calculation with
521: spreading width $\Gamma_{spr}$ $\sim 0.15$ a.u, does not show enough
522: resonance structures as have been observed in the experiment. Hence, we
523: performed another calculation with $\Gamma_{spr}$ $\sim
524: 0.05 $ a.u. and compared with the results obtained using spreading with
525: 0.15 a.u. shown in Figure 8. It is clear that the results do not show
526: any change in magnitude. However, the calculation with $\Gamma_{spr}$
527: $\sim 0.05$ interprets the resonance peaks
528: very well which are in reasonable agreement with the experiment, though
529: the position of peaks are different from the experiment. This is because the
530: present energies of the doorway states are approximate and a few eV
531: relative error is expected in this approach. The presence of narrow peaks in
532: rate coefficients can be interpreted as: the dielectronic states which
533: play the role doorway to the electron capture process weakly mix with
534: each other as has been discussed in previous section and appear as
535: single peaks. Because in statistical calculations one peak corresponds
536: to one doorway state. It may be recalled that the
537: distorted wave calculation\cite{Mitnik:98}, predicts the recombination
538: rates in agreement with the experiment above 80 eV and below this energy
539: the results are smaller in magnitude.
540: \begin{figure}[h]
541: \epsfxsize=10cm
542: \centering\leavevmode\epsfbox{Graph_final_rate_U28+.eps}
543: %\vspace{8pt}
544: \caption{Electron recombination rate on U$^{28+}$. Open circles
545: connected by solid line: Calculation 2 (using configuration energy and
546: $\Gamma_{spread}=0.15$ a.u.), Dot-dashed line: Calculation 1 (using HFD
547: energy and $\Gamma_{spread}=0.15$ a.u.), Solid lines : Experimental
548: data, Solid line connected by plus sign : Exchange contribution (using
549: HFD energy and $\Gamma_{spread}=0.15$ a.u.), Dotted line: Radiative rate.}
550: \end{figure}
551: \begin{table}
552: \caption{Electron orbitals which give the leading contribution to the low-energy (around 1 eV) electron recombination on U$^{28+}$.}
553: \label{tab:door}
554: \begin{tabular}{ccccccc}
555: %\multicolumn{4}{c}{Orbitals\tablenotemark[1]} & Direct &
556: %$\Delta E$\tablenotemark[3] & $\Delta E$\tablenotemark[4]\\
557:
558: \multicolumn{4}{c}{Orbitals[1]} & Direct &
559: $\Delta E$[3] & $\Delta E$[4]\\
560: \cline{1-4}
561: $\alpha $ & $\beta $ & $\gamma $ & $\varepsilon lj$ &
562: contribution[2] & (a.u.) & (a.u)\\
563: \hline
564: $6p_{1/2}$ & $5f_{7/2}$ & $4f_{7/2}$ & $p_{1/2}$ & 0.0117 & $-$0.132 & $-$0.114\\
565: $6p_{1/2}$ & $5f_{5/2}$ & $4f_{5/2}$ & $p_{1/2}$ & 0.0145 & 0.136 & 0.146 \\
566:
567: $5p_{3/2}$ & $7f_{7/2}$ & $4f_{7/2}$ & $p_{3/2}$ & 0.0292 & 0.085 & 0.036 \\
568:
569: $8s_{1/2}$ & $5d_{5/2}$ & $5s_{1/2}$ & $d_{3/2}$ & 0.0090 & 0.065 & 0.064 \\
570:
571: $5d_{3/2}$ & $6f_{7/2}$ & $4f_{7/2}$ & $d_{3/2}$ & 0.0707 & $-$0.033 & 0.056 \\
572:
573: $5d_{3/2}$ & $6f_{5/2}$ & $4f_{5/2}$ & $d_{3/2}$ & 0.0019 & 0.341 &0.438 \\
574:
575: $8s_{1/2}$ & $5d_{5/2}$ & $5s_{1/2}$ & $d_{5/2}$ & 0.0134 & 0.065 & 0.064 \\
576:
577: $5d_{5/2}$ & $6f_{7/2}$ & $4f_{7/2}$ & $d_{5/2}$ & 0.0019 & 0.627 & 0.602 \\
578:
579: $6p_{1/2}$ & $6f_{5/2}$ & $5p_{1/2}$ & $f_{5/2}$ & 0.0540 & $-$0.226 & $-$0.004 \\
580: $8p_{1/2}$ & $5d_{3/2}$ & $5s_{1/2}$ & $f_{5/2}$ & 0.0024 & $-$0.204 & $-$0.110\\
581:
582: $7f_{5/2}$ & $5d_{5/2}$ & $5s_{1/2}$ & $f_{5/2}$ & 0.0015 & $-$0.089 &$-$0.096 \\
583:
584: $8g_{7/2}$ & $5d_{3/2}$ & $5p_{1/2}$ & $f_{5/2}$ & 0.0083 & $-$0.087 & 0.133 \\
585:
586: $8p_{3/2}$ & $5d_{3/2}$ & $5s_{1/2}$ & $f_{7/2}$ & 0.0018 & 0.155 & 0.264 \\
587:
588: $6f_{7/2}$ & $6p_{1/2}$ & $5p_{1/2}$ & $f_{7/2}$ & 0.0758 & $-$0.158 & 0.073 \\
589:
590: $7f_{7/2}$ & $5d_{5/2}$ & $5s_{1/2}$ & $f_{7/2}$ & 0.0033 & $-$0.052 &$-$0.060 \\
591:
592: $5g_{9/2}$ & $6d_{3/2}$ & $5p_{1/2}$ & $f_{7/2}$ & 0.0011 & $-$0.159 &$-$0.119 \\
593:
594: $6d_{3/2}$ & $5g_{7/2}$ & $5p_{1/2}$ & $h_{9/2}$ & 0.0151 & $-$0.196 &$-$0.148 \\
595:
596: $6d_{5/2}$ & $5g_{7/2}$ & $5p_{1/2}$ & $h_{9/2}$ & 0.0028 & 0.040 & 0.129 \\
597:
598: $5f_{5/2}$ & $6p_{1/2}$ & $4f_{5/2}$ & $h_{9/2}$ & 0.0014 & 0.136 & 0.146 \\
599:
600: $7f_{5/2}$ & $5d_{5/2}$ & $5s_{1/2}$ & $h_{9/2}$ & 0.0094 & $-$0.090 &$-$0.096
601: \\
602:
603: $7f_{5/2}$ & $5p_{3/2}$ & $4f_{7/2}$ & $h_{9/2}$ & 0.0049 & 0.048 & $-$0.003 \\
604:
605: $5g_{7/2}$ & $6p_{3/2}$ & $5s_{1/2}$ & $h_{9/2}$ & 0.0017 & 0.447 & 0.538 \\
606:
607: $8g_{7/2}$ & $5d_{3/2}$ & $5p_{1/2}$ & $h_{9/2}$ & 0.0566 & $-$0.087 & 0.133 \\
608:
609: $5d_{5/2}$ & $7f_{5/2}$ & $5s_{1/2}$ & $h_{11/2}$ & 0.0012 & $-$0.089 & $-$0.096\\
610:
611: $6d_{5/2}$ & $5g_{7/2}$ & $5p_{1/2}$ & $h_{11/2}$ & 0.0131 & 0.040 & 0.013 \\
612:
613: $7f_{7/2}$ & $5d_{5/2}$ & $5s_{1/2}$ & $h_{11/2}$ & 0.0193 & $-$0.052 & $-$0.060 \\
614:
615: $7f_{7/2}$ & $5p_{3/2}$ & $4f_{7/2}$ & $h_{11/2}$ & 0.0084 & 0.085 & 0.036 \\
616:
617: $5g_{9/2}$ & $6d_{3/2}$ & $5p_{1/2}$ & $h_{11/2}$ & 0.0242 & $-$0.159 & $-$0.118 \\
618:
619: $5g_{9/2}$ & $6p_{3/2}$ & $5s_{1/2}$ & $h_{11/2}$ & 0.0018 & 0.485 & 0.573 \\
620:
621: $5g_{9/2}$ & $6d_{5/2}$ & $5p_{1/2}$ & $h_{11/2}$ & 0.0028 & 0.078 & 0.147 \\
622:
623: $6g_{9/2}$ & $5f_{5/2}$ & $5p_{1/2}$ & $i_{11/2}$ & 0.0053 & 0.084 & 0.157 \\
624:
625:
626: \end{tabular}
627: [1]{$\alpha $ and $\beta $ are the excited electron
628: orbitals, and $\gamma $ is the ground-state hole of the dielectronic
629: doorway state; $\varepsilon lj$ is the partial wave of the incident electron.}
630: [2]{Direct term contributions to the dimensionless sum in
631: Eq.~(\ref{eq:sig_cap2}), $\bar \sigma _c k^2/\pi ^2$, using spreading width = 0.15 a.u. and configuration energy , with magnitudes greater than $5\times 10^{-3}$.}
632: [3]{$\Delta E=\varepsilon _\alpha +\varepsilon _\beta -
633: \varepsilon _\gamma $ is the mean-field energy of the doorway state relative
634: to the threshold.}
635: [4]{$\Delta E$ is the configuration energy relative to the
636: threshold.}
637: \end{table}
638:
639: Table \ref{tab:door} lists the most important dielectronic doorway
640: contributions to the dimensionless sum in Eq.~(\ref{eq:sig_cap2}), which
641: also determines the ratio of the autoionisation width to the spacing
642: between the resonances. In total they account for about
643: two thirds of the total cross section. Although these transitions
644: have been selected according to the size
645: of their contribution, their energies from two different calculations are close to
646: the threshold, as seen in the last two columns in Table \ref{tab:door}. Indeed,
647: the spreading of configurations discussed in Sec. \ref{sec:mix} allows
648: configurations near the threshold, $|\Delta E|\lesssim \Gamma _{\rm spr}$,
649: to contribute. On the other hand, the contribution of configurations lying
650: far away from threshold, $|\Delta E|\gg \Gamma _{\rm spr}$, is suppressed. It may be pointed out that the theory of Mitnik et. al.\cite{Mitnik:98} considers the excitations from 5$s$ orbital only. However, the present calculations show that excitations involving 4$f$ orbitals are more important especially at threshold since they contribute significantly to the recombination cross sections.
651: As expected, the energy dependences of the resonance and direct recombination
652: rates are very close, although the latter is about 200 times smaller.
653: There is a good overall agreement between the resonance rate and experimental
654: data at electron energies between 1 eV and 100 eV.
655: \begin{figure}[h]
656: \epsfxsize=10cm
657: \centering\leavevmode\epsfbox{Graph_final_rateU28+_2.eps}
658: %\vspace{8pt}
659: \caption{Electron recombination rate on U$^{28+}$. Open circles
660: connected by solid line: Calculation 2 (using configuration energy and
661: $\Gamma_{spread}=.15$ a.u.), Solid line: Calculation 2 (using
662: configuration energy and $\Gamma_{spread}=0.05 $ a.u.), Dense solid line : Experiment, and Dotted line: Radiative rate.}
663: \end{figure}
664: To compare with Au$^{25+}$, we calculated the rate coefficients for the
665: electron energy ranging from 1 eV to 100 eV.
666: \begin{figure}[h]
667: \epsfxsize=10cm
668: \centering\leavevmode\epsfbox{G_final_rate_Au25+.eps}
669: %\vspace{8pt}
670: \caption{Electron recombination rate on Au$^{25+}$. Open circles
671: connected by solid line: Calculation 2 (using configuration energy and
672: $\Gamma_{spread}=0.5$ a.u., Thick Solid line (using HFD energy and $\Gamma_{spread}=0.5$ a.u.), Thick dotted line: Original calculation (as in PRA {\bf 66}, 012713 (2002)), Dense solid line : Experiment, and Dotted line : Radiative rate.}.
673: \end{figure}
674: Because the previous
675: calculation\cite{Fl:02} was restricted to the energy range below 1
676: eV. Figure 9 shows results three different calculations and the
677: experimental data. It may be seen that the present (we call our best
678: calculation) calculation uses configuration energies shows a very good
679: qualitative agreement with the experiment in comparison to the other two
680: calculations. The RR is found be of smaller in magnitude. The most
681: striking feature in this graph is that the experimental data quickly
682: departs from the theoretical values as one goes to higher energies. This
683: can be interpreted as that in this energy range the fluorescence yield
684: certainly goes down from 1 which can be well understood from the Figure
685: 10. In this figure we show the configuration energies as a function of
686: total angular momentum J. It shows that due to the presence of a lots
687: levels, many inelastic channels open up and they participate
688: strongly. This situation makes the fluorescence yield less than 1 and
689: hence the theory predicts recombination rates higher in magnitude.
690:
691: \begin{figure}[h]
692: \epsfxsize=10cm
693: \centering\leavevmode\epsfbox{J_graph_Au25+_ev.eps}
694: %\vspace{8pt}
695: \caption{Spectrum of low lying excited states of Au$^{25+}$. All
696: the levels shown belong to the 4f$^{8}$ configurations. The ground state is characterized by J=6.}
697: \end{figure}
698:
699: Which in turn just opposite to the case of U$^{28+}$ as shown in Figure 11.
700: \begin{figure}[h]
701: \epsfxsize=10cm
702: \centering\leavevmode\epsfbox{J-graph_U28+_ev.eps}
703: %\vspace{8pt}
704: \caption{Spectrum of low lying excited states of U$^{28+}$. All
705: the levels shown belong to 5s$^{2}$5p$^{2}$, 5s5p$^{3}$ and 5s$^{2}$5p5d
706: configurations (see table II). Its ground state is characterized by J=0.}
707: \end{figure}
708:
709: Finally, we compare in table II, the present energies of the lowest configurations in U$^{28+}$ with the energies calculated by using HULLAC code\cite{Mitnik:98}. As mentioned earlier that in Ref\cite{Mitnik:98}, energies have been calculated by using different codes such as multiconfiguration Hartree-Fock (MCHF), the AUTOSTRUCTURE code in the perturbative-relativistic [AS(PR)] and in the semirelativistic [AS(SR)] mode and the HULLAC code. On comparison we found the present energies are close to those obtained by Mitnik et. al. \cite{Mitnik:98} using HULLAC code since both the calculations are fully relativistic.
710: \begin{table}
711:
712: \caption{Energies of the lowest configurations in U$^{28+}$. The present calculated energies obtained by using Hartree Fock Dirac (HFD) code are compared with the energies calculated using HULLAC code [18]. The energies are in eV.}
713: \label{tab:en_con}
714: \begin{tabular}{cccc}
715: Configurations & Term(J) &HULLAC & HFD (present)\\
716: \hline
717: $5s^{2}5p^{2}$ & 0&0.00 & 0.00\\
718: $5s^{2}5p^{2}$ & 1&65.884 & 66.980\\
719: $5s^{2}5p^{2}$ & 2&68.876 & 69.904\\
720: $5s5p^{3}$ & 2& 131.80 & 132.94\\
721: $5s^{2}5p^{2}$ & 2& 137.24 & 138.80\\
722: $5s5p^{3}$ & 1&142.05 & 142.97\\
723: $5s^{2}5p^{2}$ & 0& 143.60 & 144.16\\
724: $5s^{2}5p5d$& 2& 170.33 & 171.83\\
725: $5s^{2}5p5d$& 1&182.26 & 183.27\\
726: $5s^{2}5p5d$ & 2&186.37& 188.24\\
727: $5s^{2}5p5d$ & 3& 188.07 & 189.82\\
728: $5s5p^{3}$ & 2&198.49 & 200.24\\
729: $5s5p^{3}$ & 3&202.91 & 204.78\\
730: $5s5p^{3}$ & 0&205.43 & 207.29\\
731:
732: $5s5p^{3}$ & 1&210.77 & 212.53\\
733:
734: $5s5p^{3}$ & 2&212.50 & 214.21\\
735:
736: $5s5p^{3}$ & 1&213.51 & 215.08\\
737:
738: $5s^{2}5p5d$& 2&247.30 & 249.28\\
739:
740: $5s5p^{3}$ & 1&287.25 & 250.11\\
741:
742: $5s^{2}5p5d$ &3& 248.65 & 250.66\\
743:
744: $5s^{2}5p5d$ & 0&248.40 & 250.40\\
745:
746: $5s^{2}5p5d$ & 4& 252.64 & 255.34\\
747: $5s^{2}5p5d$ & 2& 254.29 & 257.00\\
748: $5s^{2}5p5d$ & 3& 266.20 & 268.30\\
749: $5s^{2}5p5d$ & 1& 266.67 & 268.86\\
750: \end{tabular}
751: \end{table}
752: \section{Summary and outlook}
753: We have shown that the dielectronic states weakly mix with each other
754: and show a substantial mixing with complicated multiply excited
755: states. This explains the mechanism of low energy recombination of
756: electron with U$^{28+}$. The present results are found to be in good
757: agreement with the experimental data. We found that one must not ignore
758: the excitations from 4$f$ orbitals while considering the low energy
759: recombination of U$^{28+}$ ion. On detailed study, we
760: found the configuration mixing between the doubly excited states and
761: multiply excited states is not complete or uniform. Although this theory
762: is valid in a system having an extreme degree of configuration mixing
763: \cite{Fl:02}, still it predicts quite good results for other systems
764: where there is not enough strong configuration mixing, for example U$^{28+}$. This work develops further a statistical theory to study the low energy recombination process.
765: \section{acknowledgements}
766: We thank Prof. A. M\"uller for providing experimental data of Au$^{25+}$
767: and U$^{28+}$ in numerical form and for his stimulating discussions. We also
768: thank to Dr. C. Harabati for allowing us to make use of his codes as
769: well as for his useful comments. Financial help from EPSRC is highly
770: acknowledged.
771: \section{References}
772: \begin{thebibliography}{99}
773: \bibitem{Arnaud:85} M. Arnaud and R. Rsthenflug, APJ, {\bf398}, 394(1985).
774:
775: \bibitem{Holzscheiter:99} M. Holzscheiter and M. Charlton, Rep. Prog. {\bf 62},1 (1999).
776:
777: \bibitem{Massey:43}H. S. W. Massey and D. R. Bates, Rep. Prog. Phys. {\bf 9},
778: 62 (1943).
779:
780: \bibitem{Recomb:92}{\em Recombination of Atomic Ions}, NATO ASI series,
781: Series B: Physics: vol. 296, Eds. W.~G.~Graham, W. Fritsch, Y. Hahn, and
782: J. A. Tanis (Plenum Press, New York, 1992).
783:
784: \bibitem{Andersen:89}L. H. Andersen, P. Hvelplund, H. Knudsen, and
785: P. Kvistgaard, Phys. Rev. Lett. {\bf 62}, 2656 (1989).
786:
787: \bibitem{Kilgus:90}G. Kilgus {\em et al.}, Phys. Rev. Lett. {\bf 64},
788: 737 (1990).
789:
790: \bibitem{Schennach:94}S. Schennach {\em et al.}, Z. Phys. D {\bf 30},
791: 291 (1994).
792:
793: \bibitem{Gao:95}H. Gao {\em et al.}, Phys. Rev. Lett. {\bf 75}, 4381 (1995).
794:
795: \bibitem{Schuch:96}R. Shuch {\em et al.}, Hyperfine Interact. {\bf 99}, 317
796: (1996).
797:
798: \bibitem{Uwira:96}O. Uwira {\em et al.}, Hyperfine Interact. {\bf 99}, 295
799: (1996).
800:
801: \bibitem{Uwira:97}O. Uwira {\em et al.}, Hyperfine Interact. {\bf 108}, 149
802: (1997).
803:
804: \bibitem{Zong:97}W. Zong {\em et al.}, Phys. Rev. A {\bf 56}, 386 (1997).
805:
806: \bibitem{Mannervik:98} S. Mannervik {\em et al.}, Phys. Rev. Lett. {\bf 81},
807: 313 (1998).
808:
809: \bibitem{DeWitt:94}D. R. DeWitt {\em et al.}, Phys. Rev. A {\bf 50}, 1257
810: (1994); D. R. DeWitt {\em et al.}, J. Phys. B {\bf 28}, L147 (1995).
811:
812: \bibitem{DeWitt:96}D. R. DeWitt {\em et al.}, Phys. Rev. A {\bf 53}, 2327
813: (1996).
814:
815:
816: \bibitem{Hoffknecht:98} A. Hoffknecht {\em et al.}, J. Phys. B {\bf 31},
817: 2415 (1998).
818:
819: \bibitem{Mitnik:98} D. M. Mitnik {\em et al.}, Phys. Rev. A {\bf 57}, 4365
820: (1998).
821:
822: \bibitem{Au}G. F. Gribakin, A. A. Gribakina, and V. V. Flambaum, Aust. J. Phys.
823: {\bf 52}, 443 (1999); see also physics/9811010.
824:
825:
826: \bibitem{Sahoo:JPB} G. F. Gribakin and S. Sahoo, J. Phys. B:
827: At. Mol. Opt. Phys. {\bf 36} 3349 (2003).
828:
829: \bibitem{Fl:02} V. V. Flambaum, A. A. Gribakina, G. F. Gribakin and
830: C. Harabati, Phy. Rev. A {\bf 66}, 012713 (2002).
831:
832: \bibitem{Bohr:69}A. Bohr and B. Mottelson, {\em Nuclear structure}, Vol. 1
833: (Benjamin, New York, 1969).
834:
835: \bibitem{Bauche:87} J. N. Ginocchio, Phys. Rev. Lett. {\bf 31}, 1260 (1973);
836: J. Bauche and C. Bauche-Arnoult, J. Phys. B {\bf 20}, 1659 (1987).
837:
838: \bibitem{comment3}This fit is valid for $E>1$ a.u. In the Fermi-gas model
839: $a$ is related to the single-particle level density at the Fermi level
840: $g(\varepsilon _F)=3a^2/2\pi^2$. The fitted value $a=3.99$ gives
841: $g(\varepsilon _F)=2.4$ a.u., in agreement with the orbital spectrum in
842: Fig.1
843:
844:
845: \bibitem{Landau}L. D. Landau and E. M. Lifshitz, {\it Quantum Mechanics},
846: 3rd ed. (Pergamon, Oxford, 1977).
847:
848: \bibitem{comment4} Here we assume that the electron energy is below the target
849: excitation threshold.
850:
851: \bibitem{Zel}V. Zelevinsky, B. A. Brown, M. Frazier, and M. Horoi,
852: Phys. Rep. {\bf 276}, 85 (1996).
853:
854: \bibitem{Ce}V. V. Flambaum, A. A. Gribakina, G. F. Gribakin, and M. G. Kozlov,
855: Phys. Rev. A {\bf 50}, 267 (1994); V. V. Flambaum, A. A. Gribakina, and
856: G. F. Gribakin, Phys. Rev. A {\bf 54}, 2066 (1996); {\bf 58}, 230 (1998).
857:
858: \bibitem{Flambaum:93}V. V. Flambaum and O. K. Vorov, Phys. Rev. Lett.
859: {\bf 70}, 4051 (1993).
860:
861: \bibitem{Griffin:85}D. C. Griffin, M. S. Pindzola, and C. Bottcher,
862: Phys. Rev. A {\bf 31}, 568 (1985).
863:
864:
865:
866: \end{thebibliography}
867:
868: %\table
869:
870: %\begin{table}
871:
872: %\caption{Electron orbitals which give the leading contribution to
873: %the low-energy (around 1 eV) electron recombination on U$^{28+}$.}
874: %\label{tab:door}
875: %\begin{tabular}{cccccdd}
876: %\multicolumn{4}{c}{Orbitals\tablenotemark[1]} & Direct &
877: %$\Delta E$\tablenotemark[3] & $\Delta E$\tablenotemark[4]\\
878: %\cline{1-4}
879: %$\alpha $ & $\beta $ & $\gamma $ & $\varepsilon lj$ &
880: %contribution\tablenotemark[2] & (a.u.) & (a.u)\\
881: %\hline
882: %$6p_{1/2}$ & $5f_{7/2}$ & $4f_{7/2}$ & $p_{1/2}$ & 0.0117 & $-$0.132 & $-$0.114\\
883: %$6p_{1/2}$ & $5f_{5/2}$ & $4f_{5/2}$ & $p_{1/2}$ & 0.0145 & 0.136 & 0.146 \\
884:
885: %$5p_{3/2}$ & $7f_{7/2}$ & $4f_{7/2}$ & $p_{3/2}$ & 0.0292 & 0.085 & 0.036 \\
886:
887: %$8s_{1/2}$ & $5d_{5/2}$ & $5s_{1/2}$ & $d_{3/2}$ & 0.0090 & 0.065 & 0.064 \\
888:
889: %$5d_{3/2}$ & $6f_{7/2}$ & $4f_{7/2}$ & $d_{3/2}$ & 0.0707 & $-$0.033 & 0.056 \\
890:
891: %$5d_{3/2}$ & $6f_{5/2}$ & $4f_{5/2}$ & $d_{3/2}$ & 0.0019 & 0.341 &0.438 \\
892:
893: %$8s_{1/2}$ & $5d_{5/2}$ & $5s_{1/2}$ & $d_{5/2}$ & 0.0134 & 0.065 & 0.064 \\
894:
895: %$5d_{5/2}$ & $6f_{7/2}$ & $4f_{7/2}$ & $d_{5/2}$ & 0.0019 & 0.627 & 0.602 \\
896:
897: %$6p_{1/2}$ & $6f_{5/2}$ & $5p_{1/2}$ & $f_{5/2}$ & 0.0540 & $-$0.226 & $-$0.004 \\
898: %$8p_{1/2}$ & $5d_{3/2}$ & $5s_{1/2}$ & $f_{5/2}$ & 0.0024 & $-$0.204 & $-$0.110\\
899:
900: %$7f_{5/2}$ & $5d_{5/2}$ & $5s_{1/2}$ & $f_{5/2}$ & 0.0015 & $-$0.089 &$-$0.096 \\
901:
902: %$8g_{7/2}$ & $5d_{3/2}$ & $5p_{1/2}$ & $f_{5/2}$ & 0.0083 & $-$0.087 & 0.133 \\
903:
904: %$8p_{3/2}$ & $5d_{3/2}$ & $5s_{1/2}$ & $f_{7/2}$ & 0.0018 & 0.155 & 0.264 \\
905:
906: %$6f_{7/2}$ & $6p_{1/2}$ & $5p_{1/2}$ & $f_{7/2}$ & 0.0758 & $-$0.158 & 0.073 \\
907:
908: %$7f_{7/2}$ & $5d_{5/2}$ & $5s_{1/2}$ & $f_{7/2}$ & 0.0033 & $-$0.052 &$-$0.060 \\
909:
910: %$5g_{9/2}$ & $6d_{3/2}$ & $5p_{1/2}$ & $f_{7/2}$ & 0.0011 & $-$0.159 &$-$0.119 \\
911:
912: %$6d_{3/2}$ & $5g_{7/2}$ & $5p_{1/2}$ & $h_{9/2}$ & 0.0151 & $-$0.196 &$-$0.148 \\
913:
914: %$6d_{5/2}$ & $5g_{7/2}$ & $5p_{1/2}$ & $h_{9/2}$ & 0.0028 & 0.040 & 0.129 \\
915:
916: %$5f_{5/2}$ & $6p_{1/2}$ & $4f_{5/2}$ & $h_{9/2}$ & 0.0014 & 0.136 & 0.146 \\
917:
918: %$7f_{5/2}$ & $5d_{5/2}$ & $5s_{1/2}$ & $h_{9/2}$ & 0.0094 & $-$0.090 &$-$0.096
919: %\\
920:
921: %$7f_{5/2}$ & $5p_{3/2}$ & $4f_{7/2}$ & $h_{9/2}$ & 0.0049 & 0.048 & $-$0.003 \\
922:
923: %$5g_{7/2}$ & $6p_{3/2}$ & $5s_{1/2}$ & $h_{9/2}$ & 0.0017 & 0.447 & 0.538 \\
924:
925: %$8g_{7/2}$ & $5d_{3/2}$ & $5p_{1/2}$ & $h_{9/2}$ & 0.0566 & $-$0.087 & 0.133 \\
926:
927: %$5d_{5/2}$ & $7f_{5/2}$ & $5s_{1/2}$ & $h_{11/2}$ & 0.0012 & $-$0.089 & $-$0.096\\
928:
929: %$6d_{5/2}$ & $5g_{7/2}$ & $5p_{1/2}$ & $h_{11/2}$ & 0.0131 & 0.040 & 0.013 \\
930:
931: %$7f_{7/2}$ & $5d_{5/2}$ & $5s_{1/2}$ & $h_{11/2}$ & 0.0193 & $-$0.052 & $-$0.060 \\
932:
933: %$7f_{7/2}$ & $5p_{3/2}$ & $4f_{7/2}$ & $h_{11/2}$ & 0.0084 & 0.085 & 0.036 \\
934:
935: %$5g_{9/2}$ & $6d_{3/2}$ & $5p_{1/2}$ & $h_{11/2}$ & 0.0242 & $-$0.159 & $-$0.118 \\
936:
937: %$5g_{9/2}$ & $6p_{3/2}$ & $5s_{1/2}$ & $h_{11/2}$ & 0.0018 & 0.485 & 0.573 \\
938:
939: %$5g_{9/2}$ & $6d_{5/2}$ & $5p_{1/2}$ & $h_{11/2}$ & 0.0028 & 0.078 & 0.147 \\
940:
941: %$6g_{9/2}$ & $5f_{5/2}$ & $5p_{1/2}$ & $i_{11/2}$ & 0.0053 & 0.084 & 0.157 \\
942:
943:
944: %\end{tabular}
945: %\tablenotetext[1]{$\alpha $ and $\beta $ are the excited electron
946: %orbitals, and $\gamma $ is the ground-state hole of the dielectronic
947: %doorway state; $\varepsilon lj$ is the partial wave of the incident electron.}
948: %\tablenotetext[2]{Direct term contributions to the dimensionless sum in
949: %Eq.~(\ref{eq:sig_cap2}), $\bar \sigma _c k^2/\pi ^2$, using spreading width = 0.15 a.u. and configuration energy , with magnitudes greater than $5\times 10^{-3}$.}
950: %\tablenotetext[3]{$\Delta E=\varepsilon _\alpha +\varepsilon _\beta -
951: %\varepsilon _\gamma $ is the mean-field energy of the doorway state relative
952: %to the threshold.}
953: %\tablenotetext[4]{$\Delta E$ is the configuration energy relative to the
954: %threshold.}
955: %\end{table}
956: %*************************************************************************
957: %\table
958:
959: %\begin{table}
960:
961: %\caption{Energies of the lowest configurations in U$^{28+}$. The present calculated energies obtained by using Hartree Fock Dirac (HFD) code are compared with the energies calculated using HULLAC code [18]. The energies are in eV.}
962: %\label{tab:en_con}
963: %\begin{tabular}{cccc}
964: %Configurations & Term(J) &HULLAC & HFD (present)\\
965: %\hline
966: %$5s^{2}5p^{2}$ & 0&0.00 & 0.00\\
967: %$5s^{2}5p^{2}$ & 1&65.884 & 66.980\\
968: %$5s^{2}5p^{2}$ & 2&68.876 & 69.904\\
969: %$5s5p^{3}$ & 2& 131.80 & 132.94\\
970: %$5s^{2}5p^{2}$ & 2& 137.24 & 138.80\\
971: %$5s5p^{3}$ & 1&142.05 & 142.97\\
972: %$5s^{2}5p^{2}$ & 0& 143.60 & 144.16\\
973: %$5s^{2}5p5d$& 2& 170.33 & 171.83\\
974: %$5s^{2}5p5d$& 1&182.26 & 183.27\\
975: %$5s^{2}5p5d$ & 2&186.37& 188.24\\
976: %$5s^{2}5p5d$ & 3& 188.07 & 189.82\\
977: %$5s5p^{3}$ & 2&198.49 & 200.24\\
978: %$5s5p^{3}$ & 3&202.91 & 204.78\\
979: %$5s5p^{3}$ & 0&205.43 & 207.29\\
980:
981: %$5s5p^{3}$ & 1&210.77 & 212.53\\
982:
983: %$5s5p^{3}$ & 2&212.50 & 214.21\\
984:
985: %$5s5p^{3}$ & 1&213.51 & 215.08\\
986:
987: %$5s^{2}5p5d$& 2&247.30 & 249.28\\
988:
989: %$5s5p^{3}$ & 1&287.25 & 250.11\\
990:
991: %$5s^{2}5p5d$ &3& 248.65 & 250.66\\
992:
993: %$5s^{2}5p5d$ & 0&248.40 & 250.40\\
994:
995: %$5s^{2}5p5d$ & 4& 252.64 & 255.34\\
996: %$5s^{2}5p5d$ & 2& 254.29 & 257.00\\
997: %$5s^{2}5p5d$ & 3& 266.20 & 268.30\\
998: %$5s^{2}5p5d$ & 1& 266.67 & 268.86\\
999:
1000:
1001: %\end{tabular}
1002:
1003: %\end{table}
1004: %******************************************************************************
1005:
1006:
1007: %\figure
1008:
1009:
1010:
1011: \end{document}
1012:
1013: