1: \documentclass[12pt,preprint]{aastex}
2:
3: \newcommand{\kms}{km s$^{-1}\,$}
4: \newcommand{\tu}{2s$^2$2p$^3$ }
5: \newcommand{\fo}{[\ion{O}{2}] }
6: \newcommand{\oz}{O$^+$}
7: \newcommand{\Dd}[1]{$^2\textrm{D}^o_{#1}$}
8: \newcommand{\Pp}[1]{$^2\textrm{P}^o_{#1}$}
9: \newcommand{\Ss}{$^4\textrm{S}^o_{3/2}$}
10: \newcommand{\Ddn}{$^2$D$^o$}
11: \newcommand{\Ppn}{$^2$P$^o$}
12: \newcommand{\Ssn}{$^4$S$^o$}
13: \newcommand{\Ddx}{^2\textrm{D}^o}
14: \newcommand{\Ppx}{^2\textrm{P}^o}
15: \newcommand{\Ssx}{^4\textrm{S}^o}
16: \newcommand{\Ddnx}[1]{^2\textrm{D}^o_{#1}}
17: \newcommand{\Ppnx}[1]{^2\textrm{P}^o_{#1}}
18: \newcommand{\Ssnx}{^4\textrm{S}^o_{3/2}}
19: \newcommand{\oh}{OH }
20: \newcommand{\cmm}{cm$^{-1}$}
21: \newcommand{\cmc}{cm$^{-3}$}
22: \newcommand{\lam}{$\lambda$}
23: \newcommand{\lamm}{$\lambda\lambda$}
24:
25: \shorttitle{Doublet O II}
26: \shortauthors{SHARPEE ET AL.}
27: \begin{document}
28:
29: \title{Measurements of the singly ionized oxygen auroral doublet lines
30: $\lambda\lambda$7320,7330\AA\ using high resolution sky spectra}
31:
32: \author{Brian D. Sharpee, Tom G. Slanger, David L. Huestis, Philip
33: C. Cosby} \affil{Molecular Physics Laboratory, SRI International, 333
34: Ravenswood Ave, Menlo Park, CA 94025 } \email{brian.sharpee@sri.com,
35: tom.slanger@sri.com, david.huestis@sri.com, philip.cosby@sri.com}
36:
37: \begin{abstract}
38: The wavelengths of the individual \fo \tu \Dd{5/2} -- \tu \Pp{1/2,3/2}
39: \lam7320\AA\ and \fo \tu \Dd{3/2} -- \tu \Pp{1/2,3/2} \lam7330\AA\
40: auroral doublet component lines have been measured directly in the
41: nightglow for the first time, from high resolution spectra obtained
42: with the HIRES spectrograph on the Keck I telescope at the W.M. Keck
43: Observatory. Specifically, we find 7320\AA\ and 7330\AA\ doublet
44: splittings of 1.077$\pm$0.003\AA\ and 1.080$\pm$0.003\AA\
45: respectively, with the former significantly larger than the often
46: quoted and utilized value of 0.8\AA\ from Moore (1949, Atomic Energy
47: Levels, Vol 1), and in line with NIST (1.07\AA) as well as more recent
48: astrophysical observations of the lines in planetary nebulae including
49: 1.07\AA\ from DeRobertis, Osterbrock, \& McKee (1985,ApJ,293,459) and
50: 1.09\AA\ from Barnett \& McKeith (1988,MNRAS,234,241). Our results
51: suggest, however, that adjustments of +0.124\AA\ and +0.131\AA\ should
52: be made to current NIST wavelengths for the blue and red components of
53: the 7320\AA\ doublet respectively, while the wavelengths of the
54: 7330\AA\ doublet components are little changed from current NIST
55: values. The observed intensity ratio of 7320\AA/7330\AA\ from these
56: measurements agrees with the theoretical value calculated under
57: conditions of thermally populated $^2$P$^o$ levels.
58: \end{abstract}
59:
60: \keywords{atomic data -- methods:data analysis -- techniques:spectroscopic}
61:
62: \section{Introduction} \label{intro}
63:
64: The auroral \fo \tu \Dd{5/2} -- \tu \Pp{1/2,3/2} \lam7320\AA\ and \fo
65: \tu \Dd{3/2} -- \tu \Pp{1/2,3/2} \lam7330\AA\ emission line doublets
66: (hereafter referred to as ``7320,7330'') are readily visible in the
67: spectra of low ionization astrophysical plasmas such as planetary
68: nebulae \citep{B55,B60,D85,S03} and \ion{H}{2} regions
69: \citep{B00,K03}, where the levels of the \oz(\Ppn) term are populated
70: by collisions between thermal electrons and \oz\ ions in the \Ssn\
71: ground term \citep{O89}. As such, their strengths can be used to
72: determine the electron density in the moderate density regime
73: ($10^5$--$10^{6}$ \cmc) of such objects \citep{BK88}, and are commonly
74: employed in combination with the nebular \fo \tu \Ssn -- \tu
75: \Dd{3/2,5/2} \lamm3726.032,3728.815\AA\ lines (hereafter
76: ``3726,3729'') to measure electron temperature \citep{SO57,K99}. The
77: doublets are also an important component of the airglow and aurora,
78: where \oz(\Ppn) is produced primarily by photoionization of atomic O
79: by solar EUV ($\lambda<666$\AA) and electron impact ionization (with
80: energy $<1$keV), respectively. Their altitude-dependent emission
81: strength is used to calculate the quenching rate of \oz(\Ppn) ions
82: through collisions with N$_2$ and atomic O \citep{R77,C93,ST03}, the
83: altitude density profile of \oz(\Ppn) and photoionization rate of
84: atomic O \citep{M91}, and the ion convection velocity and \oz\ gas
85: temperature in the ionospheric F-layer \citep{S82,C03}. The
86: intensities and line widths of the 7320,7330 doublet components have
87: also been employed to detect the existence and characteristics of a
88: permanent ``hot'' atomic oxygen corona existing above an altitude of
89: 550 km \citep{Y80}.
90:
91: We believe two recent studies demonstrate the necessity of determining
92: and publicizing accurate energies for the \oz\ \tu levels,
93: particularly the wavelengths and splitting of the 7320,7330 doublet
94: components. In their imaging Fabry-Perot interferometric auroral
95: data, \citet{C03} measured a 7320 doublet separation of 0.796\AA\ from
96: their fit to the separable unresolvable, combined profile of their
97: observed 7320 component lines, close to the 0.8\AA\ (1.5 \cmm) which
98: they quote from \citet{S82}. However, neither study could completely
99: resolve the separate lines because of their large intensity
100: difference. Moreover, recent \textit{ab initio} and astrophysically
101: observed calculations of the splitting argue for a significantly
102: larger separation (see Table~\ref{table1}). Indeed, the newer values
103: of the separation are larger than the 0.8\AA\ quoted from \citet{S82}
104: and measured by \citet{C03} by the order of the typical 0.1--0.8 \kms
105: ion-drift velocities targeted for measurement by both studies. A
106: larger 7320 separation also calls into question both the \citet{C03}
107: interpretation of their observed line profile as the actual 7320 pair,
108: and the accuracy of their F-layer temperature determination, made from
109: a line profile fit in which the separation itself was a free
110: parameter. Meanwhile, in their recent high resolution spectra of the
111: Orion Nebula, \citet{B00} found that the difference between their
112: observed 3726,3729 and tabulated (NIST) wavelengths, after correction
113: for nebular proper motion, is in good agreement with the same for the
114: 7330 components, but is systemically larger (by $\approx$6
115: \kms=0.15\AA) than the same for the 7320 components, even though all
116: lines should originate in the same portion of the nebula. This
117: discrepancy is a significant fraction of the total range of
118: ionization-energy dependent velocities seen in the Orion Nebula
119: outflow \citep{B00}.
120:
121: High resolution spectroscopy of the night sky has been shown to be a
122: valuable tool for direct measurements of the line rest wavelengths
123: \citep{SO00}. Presented here are what we believe to be the first
124: direct measurements in the airglow of the 7320,7330 doublet component
125: wavelengths, in which the doublets can be cleanly resolved into their
126: component lines. These measurements have been used to re-evaluate the
127: energy levels of the \Ddn\ and \Ppn\ terms, with the aim of improving
128: their accuracy for use in the types of studies involving precision
129: measurements of drift and nebular velocities, such as those mentioned
130: above.
131:
132: \section{Observations and Reductions} \label{obs}
133:
134: Sky spectra were drawn from among those to be archived in the proposed
135: National Virtual Aeronomical Observatory
136: (NVAO)\footnote{\url{http://www.nvao.org}} \citep{H02} showing the
137: most prominent 7320,7330 lines, which were clearly distinguishable
138: from and/or deemed clearly stronger than other neighboring telluric
139: emission. These sky spectra were all obtained with the HIRES echelle
140: spectrograph \citep{V94} on the Keck I telescope at the W.M. Keck
141: Observatory, and were originally associated with numerous deep-sky
142: astronomical targets. The nominal instrumental resolution was in all
143: cases 45,000 (R$\approx$6.7 \kms). Only spectra which had both
144: doublets in the same order (echelle orders 48 or 49) were chosen in
145: order to minimize any wavelength and flux calibration errors that are
146: often associated with echelle spectra. An observing journal is listed
147: in Table~\ref{table2}.
148:
149: All HIRES sky spectra were delivered to the NVAO archive, extracted
150: and reduced through wavelength calibration using the
151: MAKEE\footnote{\url{http://http://spider.ipac.caltech.edu/staff/tab/makee/}}
152: reduction routines developed by T.A.\ Barlow. Line wavelengths were
153: determined by manually fitting a Gaussian and a straight line,
154: representing the local continuum, to each line profile. Fluxes were
155: determined by summing the counts under the fitted function. As can be
156: seen in Figure~\ref{figure1}, the high resolution of the HIRES spectra
157: completely separates the doublet components and neighboring \oh
158: telluric lines, and profiles were well-represented by Gaussian fits.
159:
160: Besides the HIRES data, the tabulated 7320,7330 line wavelengths (in
161: air) from \citet{D85} (planetary nebula NGC 7027: R=100,000=3 \kms),
162: and 7320,7330 lines re-measured from the spectra of \citet{B00} (Orion
163: Nebula: R=30,000=10 \kms) and \citet{S03} (planetary nebula IC 418:
164: R=33,000=9 \kms), were used to supplement the HIRES data.
165: \citet{B60}, \citet{B00}, and \citet{S03} were also used for the
166: wavelengths of the 3726,3729 lines which are too weak to
167: be seen in the airglow.
168:
169: \section{Calculations} \label{calc}
170:
171: The tabulated wavelengths of the following \oh Meinel band (8-3) lines
172: from \citet{G98}: P$_{11}$(2.5) \lam7316.282, P$_{22}$(2.5)
173: \lam7329.148, P$_{11}$(3.5) \lam7340.885, and P$_{22}$(3.5)
174: \lam7358.667, if present in the same HIRES echelle order as 7320 and
175: 7330, were compared to measured wavelengths and used to normalize each
176: HIRES spectrum to the same wavelength system. A weighted average of
177: differences between the measured and tabulated wavelengths was
178: calculated for each HIRES data-set, with the above lines receiving
179: weights of (4,10,2,1) respectively, roughly proportional to their
180: relative theoretical intensities. These corrections, ranging in
181: magnitude from 0.1--2.7 \kms (0.002--0.066\AA\ at 7325\AA), were
182: applied to all measured wavelengths in each HIRES spectrum. The
183: statistical scatter between tabulated and observed wavelengths after
184: correction was 0.5 \kms ($\approx$0.01\AA\ at 7325\AA).
185:
186: A linear chi-squared fit was then performed to simultaneously
187: determine the best set of shared 3726,3729 and 7320,7330 observed air
188: wavelengths for all data-sets in which the lines were contained, from
189: which a Doppler shift to correct for any residual internal kinematic
190: and proper motion could be determined and applied to all nebular data
191: (except for Bowen 1960, which are averages from numerous planetary
192: nebulae). The minimized merit function took the form:
193: \begin{equation}
194: \label{eq1}
195: \chi^2 = \sum_i\sum_j w_j [{\lambda^i}_j(obs)-(\lambda^i(best)+\delta_j)]^2\,,
196: \end{equation}
197: where $i$ and $j$ are sums over all lines (3726,3729 and 7320,7330)
198: and data-sets respectively, $w_j$ are arbitrary weights given to each
199: data-set (see below), ${\lambda^i}_j(obs)$ are the wavelengths of each
200: line, if observed, in each data-set, $\lambda^i(best)$ the best shared
201: wavelength for each line, and $\delta_j$ the Doppler shifts (in \AA)
202: applicable to each nebular data-set. In total 6 line wavelengths, one
203: for each 3726,3729 and 7320,7330 line, and 3 Doppler shifts, one for
204: each nebular data-set except for \citet{B60}, were determined in this
205: manner. The weights $w_j$ (0.018 for Bowen, 0.16 for DeRobertis,
206: Osterbrock \& McKee, 1.00 for Baldwin et al.\ and Sharpee et al., and
207: 10.00 for all HIRES data) were arbitrarily chosen to best represent
208: our perceived accuracy of the measurements involved, with nebular data
209: sets generally receiving lower values due to unknown amounts of
210: residual internal kinematic and proper motions.
211:
212: Following conversion of all observed line wavelengths to vacuum
213: wave-numbers (cm$^{-1}$), linear chi-squared fits were made to
214: each level/term splitting value as combinations of those wave-numbers:
215: \begin{eqnarray}
216: \label{eq8}
217: \chi^2 & = & \sum_j w_j [\Delta(\Ddx-\Ssx) - \case{1}{2}(\nu_j(3726\textrm{\AA}) + \nu_j(3729\textrm{\AA}))]^2 \,, \\
218: \chi^2 & = & \sum_j \left\{w_j [\Delta(\Ppx-\Ddx) - \case{1}{2}(\nu_j(7319\textrm{\AA}) + \nu_j(7331\textrm{\AA}))]^2\right. \nonumber \\
219: & & + w_j [\Delta(\Ppx-\Ddx) - \case{1}{2}(\nu_j(7320\textrm{\AA}) + \nu_j(7331\textrm{\AA}))]^2\left.\right\} \,, \\
220: & & \nonumber \\
221: \chi^2 & = & \sum_j \left\{w_j [\Delta(\Ddx) - (\nu_j(3726\textrm{\AA}) - \nu_j(3729\textrm{\AA}))]^2 \right.\nonumber \\
222: & & + w_j [\Delta(\Ddx) - (\nu_j(7319\textrm{\AA}) - \nu_j(7330\textrm{\AA}))]^2 \nonumber \\
223: & & + w_j [\Delta(\Ddx) - (\nu_j(7320\textrm{\AA}) - \nu_j(7331\textrm{\AA}))]^2 \left.\right\}\,, \\
224: & & \nonumber \\
225: \chi^2 & = & \sum_j \left\{w_j [\Delta(\Ppx) - (\nu_j(7319\textrm{\AA}) - \nu_j(7320\textrm{\AA}))]^2 \right.\nonumber \\
226: & & + w_j [\Delta(\Ppx) - (\nu_j(7330\textrm{\AA}) - \nu_j(7331\textrm{\AA}))]^2 \left.\right\}\,,
227: \end{eqnarray}
228: where $w_j$ are the data-set weights (here equal to one except for the
229: case noted below), $\Delta(\Ddx-\Ssx)$ and $\Delta(\Ppx-\Ddx)$ are the
230: energy differences (in cm$^{-1}$) between the \Ddn\ and \Ssn\, and the
231: \Ppn\ and \Ddn\ terms, respectively, $\Delta(\Ddx)$ and $\Delta(\Ppx)$
232: are the fine structure energy level splitting, and $\nu_j$ are the
233: observed vacuum wave-numbers of all lines from the $j$ data-sets where
234: available and corrected for any Doppler shift determined from the
235: previous fit (nebular sources only). The Bowen(1960) data was not
236: included in the $\Delta(\Ppx)$ fitting and its data was given a weight
237: ($w_1=0.5$) for purposes of the $\Delta(\Ppx-\Ddx)$ fit. The energy
238: of each term was taken to lie exactly between the fine structure level
239: energies.
240:
241: The resultant energy levels, line vacuum wave-numbers, and line air
242: wavelengths are listed in Table~\ref{table3}. Listed errors are the
243: two-sigma formal uncertainties from the correlation matrices of the
244: various fits, propagated through the equations for line vacuum
245: wave-numbers as combinations of energy levels/term splittings:
246: \begin{eqnarray}
247: \nu(3726\textrm{\AA}) & = & \Delta(\Ddx-\Ssx) + \case{1}{2}\Delta(\Ddx)\,, \\
248: \nu(3729\textrm{\AA}) & = & \Delta(\Ddx-\Ssx) - \case{1}{2}\Delta(\Ddx)\,, \\
249: \nu(7319\textrm{\AA}) & = & \Delta(\Ppx-\Ddx) + \case{1}{2}\Delta(\Ppx) + \case{1}{2}\Delta(\Ddx) \,, \\
250: \nu(7320\textrm{\AA}) & = & \Delta(\Ppx-\Ddx) - \case{1}{2}\Delta(\Ppx) + \case{1}{2}\Delta(\Ddx) \,, \\
251: \nu(7330\textrm{\AA}) & = & \Delta(\Ppx-\Ddx) + \case{1}{2}\Delta(\Ppx) - \case{1}{2}\Delta(\Ddx) \,, \\
252: \nu(7331\textrm{\AA}) & = & \Delta(\Ppx-\Ddx) - \case{1}{2}\Delta(\Ppx) - \case{1}{2}\Delta(\Ddx) \,,
253: \end{eqnarray}
254: where all symbols have the same meanings as above. The trans-auroral
255: lines at \lamm2470.2,2470.3\AA, while not observed, were also
256: calculated from the four fitted energy level and term separations.
257:
258: \section{Discussions}
259:
260: The doublet wavelength separation values determined here compare very
261: favorably with what we consider the most accurately determined
262: astrophysical values from \citet{BK88} (1.09$\pm$0.02\AA\ for
263: $\Delta(\Ppx)$ and 10.70$\pm$0.03\AA\ for $\Delta(\Ddx)$, specific
264: line wavelengths were not reported), while their formal uncertainties
265: exceed ours. The \Ddn\ and \Ppn\ term splitting values fall in the
266: middle of the range of those determined from more recent
267: experimentally inferred and astrophysically observed values (see
268: Table~\ref{table1}), although ours should be inherently more accurate
269: given our direct measurement of the 7320,7330 line wavelengths, as
270: opposed to Ritz method determinations from more easily observable UV
271: lines in laboratory experiments. The measurements made here were not
272: complicated by blending with other telluric features or significant
273: thermal broadening of the lines, and thus were not reliant upon
274: potentially error-prone deblending techniques (e.g.\ DeRobertis et
275: al.\ 1985). Our measurements of the \Ppn\ energy splitting, along
276: with other recent observationally determined values
277: (Table~\ref{table1}) are consistently and significantly smaller than
278: those determined from the recent theoretical calculations of
279: \citet{Z87} (3.0 \cmm\ for their ``A'' configuration, 2.7 \cmm\ for
280: their ``B'' configuration) and \citet{T02} (2.58 \cmm\ and 2.60 \cmm\
281: relating to \textit{ab initio} and energy corrected values
282: respectively), although the \Ddn\ energy splitting values show much
283: better agreement. Finally, the application of the new individual
284: 7320,7330 line wavelengths to the results of \citet{B00} significantly
285: reduces the scatter in velocity residuals with their observed
286: wavelengths for these lines, from 2.4--8.0 \kms to 2.2--2.7 \kms.
287: Adoption of the new 7320,7330 wavelengths would lead to an upwards
288: adjustment in the often-utilized NIST values for the 7320 lines
289: (7318.92\AA\ and 7319.99\AA) by 0.124 and 0.131\AA\ respectively, and
290: a one digit increase in the number of significant digits for which the
291: wavelengths are known.
292:
293: It appears likely that the splitting value for the \Ppn\ term of
294: 0.796\AA\ determined by \citet{C03} using their imaging Fabry-Perot
295: interferometer was not that of the 7320 components. Their
296: instrument's free spectral range of 0.107\AA\ is almost exactly 1/10
297: that of the measured 1.077\AA\ \Ppn\ splitting value determined here,
298: resulting in an unfortunate blending of the lines in their
299: observations. It is more likely that the splitting measured was
300: actually between the stronger \lam7320.121 component and a neighboring
301: strong \oh line from a different order. We speculate that the 0.8\AA\
302: splitting value quoted from \citet{S82} may date back to \citet{M49}
303: (see Table~\ref{table1}) because their resolution was also
304: insufficient to separate 7320. As noted here, \citet{M49} has been
305: superseded by more recent determinations. The large differences
306: between the wavelength and term splitting values determined here, and
307: other older and/or widely-utilized values, warrant reconsideration of
308: conclusions reached employing those standards, such as regarding
309: F-layer physical conditions \citep{S82,C03} and nebular velocity
310: structure \citep{B00}.
311:
312: Intensities from individual 7320,7330 have been measured in the four
313: HIRES spectra showing the strongest 7320,7330 lines and least amount
314: of atomic Fraunhofer absorption from scattered moon or zodiacal light.
315: The average line intensities normalized to the total 7320,7330 intensity
316: and standard deviations of the averages are listed in
317: Table~\ref{table4}. As the bulk of the atmospheric 7320,7330 emission
318: arises from altitudes between 250--400 km \citep{C03}, where quenching
319: of the \Ppn\ term via collisions with atomic O and N$_2$ (of densities
320: approximating $1\times10^9$ \cmc\ and rate coefficients of
321: $0.5\times10^{-10}$ \cmc\ s$^{-1}$ and $1.8\times10^{-10}$
322: respectively; Stephan et al.\ 2003) is more rapid than the term's
323: radiative lifetime ($\approx$5s), the \oz\ level populations are
324: thermalized. Thus, relative intensity measurements of 7320,7330 lines
325: are related only to spontaneous transition coefficients and
326: statistical weights. This provides a check on the accuracy of \oz\
327: 7320,7330 Einstein A values, which are employed in a variety of
328: astrophysical diagnostics, such as for electron density
329: \citep{D85,BK88} and temperature \citep{K99}, reddening \citep{D85},
330: and \oz\ nebular abundances \citep{L00}.
331:
332: Unfortunately, in most of the HIRES spectra, the 7320,7330 lines were
333: positioned close to the blue edge of the echelle orders in which they
334: appeared as a consequence of the original observers' chosen grating
335: tilts. Coupled with these lines' intrinsic weakness and confusion of
336: the continuum level near Fraunhofer absorption features, even in those
337: spectra with the smallest degree of contamination, the HIRES spectra
338: are not optimal for intensity measurements of these lines, as
339: reflected in the large measurement uncertainties. Still, the
340: theoretical high-density limit line intensities constructed from major
341: astrophysical sources of transition coefficients (those listed in
342: Table~\ref{table4}), are all encompassed within those standard
343: deviations, and with the possible exception of the \lam7319.044 line,
344: do compare well with the observed values. \citet{S79} has previously
345: reported a 7320/7330 auroral intensity ratio measurement of
346: 1.55$\pm0.05$, which exceeded their own calculation of the ratio's
347: theoretical limiting value at high density of 1.31 using \citet{SO57}
348: transition coefficients under maximum quenching conditions (altitude
349: $<$200 km). A re-calculation of the theoretical high-density limit
350: ratio using eqs.(1-6) of \citet{S79} with more recent coefficients
351: (sources from Table~\ref{table4}) yields values between 1.28 and 1.30,
352: virtually unchanged from the earlier estimation. However, our
353: measured 7320/7330 ratio of 1.3 is consistent with this value and with
354: the high-density limit theoretical values constructed from
355: Table~\ref{table4}.
356:
357: \citet{D85} and \citet{BK88} have proposed the use of the
358: \lam7320.121/\lam7319.044 intensity ratio as an electron density
359: diagnostic in the $\approx10^5$--$10^9$\cmc\ density regime. The
360: diagnostic takes advantage of the increase in the theoretical
361: intensity ratio, from a value of about 3.0 below $10^5$ \cmc, to a
362: high-density value of 3.8 (as calculated from Table~\ref{table4}) at
363: $10^9$ \cmc\ and above, as electron collisional excitation from the
364: \Ddn\ term and de-excitation increasingly contribute to the \Ppn\
365: levels' populations. The average value of the
366: \lam7320.121/\lam7319.044 intensity ratio measured from the HIRES
367: spectra is 3.1$\pm$0.9, while values of 3.3 and 3.5 are measured from
368: the blended profile of the 7320 doublet in two lower resolution Keck
369: II/ESI spectra in our possession, all of which are lower than the
370: composite high-density ratio value of 3.8. A lower value of the
371: high-density limit intensity ratio, as is suggested from the HIRES and
372: ESI spectra measurements, would reduce the utility of the diagnostic.
373:
374: A stronger than predicted \lam7319.044 intensity relative to the total
375: multiplet intensity, as is suggested by the HIRES data, could be
376: behind the observed smaller than predicted value of the
377: \lam7320.121/\lam7319.044 intensity ratio. However, the line's large
378: intensity measurement uncertainty in the HIRES spectra, echoed into
379: the large uncertainty in the measured intensity ratio, suggests that
380: the difference between the observed and theoretical high-density
381: values of the ratio is probably not statistically significant. The
382: ESI spectra ratio values are closer to composite theoretical
383: high-density ratio value of 3.8, and systemic errors introduced in the
384: deblending of a 7320 profile composed of unequal intensity components
385: could account for much of the difference. Significant contamination
386: of the 7320,7330 line intensities from the N$_2$ 1P(5-3) band is ruled
387: out since the band head (at 7387.2\AA) and strongest
388: features, as predicted from a DIATOM \citep{H94} simulation of the
389: band's intensity at a typical mesospheric temperature of 200K, are
390: absent from any of the utilized spectra.
391:
392: Curiously, the \lam7320.121/\lam7319.044 intensity ratios measured
393: from the IC 418 and Orion Nebula spectra of 2.7 and 2.9 respectively,
394: as calculated from Table~\ref{table4}, are slightly below the 3.0
395: low-density value, even though both objects have electron densities of
396: $\approx10^4$ \cmc. This may indicate that additional processes other
397: than electron collisional excitation may be populating the \Ppn\
398: levels not according to their statistical weights. For example,
399: electron recombination of O$^{+2}$ with subsequent cascade
400: \citep{L00}, or photoionization of neutral O \citep{F81}, are two
401: processes that may also affect the nebular 7320,7330 line intensities.
402:
403: In summary, the HIRES data generally supports the accuracy of
404: widely-used astrophysical spontaneous transition coefficients, with
405: some lingering problems which the present data-sets' large scatter and
406: small numbers are unable to resolve.
407:
408: \section{Conclusions}
409:
410: We have used high resolution HIRES sky spectra to accurately determine
411: the wavelengths and relative intensities of the 7320,7330 doublet
412: lines, measured directly for the first time from completely resolved
413: profiles in the airglow. These measurements are extremely accurate
414: given the nature and advantages of using sky spectra for such
415: measurements as opposed to deducing the wavelengths from energy levels
416: established by more easily observable transitions, or from nebular
417: lines wavelengths corrected for proper and internal kinematics. The
418: new wavelengths determined here differ by as much as 0.131\AA\ from
419: the widely-utilized NIST values, and the aeronomically-important 7320
420: doublet splitting value measured here differs by 0.277\AA\ from the
421: ``accepted'' value of 0.8\AA\ still employed as recently as
422: \citet{C03}. The magnitudes of these differences suggest that the
423: interpretation of observations made by comparing against earlier
424: standards may need to be revisited. The observed fluxes for
425: individual lines agree within their sample scatter, with theoretical
426: values calculated under the high density limit which should prevail in
427: the region of the atmosphere in which they are formed.
428:
429: \acknowledgements
430:
431: This paper is based on observations made at the W.M.\ Keck
432: Observatory, which is operated jointly by the California Institute of
433: Technology and the University of California. We thank Professor
434: Jack Baldwin for allowing us to use his spectra. This work was
435: supported by grants from the NSF Aeronomy Program and the NASA Office
436: of Space Science.
437:
438: \begin{thebibliography}{}
439:
440: \bibitem[Baldwin et al.(2000)]{B00} Baldwin, J.A., Verner, E.M.,
441: Verner, D.A., Ferland, G.J., Martin, P.G., Korista, K.T., \& Rubin,
442: R.H. 2000, \apjs, 129, 229
443:
444: \bibitem[Barnett \& McKeith (1988)]{BK88} Barnett, E.W. \& McKeith,
445: C.D. 1988, \mnras, 2234, 241
446:
447: \bibitem[Bowen(1955)]{B55} Bowen, I.S. 1955, \apj, 132, 1
448:
449: \bibitem[Bowen(1960)]{B60} Bowen, I.S. 1960, \apj, 132, 1B
450:
451: \bibitem[Chang et al.(1993)]{C93} Chang, T., Torr, D.G., Richards,
452: P.G., \& Solomon, S.C. 1993, \jgr, 98, 15589
453:
454: \bibitem[Cierpka et al.(2003)]{C03} Cierpka, K., Kosch, M.J., Holma, H.,
455: Kavanagh, A.J., \& Hagfors, T. 2003, \grl, 30,
456: 1293
457:
458: \bibitem[De Robertis, Osterbrock, \& McKee(1985)]{D85} De Robertis,
459: M.M., Osterbrock, D.E., \& McKee, C.F. 1985, \apj, 293, 459
460:
461: \bibitem[Eriksson(1987)]{E87} Eriksson, K.B.S 1987, \textit{J. Optical
462: Soc. of Amer. B}, 4, 1369
463:
464: \bibitem[Fawcett(1975)]{F75} Fawcett, B.C. 1975, \textit{Atom. Data
465: Nucl. Data Tables}, 16, 135
466:
467: \bibitem[Ferland \& Truran(1981)]{F81} Ferland, G.J. \& Truran,
468: J.W. 1981, \apj, 244, 1022
469:
470: \bibitem[Godefroid \& Froese Fischer(1984)]{GFF84} Godefroid, M. \&
471: Froese Fischer C. 1984, \textit{J. Phys. B}, 17, 681
472:
473: \bibitem[Goldman et al.(1998)]{G98} Goldman, A., Schoenfeld, W.G.,
474: Goorvitch, D., Chackerian Jr., C., Dothe, H., M\'{e}len, F., Abrams,
475: M.C., \& Selby, J.E.A. 1998, \textit{J. Quantitative. Spect. \&
476: Radiative Transfer}, 59, 453
477:
478: \bibitem[Huestis(1994)]{H94} Huestis, D.L. 1994, DIATOM Spectral
479: Simulation Computer Program, Version 7.0 (SRI International: Menlo
480: Park) \verb+http://www-mpl.sri.com/software/diatom.html+
481:
482: \bibitem[Huestis, Cosby, \& Slanger (2002)]{H02} Huestis, D.L., Cosby,
483: P.C., \& Slanger, T.G. 2002, \baas, 34, 743
484:
485: \bibitem[Keenan et al.(1999)]{K99} Keenan, F.P., Aller, L.H., Bell,
486: K.L., Crawford, F.L., Feibelman, W.A., Hyung, S., McKenna, F.C., \&
487: McLaughlin, B.M. 1999, \mnras, 304, 27
488:
489: \bibitem[Kennicutt, Bresolin, \& Garnett(2003)]{K03} Kennicutt, R.C.,
490: Jr., Bresolin, F., \& Garnett, D.R. 2003, \apj, 591, 801
491:
492: \bibitem[Liu et al.(2000)]{L00} Liu, X.W., Storey, P.J., Barlow, M.J.,
493: Danziger, I.J., Cohen, M., \& Bryce, M. 2000, \mnras, 312, 585
494:
495: \bibitem[Martin, Kaufman, \& Musgrove(1993)]{M93} Martin, W.C.,
496: Kaufman, V., \& Musgrove, A. 1993, \textit{J. Phys. Chem. Ref. Data},
497: 22, 1179
498:
499: \bibitem[McDade et al.(1991)]{M91} McDade, I.C., Sharp, W.E.,
500: Richards, P.G., \& Torr, D.G. 1991, \jgr, 96, 259
501:
502: \bibitem[Moore(1949)]{M49} Moore, C.E. 1949, \textit{Atomic Energy
503: Levels}, Vol. 1 (National Bureau of Standards Circular 467), 47
504:
505: \bibitem[Osterbrock(1989)]{O89} Osterbrock, D.E. 1989,
506: \textit{Astrophysics of Gaseous Nebulae and Active Galactic Nuclei}
507: (Mill Valley: University Science Books)
508:
509: \bibitem[Rusch, Torr, \& Hays(1977)]{R77} Rusch, D.W., Torr, D.G., \& Hays,
510: P.B. 1977, \jgr, 82, 719
511:
512: \bibitem[Seaton \& Osterbrock(1957)]{SO57} Seaton, M.J. \& Osterbrock,
513: D.E. 1957, \apj, 125, 66
514:
515: \bibitem[Slanger et al.(2000)]{SO00} Slanger, T.G., Huestis, D.L.,
516: Cosby, P.C., \& Osterbrock, D.E. 2000, \jcp, 113, 8514
517:
518: \bibitem[Sharpee et al.(2003)]{S03} Sharpee, B., Williams, R.,
519: Baldwin, J.A., \& van Hoof, P.A.M. 2003, \apjs, 149, 157
520:
521: \bibitem[Sivjee, Romick, \& Rees(1979)]{S79} Sivjee, G.G., Romick,
522: G.J., \& Rees, M.H. 1979, \apj, 229, 432
523:
524: \bibitem[Smith et al.(1982)]{S82} Smith, R.W., Sivjee, G.G., Stewart,
525: R.D., McCormac, F.G., \& Deehr, C.S. 1982, \jgr, 87. 4455
526:
527: \bibitem[Stephan et al.(2003)]{ST03} Stephan, A.W., Meier, R.R.,
528: Dymond, K.F., Budzien, S.A., \& McCoy, R.P. 2003, \jgr, 108, 1034
529:
530: \bibitem[Tachiev \& Froese Fischer(2002)]{T02} Tachiev, G.I. \& Froese
531: Fischer, C. 2002, \aap, 385, 716
532:
533: \bibitem[Vogt(1994)]{V94} Vogt, S.S. et al., 1994, Proc. SPIE, 2198, 362
534:
535: \bibitem[Wiese, Fuhr, \& Deters(1996)]{WFD96} Wiese, W.L., Fuhr, J.R.,
536: \& Deters, T.M. 1996, \textit{J. Phys. Chem. Ref. Data}, 7
537:
538: \bibitem[Yee, Meriwether, \& Hays(1980)]{Y80} Yee, J.H., Meriwether,
539: Jr., J.W., \& Hays, P.B. 1980, \jgr, 85, 3396
540:
541: \bibitem[Zeippen(1987)]{Z87} Zeippen, C.J. 1987, \aap, 173, 410
542:
543: \end{thebibliography}
544:
545: \clearpage
546: \begin{figure}
547: \plotone{f1.eps}
548: \caption{A portion of a typical HIRES echelle spectrum showing the \fo
549: \tu \Ddn--\Ppn\ 7320,7330 doublet lines and \oh P$_{22}$(2.5) (8-3)
550: \lam7329.148 line. The dashed line is a segment of the Orion Nebula
551: spectrum of \citet{B00}, scaled and shifted. \label{figure1}}
552: \end{figure}
553: \clearpage
554:
555: \input{tb1}
556:
557: \clearpage
558:
559: \input{tb2}
560:
561: \clearpage
562:
563: \input{tb3}
564:
565: \clearpage
566:
567: \input{tb4}
568:
569: \end{document}
570:
571: