-- '_.~.- -.
.- :: ~.J-
>~ ,"'- -- C- ':>
;-· ~e t"
'5 Jt .
· ' '.,
L ~.', '
' -. "
~,~, , _- ,. - -\.'r r ...
>.-. ?_ in
.... ' '~
. -. ,- ' '· ~,...
- - .
:~~ ~ ~
q ' ~ / ~"
' ~ - '
'- : '
"5 , :
C--..- '.: .7,
-".._ -- *.-.
i ~ ~
~, ', ~.
' ~ '
METALLIC IONS IN THE EQUATORIAL IONOSPHERE
A. C. Aikin
R. A. Goldberg
Laboratory for Planetary Atmospheres
NASA/Goddard Space Flight Center
Four positive ion composition measurements of the
equatorial E region at Thumba, India are presented. During
the day, the major ions between 90 and 125 km are NO+
02 . The relative concentrations are similar to those ob-
served at midlatitudes but exhibit unusual structural
behavior with altitude. A metallic ion layer centered at
92 km is observed, and found to contain Mg+, Fe+, Ca , K+
Al , and Na+
ions. The layer is explained in terms of a
similarly shaped altitude distribution of neutral atoms
which are photoionized and charge exchanged with NO+
02 . Three body reactions form molecular metallic ions
which are rapidly lost by dissociative ion-electron recom-
bination. Unfortunately reactions which create molecular
ions from Na+ and K+ using CO2
as a third body to form
cluster ions are too slow to account for the observed
alkalai metal ion distribution.
Nighttime observations show downward drifting of the
metallic ion layer caused by equatorial dynamo effects.
These ions react and form neutral metals which charge ex-
change with NO
and 0 + to produce an observed depletion
and 02 within the metallic ion region. The in-
clusion of drift effects is found to be necessary in pro-
ducing a proper description of metal ion behavior at the
magnetic equator. Similar considerations are probably
appropriate at other latitudes.
It is essential for the understanding of the equatorial
ionosphere and electrojet to have a detailed knowledge of
the altitude and temporal variation of the positive ion
species and concentrations. Several rocket flights have
aided in defining the equatorial E region electron density
distribution (Aikin and Blumle, 1968); however, prior to
the work reported here, no measurements had been made of the
ion composition in this region. This paper gives the results
of two daytime and two nighttime measurements of the ion
composition and electron density of the E region at the
magnetic equator, using rocket payloads launched from the
Thumba Equatorial Rocket Launch Site (dip latitude, 0.85 0 S).
Under most circumstances the major ionic constituents
in the E region are observed to be NO+
and 02 +
sured distribution of these molecular ions can be compared
with theoretical profiles based on a photochemical model
such as that of Keneshea et al, (1970). At night we ob-
serve that below 100 km the ions NO+ and 02+ are sufficiently
depleted to permit metallic atomic ions to become the major
Metal ions are considered to be a permanent feature
of the lower E region and have been observed over a large
range of latitudes and times (e.g. Istomin, 1963; Narcisi,
Because of several factors their distribution is
difficult to explain. The metal ions result in part from
the photoionization of neutral metals and charge exchange
between neutral metals and the principal ions of the E
region, NO and 02o The distribution of neutral metals
varies seasonally and geographically in an ill-defined man-
ner (Hunten, 1964).
The reactions and rate coefficients
for the interaction of metal ions with neutrals are very
uncertain (Ferguson, 1972). Because metal ions are
atomic and thus have lifetimes of order 105 times that
of molecular ions, they are more subject to dynamical
effects. For example, at midlatitudes metallic ions are
found in sporadic E layers (Es) formed by interaction of
wind shears with the vertical component of the terrestrial
magnetic field (Whitehead, 1961).
Electric field induced
vertical motion of such a metal ion layer at temperate
latitudes has been treated by Chimonas and Axford (1968).
At the geomagnetic equator, dynamo effects induce
vertical drifting of the ionospheric plasma. The diurnal
variation of the dynamo electric field and its relation
to the F region has been detailed by measurements at
Jicamarica, Peru (Balsley, 1969 a, b; Balsley and Woodman
Based on ionograms a similar behavior is deduced
for Thumba. The rocket observations show that the distri-
bution of metal ions in the E region is also affected by
this electric field. This paper presents measured evidence
for equatorial electric field modification of the metal
ion distribution. The behavior of all species are discussed
although large uncertainties in rate coefficients, source
and loss functions of the neutral distribution, and other
parameters which determine the metallic ion structure limit
the accuracy of the quantitative results. These results have
important implications for the metal ion distribution at mid-
Information pertinent to the four rocket firings is
listed in Table Io.
Each payload contained a quadrupole
ion mass spectrometer housed in a titanium getter pumped
system assisted by two triode ion pumps for inert gas
pumping (Goldberg and Blumle, 1970). The details of this
system plus the daytime performance characteristics are
described in Goldberg and Aikin (1971). The nighttime
details will be discussed elsewhere.
The daytime flights also contained radiowave absorption
experiments operated at 1.865 and 3.030 MHz to determine
electron density, and Gerdien probes for total positive
ion density (Aikin et al, 1972),
The electron density
profiles at night were determined by continuous wave dis-
persive Doppler experiments (Seddon, 1953; Bauer and Jackson,
1962; Jackson, 1971) operated at frequencies of 73.6 and
24.53 MHz. Electron density profiles determined by the
above techniques were used to normalize the spectrometer
currents to absolute values. Free molecular flow effects
were taken into account in obtaining the correct relative
All rocket trajectories were determined using a tone
range/telemetry interferometer tracking system developed
by Hudgins and Lease (1969).
Figure 1 illustrates typical spectra observed at two
altitudes, 89 km and 148 km, for nighttime conditions at
the magnetic equator. The mass range swept is 1 to 41 amu.
Two high pass filter modes were also employed to deduce the
total ion density > 41 amu. These are shown in the figure
for the ranges 1 -
- and 32 -
o amu. The use of'two mass
ranges in the high pass filter mode permits inflight
calibration of the high pass filter mode data as outlined
in Goldberg and Aikin (1971).
The measured ionic constituents
and their identification are 30+ , NO+; 32+, 2+; 28+, N2
and 16 , O ; which are gaseous ions resulting from ionization
of the principal gases of the thermosphere. In addition, we
identify the metallic ions 23 , Na+; 24+, Mg+; 27+ , Al; 28, Si+;
39+, K ; 40 , Ca ; and 56 , Fe+ ; the last from high pass
filter mode data during the nighttime flights. Since the night-
time high pass filter mode data is known to contain a fractional
contribution from 02 (32+), the Fe+ deduced from this data
must be considered an upper quantitative limit. We note
that 28+ can either be N2+
or Si+, with no apparent method
of separating the two at this time.
THE DISTRIBUTION OF NO+AND 02+
The distribution of the major ionic constituents, as
illustrated in Figure 2 for the daytime equatorial E region,
is characterized principally by NO and 02+
of the equatorial electrojet, which is centered near 109 km
(Maynard and Cahill, 1965), has no apparent effect on the
ion composition. Above 95 km the major ionic constituents
exhibit an unexplained oscillation that slowly dampens in
amplitude and wavelength. No such behavior is apparent
in the nighttime data shown in Figure 3, but here the
height resolution is greatly reduced because of the greater
velocity of the rocket.
The NO /°2 +
ratio for the four flights is shown in
Figures 4 and 5 for daytime and nighttime, respectively.
There have been few previous measurements of this ratio
below 100 km. The data of Narcisi (1968) for a zenith
angle of 49° showed that between 95 and 115 km the NO +/02
ratio was less than unity. None of our daytime data exhibits
a ratio less than unity for any extended altitude. For a
zenith angle of 27.80 we find NO+/02+ ratios of about 2 for
the altitude range 100 to 125 km. Our results are well
within the predicted values as deduced by Danilov (1972).
At night the ratio ranges between 5 and 150 over the
entire altitude region. Narcisi et a1l.(1967) give
a ratio of 15 + 2 for 95 to 103 km at midnight. Our com-
parable measurement is at 1:08 LMT and a ratio ranging
between 45 and 150 is obtained.
Keneshea et al. (1970) have conducted a complete diurnal
study and their computed results are shown in Figures 4 and 5..
We have performed a similar computation and our daytime re-
sults are shown for comparison in Figure 4. The agreement
of our model is better with the observed daytime data because
of the larger calculated ratio for NO+/O
between the two computations is caused in part by the dif-
ference in values of the rate coefficient kl for the re-
7.7 x 10
+ NO NO
+ 02 (6 x 10
for Keneshea et al,vs
for our calculation) and by the use of
a larger value for neutral NO above 100 km.
The NO/O 2 + ratio is a sensitive function of the solar
flux, since this factor also enters into the equation through
the electron density, Ne
There is considerable uncertainity
in the solar X-ray flux in the 30 to 60 A range. This is
particularly true of the 30 to 40 A interval which includes
the strong CVI line at 33.6 A and which Bourdeau et al
(1965) noted was an important ionization source in the 90
to 100 km altitude region. Values for the measured flux at
33.6 A range from 2-3 x 10
ergs/cm sec (Manson, 1967, 1968)
to 6 x 10- 2ergs/cm
sec (Argo et al, 1970).
putations of an E region model used 3 x 10-3
For the Keneshea et al, (1970) model it
was found necessary to increase the flux in the 30 to 60 A
range by a factor of 4 over the earlier computation. Our
computations employ a flux in this spectral range which is
a factor of 2 less than the Keneshea et al,(1970) values.
The measured X-ray flux in the 1-8 A band was enhanced at the
time of the daytime flights (Aikin et al, 1972). No measured
data were available concerning the 30 to 60 A flux.
Meira (1971) nitric oxide profile, which is utilized in our
computations, leads to unacceptably large values of NO /O2
at night so the ratios are not shown. The better agreement
of Keneshea et al,(1970) at night implies that NO may have
a diurnal variation.
THE DISTRIBUTION OF METALLIC IONS
The daytime metallic ion belt observed by rockets 14.425
and 14.424 is illustrated in Figure 2. The metallic ions are
contained in a layer which maximizes near 92 km with a density
of 400 cm
The principal ions are Mg and Fe+
concentrations of Na+, Al , K+, Ca+
and possibly Si+
also observed. Important minor isotopes of the above con-
stituents are.observed but not shown. The minor constituents
are less apparent in the data of 14.425 because of inferior
operating conditions for this instrument. This may also ac-
count in part for the apparent depletion in 24+ below 92 km
in this flight.
From the daytime data it is seen that the metallic layers
are well defined at the magnetic equator, although of broader
peak width than at midlatitudes, (Narcisi and Bailey, 1965;
Narcisi, 1968; Young et al, 1967). Above 100 km a low
background of metallics occurs to apogee, with no secondary
ledges of metal ions in the 110 km range as observed at
The nighttime data are illustrated in Figure 3. The
height resolution on these flights is poorer because of
sampling under conditions of larger velocities. The
spectrometer sweep range for the nighttime flights did not
exceed 42 + , so that Fe + must be approximated by the TB
curve. Although the nighttime estimate of Fe+ is based
on high pass filter mode data which includes an unknown
fractional contribution of 02 , this does not significantly
alter the magnitude of Fe+
at the lower altitudes (below
102 km on 18.97 and below 97 km on 18.98) where 02
trace constituent relative to TB.
Finally, we note that
the spectrometer aboard 18.97 had less sensitivity than
that aboard 18.98.
The observed metallic belt in the data of 18.97
exhibits a broad maximum centered near 98 km with a peak
density near 103/cm
The measurement was made at post-
sunset, when the F layer had reached an altitude of 550 km
as determined from ionogram analysis. The data of 18.98
were obtained after the F layer maximum had drifted downward
to 300 km. At this time the metallic layer is observed to
be more pronounced, with properties quite similar to the
daytime case, but for a peak density over twice as large.
There is no evidence for sporadic E on ionograms recorded
during either of the two night flights.
Awe (1971) has recently reported that the equatorial
sporadic E virtual height (h'Es) exhibits an in-phase
correlation with the postsunset drift induced height shift
of the equatorial F layer. Typical Es
are 1-3 m/sec. The data presented here illustrate that
the equatorial metallic layer is subjected to similar
drifts, induced by dynamo electric fields. The magnitude
of the observed displacement (6 km) can be accounted for
by an average downward drift of approximately 1 m/sec between
The height of the nighttime metallic layer coincides
with a depletion of NO
and 02 , an effect most pronounced
in the data of 18.98. Similar type depletions at midlatitude
have been observed by Narcisi (1968) and he has suggested
that charge exchange between NO and neutral metallics is
the reason for this effect. In the following sections the
theory of the metallic ion distribution will be examined
The temporal variation of the metals depends on several
1) the distribution of the neutral metals in space and
2) ionic reactions as detailed in Table IIo,
3) ion motion imposed by electric fields, diffusion
effects, and other external forces.
a. Photochemistry of the Metals
The reactions and rate coefficients describing the
behavior of metals in the upper atmosphere are poorly
known. Nevertheless reasonable estimates can be made for
most of the important processes. The reactions which have
been included in the computations together with the assumed
rate coefficients are listed in Table II. For comparison
purposes the measured rate coefficient is also given and
utilized wherever possible.
The metal ions can be placed into two categories.
The first group contains Mg+
Al , Fe and Ca+. These ions
react with 02 and 03 to form molecular ions such as XO2
and XO+ where X is any of the above constituents. Such ions
recombine dissociatively with electrons at a much greater
rate than would be the case if the ions were atomic. The
second group contains alkalai metals such as Na+
There is apparently no molecular ion formation for these
ions in a manner akin to the group I ions. However,
laboratory observations (Keller and Beyer, 1971 b) have
shown that 3 body reactions such as
+ C0 2
do occur with large rate coefficients.
The present computations do not consider the chemical
behavior of the neutral metals. No metallic oxide processes
and no diurnal variation of the neutral metals as the result
of either photodissociation of metal oxide compounds or
changes in the concentration of minor constituents (as for
example, ozone) are considered here. Similarly there has
been no attempt to include the diffusion of the neutral
metals either from the top of the layer as was done by Hanson
and Donaldson (1967) and Gadsden (1970) or as loss out the
bottom. In the model, neutral metals are created only
through dissociative recombination of molecular metal ions.
None of the above processes or values are sufficiently
understood or known to warrant inclusion at this time.
The initial altitude distribution of the neutral
metals assumed for the model is given in Table IIo. Little
is known about the neutral metal distribution in this region.
However, by the use of resonant scattering of a laser beam
(Gibson and Sanford 1971) the sodium distribution has been
studied. These observations are consistent with determina-
tions from airglow monitoring performed with rockets (Hunten
and Wallace, 1967) in that they show a neutral sodium layer
centered between 90 and 95 km with a maximum concentration of
The uncertainties of distribution are even greater
for the other remaining metals and most have never been mea-
Hunten (1964), Vallance-Jones (1966), and Gadsden
(1970) have reviewed the distribution for other metals such
as potassium. With the exception of neutral sodium at 93 km,
all neutral metallic distributions in Table III are those re-
quired to produce the measured ion distribution.
Under the above considerations and restrictions, the
computed distribution of metallic ions for daytime conditions
is shown in Figure 6.
The agreement between theory and
experiment is dependent not only on the rate coefficients
but also on the assumed concentrations of neutral metals.
In this instance the most obvious lack of agreement is
between the observed and calculated distribution of Na+
The number density of Na+ observed is nearly two orders of
magnitude less than that required by Keller and Beyer (1971 b)
in suggesting that clustering of Na+ to CO 2 and 02 could
explain the distribution of Na+
The predicted distribution of Mg is in much better agree-
ment with observations. The ratio Mg+/Mg is 1.8 at 93 km and
the total column density for Mg+
2x10 cm .
Anderson and Barth (1971) have shown Mg /Mg -
is estimated to be near
within a sporadic E layer but with a similar Mg+
of 2x10 cm
Gadsden (1972) has suggested a factor of 20
higher for this quantity. The agreement of our value with
Anderson and Barth implies a relatively constant total content
for metal ions, which redistributes under the influence of
dynamo effects. No significant diurnal variation of the metal
ions is predicted (see Figure 7 for Mg+). Hence, another mechanism
must be introduced to compensate for the observed diurnal variablity.
The data indicate that between the first and second
nighttime measurement an increase occurred in the metal ions
centered at 92 km. The increase was coupled with a sharp
decrease in NO+. The data suggest a downward moving layer