JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 110, D23302, doi:10.1029/2005JD006078, 2005
Variability of the mesospheric nightglow sodium D2/ D1 ratio
T. G. Slanger,1 P. C. Cosby,1 D. L. Huestis,1 A. Saiz-Lopez,2 B. J. Murray,2,3
D. A. O’Sullivan,2 J. M. C. Plane,2 C. Allende Prieto,4 F. J. Martin-Torres,5
and P. Jenniskens6
Received 13 April 2005; revised 26 July 2005; accepted 31 August 2005; published 1 December 2005.
[1] Measurements of the intensity ratio of the 589.0/589.6 nm sodium doublet in the
terrestrial nightglow over an 8-year period, involving >300 separate determinations, have
established that it is variable, the value RD = I(D2)/I(D1) lying between 1.2 and 1.8. Sky
spectra from the Keck I telescope with the High-Resolution Échelle Spectrometer
(HIRES) échelle spectrograph and the Keck II telescope with the Échellette Spectrograph
and Imager (ESI) échelle spectrograph were used in this analysis. The result contrasts with
the accepted view, from earlier measurements at midlatitude, that the ratio is 2.0, as
expected on statistical grounds. The lack of dependence of the ratio on viewing elevation
angle, and hence Na slant column, allows self-absorption to be ruled out as a cause of
the variability. The data suggest a semiannual oscillation in the ratio, maximum at the
equinoxes and minimum at the solstices. Airborne measurements over the North
Atlantic (40–50N) in 2002 show an even larger range in the nightglow ratio and no
correlation with the upper mesospheric temperature determined from the OH 6–2 bands.
A laboratory study confirms that the ratio does not depend on temperature; however, it is
shown to be sensitive to the [O]/[O2] ratio. It is therefore postulated that the variable
ratio arises from a competition between O reacting with NaO(A3S+), produced from the
reaction of Na with O3, to yield D-line emission with a D2/D1 ratio greater than about
2.0, and quenching by O2 to produce NaO(X2), possibly with vibrational excitation,
which then reacts with O to produce emission with a ratio of less than 1.3.
Citation: Slanger, T. G., P. C. Cosby, D. L. Huestis, A. Saiz-Lopez, B. J. Murray, D. A. O’Sullivan, J. M. C. Plane, C. Allende Prieto,
F. J. Martin-Torres, and P. Jenniskens (2005), Variability of the mesospheric nightglow sodium D2/D1 ratio, J. Geophys. Res., 110,
D23302, doi:10.1029/2005JD006078.
ðR2bÞ
1. Introduction
[2] Radiation in the nightglow spectrum at 589 nm was
first reported in 1929 [Slipher, 1929], and 10 years later it
was established that this radiation was due to the transition
Na(32PJ 32S1/2) and that the source was located within the
Earth’s atmosphere [Bernard, 1938]. Chapman [1939] then
postulated the following sequence of reactions to account
for the Na D-line emission from Na(2PJ) in the nightglow:
ðR1Þ
ðR2aÞ
Na þ O3 ! NaO þ O2
NaO þ O ! Na 2 PJ þ O2
Branching ratio f
1
Molecular Physics Laboratory, SRI International, Menlo Park,
California, USA.
2
School of Environmental Sciences, University of East Anglia,
Norwich, UK.
3
Now at Department of Chemistry, University of British Columbia,
Vancouver, British Columbia, Canada.
4
McDonald Observatory and Department of Astronomy, University of
Texas at Austin, Austin, Texas, USA.
5
AS&M, Inc., NASA/Langley Research Center, Hampton, Virginia,
USA.
6
NASA Ames Research Center, Mountain View, California, USA.
Copyright 2005 by the American Geophysical Union.
0148-0227/05/2005JD006078$09.00
ðR3Þ
NaO þ O ! Na 2 S þ O2
Branching ratio 1 f
Na 2 P ! Na 2 S þ hu
where the source of sodium atoms is ablation of the
approximately 20 tons of interplanetary dust that enters the
atmosphere each day [Plane, 2003]. The Na nightglow layer
typically peaks around 90 km, and has a full width at halfmaximum of 5 km [Clemesha et al., 1993].
[3] Laboratory measurements have shown that reactions
(R1), (R2a), and (R2b) are fast enough to generate the
observed D-line emission intensity of 30– 150 R. Reaction
(R1) is rate determining because [O3] [O] above 84 km
[Plane, 2003]. An estimate of f, the branching ratio for
production of Na(2PJ) in reaction (R2a), has been derived
from a field experiment in northern Brazil, where a rocket
measured the D-line emission profile while a ground-based
lidar simultaneously observed the Na atom concentration
[Clemesha, 1995]. Analysis of the data using a model of the
Na layer showed that f lay in the range 0.05– 0.2, with a best
estimate of 0.1 [Clemesha, 1995]. However, a more recent
combined rocket and lidar experiment at Arecibo (Puerto
Rico) produced a lower estimate, f = 0.02– 0.04, although
D23302
1 of 8
D23302
SLANGER ET AL.: SODIUM NIGHTGLOW
the mesopause on this night was very cold (165 K) and the
Na atom and nightglow layer peaked around 95 km [Hecht
et al., 2000].
[4] These estimates of f are considerably larger than an
early laboratory measurement of less than 0.01 [Plane and
Husain, 1986]. This apparent disagreement seems to have
been resolved through a series of elegant laboratory experiments. First, it was shown that reaction (R1) produces NaO
almost entirely in the low-lying NaO(A2S+) excited electronic state, rather than the 2 ground state [Shi et al., 1993;
Wright et al., 1993]. The NaO(A) state has a long radiative
lifetime and is not quenched efficiently, so that in the
mesosphere reactions (R2a) and (R2b) should involve
mostly NaO(A) rather than NaO(X) [Joo et al., 1999].
Second, a recent flow tube experiment showed that f for
NaO(A) + O is 0.14 ± 0.04 [Griffin et al., 2001]. This
measurement thus reconciled the field observations with the
underlying chemical physics, and showed that the earlier
laboratory estimate of f was probably so small because it
involved the reaction between ground state NaO(X) + O.
[5] The ratio between the intensity of the D2 line at
589.0 nm and that of the D1 line at 589.6 nm, hereafter
termed RD, should be 2.0 if Na(2P3/2) and Na(2P1/2) are
produced according to their spin orbit statistical weights
(note that at the pressures below 105 bar in the upper
atmosphere, the frequency of collisions with surrounding
air molecules is too low to equilibrate the Na(2P) spin orbit
states before emission occurs, since the Na(2P) lifetime
is only 16 ns). Chamberlain [1995] refers to various early
investigations in 1935 –1960, in which RD was estimated
to be 2. In a paper by Chamberlain and Negaard [1956],
three scattering models were used to conclude that in the
nightglow RD should be 2.0 or higher.
[6] Rees et al. [1975] carried out a study in 1975, to
determine whether auroral excitation could affect the D-line
ratio. They concluded that the presence of aurora had no
discernible effect on the twilight glow, but they also
presented data on RD for solar depression angles as large
as 18. This showed that the enhanced intensity that is
characteristic of the twilight glow ceases when the solar
depression angle (SDA) exceeds 11, and that between 11
and 18 the ratio is very variable. Rees et al. did not
comment on this fact, perhaps because the signal-to-noise
ratio of their measurements was low. A later study by Sipler
and Biondi [1978], made with a Fabry-Perot instrument at
Laurel Hill, reported that RD was 1.98 ± 0.1. Following this
paper and up until the present study, RD has been assumed
to have a constant value of 2.0.
[7] The emission intensity of the sodium nightglow is
seasonal. Although earlier publications reported that the
emission is much more intense in winter than in summer
[Chamberlain, 1995], more recent studies show a strong
semiannual oscillation (SAO), with peaks at the equinoxes
and troughs at the solstices. Observations made at the Cariri
site in Brazil in 1998 – 1999 show that the sodium SAO is
very strong, with a large high-to-low-amplitude ratio,
stronger than that of any of the other species measured
(H. Takahashi, private communication, 2004). The data
from Brazil in 1976 – 1982 [Kirchhoff, 1986] are in excellent
agreement with the 1998– 1999 Cariri data, both showing
April and October peaks of 100 – 150 R, with solstice
minima of 20– 40 R.
D23302
[8] The SAO is also characteristic of other nightglow
systems in the upper mesosphere/lower thermosphere. Buriti
et al. [2001] have shown that it can be observed in the
557.7 nm oxygen green line, in the O2(b-X) 0 – 1 band, in
the OH 6 – 2 band, and also in the OH 6 – 2 rotational
temperature. The phase is the same in each case, and
observations on the SAO in sodium have been made from
a number of sites [Kirchhoff, 1986; Kirchhoff et al., 1979,
1981; Kirchhoff and Takahashi, 1985; McNeil et al., 1995].
The explanation centers on the role of O atoms in these
disparate systems, and thus it is perhaps not surprising that
all these intensities behave in a similar manner. However, it
is not obvious that RD should exhibit the same behavior,
which we will show to be the case.
[9] The presence of the sodium layer has important
applications. Sodium lidar measurements sample the atomic
sodium density in the nightglow region by resonance
fluorescence, and return information on the absolute concentration, layer thickness, temperature and winds. In
recent years, it has been demonstrated that the flash of
light associated with laser excitation of mesospheric sodium
atoms can be used as a guide star for astronomical
purposes [Happer et al., 1994], leading to a cancellation
of atmospheric turbulence and greatly improved viewing of
stellar objects. Occasionally, incoming meteors leave
persistent trains, whose luminosity is fueled by reactions
involving O3 and atomic O, efficiently catalyzed by metals
from the freshly ablated meteoroid [Jenniskens et al.,
2000; Kruschwitz et al., 2001] The most intense emission
arises from excited FeO, with possibly a small contribution
from the Na D-lines. Another important subject is the
presence of ‘‘sudden’’ sodium layers, where very narrow
layers of sodium atoms are seen sporadically, such behavior
also having been noted for iron emission [Bills and
Gardner, 1990; Clemesha, 1995].
2. Atmospheric Observations of the Na D-line
Ratio
2.1. Data Collection
[10] The work to be described has made use of sky
spectra generated at the Keck Observatory on Mauna Kea,
from the two 10-m telescopes. Keck I is coupled to the
High-Resolution Échelle Spectrometer (HIRES) échelle
spectrograph, which has been operated at a resolution of
40,000; Keck II is coupled to the Échellette Spectrograph
and Imager (ESI) échelle spectrograph, with a resolution of
about 7000. In either case, the sodium lines are well
resolved [Slanger and Copeland, 2003; Slanger et al.,
2000a, 2000b; Slanger and Osterbrock, 2000].
[11] The typical viewing protocol of the astronomers is to
collect light from their celestial object for 40– 70 min, the
telescope tracking the object so that the star field is
stationary. Sometimes there are two or three consecutive
data collection periods for the same object, and during a
typical night 3 –4 objects are viewed. Thus the telescope
wanders across the sky during the night, but at all times
nightglow data are collected, then archived after the astronomers have subtracted the so-called sky spectrum from the
spectrum of their object.
[12] Sky spectra have been obtained from observations
made at the Keck Observatory back to 1993. The early data
2 of 8
D23302
SLANGER ET AL.: SODIUM NIGHTGLOW
D23302
relevant unless they signify signal saturation, which appears
not to be the case. The inset to Figure 1 shows the resolved
sodium line pair as measured by the 0.5-m Czerny-Turner
spectrometer used on NASA’s 2002 Leonid MAC mission,
to be described [Jenniskens, 2002].
[15] Figure 2 shows the line ratios obtained from the
entirety of data of both observing systems, regardless of
the year; there are more than 300 data points. One sees
that the ratios lie between 1.2 and 1.8, with the largest
fraction at 1.5– 1.7. The precision of each measurement is
high, as can be seen from the signal-to-noise ratio in
Figure 1, so that the variations seen in a single vertical
group of points are real changes: The spread is not
statistical. It appears that there are minima and maxima
in the overall distribution: The ratios are higher near the
equinoxes, and there is an indication of a summer dip.
[16] For any given year there is not enough sampling
throughout the year to clarify the entire picture, but in
Figure 3 we show HIRES data for 1994 and 1995 alone.
In this case, it is evident that there is a substantial difference
between the ratios found in June and in October, and the
October data for the two years are very consistent. Also
shown in Figure 3 are Brazilian data (H. Takahashi, private
communication, 2004) of the semiannual oscillation of the
sodium D1 + D2 intensity, and it may be seen that the phase
is the same, although it is not yet evident how the line ratio
and the line intensity are related. These data are shown only
Figure 1. UVES/ESI sodium spectra, 588.5 – 590.0 nm;
weak lines belong to the OH Meinel 8 – 2 band. Solid line,
HIRES; dashed line, ESI; inset, data from NASA’s 2002
Leonid MAC mission.
are all from Keck I/HIRES, while since 2000 spectra have
become available from Keck II/ESI. We have not calibrated
the HIRES data with respect to intensity, but the 2000 –
2001 ESI data have been absolutely calibrated against
standard stars. However, absolute calibration is not necessary in obtaining the line ratios, as the lines are separated by
only 0.6 nm. It is important to realize that the data were
produced with two quite different physical systems, so that
experimental artifacts that may arise in one system are
unlikely to affect both sets of measurements.
[13] There are additional data that we have evaluated
from McDonald Observatory and from the UVES échelle
spectrograph at the Very Large Telescope (VLT) in Chile
[Hanuschik, 2003]. It should be pointed out that whereas the
preponderance of data are from an equatorial site, Mauna
Kea, it is not evident that the sparser data from higher
latitudes (McDonald Observatory, Alaska, the North Atlantic) show systematically different results.
2.2. Ratio Measurements, Seasonal
[14] In Figure 1 are shown typical sodium D2(589.0 nm)
and D1(589.6 nm) line spectra from two telescopes, the VLT
8-m telescope in Chile, using the UVES échelle system, and
the Keck II 10-m telescope on Mauna Kea, using the ESI
échelle system. Here we see the difference both in resolution and in line shapes. As the line ratios are obtained by
integrating the total signal, the line shapes are not directly
Figure 2. Sodium D2/D1 ratios, all data. Open circles,
HIRES, 1993– 1997; solid circles, HIRES, 1999; open
triangles, ESI, 2000; open squares, ESI, 2001; open
diamonds, ESI, 2001/2002.
3 of 8
D23302
SLANGER ET AL.: SODIUM NIGHTGLOW
D23302
great as on 4 March 2000, suggesting a more dynamic
atmosphere.
Figure 3. Seasonal dependence of sodium D2/D1 intensity
ratios from HIRES 1993 –1994 data; 1993, open circles;
1994, triangles. Additional data from H. Takahashi (solid
points) showing the SAO for the total sodium intensity
measured at Cariri, Brazil (7.4S, 36.5W).
to display the phase; the equinox/solstice intensity ratio is
on the order of 5 – 10.
[17] In addition to the Keck data, Figure 2 shows the
range of ratios measured in an auroral zone [Rees et al.,
1975]. The limits shown are for 90% of the 42 points
displayed in this paper. We see that in spite of the fact that
the data were taken in Alaska, the range of values are in
agreement with what was measured for the same season in
Hawaii. The range of values obtained from McDonald
Observatory in 2004 exhibit a particularly low range of
1.15 – 1.4, and were taken near summer solstice.
2.3. Ratio Measurements, Hourly and Nightly
[18] Table 1 shows the nightly averages and the 2-s
standard deviations of the line ratios for March and October
2000, along with the number of spectra recorded each night,
and the total observed range of the ratios. The estimated
error in measuring peak areas is on the order of 1%.
Sometimes the ratios are extremely stable, e.g., for the
nights of 4 March 2000 and 23 October 2000, the 2-s SD
is ±2.5% of the measured ratios. The data for those nights
were collected over a 9 – 10 hour period, with the elevation angle variation being 47– 74 degrees in March and
50– 71 degrees in October, with an azimuth variation of
60 – 284 degrees in March and 120 – 246 degrees in
October. Thus a large portion of the sky was sampled
in a night with little ratio variation. On the other hand,
on 3 March 2000 the standard deviation was 3 times as
2.4. Airborne Measurements of RD
[19] Figure 2 includes results obtained during the Leonids
campaign on 17 November 2002, when the NASA DC8
research aircraft crossed the North Atlantic (40– 50N)
while making spectrographic observations of the nightglow.
The instrument consisted of a 5 field-of-view telescope
pointed at an elevation angle of 23, coupled via a fiberoptic bundle to a 0.5 m Czerny-Turner spectrometer
(1200 groove mm1 grating) and cooled CCD detector
(1024 256 pixels, maintained at 70C). This combination provided a resolution of 0.2 nm, sufficient to resolve
the D-lines (Figure 1). A total of 179 measurements of RD
were made on this night, yielding an average RD of 1.79
and a range from 1.55 to 2.01. Two other flights were made
with the same spectroscopic system, on 15 and 18 November. These flights produced similar average RD values, 1.81
and 1.78, but greater ranges, 1.43 – 2.41 and 1.33– 2.41,
from 156 and 52 measurements on the two nights, respectively. These results are compatible with the Keck data for
the same season.
[20] Figure 4 shows that RD was varied over horizontal
distances of 50– 100 km (corresponding to flight times of
3– 6 min), which are typical of the horizontal wavelengths
of atmospheric gravity waves. Note that there is possibly
some correlation between RD and the Na D-line intensity in
the early segment of the flight, but this was not consistently
observed. The airborne spectrometer was also used to
measure 20 rotational lines in the Meinel OH 6 –2 band
between 826 and 862 nm, from which the temperature could
be fitted to within ±7 K. Figure 5 shows that there is no
significant correlation between R D and temperature,
although it should be noted that the Meinel emission is
centered in a layer about 3 km below the Na nightglow layer
[Plane, 2003].
3. Laboratory Study of the Na Nightglow
Mechanism
[21] In order to interpret the atmospheric observations, a
laboratory investigation of the Chapman mechanism was
performed using a fast flow tube reactor at the University
of East Anglia. The flow tube employs liquid N2 cooling
to operate over the temperature range 90– 300 K [Murray
and Plane, 2003]. At the upstream end, gas-phase Na
atoms were produced in an oven heated to about 620 K,
entrained in a flow of N2 and introduced into the cooled
Table 1. Nightly Averages of Na Line Ratios: ESI
Date
1 March 2000
2 March 2000
3 March 2000
4 March 2000
5 March 2000
21 Oct 2000
22 Oct 2000
23 Oct 2000
24 Oct 2000
25 Oct 2000
4 of 8
RD Average
and 2-s SD
1.582
1.569
1.568
1.651
1.594
1.582
1.604
1.664
1.613
1.563
±
±
±
±
±
±
±
±
±
±
0.068
0.132
0.118
0.042
0.090
0.098
0.092
0.040
0.068
0.054
Number
of Spectra
RD Range
8
5
9
13
9
11
5
8
9
11
0.140
0.204
0.185
0.058
0.151
0.160
0.131
0.064
0.121
0.090
D23302
SLANGER ET AL.: SODIUM NIGHTGLOW
D23302
4.1. A Temperature-Dependent Mechanism
[25] Although it seems unlikely that the propensity for
producing Na(2P3/2) and Na(2P1/2) in reactions (R2a) and
(R2b) could be temperature dependent, the seasonal variation of RD observed in the Keck data (Figure 2) suggested
that temperature might play a role. This could happen, for
instance, if the propensity was sensitive to the rotational
distribution of NaO(A), which would change with temperature. However, there is no consistent correlation in the
Keck data between RD and the Meinel OH 3 – 1 temperature,
just as in the 2002 Leonid MAC mission (Figure 5). The
laboratory experiments confirm that RD is not temperature
dependent over the temperature range appropriate to the
upper mesosphere.
Figure 4. Sodium D2/D1 ratios and intensities from
NASA’s 2002 Leonid MAC mission.
section of the flow tube via a movable injector, heated
internally but with a cooled outer jacket [Helmer and Plane,
1993]. Flows of O and O3 were then introduced into the flow
tube to generate Na chemiluminescence via reactions
(R1)(R2b)– (R3). RD was measured through a glass window
on a crosspiece at the downstream end of the tube, by
using an optic fiber to collect a fraction of the chemiluminescence and dispersing this through an identical spectrometer that was deployed on the NASA aircraft.
[22] This apparatus could then be used to test the dependence of RD on temperature, the relative proportions of the
reactants (O3 and O), and possible quenching molecules (O2
and N2). To measure the temperature dependence of RD, O
was generated by microwave discharge of O2, which also
produced a small concentration of O3. This flow was
introduced to the flow tube 10 cm upstream of the crosspiece, and the Na atoms through the injector 45 cm further
upstream. RD was shown to be essentially independent of
temperature between 153 and 295 K.
[23] All other experiments were therefore performed at
room temperature. For the first set of experiments, O and O3
were generated in an O2 discharge and introduced into the
flow upstream of the Na inlet. By varying the flow of O2
through the microwave cavity, it was shown that RD did not
depend significantly on the ratios [O]/[M], [O2]/[M], [O3]/
[M] and [O]/[O3] (where M = bath gas, in this case N2). The
second set of experiments were designed to test the efficiency
of O2 as a quenching gas. O was now generated via the
microwave discharge of N2, and the resulting N atoms titrated
with NO to produce O free of O3. This flow was introduced
10 cm upstream of the crosspiece. O3 was generated in a
separate flow of O2 passing through a commercial ozoniser,
and introduced at the upstream end of the flow tube. Figure 6
shows that RD is a function of the ratio of O to O2 in the tube:
RD is 1.65 when [O]/[O2] is 1 104, and increases to
above 2.0 when [O]/[O2] approaches 1 102. These ratios
correspond to atmospheric nighttime ratios at about 83 and
90 km, respectively.
4.2. Self-Absorption in the Na Layer
[26] A second possibility is that the density of Na atoms
within and below the emitting layer is large enough to cause
self-absorption. Since the absorption cross section of the D2
line is twice that of D1, significant self-absorption would
reduce RD. We have used the UEA Na model [Plane, 2004]
to explore this possibility, for the conditions of midwinter (at
40N) when the Na layer abundance is largest [Plane, 2003].
Assuming that Na(2P3/2) and Na(2P1/2) are produced by
reactions (R2a) and (R2b) in the ratio of 2.0, then RD would
be 1.93 if measured from the ground with a zenith-pointing
instrument, and 1.86 if viewed at an elevation angle of 30
above the horizon. This modest reduction in RD, even at a
low elevation angle, indicates that self-absorption is not an
important factor in causing the variability in RD.
[27] In fact, we can show empirically that there is no
evidence for a path length effect. The Keck data encompass
a variety of elevation angles and therefore path lengths.
Figure 7 presents a plot of line ratio versus air mass,
corresponding to a variation in elevation angle of 38–
84 degrees. There is no correlation evident, and the system
appears to be optically thin to the radiation for the
conditions investigated.
4.3. Other Mechanisms for Producing D-line Emission
[28] A third possibility is that D-line emission is produced
by a completely different chemistry that competes with the
Chapman mechanism. Besides reactions (R2a) and (R2b),
4. Discussion
[24] The variability of the Na D-line ratio in the nightglow is a surprising phenomenon, and we now explore a
number of possible explanations.
Figure 5. Sodium D2/D1 ratios versus OH 6 – 2 temperature from NASA’s 2002 Leonid MAC mission.
5 of 8
D23302
SLANGER ET AL.: SODIUM NIGHTGLOW
Figure 6. Sodium D2/D1 laboratory ratios versus [O]/[O2].
the only gas-phase chemical reactions that are sufficiently
exothermic to produce Na(2P) in the upper atmosphere are
dissociative electron recombination reactions of Na+ cluster
ions [Plane, 2003]. However, the rate of cluster ion formation is too slow, even in an intense sporadic E layer, for this
mechanism to compete significantly with the Chapman
mechanism [Cox and Plane, 1998]. Another possibility
is exothermic heterogeneous reactions on the surface of
meteoric smoke particles [Kelley et al., 2003]. However,
there are two difficulties with this hypothesis: finding a
mechanism that will sputter Na(2P) atoms from the particle
surface, and the limited volumetric surface area of smoke
particles that is predicted by meteor smoke models
[Gabrielli et al., 2004].
4.4. A Modified Chapman Mechanism
[29] The most likely explanation for the variable RD is
that the NaO(A) produced by reaction (R1) can either react
with O by reaction (R4), or be quenched to the NaO(X) state
which then reacts with O atoms:
ðR4Þ
ðR5Þ
ðR6Þ
NaO A2 Sþ þ O 3 P ! Na 2 PJ þ O2
D23302
[30] The observed value of RD could then be directly
translated into a composition ratio at the emitting altitude,
with the variability of RD throughout a night showing the
effects of dynamics and gravity waves. This hypothesis
would also explain the semiannual variability of RD in the
Keck data, the high values of RD during the equinoxes
corresponding to an increase in the [O]/[O2] ratio due to
downward transport of O [Thomas, 1990]. This would
imply that RD(A) was greater than 2.0 and RD(X) was less
than about 1.3.
[31] There is, however, one difficulty with this explanation. It implies that reaction (R4) and reaction (R6) produce
D-line emission with roughly equal probabilities, whereas
considerations of electronic correlation indicate that reaction
(R6) should be unimportant [Herschbach et al., 1992], in
accord with earlier laboratory measurements of the efficiency
of that reaction [Plane and Husain, 1986]. Nevertheless, if
there is only a single channel generating the excited sodium,
it is difficult to explain how RD is so variable on the observed
timescale and space scale.
[32] One possible way of increasing the yield of
reaction (R6) is to invoke vibrational excitation in the
NaO(X) product of reaction (R5). If the molecule is left in
the v00 = 1 level, then reaction (R6) might become
atmospherically more probable, whereas under higherdensity laboratory conditions, only reaction of the v00 = 0
level would have been tested. Moreover, reaction (R5) with
M = O2 would be energetically quite favorable (106 cm1
endothermic) for X(v00 = 1) production compared to
363 cm1 exothermic for X(v = 0).
[33] The measurements of Sipler and Biondi [1978]
present an interesting enigma. These data, taken at 40N
on 21– 22 October 1974, show a constant ratio throughout
the night (RD = 1.98 ± 0.1). A decrease in RD is seen only at
evening and morning twilight, as solar illumination of the
upper atmosphere changes the emission source from chemical excitation to resonance fluorescence. This is clearly
contrary to what is shown in Figure 2, which shows observations from five different locations. There is no obvious
RDð AÞ
NaO A2 Sþ þ M ! NaO X 2 þ M
NaO X 2 þ O 3 P ! Na 2 PJ þ O2
RDð X Þ
where M is a quenching molecule (e.g., O2 or N2).
Reactions (R4) and (R6) are likely to produce Na(2P1/2)
and Na(2P3/2) with different D-line ratios (RD(A) and RD(X),
respectively). These reactions will also likely produce
Na(2P) and Na(2S) with different branching ratios. The
quenching molecule in reaction (R6) is probably O2,
following the laboratory experiments described above,
which showed that RD varies with the ratio [O]/[O2], but
seems to be independent of the [O]/[N2] ratio. This may
be because the quenching of NaO(A2S+) to NaO(X2)
requires the transfer of 1919 cm1 of electronic energy;
the electronic-to-vibrational process is exothermic by
363 cm1 for O2, but endothermic by 411 cm1 for N2
[Joo et al., 1999].
Figure 7. Sodium D2/D1 ratios versus air mass, ESI data
of March/October 2000; air mass of unity refers to the
zenith.
6 of 8
D23302
SLANGER ET AL.: SODIUM NIGHTGLOW
reason for the difference between these earlier results and
those in the present study. However, an interesting observation was made in these FPI measurements. The high spectral
resolution made it possible to measure the sodium linewidths
and thus the kinetic energy of the sodium fragments. These
increased significantly between February and May, and again
between September and November, from 40 to 50 meV. The
authors’ conclusion was that this ‘‘. . .may indicate that more
than one process contributes to the sodium nightglow.’’ This
is in accord with our hypothesis that the emitting sodium is
produced by two separate reactions.
5. Conclusions
[34] The sodium D2/D1 line intensity ratio has been investigated in different ways, and recent measurements from
astronomical sky spectra at Mauna Kea, measurements over
the North Atlantic, and older studies in Alaska have shown
that this ratio is quite variable. The values typically range
from 1.2 to 2.0, with most determinations in the 1.4– 1.9
range. The ratio shows seasonal oscillations, and is consistent
with the SAO behavior of the total sodium nightglow
intensity, exhibiting peaks in April and October. There is no
apparent correlation with optical depth in the sodium layer.
The variability of the ratio suggests that there are competing
branching reactions generating the excited sodium, which is
not consistent with the current view of the reactivity of the
parent species, NaO. Further laboratory work can answer
some of the questions raised by these observations.
[35] Acknowledgments. We express our appreciation to the NSF
Aeronomy Program and the Natural Environment Research Council (UK)
for support of this work. We gratefully acknowledge discussions with Hisao
Takahashi concerning measurements of nightglow features from Cariri,
Brazil, as well as the SAO sodium data. The W. M. Keck telescopes are
operated jointly by the California Institute of Technology and the University of California. The 2002 Leonid Multi-Instrument Aircraft Campaign
was sponsored by NASA’s Astrobiology and Planetary Astronomy programs. The mission was executed by NASA’s Dryden Flight Research
Center and the SETI Institute. Participation of the University of East Anglia
team was made possible by the University of East Anglia. We thank
Stephen Ashworth (UEA) for assistance with the OH spectral analysis.
References
Bernard, R. (1938), Das Vorhandensein von Natrium in der Atmosphare auf
Grund von interferometrischen Untersuchungen der D-line im Abendund Nachhimmelslicht, Z. Phys., 119, 291 – 302.
Bills, R. E., and C. S. Gardner (1990), Lidar observations of mesospheric
Fe and sporadic Fe layers at Urbana, Illinois, Geophys. Res. Lett., 17,
143 – 146.
Buriti, R. A., H. Takahashi, and D. Gobbi (2001), First results from mesospheric airglow observations at 7.5 degrees south, Braz. J. Geophys.,
19(2), 169 – 176.
Chamberlain, J. W. (1995), Physics of the Aurora and Airglow, AGU,
Washington D. C.
Chamberlain, J. W., and B. J. Negaard (1956), Resonance scattering by
atmospheric sodium: II. Nightglow theory, J. Atmos. Terr. Phys., 9,
169 – 178.
Chapman, S. (1939), Notes on atmospheric sodium, J. Astrophys., 90,
309 – 316.
Clemesha, B. R. (1995), Sporadic neutral metal layers in the mesosphere
and lower thermosphere, J. Atmos. Terr. Phys., 57, 725 – 736.
Clemesha, B. R., D. M. Simonich, H. Takahashi, and S. M. L. Melo (1993),
A simultaneous measurement of the vertical profiles of sodium nightglow
and atomic sodium density in the upper atmosphere, Geophys. Res. Lett.,
20, 1347 – 1350.
Cox, R. M., and J. M. C. Plane (1998), An ion-molecule mechanism for the
formation of neutral sporadic Na layers, J. Geophys. Res., 103, 6349 – 6359.
Gabrielli, P., et al. (2004), Meteoric smoke fallout over the Holocene epoch
revealed by iridium and platinum in Greenland ice, Nature, 432, 1011 –
1014.
D23302
Griffin, J., D. R. Worsnop, R. C. Brown, C. E. Kolb, and D. R. Herschbach
(2001), Chemical kinetics of the NaO(A) + O(3P) reaction, J. Phys. Chem.
A, 105(9), 1643 – 1648.
Hanuschik, R. W. (2003), A flux-calibrated high-resolution atlas of optical
sky emission from UVES, Astron. Astrophys., 407, 1157 – 1164.
Happer, W., G. J. MacDonald, C. E. Max, and F. J. Dyson (1994), Atmospheric turbulence compensation by resonant optical backscattering from
the sodium layer in the upper atmosphere, J. Opt. Soc. Am. A Opt. Image
Sci., 11, 263 – 276.
Hecht, J. H., S. Collins, C. Kruschwitz, M. C. Kelley, R. G. Roble, and
R. L. Walterscheid (2000), The excitation of the Na airglow from
Coqui Dos rocket and ground-based observations, Geophys. Res. Lett.,
27, 453 – 456.
Helmer, M., and J. M. C. Plane (1993), A study of the reaction NaO2 +
O ! NaO + O2 - Implications for the chemistry of sodium in the
upper atmosphere, J. Geophys. Res., 98, 23,207 – 23,222.
Herschbach, D. R., C. E. Kolb, D. R. Worsnop, and X. Shi (1992), Excitation mechanism of the mesospheric sodium nightglow, Nature, 356,
414 – 416.
Jenniskens, P. (2002), The 2002 Leonid MAC airborne mission: First
results, WGN, J. Int. Meteor Organ., 30(6), 218 – 224.
Jenniskens, P., M. Lacey, B. J. Allan, D. E. Self, and J. M. C. Plane (2000),
FeO ‘‘orange arc’’ emission detected in optical spectrum of Leonid persistent train, Earth Moon Planets, 82/83, 429 – 438.
Joo, S., D. R. Worsnop, C. E. Kolb, S. K. Kim, and D. R. Herschbach
(1999), Observation of the A-X electronic transition of NaO, J. Phys.
Chem. A, 103, 3193 – 3199.
Kelley, M. C., C. Kruschwitz, J. Drummond, C. Gardner, L. Gelinas,
J. Hecht, E. Murad, and S. Collins (2003), A new explanation of
persistent double meteor trains, Geophys. Res. Lett., 30(23), 2202,
doi:10.1029/2003GL018312.
Kirchhoff, V. W. J. H. (1986), Theory of the atmospheric sodium layer: A
review, Can. J. Phys., 64, 1664 – 1672.
Kirchhoff, V. W. J. H., and H. Takahashi (1985), First sodium nightglow
results from Natal, Planet. Space Sci., 33, 757 – 760.
Kirchhoff, V. W. J. H., B. R. Clemesha, and D. M. Simonich (1979),
Sodium nightglow measurements and implications on the sodium photochemistry, J. Geophys. Res., 84, 1323 – 1327.
Kirchhoff, V. W. J. H., B. R. Clemesha, and D. M. Simonich (1981),
Average nocturnal and seasonal variations of sodium nightglow at 23S,
46W, Planet. Space Sci., 29, 765 – 766.
Kruschwitz, C. A., M. C. Kelley, G. Swenson, A. Z. Liu, X. Chu, J. D.
Drummond, B. W. Grime, W. T. Armstrong, J. M. C. Plane, and
P. Jenniskens (2001), Observations of persistent Leonid meteor trails:
2. Photometry and numerical modeling, J. Geophys. Res., 106,
21,525 – 21,541.
McNeil, W. J., E. Murad, and S. T. Lai (1995), Comprehensive model for
the atmospheric sodium layer, J. Geophys. Res., 100, 16,847 – 16,855.
Murray, B. J., and J. M. C. Plane (2003), The uptake of atomic oxygen on
ice films, Phys. Chem. Chem. Phys., 5, 4129 – 4138.
Plane, J. M. C. (2003), Atmospheric chemistry of meteoric metals, Chem.
Rev., 103(12), 4963 – 4984.
Plane, J. M. C. (2004), A time-resolved model of the mesospheric Na layer:
Constraints on the meteor input function, Atmos. Chem. Phys., 4, 627 –
638.
Plane, J. M. C., and D. Husain (1986), Determination of the absolute rate
constant for the reaction O + NaO ! Na + O2 by time-resolved atomic
chemiluminescence at l = 589 nm (Na(32PJ) ! Na(32S1/2) + hn),
J. Chem. Soc. Faraday Trans. 2, 82, 2047 – 2052.
Rees, M. H., G. J. Romick, and A. E. Belon (1975), The intensity of the
sodium D lines in the auroral zone, Ann. Geophys., 31(2), 311 – 320.
Shi, X., D. R. Herschbach, D. R. Worsnop, and C. E. Kolb (1993), Molecular beam chemistry: Magnetic deflection analysis of monoxide electronic states from alkali-metal atom plus ozone reactions, J. Phys. Chem.,
97(10), 2113 – 2122.
Sipler, D. P., and M. A. Biondi (1978), Interferometric studies of the twilight
and nightglow sodium D-line profiles, Planet. Space Sci., 26, 65 – 73.
Slanger, T. G., and R. A. Copeland (2003), Energetic oxygen in the upper
atmosphere and the laboratory, Chem. Rev., 103(12), 4731 – 4765.
Slanger, T. G., and D. E. Osterbrock (2000), Investigation of potassium,
lithium, and sodium emissions in the nightglow, and OH cross-calibration, J. Geophys. Res., 105, 1425 – 1429.
Slanger, T. G., D. L. Huestis, and P. C. Cosby (2000a), Sodium D2/D1 ratio
in the terrestrial nightglow, from Keck/HIRES sky spectra, Bull. Am.
Astron. Soc., 33, 1599.
Slanger, T. G., D. L. Huestis, P. C. Cosby, and D. E. Osterbrock
(2000b), Accurate atomic line wavelengths from astronomical sky
spectra, J. Chem. Phys., 113, 8514 – 8520.
Slipher, V. M. (1929), Emissions in the spectrum of the light of the night
sky, Publ. Astron. Soc. Pac., 41, 262 – 264.
7 of 8
D23302
SLANGER ET AL.: SODIUM NIGHTGLOW
Thomas, R. J. (1990), Atomic hydrogen and atomic oxygen density in the
mesopause region: Global and seasonal variations deduced from Solar
Mesosphere Explorer near-infrared emissions, J. Geophys. Res., 95,
16,457 – 16,470.
Wright, T. G., A. M. Ellis, and J. M. Dyke (1993), A study of the products
of the gas-phase reaction M + N2O and M + O3, with ultraviolet photoelectron spectroscopy, J. Chem. Phys., 98, 2891 – 2907.
C. Allende Prieto, McDonald Observatory, University of Texas at Austin,
Austin, TX 78712-1083, USA.
D23302
P. C. Cosby, D. L. Huestis, and T. G. Slanger, Molecular Physics
Laboratory, SRI International, Menlo Park, CA 94025, USA. (tom.slanger@
sri.com)
P. Jenniskens, NASA Ames Research Center, Mountain View, CA 94043,
USA.
F. J. Martin-Torres, AS&M, Inc., NASA/Langley Research Center,
Hampton, VA 23666, USA.
B. J. Murray, Department of Chemistry, University of British Columbia,
2036 Main Mall, Vancouver, BC, Canada V6T 1Z1.
D. A. O’Sullivan, J. M. C. Plane, and A. Saiz-Lopez, School of
Environmental Sciences, University of East Anglia, Norwich NR4 7TJ,
UK.
8 of 8