1: \documentclass[12pt,preprint]{aastex}
2:
3: \shorttitle{K-shell X-ray Lines from Oxygen Ions}
4: \shortauthors{Gu et al.}
5: \received{}
6: \revised{}
7: \accepted{}
8: \ccc{AAS}
9:
10: \begin{document}
11:
12: \title{Laboratory Measurement and Theoretical Modeling of K-shell X-ray Lines
13: from Inner-shell Excited and Ionized Ions of Oxygen}
14: \author{
15: Ming Feng Gu\altaffilmark{1},
16: Mike Schmidt\altaffilmark{2,3},
17: Peter Beiersdorfer\altaffilmark{2},
18: Hui Chen\altaffilmark{2},
19: Daniel B. Thorn\altaffilmark{2},
20: Elmar Tr\"{a}bert\altaffilmark{2,4},
21: Ehud Behar\altaffilmark{5}, and
22: Steven M. Kahn\altaffilmark{1}
23: }
24: \altaffiltext{1}{Kavli Institute for Particle Astrophysics and Cosmology,
25: and \\ Department of Physics, Stanford University, CA 94305, USA}
26: \altaffiltext{2}{High Temperature and Astrophysics Division,
27: Physics and Advanced Technologies, \\ Lawrence
28: Livermore National Laboratory, Livermore, CA 94550-9234, USA}
29: \altaffiltext{3}{Technische Universit\"{a}t Dresden, Germany}
30: \altaffiltext{4}{Ruhr-Universit\"{a}t Bochum, Germany}
31: \altaffiltext{5}{Department of Physics, Technion, Haifa 32000, Israel}
32:
33: \begin{abstract}
34: We present high resolution laboratory spectra of K-shell X-ray lines from
35: inner-shell excited and ionized ions of oxygen, obtained with a reflection
36: grating spectrometer on the electron beam ion trap (EBIT-I) at the Lawrence
37: Livermore National Laboratory. Only with a multi-ion model including all major
38: atomic collisional and radiative processes, are we able to identify the
39: observed K-shell transitions of oxygen ions from \ion{O}{3} to \ion{O}{6}. The
40: wavelengths and associated
41: errors for some of the strongest transitions are given, taking into
42: account both the experimental and modeling uncertainties. The present data
43: should be useful in identifying the absorption features present in
44: astrophysical sources, such as active galactic nuclei and X-ray
45: binaries. They are also useful in providing benchmarks for the testing of
46: theoretical atomic structure calculations.
47: \end{abstract}
48:
49: \keywords{atomic data, galaxies: active, galaxies: outflow velocity}
50:
51: \section{INTRODUCTION}
52:
53: The powerful diffraction grating instruments on board {\it Chandra}
54: \citep{Canizares00,Brinkman00} and {\it XMM-Newton}
55: \citep{den Herder01} have enabled the observation of K-shell X-ray absorption
56: lines of the ionized plasma surrounding Active Galactic Nuclei (AGNs).
57: Absorption features attributed to various ionization stages of oxgygen
58: ranging from neutral to as high as \ion{O}{8} have been seen,
59: for example, in NGC 5548, NGC 3783, NGC 4593, and NGC 7469
60: \citep{Behar03, Steenbrugge03a, Steenbrugge03b, Blustin03}.
61: These new K-shell X-ray absorption measurements complement
62: those in the ultraviolet band (e.g., \citet{Arav03, Crenshaw03}).
63:
64: One important physical property that can be derived from such
65: measurements is the outflow velocity, which is based on the
66: shift of a given line from its wavelength at rest. Clearly,
67: the accuracy of this procedure depends strongly on the accuracy
68: with which the rest frame wavelength is known \citep{Behar02}.
69: Calculations offer only limited guidance, as calculations have a
70: rather high uncertainty. A comparison showed that K-shell X-ray line
71: positions of \ion{O}{6} and \ion{O}{5} calculated by several
72: authors differ by as much as 60~m{\AA} and 80~m{\AA}, respectively
73: \citep{Schmidt04}. These differences correspond to uncertainties
74: between 800 and 1100~km~s$^{-1}$, respectively. The spread in the
75: calculated rest frame wavelengths is thus significantly larger
76: than the wavelength shift expected for typical outflow velocities,
77: making most theoretical values nearly useless as rest frame reference
78: standards for a given line. Because of the difficulty of accurately predicting
79: electron-electron correlations, the uncertainty is expected to
80: increase further for calculations of the K-shell X-ray lines of
81: the lower charge states of oxygen. Only laboratory measurements
82: can establish the rest frame wavelengths of a given line and provide
83: an associated uncertainty estimate.
84:
85: Recently, \citet{Schmidt04} reported a laboratory measurement of the rest
86: frame wavelengths of the K-shell X-ray resonance lines in \ion{O}{6} and
87: \ion{O}{5}. The experimental precision was sufficient to determine flow
88: velocities to within 20 -- 40~km~s$^{-1}$, exceeding the typical measurement
89: accuracy of any current astrophysical X-ray instrument. These measurements
90: were conducted by emission spectroscopy in the interaction of oxygen ions with
91: electrons of sufficient energy to excite the respective K-shell
92: transition. Because the energy needed to excite a K-shell transition
93: is significantly larger than the energy required to ionize all but
94: heliumlike and hydrogenlike oxygen, the measurements had to be
95: performed in a non-equilibrium, ionizing plasma.
96:
97: Here we report on measurements of K-shell X-ray lines from charge
98: states of oxygen as low as \ion{O}{3}. These measurements again
99: utilize emission spectroscopy of oxygen ions excited by high-energy
100: electrons in a non-equilibrium, ionizing plasma. The lines observed
101: under such conditions may be similar to those observed in a photoionized
102: spectrum, where ionization by an electron is replaced by ionization
103: by X rays.
104:
105: The interpretation of non-equilibrium, ionizing plasmas is difficult.
106: This has recently been shown in various analyses of K-shell iron
107: spectra \citep{Decaux97, Decaux03, Jacobs97}.
108: Difficulties arise (1) because the wavelengths of the lines are not
109: well known so that the line assignment may be uncertain when
110: clusters of lines are involved and (2) because the excitation
111: processes may be complex, involving an interplay of various processes
112: that are often not fully modeled (if at all),
113: e.g., inner-shell ionization, inner-shell excitation, autoionization,
114: and radiative cascades.
115:
116: In the simplest case, it is necessary to calculate all excitation,
117: radiative, and autoionization rates affecting a given charge state. Because
118: levels in two neighboring charge states are involved in the formation of
119: spectra due to the presence of autoionization transitions, we denote this
120: model as a ``two-ion'' model. This model is insufficient to predict the
121: emission in an
122: ionizing plasma where, for example, inner-shell ionization of the
123: neighboring, lower charge state can produce an excited state that
124: in turn contributes to the observed spectrum when undergoing radiative
125: decay. A complex ``three-ion" model is typically used to describe the
126: emission from plasmas in which significant ionization and recombination
127: processes take place. Such a model was used, for example, by \citet{Doron02}
128: to examine the effect of ionization of \ion{Fe}{16} and the recombination
129: from \ion{Fe}{18} on the \ion{Fe}{17} X-ray emission. We show that
130: the three-ion model is also insufficient to describe the K-shell
131: oxygen emission observed in our ionizing plasma. In order to describe
132: our observations, we present an ``all-ion" model. In this model,
133: the emission of a given ion is coupled to more than just its nearest
134: neighbors. Only this model is found to be able to satisfactorily
135: reproduce the observed oxygen K-shell emission.
136:
137: The combination of the laboratory data and the all-ion model
138: allows us to identify lines from \ion{O}{3}, \ion{O}{4},
139: \ion{O}{5}, and \ion{O}{6} that, to the best of our knowledge,
140: have not yet been measured before. Although many features are
141: likely to be blends of several lines, the features are sufficiently
142: narrow so that we can readily assign wavelengths with small
143: uncertainties. In addition to aiding the identification of lines
144: in absorption spectra, the measured line positions are useful for
145: establishing the quality of wavelength computations and as empirical
146: input for optimizing calculational schemes, such as those carried out recently
147: by \citet{Garcia04}.
148:
149:
150:
151: \section{MEASUREMENT}
152:
153: The measurement was performed at the University of California Lawrence
154: Livermore National Laboratory using the EBIT-I electron beam ion trap.
155: The device has been used for laboratory astrophysics measurements for
156: over a decade \citep{Beiersdorfer03}.
157:
158: The X-ray emission of EBIT-I was detected by a high-resolution grazing
159: incidence spectrometer \citep{Beiersdorfer04} with an angle
160: of incidence of 88.5$^{\circ}$.
161: The spectrometer features a variable line-spaced reflective grating with
162: an average line spacing of 2400~grooves/mm and a radius of curvature
163: of 44.3~m. The spectral image of the ion cloud in the ion trap
164: is nearly flat and permits the use of a
165: two-dimensional charge-coupled device (CCD) as a multichannel detector.
166: The detector is cooled by liquid nitrogen and employs a
167: 27~mm~$\times$~26~mm, thinned,
168: back-illuminated CCD chip with 1340~$\times$~1300 pixels of
169: 20~$\mu$m $\times$ 20~$\mu$m nominal size; the quantum efficiency is
170: about 45\% at 22~{\AA}.
171: In the present setting the detector spans a wavelength interval about
172: 7~{\AA} wide,
173: which is ample to cover the present 21-24~{\AA} range of interest. The
174: resolving power of this setup, drawn from the measured spectra, is about 1100.
175:
176: Oxygen was introduced to the electron beam ion trap by injection
177: of carbon dioxide. This option
178: provided {\it in situ} calibration lines in form of helium-like oxygen
179: (\ion{O}{7}) as well as hydrogen-like carbon (\ion{C}{6}). We observed the
180: resonance, intercombination, and forbidden lines of \ion{O}{7}, denoted $w$,
181: $y$, and $z$, respectively, following the labeling convention introduced by
182: \citet{Gabriel72}. The wavelengths of these lines are well known from
183: the calculations of \citet{Drake88} and some also from measurements by
184: \citet{Engstroem95}.
185: We also observed the {\sc Lyman} series in the hydrogen-like spectrum of
186: \ion{C}{6}, {$\rm Ly\beta$} through {$\rm Ly\varepsilon$}, for which
187: \citet{GarciaMack64} provide calculated wavelengths. All these calibration
188: line wavelengths are presumed to be accurate to better than 1~m{\AA}.
189:
190: Figure~\ref{fig:mspec} shows features from 21.7--23.5~{\AA}. Ten superimposed
191: spectra, each with 120 min exposure time, show the reproducibility of weak
192: emission features. The black trace in Figure~\ref{fig:spec}
193: represents the sum of the ten spectra in the wavelength range of 21.4 --
194: 23.2~{\AA}, from 20~h of total observation time.
195: Each of the ten spectra was individually filtered against cosmic ray events,
196: calibrated, and analyzed. The reproducibility of these ten measurements
197: was used to assess statistical and systematic errors.
198:
199: The measured spectrum in Figure~\ref{fig:spec} is dominated by the He-like
200: lines $w$, $y$, and $z$. We also readily identify the Li-like line pair $q$
201: and $r$. These constitute the four strongest lines seen in the spectrum. In
202: addition, the spectrum contains nine features, labeled $A$ -- $I$. These
203: satellite lines belong to charge states between \ion{O}{3} and \ion{O}{6}. They
204: are situated at the long-wavelength side of the spectrum, and
205: their intensities are considerably weaker than those of the \ion{O}{7}
206: lines. This is partly due to the fact that the upper levels of these
207: transitions have high autoionization rates, and therefore small radiative
208: branching ratios. Another reason lies in the ionization balance; at the
209: electron energies that are needed to excite K-shell lines, the charge state
210: distribution strongly favors \ion{O}{7},
211: leaving only a small fraction of \ion{O}{6} and a minute abundance of
212: \ion{O}{5} and lower charge states (\ion{O}{4} and \ion{O}{3}).
213: In fact, the ionization potential of \ion{O}{4} and \ion{O}{3} is about
214: 114~eV and 77~eV, respectively, while for
215: technical reasons we used an electron beam energy of 4~keV. We therefore
216: continually supplied neutral oxygen and then detected the different charge
217: states lines (\ion{O}{3} through \ion{O}{6}) during the ionizing phase,
218: before the ionization equilibrium was reached.
219: This procedure is similar to that used in earlier measurements of the
220: K-shell emission of low charge state ions of iron \citep{Decaux95,Decaux97}.
221:
222: The wavelengths of all features, $A$ -- $I$,
223: are determined relative to the reference lines from \ion{O}{7} and \ion{C}{6},
224: and are tabulated in Table~\ref{tab:id}. The uncertainty of the reference
225: lines includes two contributions, one from the errors of their observed
226: positions and the other from their wavelength errors.
227: The uncertainty of reference line positions was added in quadrature
228: to the uncertainty of the line positions of the measured line features. The
229: uncertainty of the reference line wavelengths, which is assumed to be less
230: than 1~m{\AA}, was added linearly.
231: The overall wavelength uncertainty ranges from 2 to 4~m{\AA} for line features
232: $C$ -- $G$ and $I$, and about 10 -- 15~m{\AA} for $A$, $B$ and $H$.
233:
234: It is not trivial to unambiguously identify the
235: emission features $A$ -- $I$ with
236: specific transitions in various oxygen ions. After examining several
237: theoretical models, we find that only a complex multi-ion
238: model including major collisional and radiative processes is able to
239: account for all line features. In the following we explain the details of these
240: theoretical models, and the most likely identifications of the observed emission
241: lines.
242:
243:
244: \section{THEORETICAL MODELS}
245: The modeling of K-shell emission spectra of L-shell ions under the
246: EBIT-I plasma
247: conditions requires the inclusion of various atomic processes, such
248: as electron collisional
249: excitation, radiative decay, autoionization, and direct electron collisional
250: ionization. With the present beam energy, recombination processes are
251: generally insignificant in populating the excited levels. Because no external
252: radiation field is present in EBIT, we do not include photoionization process
253: in the model.
254: The Flexible Atomic Code (FAC) is used in
255: the present analysis for the basic atomic parameters. FAC is a relativistic
256: configuration interaction atomic code, and uses the distorted-wave
257: approximation to treat electron-ion collision processes \citep{Gu03}.
258:
259: With the given non-equilibrium charge balance for O ions, inner-shell
260: direct electron collisional ionization is
261: expected to be a major process in forming the K-shell lines in addition to the
262: collisional excitation. Moreover, with an electron density of
263: $\approx 10^{12}$~cm$^{-3}$ provided by the electron beam current of 120~mA in
264: EBIT-I, many levels in the $1s^22l^q$ configurations are
265: significantly populated, and inner-shell excitation as well as ionization from
266: these excited states also play important roles in forming the observed
267: spectrum. To illustrate these effects, we have constructed three theoretical
268: collisional-radiative models with increasing sophistication, which we denote
269: as two-ion,
270: three-ion, and all-ion models. The electron density used in these models,
271: $1.2\times 10^{12}$~cm$^{-3}$, is determined by matching the observed
272: \ion{O}{7} line ratio $y/z$, which is density sensitive,
273: to the theoretical result of the all-ion model. In all three models, we only
274: consider the atomic states belonging to the $1s^22l^q$, $1s^22l^{q-1}nl$ and $1s2l^qnl$
275: ($2\le n\le 4$) configuration complexes for each ion. In
276: the calculation of atomic data, configuration interaction within the same
277: complex is included.
278:
279: In the two-ion model, we ignore the direct electron collisional ionization
280: processes, and only include collisional excitation followed by
281: radiative cascades and autoionization
282: transitions into the next higher charge state. Therefore, only levels of two
283: neighboring charge states are involved in determining the spectrum of a
284: particular ion. The emission lines from \ion{O}{3} -- \ion{O}{7} are
285: calculated separately,
286: with the fractional abundance of each charge state adjusted to match the
287: brightest lines observed. The result of this model is shown in the top panel
288: of Figure~\ref{fig:spec}. The relative abundances of \ion{O}{3} --
289: \ion{O}{7} ions are
290: 0.36, 0.42, 1.44, 0.77, 1.00, respectively. With this model, it is clear that
291: the intensities of \ion{O}{7} lines $y$ and $z$, and the line features
292: labeled as $D$, $F$, $G$, and $H$ cannot be explained. Feature $C$ is
293: attributed to the \ion{O}{5} line $\beta $, and features $E$ and $I$ are
294: attributed to the \ion{O}{4}
295: and \ion{O}{3} lines, respectively. Features $A$ and $B$ do not seem to
296: belong to the KLL satellite transitions, and their nature is discussed later
297: in this section. The failure of this model is due to the lack of ionization
298: processes that make contributions to certain lines.
299:
300: In the three-ion model, we add the next lower
301: charge state to the modeling of the
302: spectrum of a particular ion, and therefore we include the inner-shell
303: electron collisional ionization processes
304: forming K-shell lines in addition to the collisional excitation and
305: autoionization. However, the spectral contributions from
306: individual ions are
307: again treated in separate calculations, with the fractional abundance of each
308: charge state adjusted to match the brightest lines, which are the same as
309: those used in the two-ion model, except that one more ion, \ion{O}{2} with a
310: fractional abundance of 0.12 is added to the spectral model of \ion{O}{3}. The
311: result of this model is shown in the middle panel of
312: Figure~\ref{fig:spec}. The three-ion model represents a significant
313: improvement over the two-ion model. The \ion{O}{7} lines $y$ and $z$ and the
314: line features D and F are now
315: accounted for satisfactorily. However, feature $H$, which is attributed to
316: the \ion{O}{6} lines $o$ and $p$, is now overpredicted. Lines $o$ and $p$ are
317: mainly produced by the
318: ionization of \ion{O}{5}, whose abundance is fixed by matching the
319: $\beta$ line, which is mainly produced by collisional excitation, to the line
320: feature $C$. Therefore the ratio of
321: ($o+p$)/$\beta $ in this model does not depend
322: on the relative abundances of the \ion{O}{6} and \ion{O}{5} ions, and is in
323: disagreement with the observed spectrum.
324: Moreover, the observed line feature $G$
325: cannot be explained satisfactorily in this model. It can be tentatively
326: attributed to the \ion{O}{3}
327: lines in its vicinity, which are mainly due to the ionization of \ion{O}{2}
328: ions. However, the ionization of \ion{O}{2} ions also contributes to
329: the intensity of line feature $I$, and with the chosen abundance of 0.12 for
330: \ion{O}{2}, line $I$ is already
331: overpredicted, while the \ion{O}{3} lines in the vicinity of $G$ are
332: not nearly strong enough to account for its intensity.
333:
334: The problems with the three-ion model arise from the fact that only
335: levels in three adjacent charge states are allowed to influence the spectrum
336: of the ion. For example,
337: when calculating the spectrum of \ion{O}{5}, only levels in \ion{O}{6},
338: \ion{O}{5}, and \ion{O}{4} enter the model. At very low electron
339: densities, where only the
340: ground state of each charge state is significantly populated, this is a good
341: approximation. At the present electron density, many states in the $1s^22l^q$
342: configurations have large populations relative to the ground state. The
343: ionization of these excited states has an important role in populating the
344: excited states of the next higher charge state, which in turn makes significant
345: contributions to the K-shell lines of even higher charge states. In the
346: example of \ion{O}{5}, therefore, all ions having
347: lower charge than \ion{O}{5} can
348: indirectly influence the level population of \ion{O}{5} through successive
349: ionization. With a three-ion model, such indirect effects are lost. In order
350: to address such far-reaching line formation processes, we have constructed an
351: all-ion model, which includes \ion{O}{1}
352: -- \ion{O}{7} charge states in a single collisional radiative model. The relative
353: abundances of
354: \ion{O}{1} -- \ion{O}{7} are adjusted to be 0.01, 0.03, 0.10, 0.15,
355: 0.36, 0.77, and 1.00, respectively, in order to match the brightest lines in
356: the observed spectrum. The small abundances of \ion{O}{1} and \ion{O}{2} are
357: required by the lack of any
358: significant lines at wavelengths longer than 23.2 {\AA} (see
359: Figure~\ref{fig:mspec}). The result of this
360: model is shown in the bottom panel of Figure~\ref{fig:spec}. Unlike the
361: two and three-ion models, this model clearly accounts for all major emission
362: lines in the observed spectrum. An interesting
363: fact is that now line feature $C$
364: is mainly attributed
365: to the \ion{O}{6} lines $u$ and $v$ instead of \ion{O}{5} $\beta $, and
366: feature $G$ is attributed to \ion{O}{5} lines, both of which arise
367: due to the step-wise ionization processes from the lower charge states.
368:
369: The observed spectrum also shows two significant features, $A$ and $B$, on the
370: long wavelength shoulder of the $w$ and $y$
371: lines. The theoretical models indicate
372: that all $1s^22l-1s2l^2$ transitions of \ion{O}{6}, notably, $s$, $t$, $n$,
373: and $m$, are too
374: weak to explain them, and no lines from lower charge states appear in this
375: region. However, the models do give several \ion{O}{6} lines resulting from the
376: $1s^23l^\prime-1s2s3l$ ($l^\prime=l\pm 1$) transitions in the vicinity of this
377: region, and they have intensities comparable to those observed. These lines are
378: mainly produced
379: by the inner-shell excitation of the \ion{O}{6} ground state, and are
380: enabled only
381: through configuration interaction effects, since the radiative decay involves
382: two-electron transitions. However, due to the difficulties in calculating
383: the radiative and autoionization rates associated with these levels,
384: the confidence in these identifications is somewhat limited.
385:
386: In Table~\ref{tab:id}, we list possible line identifications of all
387: significant features from \ion{O}{3} -- \ion{O}{6} in the observed
388: spectrum. All of
389: these features consist of multiple closely spaced lines, and the theoretical
390: relative intensities of individual lines are given according to the all-ion
391: model. We also compare the measured wavelengths with various theoretical
392: calculations and astrophysical observations, wherever available. In the
393: present calculation, wavelengths are obtained with the configuration
394: interaction (CI) method for all ions, and with a
395: combined CI and second-order many-body perturbation theory method (MBPT) for
396: the KLL transitions of \ion{O}{5} and \ion{O}{6} \citep{Gu04}.
397: A second-order MBPT calculation is found to
398: have no improvements over the CI method
399: for charge states lower than \ion{O}{5}, and therefore the corresponding
400: results are not shown. For \ion{O}{5} and \ion{O}{6}, the MBPT correction
401: does seem to improve the theoretical wavelengths significantly.
402:
403: \section{DISCUSSION}
404:
405: Based on the identifications presented in the previous section, we assign
406: measured wavelengths and associated uncertainties to one or two strongest
407: lines in each blended feature, as listed in Table~\ref{tab:lines}. The
408: uncertainties not only reflect the measurement errors, but also take into
409: account the fact that each observed feature contains contributions from weaker
410: lines.
411:
412: One result of this modeling effort is the recognition that
413: what we had taken to represent line $\beta $ of \ion{O}{5} at 22.374(3)~{\AA}
414: \citep{Schmidt04} is now seen
415: as a contribution to a line blend that is dominated by the \ion{O}{6}
416: satellite lines $u$ and $v$. The line blend has the small wavelength uncertainty
417: determined previously \citep{Schmidt04}. The individual, unresolved line
418: constituents now have to be
419: assigned larger errors. As we do not see any distortion of the joint line
420: profile from the instrumental profile, line $\beta $ may be assumed to lie
421: within the half width of the line, giving it a line position uncertainty of
422: about 30~m{\AA}, or ten times more than that of the full line. By comparing
423: the model calculations with the observed spectra, we notice that our new MBPT
424: wavelenghts for other \ion{O}{5} and \ion{O}{6} lines agree very well with the
425: measured values. It is therefore reasonable to expect that the MBPT wavelength
426: for line $\beta$ is reliable to within 10~m{\AA}, and we assign it a wavelength
427: of 22.370(10)~{\AA}. This uncertainty of 10~m{\AA} is about three times larger
428: than previously assumed, but still significantly better than that of
429: {\it ab initio} calculations. The uncertainty corresponds to outflow
430: velocity uncertainties of about 130~km~s$^{-1}$ and is comparable or better than
431: the measurement uncertainty of any current astrophysical measurement.
432:
433: Line feature $D$ is mainly comprised of six \ion{O}{5} transitions with
434: lower and upper configurations of $1s^22s2p$ and $1s2s2p^2$, respectively. Of
435: these six lines, the $J=2\to J=2$ transition is the strongest with
436: roughly 40\% of the total intensity. All six transitions have theoretical
437: wavelengths within 4~m{\AA} of each other. Based
438: on the representative wavelength measurement of the complex, we assign an
439: uncertainty of 8~m{\AA} to the lines in
440: Table~\ref{tab:lines}.
441:
442: Feature $E$ has two dominant components with 80\% of the total
443: intensity. The two transitions have similar
444: theoretical wavelengths. We therefore
445: expect the measured wavelength to be an accurate representation of these two
446: transitions, and we assign a small uncertainty of 5~m{\AA} to them.
447:
448: Feature $F$ is almost entirely due to two \ion{O}{4} transitions
449: between $1s^22s2p^2$($\frac{5}{2}$,$\frac{3}{2}$) and
450: $1s2s2p^3$($\frac{3}{2}$) states. The two transitions have very close
451: theoretical wavelengths, thereofore, we assign the measured wavelength to this
452: blended feature with a small uncertainty.
453:
454: Feature $G$ is mainly due to the \ion{O}{5} transition $1s^22s2p$(2) --
455: $1s2s2p^2$(2) with 75\% of the total intensity. The remaining blending have
456: the calculated wavelengths very close to the main component. Accordingly, a
457: small uncertainty is assigned to the main component.
458:
459: Feature $H$ is mainly due to the \ion{O}{6} line pair $o$ and
460: $p$. However, this feature is quite weak and broad. The calculated intensity
461: of $o$ and $p$ does not agree with the observed intensity as well as
462: other features do. It
463: is possible that many weak transitions from other charge states contribute to
464: this feature. We therefore assign a relatively large uncertainty of 20~m{\AA}
465: to this line pair.
466:
467: The major component of feature $I$ is the blend of the \ion{O}{3} lines
468: $1s^22s^22p^2$(1,2) -- $1s2s^22p^3$(1) with 70\% of the total intensity, and
469: having nearly identical theoretical wavelengths. We therefore assign the
470: measured wavelength of this feature as belonging
471: to these two transitions, with a
472: small uncertainty of 6~m{\AA}.
473:
474: The wavelengths determined for these transitions in the present work are all
475: significantly better than what is achievable with {\it ab initio}
476: calculations, and are better than or at least comparable to the accuracy of the
477: \textit{Chandra} and \textit{XMM-Newton} grating instruments. Our data provide
478: reliable reference lines for outflow velocity measurements in the absorption
479: spectroscopy of AGNs and X-ray binaries.
480:
481: \acknowledgments
482: The work at Lawrence Livermore National Laboratory was performed under
483: the auspices of the Department of Energy under Contract No. W-7405-Eng-48
484: and was supported by the National Aeronautics and Space
485: Administration under work order W19,878 issued by the Space
486: Astrophysics Research and Analysis Program. M.F.G and S.M.K acknowledge the
487: support by the NASA grant NAG5-5419.
488: E.T. acknowledges travel support from the German Research Association (DFG).
489: E.B. was supported by grant No. 2002111 from the United States Israel
490: Binational Foundation.
491:
492:
493: \begin{thebibliography}{}
494:
495: \bibitem[Arav et al.(2003)]{Arav03}
496: Arav, N., Kaastra, J., Steenbrugge, K., Brinkman, B., Edelson, R.,
497: Korista, K. T., \& De Kool, M., 2003, \apj, 590, 174
498:
499: \bibitem[Behar \& Kahn(2002)]{Behar02}
500: Behar, E., \& Kahn, S. M., 2002, NASA Laboratory Astrophysics Workshop,
501: held May 1-3 2002 at NASA Ames Research Center, Moffett Field,
502: CA 94035-1000, NASA/CP-2002-21186, 23, also astro-ph/0210280
503:
504: \bibitem[Behar et al.(2003)]{Behar03}
505: Behar, E., Rasmussen, A. P., Blustin, A. J., Sako, M., Kahn, S. M.,
506: Kaastra, J. S., {Branduardi-Raymont}, G., Steenbrugge, K. C.,
507: 2003, \apj, 598, 232
508:
509: \bibitem[Beiersdorfer(2003)]{Beiersdorfer03}
510: Beiersdorfer, P., 2003, \araa, 41, 343
511:
512: \bibitem[Beiersdorfer et al.(2004)]{Beiersdorfer04}
513: Beiersdorfer, P., Magee, E. W., Tr\"{a}bert, E., Chen, H., Lepson, J. K.,
514: Gu, M.-F., \& Schmidt, M., 2004, Rev. Sci. Instrum., 75, 3723
515:
516: \bibitem[Blustin et al.(2003)]{Blustin03}
517: Blustin, A. J., {Branduardi-Raymont}, G., Behar, E., Kaastra, J. S.,
518: Kriss, G. A., Page, M. J., Kahn, S. M., Sako, M., \& Steenbrugge, K. C.,
519: 2003, \aap, 403, 481
520:
521: \bibitem[Brinkman et al.(2000)]{Brinkman00}
522: Brinkman, A. C., Gunsing, C. J. T., Kaastra, J. S., van der Meer, R. L. J.,
523: Mewe, R., Br\"{a}uninger, H., Burkert, W., Burwitz, V.,
524: Hartner, G., Predehl, P., Ness, J.-U., Schmitt, J. H. M. M., Drake, J. J.,
525: Johnson, O., Juda, M., Kashyap, V., Murray, S., Pease, D., Ratzlaff, P.,
526: \& Wargelin, B., 2000, \apjl, 530, L111
527:
528: \bibitem[Canizares et al.(2000)]{Canizares00}
529: Canizares, C.R, Huenemoerder, D. P., Davis D. S., Dewey, D., Flanagan, K. A.,
530: Houck, J., Markert, T. H., Marshall, H. L., Schattenburg, M. L., Schulz,
531: N. S., Wise, M., Drake, J. J., \& Brickhouse, N. S., \apjl, 539, 41
532:
533: \bibitem[Chen(1985)]{Chen85}
534: Chen, M. H., 1985, \pra, 31, 1449
535:
536: \bibitem[Chen(1986)]{Chen86}
537: Chen, M. H., 1986, ADNDT, 34, 301
538:
539: \bibitem[Chen(1988)]{Chen88}
540: Chen, M. H., 1988, ADNDT, 38, 381
541:
542: \bibitem[Crenshaw et al.(2003)]{Crenshaw03}
543: Crenshaw, D. M., Kraemer, S. B., Gabel, J. R., Kaaastra, J. S.,
544: Steenbrugge, K. C., Brinkman, A. C., Dunn, J. P., George, I. M.,
545: Liedahl, D. A., Paerels, F. B. S., Turner, T. J., \& Yaqoob, T.,
546: 2003, \apj, 594, 116
547:
548: \bibitem[Decaux et al.(1995)]{Decaux95}
549: Decaux, V., Beiersdorfer, P., Osterheld, A., Chen, M., \& Kahn, S. M.,
550: 1995, \apj, 443, 464
551:
552: \bibitem[Decaux et al.(1997)]{Decaux97}
553: Decaux, V., Beiersdorfer, P., Kahn, S. M., \& Jacobs, V. L.,
554: 1997, \apj, 482, 1076
555:
556: \bibitem[Decaux et al.(2003)]{Decaux03}
557: Decaux, V., Jacobs, V. L., Beiersdorfer, P., Liedahl, D. A., \& Kahn, S. M.,
558: 2003, \pra, 68, 012509
559:
560: \bibitem[den Herder et al.(2001)]{den Herder01}
561: den Herder, J. W., Brinkman, A. C., Kahn, S. M., Branduardi-Raymond, G.,
562: Thomson, K., Aarts, H., Audard, M., Bixler, J. V., den Boggende, A. J.,
563: Gottam, J., Decker, T., Dubbeldam, L., Erd, C., Goulooze, H.,
564: G\"{u}del, M., Guttridge, P., Hailey, C. J., Al Janabi, K., Kaastra, J. S.,
565: de Korte, P. A. J., van Leeuwen, B. J., Mauche, C., McCalden, A. J.,
566: Mewe, R., Naber, A., Paerels, F. B., Peterson, J.R., Rasmussen, A. P.,
567: Rees, K., Sakelliou, I., Sako, M., Spodek, J., Stern, M., Tamura, T.,
568: Tandy, J., de Vries, C. P., Welch, S., \& Zehnder, A., 2001, \aap,
569: 365, L7
570:
571: \bibitem[Doron \& Behar(2002)]{Doron02}
572: Doron, R., \& Behar, E., 2003, \apj, 574, 2002
573:
574: \bibitem[Drake(1988)]{Drake88}
575: Drake, G. W. F., 1988, Can. J. Phys., 66, 586
576: %Theoretical energies for the n=1 and 2 states of helium
577: %isoelectronic sequence %up to Z=100,
578:
579: \bibitem[Engstr\"{o}m \& Litz\'{e}n(1995)]{Engstroem95}
580: Engstr\"{o}m, L., \& Litz\'{e}n, U., 1995, J. Phys. B: At. Mol. Opt.
581: Phys., 28, 2565
582:
583: \bibitem[Gabriel(1972)]{Gabriel72}
584: Gabriel, A. H., 1972, Mon. Not. R. Astr. Soc., 160, 99
585:
586: \bibitem[Garcia \& Mack(1964)]{GarciaMack64}
587: Garcia, J. D., \& Mack, J. E., 1964, J. Opt. Soc. Am., 55, 654
588:
589: \bibitem[Garcia et al.(2004)]{Garcia04}
590: Garcia, J., Mendoza, C., Bautista, M. A., Gorczyca, T. W., Kallman, T. R., \&
591: Palmeri, P., 2004, astro-ph/0411374
592:
593: \bibitem[Gu(2003)]{Gu03}
594: Gu, M. F., 2003, \apj, 582, 1241
595:
596: \bibitem[Gu(2004)]{Gu04}
597: Gu, M. F., 2004, \apjs, in press
598:
599: \bibitem[Jacobs et al.(1997)]{Jacobs97}
600: Jacobs, V. L., Decaux, V., \& Beiersdorfer, P., 1997,
601: J. Quant. Spect. Rad. Transf., 58, 645
602:
603: \bibitem[Kaastra et al.(2004)]{Kaastra04}
604: Kaastra, J. S., Raassen, A. J. J., Mewe, R.,
605: Arav, N., Behar, E., Costantini, E.,
606: Gabel, J. R., Kriss, G. A., Proga, D., Sako, M., \& Steenbrugge, K. C., 2004,
607: \aap, 428, 57
608:
609: \bibitem[Nicolosi \& Tondello(1977)]{Nicolosi77}
610: Nicolosi, P., \& Tondello, G., 1977, J. Opt. Soc. Am., 67, 1033
611:
612: \bibitem[Pradhan et al.(2003)]{Pradhan03}
613: Pradhan, A. K., Chen, G. X., Delahaye, F., Nahar, S. N.,
614: \& Oelgoetz, J., 2003, Mon. Not. R. Astron. Soc., 341, 1268
615:
616: \bibitem[Safronova \& Lisina(1979)]{Safronova79}
617: Safronova, U. I., \& Lisina, T. G., 1979, ADNDT, 24, 49
618:
619: \bibitem[Schmidt et al.(2004)]{Schmidt04}
620: Schmidt, M., Beiersdorfer, P., Chen, H., Thorn, D. B., Tr\"{a}bert,
621: E., \& Behar, E., 2004, \apj, 604, 562
622:
623: \bibitem[Steenbrugge et al.(2003a)]{Steenbrugge03a}
624: Steenbrugge, K. C., Kaastra, J. S., de Vries, C. P., \& Edelson, R.,
625: 2003, \aap, 402, 477
626:
627: \bibitem[Steenbrugge et al.(2003b)]{Steenbrugge03b}
628: Steenbrugge, K. C., Kaastra, J. S., Blustin, A. J., {Branduardi-Raymont}, G.,
629: Sako, M., Behar, E., Kahn, S. M., Paerels, F. B. S., \& Walter, R., 2003,
630: \aap, 408, 921
631:
632: \bibitem[Vainshtein \& Safronova(1978)]{Vainshtein78}
633: Vainshtein, L. A., \& Safronova, U. I., 1978, ADNDT, 21, 49
634:
635: \end{thebibliography}
636:
637: \clearpage
638:
639: \newpage
640: \begin{deluxetable}{ *{14}{c} }
641: \rotate
642: \tabletypesize{
643: \scriptsize
644: }
645: \tablecaption{\label{tab:id}Line identifications and wavelength comparisons
646: for K-shell lines of \ion{O}{3} -- \ion{O}{6}.}
647: \tablehead{
648: \colhead{Index} &
649: \colhead{$\lambda_{\mbox{exp}}$\tablenotemark{a}} &
650: \colhead{Ion} &
651: \colhead{Lower($J$)} &
652: \colhead{Upper($J$)} &
653: \colhead{$f_{ij}$\tablenotemark{b}} &
654: \colhead{$\lambda_{\mbox{CI}}$\tablenotemark{c}} &
655: \colhead{$\lambda_{\mbox{MBPT}}$\tablenotemark{d}} &
656: \colhead{$\lambda_{\mbox{G}}$\tablenotemark{e}} &
657: \colhead{$\lambda_{\mbox{VS}}$\tablenotemark{f}} &
658: \colhead{$\lambda_{\mbox{C}}$\tablenotemark{g}} &
659: \colhead{$\lambda_{\mbox{BK}}$\tablenotemark{h}} &
660: \colhead{$\lambda_{\mbox{K}}$\tablenotemark{i}} &
661: \colhead{Intensity}
662: }
663: \startdata
664: $A$ & 21.672(15) & \ion{O}{6} & $1s^23s$($\frac{1}{2}$) &
665: $1s2p3d$($\frac{3}{2}$) & 1.4[-2] & 21.600 & & & & & & & 0.19 \\
666: & & \ion{O}{6} & $1s^23d$($\frac{5}{2}$) &
667: $1s2p3d$($\frac{3}{2}$) & 4.5[-3] & 21.763 & & & & & & & 0.18 \\
668: $B$ & 21.845(10) & \ion{O}{6} & $1s^23s$($\frac{1}{2}$) &
669: $1s2s3p$($\frac{3}{2}$) & 1.6[-2] & 21.846 & & & & & & & 0.58 \\
670: & & \ion{O}{6} & $1s^23s$($\frac{1}{2}$) &
671: $1s2s3p$($\frac{1}{2}$) & 7.5[-3] & 21.846 & & & & & & & 0.27 \\
672: $C$ & 22.374(3) & \ion{O}{6} &
673: $1s^22s$($\frac{1}{2}$) & $1s2s2p$($\frac{1}{2}$)
674: & 2.4[-6] & 22.397 & 22.377 & 22.37& & 22.42
675: & & 22.377(12) & 5.49 \\
676: & & \ion{O}{6} & $1s^22s$($\frac{1}{2}$)&
677: $1s2s2p$($\frac{3}{2}$) & 1.2[-5] &
678: 22.397 & 22.376 & &22.373 & & & & 4.18 \\
679: & & \ion{O}{5} & $1s^22s^2$(0) & $1s2s^22p$(1) &
680: 5.5[-1] & 22.363 & 22.374 & &
681: 22.41 & 22.33 & 22.35 & & 1.25 \\
682: %\hline
683: $D$ & 22.449(3) & \ion{O}{5} & $1s^22s2p$(2) & $1s2s2p^2$(2) & 2.0[-1] & 22.465 &
684: 22.451 & & & & & 22.501(19) & 1.65 \\
685: & & \ion{O}{5} & $1s^22s2p$(1) & $1s2s2p^2$(2) & 1.2[-1] & 22.464 &
686: 22.449 & & & & & & 0.60 \\
687: & & \ion{O}{5} & $1s^22s2p$(2) & $1s2s2p^2$(1) & 6.6[-2] & 22.467 &
688: 22.453 & & & & & & 0.54 \\
689: & & \ion{O}{5} & $1s^22s2p$(0) & $1s2s2p^2$(1) & 2.8[-1] & 22.465 &
690: 22.451 & & & & & & 0.46 \\
691: & & \ion{O}{5} & $1s^22s2p$(1) & $1s2s2p^2$(0) & 9.0[-2] & 22.466 &
692: 22.451 & & & & & & 0.45 \\
693: & & \ion{O}{5} & $1s^22s2p$(1) & $1s2s2p^2$(1) & 6.3[-2] & 22.467 &
694: 22.452 & & & & & & 0.32 \\
695: %\hline
696: $E$ & 22.741(4) & \ion{O}{4} & $1s^22s^22p$($\frac{3}{2}$) &
697: $1s2s^22p^2$($\frac{3}{2}$) & 2.0[-1] & 22.741 & & & 22.75 & 22.73 & 22.73
698: & 22.740(20) & 0.42 \\
699: & & \ion{O}{4} & $1s^22s^22p$($\frac{1}{2}$) &
700: $1s2s^22p^2$($\frac{1}{2}$) & 1.6[-1] &
701: 22.741 & & & & & & & 0.19 \\
702: & & \ion{O}{4} & $1s^22s^22p$($\frac{3}{2}$) &
703: $1s2s^22p^2$($\frac{1}{2}$) & 4.0[-2] &
704: 22.743 & & & & & & & 0.09 \\
705: & & \ion{O}{4} & $1s^22s^22p$($\frac{1}{2}$) &
706: $1s2s^22p^2$($\frac{3}{2}$) & 6.0[-2] &
707: 22.739 & & & & & & & 0.06 \\
708: %\hline
709: $F$ & 22.836(4) & \ion{O}{4} & $1s^22s2p^2$($\frac{5}{2}$) &
710: $1s2s2p^3$($\frac{3}{2}$) & 1.0[-1] & 22.843 & & & & & & & 0.20 \\
711: & & \ion{O}{4} & $1s^22s2p^2$($\frac{3}{2}$) &
712: $1s2s2p^3$($\frac{3}{2}$) & 1.1[-1] & 22.842 & & & & & & & 0.14 \\
713: & & \ion{O}{4} & $1s^22s2p^2$($\frac{1}{2}$) &
714: $1s2s2p^3$($\frac{3}{2}$) & 1.0[-1] & 22.841 & & & & & & & 0.06 \\
715: %\hline
716: $G$ & 22.871(4) & \ion{O}{5} & $1s^22s2p$(2) & $1s2s2p^2$(2) & 3.6[-6] & 22.882 &
717: 22.877 & & & & & & 0.30 \\
718: & & \ion{O}{5} & $1s^22s2p$(2) & $1s2s2p^2$(3) & 4.2[-6] & 22.881 &
719: 22.875 & & & & & & 0.05 \\
720: & & \ion{O}{5} & $1s^22s2p$(2) & $1s2s2p^2$(1) & 3.2[-7] & 22.882 &
721: 22.874 & & & & & & 0.04 \\
722: %\hline
723: $H$ & 23.017(10) & \ion{O}{6} &
724: $1s^22p$($\frac{3}{2}$) & $1s2s^2$($\frac{1}{2}$)
725: & 3.6[-3] &
726: 23.055 & 23.015 & 23.00 & 23.031 & 23.08 & & & 0.06 \\
727: & & \ion{O}{6} &
728: $1s^22p$($\frac{1}{2}$)& $1s2s^2$($\frac{1}{2}$) & 3.6[-3] &
729: 23.053 & 23.012 && 23.028 & & & & 0.03 \\
730: %\hline
731: $I$ & 23.071(4) & \ion{O}{3} & $1s^22s^22p^2$(2) & $1s2s^22p^3$(1) & 8.8[-2] &
732: 23.066 & & & & & & 23.170(10) & 0.13\\
733: & & \ion{O}{3} & $1s^22s^22p^2$(1) & $1s2s^22p^3$(1) & 8.0[-2] &
734: 23.065 & & & & & & & 0.07\\
735: & & \ion{O}{3} & $1s^22s^22p^2$(0) & $1s2s^22p^3$(1) & 1.2[-1] &
736: 23.065 & & & & & & & 0.03\\
737: & & \ion{O}{3} & $1s^22s^22p^2$(2) & $1s2s^22p^3$(2) & 2.0[-1] &
738: 23.074 & & & & & & & 0.03\\
739: & & \ion{O}{3} & $1s^22s^22p^2$(2) & $1s2s^22p^3$(3) & 9.4[-2] &
740: 23.086 & & & & & & & 0.02\\
741: \enddata
742: \tablenotetext{a}{Present measurement, numbers in parentheses
743: are uncertainties in m{\AA}.}
744: \tablenotetext{b}{Theoretical absorption oscillator strength of the transition
745: with FAC, $a$[$b$] denotes $a\times 10^{b}$.}
746: \tablenotetext{c}{Configuration interaction calculation with FAC.}
747: \tablenotetext{d}{Configuration interaction with second-order MBPT
748: correction.}
749: \tablenotetext{e}{\citet{Gabriel72}}
750: \tablenotetext{f}{\citet{Vainshtein78, Safronova79}}
751: \tablenotetext{g}{\citet{Chen85, Chen86, Chen88}}
752: \tablenotetext{h}{\citet{Behar02}}
753: \tablenotetext{i}{Values for \ion{O}{5} and \ion{O}{6} are \textit{Chandra}
754: measurements adjusted for flow velocities from
755: \citet{Kaastra04}, those for lower charge states are from Kaastra (private
756: communications).}
757: \end{deluxetable}
758:
759: \clearpage
760: \begin{deluxetable}{ *{6}{c} }
761: %\tabletypesize{
762: \scriptsize
763: %}
764: \tablecaption{\label{tab:lines} Best estimate of wavelengths for the strongest
765: transitions of \ion{O}{3} -- \ion{O}{6}.}
766: \tablehead{
767: \colhead{Ion} &
768: \colhead{Lower($J$)} &
769: \colhead{Upper($J$)} &
770: \colhead{$\lambda_{\mbox{exp}}$({\AA}) \tablenotemark{a}} &
771: \colhead{Label\tablenotemark{b}} &
772: \colhead{Index\tablenotemark{c}}
773: }
774: \startdata
775: \ion{O}{6} & $1s^22s$($\frac{1}{2}$) & $1s2s2p$($\frac{1}{2}$,$\frac{3}{2}$) &
776: 22.374(8) & $u$,$v$ & $C$ \\
777: \ion{O}{6} & $1s^22p$($\frac{1}{2}$,$\frac{3}{2}$) & $1s2s^2$($\frac{1}{2}$) &
778: 23.017(20) & $o$,$p$ & $H$ \\
779: \ion{O}{5} & $1s^22s^2$(0) & $1s2s^22p$(1) &
780: 22.370(10) & $\beta$ & $C$ \\
781: \ion{O}{5} & $1s^22s2p$(0,1,2) & $1s2s2p^2$(0,1,2) &
782: 22.449(8) & & $D$ \\
783: \ion{O}{5} & $1s^22s2p$(2) & $1s2s2p^2$(2) &
784: 22.871(5) & & $G$ \\
785: \ion{O}{4} & $1s^22s^22p$($\frac{1}{2}$,$\frac{3}{2}$) &
786: $1s2s^22p^2$($\frac{1}{2}$,$\frac{3}{2}$) &
787: 22.741(5) & & $E$\\
788: \ion{O}{4} & $1s^22s2p^2$($\frac{5}{2}$,$\frac{3}{2}$) & $1s2s2p^3$($\frac{3}{2}$)&
789: 22.836(5) & & $F$\\
790: \ion{O}{3} & $1s^22s^22p^2$(1,2) & $1s2s^22p^3$(1) &
791: 22.071(6) & & $I$\\
792: \enddata
793: \tablenotetext{a}{Present estimate, numbers in parentheses
794: are uncertainties in m{\AA}.}
795: \tablenotetext{b}{{L}ine labeling convention introduced by \citet{Gabriel72}.}
796: \tablenotetext{c}{The line feature index in Table~\ref{tab:id} to which the
797: transitions belong.}
798: \end{deluxetable}
799:
800: \clearpage
801: \begin{figure}
802: \epsscale{0.7}
803: \plotone{f1.eps}
804:
805: \caption{\label{fig:mspec} Ten superimposed spectra of oxygen K-shell
806: lines observed in the wavelength range of 21.7 -- 23.5~{\AA}. Each spectrum
807: is the result of 120 min observation.}
808: \end{figure}
809:
810: \clearpage
811: \begin{figure}
812: \epsscale{0.7}
813: \plotone{f2.eps}
814:
815: \caption{\label{fig:spec}Comparison of 2-ion, 3-ion, and all-ion models
816: with the observed spectrum. Black: experiment, red: \ion{O}{7} lines,
817: green: \ion{O}{6}
818: lines, blue: \ion{O}{5} lines, magenta: \ion{O}{4} lines, and cyan:
819: \ion{O}{3} lines. The
820: theoretical wavelengths in the figure are the configuration interaction
821: results obtained with FAC. The \ion{O}{6} lines have been shifted to
822: the left by 28~m{\AA}, and the \ion{O}{5} lines have been shifted to
823: the left by 10~m{\AA}. Labels $A$ -- $I$ serve to identify the spectral
824: features discussed in the text and listed in Table~\ref{tab:id}. }
825: \end{figure}
826:
827: \end{document}
828: