1: \documentclass[epj]{svjour}
2: %\documentclass[epj,referee]{svjour}
3: % Remove option referee for final version
4: %
5: % Remove any % below to load the required packages
6: %\usepackage{latexsym}
7: \usepackage{graphics}
8: % etc
9: %
10: \begin{document}
11: %
12: %introduction des macros
13: %
14:
15: %
16: \title{A collision-induced satellite in the Lyman~$\beta$ profile
17: due to H-H collisions}
18: \subtitle{Lyman~$\beta$ satellites}
19: \author{N.F. Allard \inst{1}
20: \and J. Kielkopf \inst{2}
21: \and I. Drira \inst{3}
22: \and P. Schmelcher \inst{4}
23: } % Do not remove
24: %
25: \offprints{N.F. Allard} % Insert a name or remove this line
26: %
27: \institute{
28: Observatoire de Paris-Meudon, D\'epartement Atomes et
29: Mol\'ecules en Astrophysique, 92195 Meudon Principal Cedex,
30: France \\
31: CNRS Institut d'Astrophysique, 98 bis Boulevard Arago, 75014
32: Paris, France \\
33: email: allard@iap.fr
34: \and
35: Department of Physics, University of Louisville,
36: Louisville, Ky 40292 U.S.A. \\
37: email: john@aurora.physics.louisville.edu
38: \and
39: Observatoire de Paris-Meudon, D\'epartement Atomes et
40: Mol\'ecules en Astrophysique, 92195 Meudon Principal Cedex,
41: France \\
42: email:Ikhlas.Drira@obspm.fr
43: \and
44: Institut f\"{u}r Physikalische Chemie, Universit\"{a}t Heidelberg,
45: Im Neuenheimer Feld 229, 69120 Heidelberg, Germany \\
46: email: peter@tc.pci.uni-heidelberg.de
47: }
48: %
49: \date{Received: date / Revised version: date}
50: % The correct dates will be entered by Springer
51: %
52: \abstract{
53: We present a theoretical profile of the
54: Lyman~$\beta$ line of atomic hydrogen perturbed by collisions
55: with neutral hydrogen atoms and protons.
56: We use a general unified theory in which the electric dipole moment
57: varies during a collision.
58: A collision-induced satellite appears on Lyman~$\beta$, correlated to the
59: $\mathrm{B}''\bar{\mathrm{B}}\;^1\Sigma^+_u - \mathrm{X}\;^1\Sigma^+_g$
60: asymptotically forbidden transition of H$_2$.
61: As a consequence,
62: the appearance of the
63: line wing between Lyman~$\alpha$ and Lyman~$\beta$ is shown to be
64: sensitive to the relative abundance of hydrogen ions and neutral atoms,
65: and thereby to provide a temperature diagnostic for
66: stellar atmospheres and laboratory plasmas.
67: %
68: \PACS{{32.70.Jz}{Line shapes, widths, and shifts} \and
69: {52.25.Qt}{Emission, absorption, and scattering
70: of ultraviolet radiation} \and
71: {95.30.Dr}{Atomic processes and interactions} \and
72: {95.30.Ky}{Atomic and molecular data, spectra,
73: and spectral parameters} \and
74: {97.20.Rp}{Faint blue stars (including blue
75: stragglers), white dwarfs, degenerate
76: stars, nuclei of planetary nebulae }
77: } % end of PACS codes
78: } %end of abstract
79: %
80: \maketitle
81: %
82: \section{Introduction}
83: \label{intro}
84: In Allard {\it et al.}\/ 1999~\cite{allard99} we
85: derived a classical path expression for a pressure-broadened atomic
86: spectral line shape that allows for a radiative electric dipole transition
87: moment which
88: depends on the position of the perturbers. This factor is not included in the
89: more usual approximation of Anderson \& Talman 1956~\cite{anderson}
90: and Baranger 1958~\cite{baranger58a,baranger58b}.
91: We used this theory to study the influence of the
92: variation of the dipole moment on the satellites present in the far
93: wing profiles of the Lyman series lines of atomic hydrogen seen in stars and
94: in laboratory plasmas.
95:
96: Satellite features at 1600~\AA\/ and 1405~\AA\/ in
97: the Lyman~$\alpha$ wing associated with free-free quasi-molecular
98: transitions of H$_2$ and H$_2^+$ have been observed in
99: ultraviolet (UV) spectra of certain stars obtained with the
100: International Ultraviolet Explorer (IUE) and the
101: Hubble Space Telescope
102: (HST)~\cite{koester93,koester94,holweger94,bergeron95}.
103: The stars which show Lyman~$\alpha$ satellites are
104: DA white dwarfs, old {\it Horizontal} {\it Branch}\/ stars of spectral
105: type A, and the $\lambda$~Bootis stars. The last two have the
106: distinctive property of poor {\it metal}\/ content, that is, low
107: abundances of elements other than H and He.
108: Satellites also have been observed in the laboratory spectra of
109: laser-produced hydrogen plasmas~\cite{kielkopf95,kielkopf98}.
110:
111: Satellite features in hydrogen lines are not limited to Lyman~$\alpha$, which
112: is the only Lyman-series line accessible to IUE as well as HST.
113: Observations
114: with HUT (Hopkins Ultraviolet Telescope)
115: and with ORFEUS (Orbiting Retrievable Far and Extreme Ultraviolet
116: Spectrograph)
117: of
118: Lyman~$\beta$ of DA white dwarfs with
119: T$_{\rm eff}$ close to 20000~K have revealed a line shape
120: very different from the
121: expected simple Stark broadening, with line satellites near 1078 and
122: 1060~\AA\/ \cite{koester96,koester98}.
123: The satellites in the red wing of Lyman~$\beta$ are in the 905
124: to 1187~\AA\/ spectral region covered by the Far Ultraviolet Spectroscopic
125: Explorer (FUSE) launched in June 1999. Furthermore,
126: Lyman~$\beta$ profiles are also the subject of an ongoing study of
127: the far ultraviolet spectrum of dense hydrogen plasmas.
128: The strengths of these satellite features and indeed the entire shape
129: of wings in the Lyman series are very sensitive to the degree of
130: ionization in the stellar atmosphere and laboratory plasmas, because that
131: determines the relative importance of broadening by ion and neutral
132: collisions.
133:
134: In a previous work~\cite{allard98a} we
135: presented theoretical profiles of Lyman~$\beta$ perturbed solely by
136: protons. The calculations were based on the
137: accurate theoretical H$_2^+$ molecular potentials of
138: Madsen and Peek~\cite{madsen71}
139: to describe the interaction between radiator and perturber,
140: and dipole transition moments of
141: Ramaker and Peek~\cite{ramaker72}. The line profiles were included
142: as a source of opacity in model atmospheres for hot white dwarfs,
143: and the predicted spectra compared very well with the observed
144: ORFEUS spectra \cite{koester98}.
145:
146: The very recent {\it ab initio}\/ calculations of Drira 1999~\cite{drira99}
147: of electronic transition
148: moments for excited states of the H$_2$ molecule
149: and very accurate molecular potentials of Schmelcher
150: \cite{schmelcher,detmer98}
151: now allow us to compute Lyman~$\beta$ profiles simultaneously perturbed
152: by neutral atomic hydrogen and by protons.
153: The aim of our paper is to point out a collision-induced satellite
154: correlated to the
155: $\mathrm{B}''\bar{B}\;^1\Sigma^+_u - \mathrm{X}\;^1\Sigma^+_g$
156: asymptotically forbidden transition of H$_2$.
157: We show that the shape of the wing in
158: the region between Lyman~$\beta$ and Lyman~$\alpha$ is
159: particularly sensitive
160: to the relative abundance of the neutral and ion perturbers responsible
161: for the broadening of the lines.
162:
163: \section{Theory}
164:
165: The classical path theory for the shape of
166: pressure-broadened atomic spectral lines which takes into account the variation
167: of the electric dipole moment during a collision is only briefly
168: outlined here. The theory has been described in detail in~\cite{allard99}.
169: Our approach is based on the quantum theory of
170: spectral line shapes of Baranger~\cite{baranger58a,baranger58b} developed in
171: an {\it adiabatic representation}\/ to include the degeneracy of atomic
172: levels~\cite{royer74,royer80,allard94}.
173:
174:
175: \subsection{General expression for the spectrum in an adiabatic
176: representation }
177:
178: \label{sec:theory}
179: The spectrum $I(\omega)$ can be written as the Fourier transform
180: of the dipole autocorrelation function $\Phi$(s),
181: \begin{equation}
182: I(\omega)=\frac{1}{\pi} \, Re \, \int^{+\infty}_0\Phi(s)e^{-i \omega s} ds \; .
183: \end{equation}
184: Here,
185: \begin{eqnarray}
186: \Phi(s)
187: & = & {\bf Tr} \, \rho {\bf D}^{\dagger}
188: e^{ \frac{ is{\bf H} }{ \hbar} }
189: {\bf D} e^{ \frac{-is{\bf H} }{ \hbar} } \\
190: & = & \langle {\bf D}^{\dagger}(0)
191: {\bf D}(s) \rangle \label{eq:autocorr}
192: \end{eqnarray}
193: is the autocorrelation function of \mbox{ ${\bf D}(s)$ },
194: the dipole moment of the radiator in the Heisenberg
195: representation (we use bold notation for operators)~\cite{allard82}.
196: ${\bf H}$ is the total Hamiltonian
197: \begin{equation}
198: {\bf H} = {\bf T}_{nucl} + {\bf T}_{elec} + V({\bf x},{\bf R}),
199: \end{equation}
200: where ${\bf T}_{nucl}$ and ${\bf T}_{elec}$ are sums of nuclear and electronic
201: kinetic energy operators respectively, and $V({\bf x},{\bf R})$ is the
202: interaction between particles.
203: Here ${\bf x}$ denotes collectively the set of electronic coordinates
204: (position and spin) plus spin coordinates of the nuclei, while ${\bf R}$
205: denotes the set of position coordinates of the nuclei.
206: We assume that the radiating atom is immersed in a
207: perturber bath in thermal equilibrium.
208: The density matrix $\rho$ is
209: \begin{equation}
210: \rho \equiv \frac{ e^{-\beta {\bf H} } }
211: { {\bf Tr} \, e^{-\beta {\bf H} } } \; ,
212: \end{equation}
213: where $\beta$ is the inverse temperature ($1/kT$).
214:
215: We use the notation
216: \begin{equation}
217: \langle (\; ) \rangle \equiv {\bf Tr} \, \rho (\; ) \; ,
218: \end{equation}
219: where $ {\bf Tr} $ denotes the trace operation.
220:
221: The adiabatic or Born-Oppenheimer representation
222: comprises expanding states of the gas in terms of electronic states
223: $\chi_e(x;R)$ corresponding to frozen nuclear configurations.
224: In the Schr\"{o}dinger equation
225: \begin{eqnarray}
226: ( {\bf T}_{elec} + V({\bf x},{\bf R}) ) \chi_e(x;R) &=&
227: {\bf H}_{elec}(R)\chi_e(x;R) \\
228: &=& E_e(R) \chi_e(x;R) \; .
229: \end{eqnarray}
230: $R$ appears as a parameter, and the
231: eigenenergies $E_e(R)$ play the role of potential energies for the
232: nuclei.
233: Any total wave function $\Psi(x,R)$ can be
234: expanded as
235: \begin{equation}
236: \Psi(x,R) = \sum_e \psi_e(R) \, \chi_e(x;R) \; .
237: \end{equation}
238:
239: As the nuclei get far from each other, which we denote by
240: \mbox{ $R \rightarrow \infty$ },
241: the electronic energies $E_e(R)$ tend to asymptotic values
242: $E_e^{\infty}$ which are sums of individual atomic energies.
243: Since atomic states
244: are usually degenerate, there are in general several different
245: energy surfaces which tend to a same asymptotic energy as
246: \mbox{ $R \rightarrow \infty$ }.
247: We will consider specifically a single radiating atom, the
248: {\em radiator}, immersed in a gas of optically inactive atoms,
249: the {\em perturbers}.
250: For a transition $\alpha =(i,f)$ from
251: initial state $i$ to final state $f$, we have
252: $R$-dependent frequencies
253: \begin{eqnarray}
254: \omega_{e'e}(R) \equiv (E_{e'}(R)-E_e(R))/\hbar \; , & \; \; \; \; e \; \in
255: \varepsilon_i \; , \; e' \; \in \varepsilon_f
256: \end{eqnarray}
257: which tend to the isolated radiator frequency
258: \begin{equation}
259: \omega_{\alpha} \equiv \omega_{fi} \equiv (E_f^{\infty}-E_i^{\infty})/\hbar
260: \end{equation}
261: as the perturbers get sufficiently far from the radiator:
262: \begin{eqnarray}
263: \omega_{e'e}(R) \rightarrow \omega_{fi} & \; {\rm as}
264: & \; R \rightarrow \infty \;, \; \; \; \; e \; \in \varepsilon_i \; , \;
265: e' \; \in \varepsilon_f \; .
266: \end{eqnarray}
267: Let us introduce projectors ${\bf P}_e$ which select the $e^{th}$
268: adiabatic component of any $\Psi(x,R)$ according to~\cite{royer80}
269: \begin{equation}
270: {\bf P}_e\Psi(x,R) = \psi_e (R)\chi_e(x;R) \; .
271: \end{equation}
272: We write the dipole moment as a sum over transitions
273: \begin{equation}\label{eq:expdal}
274: {\bf D} = \sum_{\alpha} {\bf D}_{\alpha} \; ,
275: \end{equation}
276: \begin{equation}
277: {\bf D}_{\alpha} \equiv
278: \sum_{e,e'} \, \! ^{(\alpha)} \; {\bf P}_{e'} {\bf D} {\bf P}_{e} \; .
279: \end{equation}
280: In the Heisenberg representation
281: \begin{eqnarray}
282: {\bf D}_{\alpha}(t) & \equiv & \sum_{e,e'} \, \! ^{(\alpha)} \;
283: e^{\frac{i t {\bf H} }{\hbar }} {\bf P}_{e'}{\bf D}{\bf P}_e
284: e^{\frac{-i t{\bf H}}{\hbar }} \; , \\
285: & \equiv &\sum_{e,e'} \, \! ^{(\alpha)} \; {\bf D}_{e'e}(t) \; .
286: \label{eq:dalpha}
287: \end{eqnarray}
288: The sum $\sum_{e,e'} ^{(\alpha)}$ is over all pairs ($e,e'$) such that
289: \mbox{$\omega_{e',e}(R) \rightarrow \omega_{\alpha}$} as
290: \mbox{$R \rightarrow \infty$}.
291: Thus ${\bf D}_{\alpha}$ connects all pairs of adiabatic states whose
292: electronic energy differences become equal to $\omega_{\alpha}$
293: as $R \rightarrow \infty$. In the absence of perturbers,
294: ${\bf D}_{\alpha}$ would be the component of ${\bf D}$ responsible for
295: the radiative transitions of frequency $\omega_{\alpha}$.
296: We note that the projection operators account for the
297: weighting factors discussed in Ref.~\cite{allard94}.
298:
299: Introducing the expansion Eq.~(\ref{eq:expdal})
300: for ${\bf D}$ into the expression Eq.~(\ref{eq:autocorr}) for
301: $\Phi(s)$, we obtain
302: \begin{equation}
303: \Phi(s) = \sum_{\alpha,\beta} \Phi_{\alpha,\beta}(s)
304: \end{equation}
305: where
306: \begin{eqnarray}
307: \Phi_{\alpha,\beta}(s)
308: & = & {\bf Tr} \, \rho \, {\bf D}^{\dagger}_{\alpha} \,
309: e^{ \frac{i s {\bf H} }{\hbar } } \,
310: {\bf D}_{\beta} \,
311: e^{ \frac{-i s {\bf H} }{\hbar } } \label{eq:phi_trace} \\
312: & = & \langle \,
313: {\bf D}^{\dagger}_{\alpha}(0) \, {\bf D}_{\beta}(s) \,
314: \rangle \; .
315: \end{eqnarray}
316: The line shape is then
317: \begin{equation} \label{eq:shape1}
318: I(\omega) = \sum_{\alpha,\beta} I_{\alpha,\beta}(\omega) \; .
319: \end{equation}
320: The terms \mbox{$I_{\alpha,\beta}(\omega)$ }, \mbox{ $\alpha \neq \beta$},
321: represent interference between different spectral lines~\cite{baranger58b}.
322: When these interference terms are neglected, we get
323: \begin{equation} \label{eq:shape2}
324: I(\omega) = \sum_{\alpha} I_{\alpha}(\omega)
325: \end{equation}
326: and
327: \begin{equation} \label{eq:sumphi}
328: \Phi(s) = \sum_{\alpha} \Phi_{\alpha}(s)
329: \end{equation}
330: where
331: \begin{equation}
332: \Phi_{\alpha}(s)=
333: \langle \, {\bf D}^{\dagger}_{\alpha}(0) \, {\bf D}_{\alpha}(s) \,
334: \rangle \; .
335: \end{equation}
336: The time dependence of $\Phi_{\alpha}(s)$ is determined by
337: ${\bf D}_{\alpha}(s)$, the part of the dipole moment which,
338: in the absence of perturbers, oscillates at the frequency $ \omega_{\alpha} $.
339: Let us now denote
340: \begin{equation}
341: {\bf {d} }_{\alpha}(s) \equiv
342: {\bf D}_{\alpha} (s)e^{-i\omega_\alpha s}
343: \end{equation}
344: wherein the free evolution $e^{-i\omega_\alpha s}$ is factored out.
345:
346: For an isolated line, such as Lyman~$\alpha$, we have shown
347: (Allard {\it et al.} \cite{allard99}) that the normalized
348: line shape $J_{\alpha}(\Delta \omega)$, in the uncorrelated perturbers
349: approximation, is given by
350: \begin{equation}
351: J_{\alpha}(\Delta \omega) = {\bf FT} [e^{ng_{\alpha}(s)}]
352: \label{eq:J}
353: \end{equation}
354: In the classical path approximation, where we assume that the
355: perturber follows a rectilinear trajectory
356: at a single mean velocity $\bar{v}$, we get
357: from~\cite{allard94,allard99}
358: that $g_{\alpha}(s)$ can be written as
359: \begin{eqnarray}
360: g_{\alpha}(s) && \,= \frac{1}
361: {\sum_{e,e'} \, \! ^{(\alpha)} \, |d_{ee'}|^2 }
362: \sum_{e,e'} \, \! ^{(\alpha)} \; \; \nonumber \\
363: && \int^{+\infty}_{0}\!\!2\pi\rho d\rho
364: \int^{+\infty}_{-\infty}\!\! dx \;
365: \tilde{d}_{ee'}[ \, r(0) \, ] \, \nonumber \\
366: &&[ \, e^{\frac{i}{\hbar}\int^s_0 \, dt \;
367: V_{e'e }[ \, r(t) \, ] } \,
368: \, \tilde{d^{*}}_{ee'}[ \, r(s) \, ] \, - \, \tilde{d}_{ee'}[ \, r(0) \, ] \, ]
369: \; .
370: \label{eq:final}
371: \end{eqnarray}
372: The separation of the radiator and perturber is
373: \newline
374: \mbox{ $r(t) = [ \, \rho ^2 + (x+\bar{v} t)^2 \, ]^{1/2}$ } with
375: $\rho$ the impact parameter of the perturber trajectory and
376: $x$ is the position of the perturber along its trajectory at time
377: $t=0$.
378: The total line strength of the transition is
379: $ \sum_{e,e'} \, \! ^{(\alpha)} \, |d _{ee'}|^2 $.
380: The potential energy for a state $e$ is
381: \begin{equation}
382: V_{e}[r(t)] = E_e[ r(t) ]-E_e^{\infty} \; ;
383: \end{equation}
384: the difference potential is
385: \begin{equation}
386: V_{e' e}[r(t)] = V_{e' }[r(t)] - V_{ e}[r(t)] \; ;
387: \end{equation}
388: and we defined a {\em modulated} dipole \cite{allard99}
389: \begin{equation}
390: \tilde{d}_{ee'}[r(t)] =
391: d_{ee'}[r(t)]e^{-\frac{\beta}{2}V_{e}[r(t)] } \; ,
392: \end{equation}
393: where we denoted
394: \begin{equation}
395: d_{ee'}( {\bf \vec{r}} )=\langle \,\chi_e ( {\bf \vec{r}} )|{\bf d}|
396: \chi_{e'} ( {\bf \vec{r}} )\rangle \; .
397: \end{equation}
398: In the above, we neglected the influence of the potentials
399: $V_e (r)$ and $V_{e'} (r)$ on the perturber trajectories, which remain
400: straight lines.
401: Although we should
402: drop the Boltzmann factor $e^{-\beta V_e (r)}$ for consistency with our
403: straight trajectory approximation, by keeping it we improve the
404: result in the wings. Note that over regions where $ V_e (r) <0 $,
405: the factor $e^{-\beta V_e (r)}$ accounts for bound states of the
406: radiator-perturber pair, but in a classical approximation wherein the
407: discrete bound states are replaced by a continuum; thus any band
408: structure is smeared out.
409:
410:
411: \section{Theoretical analysis}
412: \label{sec:3}
413:
414: \subsection{Formation of line satellites}
415: \label{satform}
416: Close collisions between a radiating atom and a perturber
417: are responsible for transient quasi-molecules
418: which may lead to the appearance of satellite features in
419: the wing of an atomic line profile.
420:
421:
422: When the difference $\Delta V(R)$ between the
423: upper and lower interatomic potentials for a given transition goes through an
424: extremum, a relatively wider range of interatomic distances contribute to the
425: same spectral frequency, resulting in an enhancement, or {\it satellite}, in
426: the line wing. The unified theory~\cite{anderson,allard82}
427: predicts that there will be satellites centered periodically at
428: frequencies corresponding to the extrema of the difference potential
429: between the upper and lower states,
430: $\Delta\omega = k \Delta V_{\mathrm{ext}}, ( k = 1,2, \ldots) $
431: \cite{allard78,royer78,kielkopf79}.
432: Here $\Delta\omega$ is the frequency difference between the center of
433: the unperturbed spectral line and the satellite feature, measured for
434: convenience in the same units as the potential energy difference.
435: This series of satellites is due to many-body interactions.
436:
437: \subsection {Diatomic potentials}
438: \label{pot}
439: The adiabatic interaction of the neutral hydrogen atom with a proton
440: or another hydrogen atom
441: is described by potential energies $V_e(R)$ for each electronic state of
442: the H$_2^+$ or H$_2$ molecule ($R$ denotes the internuclear distance between
443: the radiator and the perturber).
444: For H-H$^+$ collisions we have used the potentials of H$_2^+$ calculated
445: by Madsen and Peek \cite{madsen71}.
446: For H-H collisions we have used the potentials of H$_2$ calculated
447: by Schmelcher~\cite{schmelcher}.
448:
449: \begin{figure*}
450: \resizebox{0.75\textwidth}{!}
451: { \includegraphics{cis_fig1.eps} }
452: \caption{Difference potential energies expressed in
453: cm$ ^{-1}$ for the singlet states of H$_2$ contributing to Lyman~$beta$. }
454: \label{singlets}
455: \end{figure*}
456:
457: In Fig.~\ref{singlets} we have plotted the H$_2$ potential differences
458: $\Delta V (R)$ for the
459: singlet states which contribute to Lyman~$\beta$.
460: We have used letters here to label the states.
461: B3 is the well known $\mathrm{B}''\bar{\mathrm{B}}\;^1\Sigma^+_u$ state.
462: At small internuclear separation, $R$, the state correlates with the
463: $\mathrm{B}''\;^1\Sigma^+_u $ state, the third $^1\Sigma^+_u $ state
464: of the Rydberg series. At larger internuclear separation the state has
465: an ionic character until $R$= 19 \AA\/ where the potential energy curve of
466: the $\mathrm{H}^+ + \mathrm{H}^- \; (1s)^2$ state crosses the
467: $\mathrm{H} (n=1) + \mathrm{H} (n=3)$ energy levels, because of an
468: avoid crossing $\bar{B}$ looses its ionic character.
469: Dabrowski and Herzberg~\cite{dabrowski76}
470: predicted the existence of the $\bar{B}$, the first calculations were
471: done by Kolos~\cite{kolos76,kolos81}. More recent calculations
472: are reported in \cite{detmer98,reinhold99}.
473:
474: We use B4 and B5 respectively to label the {\hbox{4 $^1\Sigma^+_u$} and
475: {\hbox{5 $^1\Sigma^+_u$} states, and
476: V and D to label the {\hbox{3 $^1\Pi_u$}} and {\hbox{2 $^1\Pi_u$}} states.
477:
478: Each difference potential exhibits at least one extremum which,
479: in principle, leads to a corresponding
480: satellite feature in the wing of Lyman~$\beta$ (see Sect.~\ref{satform}).
481: The present approach now allows us also to take into account
482: the asymptotically forbidden transitions of quasi-molecular
483: hydrogen which dissociate into (1s,3s) and (1s,3d) atoms.
484: The satellite amplitude depends on the value of the dipole moment
485: through the region of the potential extremum responsible of the
486: satellite and on the position of this extremum.
487: We have shown~\cite{allard99,allard98a,allard98b} that a large
488: enhancement in the amplitude of a spectral line satellite occurs
489: whenever the dipole moment increases through the region of
490: internuclear distance where the satellite is formed.
491:
492: The potential differences of the B3-X and B4-X transitions exhibit
493: double wells. The maximum at 3.0~\AA\/ of the B3-X potential and
494: minimum of the B4-X potential are due to
495: an avoided-crossing.
496:
497: The most significant characteristic of Fig.~\ref{singlets} is the
498: existence of the deep outer well at 6~\AA\/ of the B3-X potential.
499: The ionic interaction decays slowly making the potential energy difference
500: very broad compared to the very steep wells of the other transitions.
501: This is very important as the position of the extremum and the
502: functional dependence of the potential difference on internuclear
503: separation determine the amplitude and shape of the
504: satellites~\cite{allard94}.
505:
506: \subsection{Electronic transition dipole moments}
507:
508: The dipole moment taken between the initial and
509: final states of a radiative transition determines the transition
510: probability, but for two atoms in collision, the moment depends
511: on their separation. This modifies relative contributions to the
512: profile along the collision trajectory.
513: Dipole moments for H$^+_2$ and H$_2$, calculated as a function of
514: internuclear distance respectively by Ramaker and Peek
515: \cite{ramaker72} and by Drira \cite{drira99}, were used for
516: the transitions contributing to Lyman~$\alpha$ and Lyman~$\beta$.
517: For Lyman~$\beta$, the four components which correspond to $1s-3p$ atomic
518: transitions are dipole allowed \cite{drira99}:
519: the two singlet B4-X and D-X transitions, and the two triplet
520: {4 \hbox{$^3\Sigma_g^{+}$}}-{\hbox{$^3\Sigma_u^{+}$}}
521: and {\hbox{2 $^3\Pi_g$}}-{\hbox{$^3\Sigma_u^{+}$}} transitions .
522:
523: If electronic states $i$ and $f$ of an isolated radiator are not
524: connected by the dipole moment operator, that is if
525: $D_{if}(R \rightarrow \infty) = 0$, allowed radiative transitions
526: cannot occur between these two states. This happens for
527: the other transitions which correspond to $1s-3s$ and $1s-3d$ atomic
528: transitions.
529: Although these transitions should not contribute to the unperturbed line
530: profile, $D_{if}(R)$ may differ
531: from zero when a perturber passes close to the radiator. In this instance
532: radiative transitions are induced by collisions, but not at the unperturbed
533: line frequency.
534:
535: It often arises that an extremum in the potential difference
536: occurs when the final (or initial) potential energy curve exhibits an
537: avoided crossing, the corresponding wavefunctions exchange their
538: characteristics and the radiative dipole transition moment varies
539: dramatically with $R$.
540: This is exactly what happens for the B3-X transition.
541: To point out the importance of variation of dipole moment
542: on the formation of a collision-induced (CI) satellite,
543: we have displayed in Fig.~\ref{forbidden}
544: $D(R)$ together with the corresponding $\Delta V (R)$ for the $B3-X$
545: asymptotically forbidden transition.
546: The dipole transition is extremely small for the isolated radiating
547: atom (\mbox{ $R \rightarrow \infty$ }) but it goes through a maximum
548: at the value of $R$ where the avoid-crossing occurs and
549: remains quite important at the internuclear distance where the
550: potential difference of the outer well goes through a minimum.
551:
552: In such a case we expect a contribution from this transition in the
553: wing and the formation of a CI line satellite, if
554: it is not smeared out by larger dipole-allowed contributions.
555:
556: \begin{figure*}
557: \resizebox{0.75\textwidth}{!}
558: { \includegraphics{cis_fig2.eps} }
559: \caption{Difference potential energy $\Delta V$ in cm$^{-1}$
560: and the corresponding $D(R)$ in atomic units
561: for the dipole forbidden B3-X transition which gives the
562: CI satellite of Lyman~$\beta$. }
563: \label{forbidden}
564: \end{figure*}
565:
566:
567: \section { Lyman~$\beta$ profile }
568:
569: The Lyman profiles and satellites shown here are calculated at the low
570: densities met in the atmospheres of stars. The typical particle
571: densities from 10$^{15}$ to 10$^{17}$ cm$^{-3}$ allows
572: us to use an expansion of the autocorrelation function in powers of
573: density as described in~\cite{allard94,royer71}. Line profiles are
574: normalized so that over $\omega$ they integrate to 1.
575:
576: \subsection {Collisional profile perturbed by neutral H }
577: \label{lybh}
578: We will consider the two following mechanisms which contribute to the
579: Lyman~$\beta$ wing.
580: \begin{itemize}
581: \item The far wing of allowed dipole lines, due to the free-free
582: transition in a pair of colliding atoms.
583: \item The collision-induced absorption due to the free-free transition
584: involving the transient dipole moment existing during a binary
585: collision.
586: \end{itemize}
587: The line profile calculation shown Fig.~\ref{betatot}
588: has been done at a temperature of 10000~K for
589: a perturber density of $10^{16}$~cm$^{-3}$ of neutral hydrogen.
590: The only line feature is a broad CI satellite
591: situated at 1150~\AA\/ in the far wing, due to the B3-X dipole
592: forbidden transition.
593: Normally such an effect would be overshadowed by the allowed transition wing,
594: but in this case there is no large contribution of the dipole
595: allowed transitions in this region,
596: as can be easily predicted by the examination of
597: Fig.~\ref{singlets}.
598: The extrema of the allowed B4-X and D-X
599: transitions occur for very short distances, and are much smaller
600: compared to the position and depth of the outer well in the
601: B3-X transition (see Sect.~\ref{satform} and ~\ref{pot}).
602:
603: The collision-induced absorption depends on the internuclear
604: separation and produces very broad spectral lines with a
605: characteristic width of the order of the inverse
606: of the duration of the close collision.
607: It is strongly dependent on
608: \begin{itemize}
609: \item the temperature
610: \item the amplitude of the dipole for $R_{\rm min}$ when the
611: potential difference presents a minimum.
612: \end{itemize}
613: This emphasizes the importance
614: of the accuracy of both the potential energies {\em and} the dipole moments
615: for the line shape calculations.
616:
617: \begin{figure*}
618: \resizebox{0.75\textwidth}{!}
619: { \includegraphics{cis_fig3.eps} }
620: \caption{Line profile of Lyman~$\beta$ perturbed by neutral hydrogen
621: and protons. The dashed line ($---$) shows only the contribution from neutral
622: perturbers.}
623: \label{betatot}
624: \end{figure*}
625:
626:
627:
628: Oscillatory structures appear near the 1150~\AA\/ satellite as they
629: appear between the line and the 1600~\AA\/ satellite in both
630: theory and experiment~\cite{allard99}.
631: These oscillations were predicted by Royer \cite{royer71b} and
632: Sando {\it et al.}\/ \cite{sando69,sando73}.
633:
634: \subsection {Simultaneous perturbations by H and H$^+$}
635:
636: The complete Lyman~$\beta$ profile perturbed by collisions with
637: neutral hydrogen and protons is shown in Fig~\ref{betatot}.
638: We notice that the collision-induced H-H satellite
639: is much broader than the allowed H-H$^+$ satellite since the dipole
640: moment differs from zero only over a short range of internuclear
641: distances (see Sect.~\ref{lybh}).
642: The CI satellite of Lyman~$\beta$ is quite far from the
643: unperturbed Lyman~$\beta$ line center, actually closer
644: to the Lyman~$\alpha$ line. It is therefore necessary to take into proper
645: account the total
646: contribution of both the Lyman~$\alpha$ and Lyman~$\beta$ wings of H
647: perturbed simultaneously by neutrals and protons and to study the
648: variation of this part of the Lyman series
649: with the relative density of ionized and neutral atoms.
650:
651: In~\cite{allard99} we evaluated both the Lyman~$\alpha$ and Lyman~$\beta$ wings of H
652: perturbed by protons. However, we neglected interference terms
653: between the two lines.
654: Equation~(51) of \cite{allard99}, which gives the profile for a pair of lines
655: such as Lyman $\alpha$ and $\beta$, is
656: \begin{eqnarray}
657: I(\omega)= \phi_{\alpha}^{(0)} \, e^{nf_{\alpha}(0)} \,
658: J_{\alpha}(\omega-\omega_{\alpha}) \nonumber \\
659: + \,\phi_{\beta}^{(0)} \, e^{nf_{\beta}(0)} \,
660: J_{\beta}(\omega-\omega_{\beta}) \; .
661: \end{eqnarray}
662: The perturbed line strength $\phi_{\alpha}^{(0)} e^{nf_{\alpha}(0)}$
663: differs from the {\it free}\/ line strength $\phi_{\alpha}^{(0)}$ by the
664: factor $e^{nf_{\alpha}(0)}$. This density-dependent factor
665: expresses the fact that the total intensity radiated increases or
666: decreases when the dipole moment is increased or decreased, on
667: average, by the proximity of perturbers.
668: Because of the low densities we consider, we have neglected this factor here.
669:
670: We show the sum of the profiles of Lyman~$\alpha$ and
671: Lyman~$\beta$ in Fig.~\ref{lytotal}. We can see that a ratio of 5
672: between the neutral and proton density is
673: enough to make the CI satellite appear in the far wing.
674: The CI satellite appearance is then very sensitive to the degree of
675: ionization and may be used as a temperature diagnostic.
676:
677: \begin{figure*}
678: \resizebox{0.75\textwidth}{!}
679: { \includegraphics{cis_fig4.eps} }
680: \caption{Total profile of Lyman~$\alpha$ and
681: Lyman~$\beta$ perturbed by neutral hydrogen
682: and protons. Three different neutral densities ($1\times10^{16}$,
683: $5\times10^{16}$, $1\times10^{17}$ cm$^{-3}$) are compared for a fixed
684: ion density ($1\times10^{16}$ cm$^{-3}$).}
685: \label{lytotal}
686: \end{figure*}
687:
688:
689: \section{Conclusions}
690:
691: In the case of Lyman~$\alpha$ and the H-H$^+$ Lyman~$\beta$ satellites,
692: the potential shape played a dominant role in the large difference
693: in the broadening of the quasi-molecular features~\cite{allard98b}.
694: The width of the collision-induced absorption is determined, for the most
695: part, by the short range over which the corresponding
696: transition dipole moment is significant.
697: In the CI satellite, it is the dipole moment which is very important
698: and which is responsible of the shape of the satellite, the
699: observation of such a satellite would be a test of
700: the accuracy of the dipole moment calculation.
701: Satellites due to allowed and forbidden transitions
702: depend linearly on density. The CI satellite is very sensitive to
703: the temperature of the absorption, and it may also be used
704: as a diagnostic tool for temperature.
705: We emphasize that the effect of finite collision
706: duration does play a role in the shape of the far wing.
707: The present calculations are done in an adiabatic approximation
708: using a rectilinear trajectory. This should affect slightly the
709: shape of the satellite, although no great error is expected. We are
710: developing methods to include trajectory effects in the evaluation of the
711: line shape.
712:
713: \section*{Acknowledgements}
714:
715: The computations of dipole transition moments were
716: performed on the CRAY of the computer center IDRIS.
717: The work at the University of Louisville is supported by a grant
718: from the U.S. Department of Energy, Division of Chemical Sciences,
719: Office of Basic Energy Sciences, Office of Energy Research.
720:
721:
722:
723: %
724:
725: %
726: % BibTeX users please use
727: % \bibliographystyle{}
728: % \bibliography{}
729: %
730: % Non-BibTeX users please use
731: \begin{thebibliography}{}
732: %
733: % and use \bibitem to create references.
734: %
735: % Format for Journal Reference
736: %Author, Journal \textbf{Volume}, (year) page numbers.
737: % Format for books
738: %\bibitem{RefB}
739: %Author, \textit{Book title} (Publisher, place year) page numbers
740: %
741: \bibitem{allard99}
742: N.F. Allard, A. Royer, J.F. Kielkopf, N. Feautrier, Phys. Rev. A
743: \textbf{60}, 1021 (1999).
744: \bibitem{anderson}
745: P.W. Anderson, J.D. Talman , Bell System Technical
746: Publication No. 3117 (Murray Hill, NJ, 29 1956).
747: \bibitem{baranger58a}
748: M. Baranger, Phys. Rev. \textbf{111}, 481 (1958).
749: \bibitem{baranger58b}
750: M. Baranger, Phys. Rev. \textbf{111}, 494 (1958).
751: \bibitem{koester93}
752: D. Koester, N.F. Allard, in \textit{ White Dwarfs: Advances in
753: Observation and Theory}, edited by M. Barstow (Kluwer, Dordrecht,1993),
754: p. 237.
755: \bibitem{koester94}
756: D. Koester, N.F. Allard, G. Vauclair, Astron. Astrophys. \textbf{291},
757: L9 (1994).
758: \bibitem{holweger94}
759: H. Holweger, D. Koester, N.F. Allard, Astron. Astrophys.
760: \textbf{290}, L21 (1994)
761: \bibitem{bergeron95}
762: P. Bergeron, F. Wesemael, R. Lamontagne, G. Fontaine, R. Saffer,
763: N.F. Allard, Astrophys. J. \textbf{449}, 258 (1995).
764: \bibitem{kielkopf95}
765: J.F. Kielkopf, N.F. Allard, Astrophys. J. \textbf{450}, L75 (1995).
766: \bibitem{kielkopf98}
767: J.F. Kielkopf, N.F. Allard, Phys. Rev. A \textbf{58}, 4416 (1998).
768: \bibitem{koester96}
769: D. Koester, D.S. Finley, N.F. Allard, J.W. Kruk, R.A. Kimble,
770: Astrophys. J. \textbf{463}, L93 (1996).
771: \bibitem{koester98}
772: D. Koester, U. Sperhake, N.F. Allard, D.S. Finley, S. Jordan,
773: Astron. Astrophys. \textbf{336}, 276 (1998).
774: \bibitem{allard98a}
775: N.F. Allard, J.F. Kielkopf, N. Feautrier, Astron. Astrophy.
776: \textbf{330}, 782 (1998).
777: \bibitem{madsen71}
778: Madsen M.M., Peek J.M., Atomic Data \textbf{2}, 1971, 171.
779: \bibitem{ramaker72}
780: Ramaker D.E., Peek J.M., J. Phys. B \textbf{5}, 1972, 2175.
781: \bibitem{drira99}
782: Drira, I., J. Mol. Spectroscopy \textbf{198}, 52 (1999).
783: \bibitem{schmelcher}
784: Schmelcher, P., private communication.
785: \bibitem{detmer98}
786: T. Detmer, P. Schmelcher, L.S. Cederbaum, J. Chem. Phys. \textbf {109},
787: 9694 (1998).
788: \bibitem{royer74}
789: A. Royer, Can. J. Phys. \textbf{52}, 1816 (1974).
790: \bibitem{royer80}
791: A. Royer, Phys. Rev. A \textbf{22}, 1625 (1980).
792: \bibitem{allard94}
793: N.F. Allard, D. Koester, Feautrier N., Spielfiedel A.,
794: Astron. Astrophy. Suppl. Ser. \textbf{108}, 417 (1994).
795: \bibitem{allard82}
796: N.F. Allard, J. Kielkopf, Rev.~Mod.~Phys. \textbf{54}, 1103 (1982).
797: \bibitem{allard78}
798: N.F. Allard, J. Phys. B \textbf{11}, 1383 (1978).
799: \bibitem{royer78}
800: A. Royer, Acta~Phys.~Pol.~A \textbf{54}, 805 (1978).
801: \bibitem{kielkopf79}
802: J.F. Kielkopf, N.F. Allard, Phys. Rev. Lett. \textbf {43}, 3 (1979).
803: \bibitem{dabrowski76}
804: I. Dabrowski, G. Herzberg, Can. J. Phys. \textbf {54}, 525 (1976).
805: \bibitem{kolos76}
806: W. Kolos, J. Mol. Spectroscopy \textbf{62}, 429 (1976).
807: \bibitem{kolos81}
808: W. Kolos, J. Mol. Spectroscopy \textbf{86}, 420 (1981).
809: \bibitem{reinhold99}
810: Reinhold E., Hogervorst W., Ubachs W., Wolniewicz L. , Phys. Rev. A
811: \textbf{60}, 1258 (1999)
812: \bibitem{allard98b}
813: N.F. Allard, I. Drira, M. Gerbaldi, J.F. Kielkopf, A. Spielfiedel,
814: Astron. Astophys. \textbf{335}, 1124 (1998).
815: \bibitem{royer71}
816: A. Royer, Phys. Rev. A \textbf{3}, 2044 (1971).
817: \bibitem{royer71b}
818: A. Royer, Phys. Rev. A, \textbf{43}, 499 (1971).
819: \bibitem{sando69}
820: Sando K.M., Doyle R.O., Dalgarno A., Astrophys. J. \textbf{157}, L143 (1969).
821: \bibitem{sando73}
822: Sando K.M., Wormhoudt J.G., Phys. Rev. A \textbf{7}, 1889 (1973).
823:
824: % etc
825: \end{thebibliography}
826:
827:
828: \end{document}
829:
830:
831:
832: