1
Influences of ice particles on the ion chemistry of the polar summer
mesosphere
J. Gumbel1,2,3, D. E. Siskind2, G. Witt4, K. M. Torkar5, and M. Friedrich6
Abstract. In the polar summer mesosphere, charge is distributed over a wide range of constituents
closely connected to phenomena like noctilucent clouds and polar mesosphere summer echoes. In this
paper we study how the presence of ice particles influences mesospheric ion chemistry, and how this
may feed back on the particle population. To this end, we present an ion-chemical model that for the
first time features close coupling with cluster growth and ice particle charging. Starting out from
molecular ion reactions, the H+(H2O)n proton hydrate chain is described using the Thomson model and
Natanson's recombination scheme. Under most mesospheric conditions electron capture by particles is
expected to enhance the lifetimes and concentrations of positive ions and clusters. This has important
consequences for the total charge density and mobility in the environment of particle layers. Extending
the proton hydrate chain to large cluster sizes, we also quantify the efficiency of ionic nucleation of
mesospheric ice particles. While ionic nucleation is not feasible as the major mesospheric nucleation
process, it can become efficient given moderate atmospheric variations as induced by gravity waves.
This leads to a scenario of rapid generation of populations with many small particles in local
temperature minima. We show that electron capture to existing particles can significantly enhance the
ionic nucleation of new particles. In summary, there are many potential connections between ion
chemistry and layered phenomena in the mesosphere that should be included in comprehensive models
of NLC/PMSE. Unfortunately, uncertainties in ionic reaction rates are a persistent problem and in
great need of laboratory measurements representative for cold summer mesopause conditions.
1. Introduction
The summer mesopause region at high latitudes is a
complex ion-chemical environment where light ions, clusters
and ice particles interact in various ways. This environment
coincides with the ionospheric D-region that features drastic
changes in charge density and ion composition over vertical
scales of a few kilometers. Comprehensive models of this ion
chemistry have been developed e.g. by Reid [1976], Kopp
[1996], and Turunen et al. [1996]. These models involve ionchemical schemes with of the order of 50 positive and negative
ions and ion clusters. Below a transition altitude of 80-90 km,
a general feature is the conversion of light ions, primarily NO+
and O2+, into oxonium clusters (or "proton hydrates")
H+(H2O)n. Reid [1989] discussed D-region ion chemistry with
particular focus on the very cold conditions near polar summer
mesopause. Here, at temperatures of 130 K and below, heavy
__________________________________________________
1
Universities Space Research Assoc., Washington D.C.
E. O. Hulburt Center for Space Research, Naval Research
Laboratory, Washington D.C.
3
NASA Goddard Space Flight Center, Greenbelt, MD
4
Department of Meteorology, Stockholm University,
Stockholm, Sweden
5
Space Research Institute, Austrian Academy of Sciences,
Graz, Austria
6
Department of Communications and Wave Propagation,
Technical University Graz, Austria
2
proton hydrates (n >> 6) can gain importance and give rise to
significant changes of the plasma properties [Björn and
Arnold, 1981; Sugiyama, 1994].
In the same region of the atmosphere, charging processes
are closely connected to a variety of layered phenomena
involving icy particles. Most prominent among these are
Noctilucent Clouds (NLC) and Polar Mesosphere Summer
Echoes (PMSE) [Thomas, 1991; Cho and Röttger 1997]. A
long list of rocket-borne and ground-based campaigns have
been dedicated to these phenomena. Related in situ
observations of layered charge features include electron biteouts [Ulwick et al., 1988], ion enhancements and bite-outs
[Balsiger et al., 1996; Lübken and Rapp, 2001], cluster ions
[Goldberg and Witt, 1977; Björn et al., 1981], charged
particles of either sign [Havnes et al., 1996], as well as smallscale fluctuations of charge [Lübken et al., 1993] and electric
fields [Pfaff et al., 2001].
Interactions between ion chemistry and the existence of icy
particles are mutual. Icy particles influence ion composition
and reaction scheme by capture of electrons and ions
[Natanson, 1960; Parthasarathy, 1976]. Charges captured by
particles, on the other hand, can influence the particle growth,
in particular coagulation processes [Jensen and Thomas, 1991;
Reid, 1997] or the particle shape resulting from condensation.
Moreover, proton hydrate clusters have been suggested as
possible condensation nuclei for ice particles [e.g., Witt, 1969;
Arnold, 1980]. A detailed mechanism for nucleation on proton
hydrates has been outlined by Sugiyama [1994]. However,
2
based on current knowledge meteoric smoke particles rather
than cluster ions are thought to serve as dominant
condensation nuclei in the mesosphere [Turco et al., 1982;
Keesee, 1989]. The groundbreaking work of Turco et al.
[1982] combined both smoke and ionic nucleation in a coupled
ice particle model for the mesosphere. This work showed that
in the direct competition between both nucleation mechanisms,
supersaturation requirements for condensation on smoke
particles are significantly less restrictive. Ionic nucleation can
only become an efficient competitor at extremely low
temperatures; the generation of large numbers of small (subvisible) particles becomes the most likely scenario under those
conditions. Many later NLC models have completely focused
on condensation onto the meteoric nuclei [e.g., Jensen and
Thomas, 1988; Klostermeyer, 1998]. This choice is also
supported by recent simulations of ice particle growth and
detailed water budget by Berger and von Zahn [2002] and von
Zahn and Berger [2002]. These authors suggest a very
efficient freeze-drying of the NLC/PMSE region, with water
ice redistribution both vertically and equatorward. This would
result in persistent dehydrated conditions near the summer
mesopause, thus further limiting the feasibility of ionic
nucleation. On the other hand, a caveat of nucleation on
meteoric material remains the fact that the existence and
properties of meteoric smoke particles as proposed by Hunten
et al. [1980] are still in lack of experimental verification in the
mesosphere.
With the present work, we study the effect of mesospheric
icy particles from an ion-chemical point of view. We
investigate how ion reactions and ion composition are changed
in the presence of particles. A model that aims to describe
these processes must combine basic ion chemistry, clustering
processes as well as charge capture by particles. Our OASIS
model (Originally Austrian Study of the IonoSphere) is based
on an ion-chemical code originally developed by Torkar and
Friedrich [1983]. The ion chemistry scheme is similar to the
standard models listed above [Reid, 1976; Kopp, 1996;
Turunen et al., 1996]. The Thomson droplet description [e.g.,
Holland and Castleman, 1982] is applied to describe cluster
growth to the larger sizes. The interaction of particles and light
charge carriers is included based on the formalism by
Natanson [1960]. Finally, the ideas developed by Arnold
[1980] and Sugiyama [1994] are used to allow for cluster
growth towards condensation nuclei for ice particles. This fills
the size gap between molecular ions on the one hand and
particles of radii exceeding 1 nm on the other.
It is important to note that our model is not an ice particle
growth model for NLC or PMSE. A complete, interactive ice
particle model for the mesosphere must include transport
processes (advection, sedimentation) and time-dependent
feedback on the neutral atmosphere, in particular on the water
budget [Turco et al., 1982; Jensen and Thomas, 1994; Rapp et
al., 2002]. This is beyond the scope of our steady-state ionchemical model. Ice particle layers are introduced only as
input to the simulations; the model is concerned with their
effect rather than with their origin. Nevertheless, the OASIS
model is able to provide important information about the first
step of particle evolution, i.e. particle nucleation. We show
that ionic nucleation rates can be enhanced drastically when
particle influences on the ion chemistry are present.
Chapter 2 of this paper provides a description of the OASIS
model and discusses relationships to earlier models of Dregion chemistry, clustering and particle charging. Sections 2.1
and 2.2 introduce the basic ion chemistry. Section 2.3 provides
parameterizations for proton hydrate reaction rates, thus
allowing us to incorporate heavy clusters into the ion-chemical
scheme. The possible growth of these clusters to condensation
nuclei for ice particles is introduced in section 2.4. Finally,
section 2.5 describes Natanson's charge capture formalism that
is used to treat both cluster / electron recombination and ice
particle / charge interaction. Chapter 3 presents simulation
results that illustrate these various model features. Chapter 4
discusses these results with regard to the summer mesopause
environment. In 4.1, results are compared to in situ charge
measurements in the vicinity of ice particles. Section 4.2
investigates particle effects on the ion composition, while
sections 4.3 and 4.4 discuss implications for ionic nucleation.
Conclusions are given in chapter 5.
When dealing with ion chemistry in the polar summer
mesosphere, it has to be kept in mind that many conclusions
must remain qualitative. Several complications arise under
these cold conditions (T < 150 K): Many reaction coefficients
are highly uncertain at these temperatures or rely only on
extrapolations; proton hydrates can grow to large sizes for
which reaction rates have not been measured; the capture of
electrons and ions by ice particles complicates the reaction
scheme. In particular, clustering simulations must currently
rely on interpolations between kinetic (molecular ion) and
thermodynamic (ice particle) approximations. Given these
quantitative limitations, many basic conclusions can still be
drawn about mechanisms in the mesospheric dusty plasma.
2. Model
2.1. Ionization sources
The neutral atmospheric input for OASIS can be taken from
MSISE-90 [Hedin, 1991] or from individual data sets. The
input used for the present work is discussed in section 3.
Figure 1 shows typical profiles of ionization in the summer
mesopause region. The solar zenith angle corresponds to midJuly at 67.9°N, Figure 1a represents quiet twilight conditions
at local midnight (χ = 91°), Figures 1b and c refer to local
noon (χ = 47°) under geomagnetically quiet and active
conditions, respectively. (Twilight conditions are important
since they are typical for many NLC observations and in situ
experiments.) During daytime, the major continuous source of
charge in the mesopause region is solar EUV radiation, in
particular Lyman-α ionization of NO. Galactic cosmic rays
become important below 80 km. During geomagnetically
perturbed conditions, electron and proton precipitation can
increase ionization rates in the NLC/PMSE region by several
orders of magnitude. The ionization profile in Figure 1c
corresponds to a precipitating electron flux of 10-7 J cm-2 s-1
distributed over a Maxwellian energy spectrum with a
characteristic energy of 10 keV. Major products of ionization
by precipitating particles are N2+, N+, O2+ and O+, which at
3
mesospheric altitudes are rapidly converted into the "primary
ions" O2+ and NO+. During twilight and at night in the absence
of particle precipitation, ionization of NO by geocoronal
Lyman-α dominates the charge production at NLC/PMSE
altitudes.
As solar input for photoionization we use the spectra given
by Torr and Torr [1985] for high and low solar activity,
extended towards longer wavelength with spectrum R74113 by
Heroux and Hinteregger [1978]. We interpolate these spectra
linearly with solar activity in accordance with the F10.7 flux
[Fritzenwallner, 1997]. An F10.7 value of 150 was used for the
current work. Solar Lyman-α is interpolated in the same way
[Brasseur and Solomon, 1986] or taken from direct
measurements [e.g., Lean, 1987]. Absorbing species in the
model are N2, O2, O, NO, O2(1∆g), and CO2. The transfer of
direct solar radiation is calculated for a spherical atmosphere;
absorption and photoionization cross section are used as
tabulated by Turunen et al. [1996]. The photon flux of
geocoronal Lyman-α is based on measurements by Meier and
Mange [1973]. A parameterization as a function of solar zenith
angle has been given by Strobel et al. [1974] with
extrapolations to lower altitudes by Thomas and Bowman
[1985a].
The role of energetic particle precipitation in mesospheric
ionization has discussed e.g. by Rusch et al. [1981] and Frahm
et al. [1997]. We describe the energy deposition and ionization
using the formalism of Rees [1969; 1982], which has been
extended to lower altitudes by Kirkwood and Osepian [1995].
Maxwellian and Gaussian energy spectra are used to represent
diffuse and discrete auroral conditions, respectively
[Strickland et al., 1983]. Alternatively, measured energy
spectra can be used as input [e.g., Rapp et al., 2002a].
Ionization by galactic cosmic rays (GCR) is parameterized
according to Heaps [1978] with the branching ratios for the
individual ions given by Rusch et al. [1981].
2.2. Ion Chemistry
OASIS is a one-dimensional steady state model for the ionchemistry of the D-region. As ion-chemical lifetimes are
generally short, transport processes are not considered. Under
the conditions of interest, reaction time constants for initial
clustering and successive proton hydrate growth are of the
order of seconds and milliseconds, respectively. Transport
times over a vertical distance of 1 km by advection or
diffusion are typically several hours [Brasseur and Solomon,
1984]. The numerical scheme for the steady state equations
builds on the Newton-Raphson method, considering overall
charge neutrality. Simulations typically comprise 25 ions and
an additional arbitrary number of proton hydrates. The basic
OASIS ideas are similar to other D-region ion models [Reid,
1976; Kopp, 1996; Turunen et al., 1996], and we refer to these
publications for more general discussions of the chemistry.
Reaction rate coefficients have largely been taken from the
compilation by Kopp [1996]; rates for the clustering processes
are discussed in the next section. Figure 2 shows the basic
positive ion reaction scheme. In order to illustrate the role of
individual reaction paths, the thickness of the arrows
resembles the individual rates. The rates correspond to quiet
daytime conditions (Figures 1b, 4, 5b).
In the mesosphere, primary ions are efficiently transformed
into proton hydrates. Starting out from O2+ and NO+, there are
two distinct reaction chains. Both are complex and indirect;
they involve three-body clustering with major neutral
constituents and fast two-body switching reactions of the
intermediates [Chakrabarty et al., 1978; Reid, 1989]. The O2+
chain was identified by Fehsenfeld and Ferguson [1969] and
Good et al. [1970]. The initial step that determines the overall
time constant is the formation of O4+. Subsequent switching
reactions lead to the first proton hydrates. Efficient hydration
in the NO+ chain involves intermediate clusters with the more
abundant species N2 and CO2 rather than direct clustering with
H2O [Ferguson, 1971]. A final switching step leads to the third
proton hydrate. Only for n > 3 is the direct hydration of
H+(H2O)n thought to dominate over indirect reactions via
intermediates with CO2 or N2 [Chakrabarty et al., 1978].
Considering the ionization sources, the NO+ chain is always
important, while the O2+ chain only gains a significant role
under perturbed conditions.
Proton hydrates largely govern the positive ion composition
below about 80-85 km. The above scheme of initial clustering
steps gives rise to a sharp decline in the proton hydrate
abundance above these altitudes. Several factors contribute to
this abrupt change: the need for three-body reactions, the
decline in water availability and enhanced recombination
connected to a ledge in the electron number density.
Laboratory data about proton hydrate reactions is available
only to about n = 6. The clustering to larger sizes will be
discussed in the next section.
The current version of the OASIS model uses only a very
crude scheme for negative ions. Only two species are
simulated: O2− and X−, the latter representing "heavy negative
ions" [Torkar and Friedrich, 1983]. This is justified for
simulations of the summer mesopause region since
contributions of negative ions to the charge budget are
negligible above 80 km under most conditions [Thomas and
Bowman, 1985b]. Neither does the current OASIS model
include metallic ions [e.g., Plane, 1991]. These ions are
usually rapidly depleted below 95 km [Kopp, 1997], due to
conversion to molecular ions and subsequent efficient
recombination [Plane et al., 1999]. Their chemistry is not
expected to be involved in the growth of proton hydrates. On
the other hand, Goldberg and Witt [1977] found families of
Fe+ hydrate clusters during mass spectrometric measurements
in NLC conditions.
OASIS does not comprise any feedback on the neutral
chemistry. Effects of ionization on the abundance of NO and
OH, and indirectly on O3, have been discussed e.g. by Rusch et
al. [1981] and Jackman et al. [1990]. As Lyman-α ionization
of NO is an important source of ions, the ion-chemical
production of NO will be incorporated into future versions of
OASIS. However, the local feedback of NO production on
ionization rates may not be large: Ion-chemical production of
NO is restricted to conditions of energetic particle
precipitation. Ionization of NO by Lyman-α, on the other
hand, plays a dominant role only under geomagnetically quiet
conditions.
4
Mesospheric ion-chemistry is complex. Many reaction rates
are uncertain or based solely on estimates. The latter applies
e.g. to many reactions involving intermediate clusters in the
NO+ chain that were originally suggested by Reid [1977] and
that are frequently used in current models. Other reactions are
still controversial. An example is the conversion of the primary
ions
O2+ + N2 → NO+ + NO
2.3. Cluster Growth
Under most mesospheric conditions, the number of ligands
in proton hydrates does not exceed n = 4-6. However, near the
cold polar summer mesopause, growth to larger clusters is
possible [Yang and Castleman, 1991]. Available in situ
measurements cover the range up to n = 21 [Björn and Arnold,
1981].
In thermodynamical terms, hydration processes are
governed by the Gibbs free energy of cluster growth
[Castleman and Tang, 1972]. Homogeneous nucleation of ice
particles is not feasible in the mesosphere as it exhibits a Gibbs
free energy barrier that cannot be overcome [Witt, 1969]. This
is illustrated for specific conditions by the dashed curves in
Figure 3. When growth takes place about an ion, Coulomb
interaction lowers the barrier, or completely removes it at
sufficiently high supersaturation [Castleman et al., 1978], as
shown by the solid curves in Figure 3.
Equivalently to the thermodynamic picture, cluster growth
can be described in terms of reaction kinetics by a balance
between the forward clustering reaction
(2)
and the unimolecular decomposition
H+(H2O)n + M → H+(H2O)n-1 + H2O + M.
k f , n −1 [H 2 O][M ]
k b, n [M ]
 ∆G n −1, n
= exp  −
kT


.


(4)
(1)
that has been suggested with a rate coefficient of 5×10-16 cm3 s1
[Rees, 1989], which would have a significant impact on
auroral production of NO. However, a laboratory measurement
by Matsuoka et al. [1981] failed to observe NO+ production
associated with the decay of O2+ and thus placed an upper limit
of 2×10-18 cm3 s-1. This low rate would render reaction 1
insignificant for atmospheric applications. A more
comprehensive discussion of critical reaction rates and a
comparison to D-region mass spectrometer measurements will
be published elsewhere.
Uncertainties are particularly serious for cold summer
mesopause conditions because of enhanced clustering
processes and limited knowledge about the temperature
dependence of many reactions. The large number of
interdependent reactions also complicates comparisons of
model results with atmospheric measurements such as radar or
in situ mass spectrometry. In particular, it is difficult to draw
unambiguous conclusions about individual rate coefficients
from these measurements. These limitations need to be kept in
mind and many conclusions must remain qualitative.
H+(H2O)n-1 + H2O + M → H+(H2O)n + M
Applying detailed balancing, the reaction coefficients for
forward reaction kf and backward reaction kb are related via the
Gibbs free energy change ∆Gn-1,n of the corresponding
clustering step [Castleman and Tang, 1972]:
(3)
In the present work, we apply the Thomson equation to
describe the Gibbs free energy [e.g., Mason, 1971; Holland
and Castleman, 1982]. Just as the Kelvin equation for
homogeneous nucleation, the Thomson equation is based on a
liquid drop formulation where the energetics of adding or
removing molecules is treated in terms of surface energy
(capillarity approximation). In this framework, the stepwise
change of Gibbs free energy is given by
 pH O
∆Gn −1, n = − kT ln  2
 pS
+
e2
32π 2ε 0
13
 32π M 2 

 σ
 + 

 3ρ 2 n 

ice



1
1 −
 ε
r

(5)
  ρice 1 3  4π  4 3

  M   3n  .

Here, pH2O is the ambient pressure of water vapor and pS is the
saturation pressure above ice. M is the mass of a water
molecule, ρice the mass density of ice. σ and εr are the surface
energy and dielectric constant of ice. The three terms in
equation 5 describe energy changes connected to evaporation,
surface area and electric field, respectively. Assuming a
spherical droplet with the mass density of bulk ice, the radius
of the nth cluster in equation 5 is
13
 3M n 

rn = 

 4π ρ ice 
(6)
We have used the parameters given by Shi et al. [1993] for 130
K (ρice = 937 kg m-3, σ = 0.00995 J m-2, εr = 132). The
temperature-dependent saturation vapor pressure by Marti and
Mauersberger [1993] is applied down to mesopause
temperatures.
The use of the classical Thomson theory obviously is an
approximation. Limitations of this approach have been
discussed by Castleman et al. [1972; 1978] and Holland and
Castleman [1982]. Chan and Mohnen [1980] have introduced
corrections to the Thomson equation for small clusters.
Sugiyama [1994] used the clustering formulation by Lothe and
Pound [1962]. Ultimately, more laboratory measurements
representative for mesospheric conditions will be needed to
better quantify the role of proton hydrates at the summer
mesopause.
In Figure 3 results of the Thomson equation have been
plotted for various temperatures. Shown is the Gibbs free
energy of formation of the nth cluster as
5
n
∆G0,n = ∑ ∆G n −1,n ,
(7)
1
with the energy of the ionic nucleus normalized to zero.
Atmospheric conditions resemble an altitude of 86 km with
Nair = 2×1014 cm-3 and a water vapor mixing ratio of 3 ppm.
The strong temperature dependence of the Gibbs free energy
barrier is obvious. For the conditions in Figure 3, the barrier
disappears at temperatures below about 130 K. At higher
temperatures, there is a distinct minimum in the free energy
curves that corresponds to the most abundant cluster sizes
under those conditions. Also shown in Figure 3 are the
significantly larger Gibbs free energy barriers for
homogeneous nucleation, i.e. cluster growth about a neutral
water molecule [Gadsden and Schröder, 1989].
Equation 4 defines the relative rates of forward and
backward reactions in the cluster chain. These equilibrium
considerations have been applied e.g. by D’Auria and Turco
[2001] to study the size distribution of proton hydrates in the
stratosphere. They find size distributions peaking at n = 6-14
in the temperature range 220-170 K.
However, in the mesosphere this equilibrium between 2 and
3 is no longer valid because the growth process must compete
with the termination of the clustering chain by efficient
recombination with free electrons
H+(H2O)n + e- → neutral products
(8)
Consequently, in order to determine cluster size distributions
in the D-region, it is not sufficient to know the relative rates of
reactions 2 and 3 as determined by detailed balancing. Rather,
absolute rate coefficients for clustering and recombination are
needed together with the local production rate of ions and
electrons. This is a serious limitation for mesospheric cluster
studies. Clustering rates have been measured in the laboratory
only up to n = 6 [Lau et al., 1982], restricted to CH4 as third
body and to temperatures above 200 K. At lower temperatures,
these measurements indicate that the forward step 2 can be
described as a quasi 2nd order reaction in H+(H2O)n and H2O,
independent of the third body density [Lau et al., 1982; Yang
and Castleman, 1991]. We follow the approach of Sugiyama
[1994] and adopt a Langevin rate of 1×10-9 cm3 s-1 for the
clustering step [Chesnavich et al., 1980]. The backward rate is
then obtained by detailed balancing.
Also laboratory results on recombination coefficients
(reaction 8) are available only up to n = 6 and at temperatures
above 200 K [Leu et al., 1973; Huang et al., 1978; Johnsen,
1993]. Sugiyama [1994] has extended these rates to larger
sizes by assuming a size-independent recombination
coefficient αn = 5×10-6 300 / T cm3 s-1. In the present work,
we deem a size-dependent recombination efficiency to be a
more realistic approach. In the limit of large clusters, the
coefficient should merge the result for ice particles. Therefore,
we apply the particle / charge capture formulation by Natanson
[1960] to cluster reaction 8 (see section 2.5). Resulting
recombination rates increase with n1/3. As discussed in section
4.1, these rates generally exceed Sugiyama’s values.
Considering the number of steps necessary for the
production of large hydrates, the extreme sensitivity of the
clustering process to temperature, water abundance and
electron density becomes evident: At each individual
clustering step, the water density enters into the forward
direction and the temperature via the Gibbs free energy into
the backward direction. Also at each step, recombination
threatens to terminate the entire clustering chain. At
temperatures around 130 K, reaction 2 and 3 are in a rough
balance and the clustering process is mathematically similar to
a "diffusion" in n-space. At lower temperatures, the backward
reaction loses importance and the growth process can be
regarded as an "advection" towards larger n, limited only by
the recombination reaction.
2.4. Ionic Nucleation
Two major candidates have been suggested to act as
embryos for ice particle nucleation near the mesopause:
meteoric material and cluster ions. As discussed in the
introduction, it is today widely assumed that meteoric smoke
particles are the dominant condensation nuclei in the
mesosphere [Turco et al., 1982, Keesee, 1989]. In the OASIS
model, we have incorporated the generation of ionic
condensation nuclei as a direct extension of the cluster growth
parameterization. This allows us to study the feasibility of
ionic contributions to nucleation under various mesopause
conditions. Arnold [1980] has described the growth of ice
particles on proton hydrates, and a detailed mechanism has
been outlined by Sugiyama [1994]. As discussed in the
previous section, the Gibbs free energy barrier for ionic
nucleation vanishes at sufficiently low temperature (solid
curves in Figure 3). However, this does imply a run-away
particle growth. Rather, the growth of the ion clusters has to
compete with efficient recombination reaction 8. Hence, a
stable ice particle can only be nucleated if the neutral particle
produced during recombination is beyond the Gibbs free
energy barrier for neutral particles (dashed curves in Figure 3).
Upon recombination, energy is released that needs to be
dissipated by sublimation of individual water molecules from
the cluster. Sugiyama [1994] suggested that 13.6 eV need to be
dissipated, which would require the sublimation of about 30
molecules. Hence, smaller clusters would be completely
destroyed by the recombination. However, while 13.6 eV is the
energy released by recombination of a free H+, significantly
less energy is released when a solvated proton recombinates.
Hence, the sublimation of as many as 30 molecules upon
cluster recombination as suggested by Sugiyama [1994] has to
be regarded as an upper limit. Nevertheless, we use this
number of 30 molecules in our model in order to arrive at a
conservative estimate of ionic nucleation efficiencies. The
above scheme defines a critical cluster size for ionic
nucleation: a proton hydrate must grow to a size of about 30
molecules above the neutral Gibbs free energy barrier before it
can successfully act as a condensation nucleus.
As can be expected from the discussion about clustering in
the previous section, the feasibility of ionic nucleation is
extremely sensitive to atmospheric conditions [Sugiyama,
1994]. Gumbel and Witt [2002] have discussed the
6
implementation of Sugiyama's nucleation scheme into the
OASIS model. For most mesospheric conditions, the critical
proton hydrate size is beyond 100. The OASIS output
comprises this critical size as well as the corresponding
nucleation rate J, which is equal to the rate of proton hydrates
reaching this critical size.
2.5. Charge Capture
In the OASIS model, parameterization of charge capture is
used in two connections: One is the recombination of proton
hydrate clusters with electrons discussed in section 2.3. The
other is the capture of charge by ice particles.
The capture of electrons or ions by aerosol particles has
Natanson
[1960].
Analytical
been
described
by
approximations for mesospheric conditions have been obtained
by Rapp [2000]. Capture rates are essentially determined by
the cross section of the particle, the thermal speed of the
electron or ion and an additional term describing the Coulomb
interaction. The capture rate of electrons or singly charged ions
by a neutral particle of radius r is

k0 = πr 2 c 1 +


e2
8ε 0 kT r




(9)
with the mean thermal speed of the electron or ion
c=
8kT
.
πm
(10)
This dependence on the thermal speed obviously makes the
capture of electrons a much more efficient process than the
capture of ions. The attractive capture of an electron or ion by
an oppositely charged particle with charge number q and
radius r is

| q | e 2 
ka = πr 2 c 1 +
.
 4πε kT r 
0


(11)
The current version of the model handles only singly
charged particles. This is reasonable for particle radii up to
about 10 nm since multiple charging only becomes feasible for
larger particles [Rapp and Lübken, 1999]. The current model
does not include photodetachement of electrons from the ice
particles.
3. Results
3.1. Ion Composition
Simulations in this paper were performed for mid-July
conditions at 67.9°N. Figure 4 shows the neutral atmospheric
input. A temperature profile was taken from the high-latitude
climatology of Lübken [1999] with a minimum of 129.5 K at
88 km. The NRL CHEM2D model [Summers et al., 1997]
provided a water profile featuring a mixing ratio of 3 ppm at
86 km. A representative NO profile was obtained from UARS
/ HALOE measurements [Siskind et al., 1998]. Odd oxygen
species were taken from NLC-related in situ measurements by
Gumbel et al. [1998].
Figure 5 shows resulting ion profiles corresponding to the
ionization conditions in Figure 1. Cluster ions have been
summarized in groups. The "intermediate" profile comprises
all intermediate clusters. Two characteristic altitudes can be
identified: the transition from NO+/O2+ dominance to
intermediate clusters, and the transition from intermediate
clusters to proton hydrates. Both altitudes are considerably
higher in the polar summer mesosphere than during other
(warmer) conditions. However, these transitions decrease in
altitude with increasing ionization. For the examples in Figure
5, the primary/intermediate transition is at 92, 91, and 85 km,
while the intermediate/hydrate transition decreases takes place
at 88, 85, and 84 km for twilight, daytime, and perturbed
conditions, respectively.
Generally, the overall number of ion species increases with
increasing ionization. However, this is not true for the larger
proton hydrates (n > 10). The reason is the larger
recombination coefficient of these clusters. This causes the
increased clustering from smaller ions to be more than
compensated by destructive recombination with the enhanced
electrons. This is also evident from the detailed cluster
distributions in Figure 6. Shown are the cluster spectra at 86
km for the three cases in Figure 5. While the small proton
hydrates correlate with the ionization just as the small ions,
this trend reverses for the larger proton hydrates. In section
3.3, the influence of ionization on the growth towards even
larger clusters and ice particles at lower temperatures will be
discussed.
3.2. Particle Effects
The impact of ice particles on the abundance of electrons
and ions has been studied e.g. by Reid [1990] and Rapp and
Lübken [2001]. With the OASIS model we have now the
possibility to look at particle effects on the detailed ion
composition. As an example, particle layers extending from 84
to 87 km have been introduced to the twilight conditions and
perturbed conditions in Figure 5a and c. We adopt a small
(2×102 cm-3) and a large (2×104 cm-3) population of particles
with radius 5 nm. Applying these particles to conditions of low
and high ionization (Figure 1a, c), some limiting cases of
particle effects can be studied.
In quiet twilight conditions with few particles (Figure 7a),
the particles cause a moderate bite-out in the electron density.
Nearly all particles carry negative charge. As the electron
density decreases, so does the recombination rate. This causes
an enhancement of the proton hydrate density. With 2×104 cm3
particles (Figure 7b), a huge electron bite-out is caused.
Given the large surplus of particles, now also the capture of
ions becomes efficient, and the proton hydrates are depleted as
well. Almost all charge is carried by the particles; there are
about equal numbers of positive and negative particles.
Nevertheless, most particles remain neutral. As compared to
the no-particle situation, the local charge density has increased
by more than an order of magnitude.
7
In the case of high ionization rates (Figure 8a, b), the
particle layers can cause only moderate depletions of the
electron density. Almost all particles become negatively
charged. And again, reduced recombination leads to an
enhancement in the proton hydrate densities.
Our findings are generally consistent with the charts given
by Rapp and Lübken [2001] that describe charging conditions
as a function of particle population and ionization. As a new
feature, the OASIS model can provide detailed insight into the
fate of individual ion and cluster species. Situations with both
enhancement and bit-outs in the ion profile have been
observed in situ in connection with ice particle layers [e.g.,
Lübken and Rapp, 2001] and will be discussed further in
section 4.1. Jensen and Thomas [1991] and Reid [1997] have
discussed the influence of particle charging on coagulation
processes. The situation in Figure 7b will enhance coagulation
by Coulomb attraction between positive and negative ice
particles. However, as in most mesospheric conditions
negative charge completely dominates the particle population,
coagulation processes in NLC and PMSE will usually be
suppressed by Coulomb repulsion.
It is important to consider the lifetime of particle charges
against recombination. This is highly dependent on the
ionization conditions. Ignoring possible contributions of
photodetachment processes, the lifetime of negatively charged
particles at 86 km in Figure 7 is several hours. This is critical
in terms of the steady state assumption, in particular when
considering sedimentation transport of the ice particles. For the
high-ionization conditions in Figure 8, on the other hand, the
lifetime of the negative particle charges reduces to tens of
seconds. This is short as compared to local transport processes.
In this situation, the concept of charged particles and their role
in mesopause phenomena must be considered as a fairly
dynamic rather than as a static idea.
3.3. Clustering and Nucleation
As pointed out in sections 2.3 and 2.4, cluster growth and
the feasibility of ionic nucleation depend strongly on
atmospheric conditions. Corresponding examples of OASIS
output are shown in Figure 9. Black lines denote cluster
spectra representative for about 86 km at quiet daytime
conditions (Figure 1b, Q = 10 cm-3 s-1, Nair = 2×1014 cm-3, 3
ppm water vapor) at various temperatures. As the temperature
decreases, the balance of the proton hydrate reactions
(equation 4) changes rapidly and gives rise to a strongly
enhanced clustering process. At 128 K, significant cluster
concentrations are found beyond the critical proton hydrate
size for nucleation (n ≈ 73) under these conditions. The ionic
nucleation rate J, i.e. the production rate of proton hydrates of
this critical size, is 0.2 cm-3 s-1. In other words, 2% of all ions
originally produced ultimately are converted into ice
condensation nuclei.
Capture of electrons by ice particles can enhance the
clustering process. This is shown by the gray curves in Figure
9. These curves represent cluster spectra under the same
conditions as above, but with a population of 103 cm-3 particles
of radius 10 nm added. The resulting decline in recombination
leads to a strong increase of the larger clusters. The nucleation
rate at 128 K increases to 1.5 cm-3 s-1, i.e. the ionic nucleation
efficiency becomes 15%. This feedback of existing particles
on the nucleation of new particles has been discussed by
Gumbel and Witt [2002] and will be considered further in
section 4.3.
4. Discussion
4.1. In Situ and Model Results
Figures 7 and 8 cover a wide range of ionospheric
conditions near the mesopause. In the presence of particles, we
generally find bite-outs in the electron density profiles. For the
less mobile positive ions, the picture is more complex as two
effects compete in the presence of particles: capture by
particles tends to reduce the ion density while the diminished
electron population leads to reduced recombination, which
tends to enhance the positive ions. Our results suggest that the
latter effect is dominant under most conditions. Ion bite-outs
are only to be expected under conditions of relatively low
ionization with large numbers of particles present.
The OASIS results are in general agreement with the
findings by Rapp and Lübken [2001]. In a comprehensive
study on particle charging, these authors find that indeed
almost all combinations of electron biteouts and ion
depletions/enhancements are achievable in the D-region
depending on particle population and ionospheric conditions
like ionization rate and recombination rate. Rapp and Lübken
[2001] show specifically that ion enhancements in the presence
of particles are most likely to occur when the ion population
exhibits heavy clusters (i.e., large effective recombination
coefficients). In the OASIS model, this effective
recombination coefficient is no longer a free parameter but
follows directly from the simulated ion composition (see
section 4.2).
A number of simultaneous in situ charge and particle
measurements are available in the summer mesosphere, as
reviewed by Lübken and Rapp [2001]. Detailed ion
composition measurements are available of rocket-borne ion
mass spectrometers flown by the Max-Planck-Institut für
Kernphysik, Germany [e.g., Björn and Arnold, 1981], the
University of Bern, Switzerland [e.g., Balsiger et al., 1996],
and NASA Goddard Space Flight Center [Goldberg and Witt,
1977]. In addition, measurements by positive ion probes can
give a coarse picture of particle influences on ion chemistry
without revealing details of the ion composition [e.g., Blix and
Thrane, 1993]. However, comparisons of model results to
these measurements are difficult. Some caveats are listed here:
Information about the particle population is usually restricted
to their pure existence (NLC, PMSE) and does not reveal
necessary details about number density or size distribution.
Also, particles are usually not restricted to optically observed
layers, but may be present in the form of sub-visible, possibly
persistent populations throughout the summer mesopause
region [von Zahn and Berger, 2002]. As opposed to model
simulations, in situ measurements obviously cannot allow for a
direct comparison of particle / no particle conditions. Hence,
the interpretation of measured structures in ion / electron
profiles often remains ambiguous in terms of the role of
8
particles or other processes like local dynamics. A limitation of
many mass spectrometer measurements is the limited mass
range (e.g., n < 10) as the behavior of smaller proton hydrates
is not necessarily representative for the entire cluster
population (Figures 6 and 9). Finally, temperature information
in the region is usually not available with sufficient altitude
resolution and accuracy.
Keeping these limitations in mind, a review of the data
collected by Lübken and Rapp [2001] reveals that electron
bite-outs are indeed a common phenomenon in connection
with NLC and PMSE. However, the data show that ion
enhancements are relatively scare under the same conditions.
Ten of the flights reviewed by Lübken and Rapp [2001] had
the capability to measure profiles of positive ions. Only two
measurements of these, both by electrostatic ion probes, reveal
clear ion enhancements [Blix and Thrane, 1993; Blix, 1999]. In
addition, the mass spectrometer measurement by Balsiger et
al. [1993] shows a narrow layer of enhanced proton hydrates
while the total positive ion density is depleted. Mass
spectrometer data presented by Kopp et al. [1985] (flight
S26/1) feature a broad layer of reduced electron densities at
80-90 km. A broad maximum of proton hydrate densities over
the same height range may be indicative of an enhancement
due to reduced recombination. On the other hand, ion
depletions are clearly observed on five flights [Pedersen et al,
1969; Johannessen and Krankowsky, 1972; Balsiger et al.,
1993; 1996; Blix, 1999]. These observations range from very
low (Q ≈ 0.006 cm-3 s-1) to moderate (Q ≈ 400 cm-3 s-1)
ionization conditions. This predominance of ion depletion in
connection with NLC or PMSE is seemingly in contrast with
the model results in section 3.2. The model simulations
suggest for most mesospheric conditions a tendency of ion
enhancement due to electron capture and reduced
recombination in the presence of ice particles. A possible
explanation of the depletions is a presence of populations of
very many (small) particles. This has also been suggested by
Balsiger et al. [1996] for the case of their ion depletion
measurement. It is interesting that the existence of large
amounts of particles has also been pointed out by Rapp et al.
[2002a] as a possible explanation for the occurrence of PMSE
up to very high ionization rates.
Some flight results are particularly interesting. Among the
measurements listed above, the strongest ion depletion
[Balsiger et al., 1996] and the clearest ion enhancement [Blix,
1999; Lübken and Rapp, 2001] in the presence of NLC were
observed only 39 minutes apart within the same rocket salvo
(NLC-93, Esrange, Sweden). Both flights took place under
very quiet twilight conditions. The substantial difference in the
particle effect on the positive ions points to the presence of
very different local particle populations during these two
flights, resulting in conditions as resembled in Figures 7a and
b. Kopp et al. [1984] presented mass spectrometer data
obtained during a simultaneous NLC and aurora event with
very high ionization rates (Q ≈ 6000 cm-3 s-1). For this flight,
no obvious depletions or enhancements were observed in the
particle layer, neither for the electrons nor for the positive ions.
This small effect is consistent with model results for high
ionization and small particle populations (Figure 8a). Finally,
Havnes et al. [1996] presented particle and electron
measurements in NLC/PMSE that showed a dominant
population of positively charged particles together with en
enhancement in the electron profile. These observations cannot
be explained in the framework of the present particle charging
model. It has been suggested that the photoeffect acting on ice
particles that contain metallic impurities is a source of large
positive particle charges. However, Rapp and Lübken [1999]
showed that this mechanism is not feasible in the mesopause
environment. The generation of positive particle populations
remains an open issue.
4.2. Effects of Heavy Clusters
Clustering processes have significant influence on charge
mobility in the mesopause region. Heavy charge carriers have
frequently been measured in situ by rocket-borne Gerdien
condensers [e.g., Croskey et al., 2001]. It is commonly
accepted that the reduction of charge mobility by ambipolar
interaction with charged particles is a crucial ingredient of
PMSE [Cho et al., 1992]. The results of the present work
suggest that the effect of particles on charge mobility is
twofold: In addition to direct ambipolar forces on the
electrons, the particles can reduce the mobility in the plasma
further by enhancing the growth of positive charge carriers to
significantly larger cluster sizes (Figure 9).
Cluster growth also has a strong effect on the overall
recombination rate. Figure 10 shows recombination
coefficients for the proton hydrates. Laboratory measurements
for n ≤ 6 have been scaled from the laboratory results to 130 K
using a T−1/2 dependence. The recombination parameterization
used in the present work [Natanson, 1960] is compared to the
value applied by Sugiyama [1994]. The use of more efficient
recombination in our simulations yields more conservative
estimates for the efficiencies of ion clustering and particle
nucleation.
Above 80 km, the effective recombination coefficient can
be defined as
α eff = ∑ N i α i ,
(13)
i
where Ni and αi are individual number densities and
recombination coefficients and i sums over all positive ion
species. (Below 80 km, a similar definition needs to take into
account contributions of negative ions [e.g., Montbriand and
Belrose, 1981].) αeff is an important parameter for the
assessment of D-region processes. It is applied e.g. in the
analysis of radar measurements in order to relate electron
densities and ion production rates [e.g. Hargreaves et al.,
1987; Kirkwood and Osepian, 1995]. It is also used in particle
charging models [e.g., Reid, 1990; Rapp and Lübken, 2001]
and in ion-chemical models that do not attempt to simulate the
entire set of mesospheric ion species [e.g., Rodriguez and
Inan, 1994; Rodger et al., 1998]. In steady state conditions, the
local electron number density Ne in the mesopause region is
determined by a balance between ion production rate Q and the
recombination of ions and electrons
9
N e = Q α eff .
(14)
The lifetime of electrons against recombination is determined
by the same parameters as
τ e = 1 Q α eff .
(15)
While equation 14 formally provides a simple relationship
between charge production and charge density, any application
to atmospheric problems is limited by the knowledge of αeff
[Gledhill, 1986]. Individual recombination coefficients vary
from a few 10-7 cm3 s-1 for light molecular ions to several 10-5
cm3 s-1 for proton hydrates (Figure 10). As a consequence,
detailed information about ion composition and, hence, about
ion chemistry is needed in order to access αeff. This is in
particular true in the summer mesopause region where cluster
formation in local temperature minima can push αeff to very
large values. The OASIS model can provide parameterizations
of αeff as a function of D-region conditions and particle
abundance.
4.3. Ionic Nucleation and Layered Structures
Based on currently available laboratory data on cluster
growth, we deem ionic nucleation not to be feasible under
average conditions near the high-latitude summer mesopause.
However, conditions required for ionic nucleation are within
the range of mesopause variability. Figure 11 summarizes the
feasibility of ionic nucleation as a function of the critical
parameters temperature, water abundance and ionization. Each
curve in the figure is representative for a specific water
concentration (0.3, 1, 3, 10, or 30 ppm). The area to the left of
a given curve specifies the range of temperature and ionization
that make ionic nucleation for this given water concentration
efficient. As "efficient" we have considered ionic nucleation
rates that exceed a threshold of J0 = 0.03 cm-3 s-1. This
corresponds to the nucleation of about 100 cm-3 ice particles
per hour, which can be regarded as a reasonable value
necessary for particle phenomena at the mesopause. Figure 11
is representative for an altitude of about 86 km (Nair = 2×1014
cm-3).
As discussed earlier by Turco et al. [1982] and Sugiyama
[1994], Figure 11 shows a dramatic sensitivity of ionic
nucleation to moderate changes in temperature or water
abundance. As for the dependence of the ionization rate, for a
given temperature and water abundance, there is a lower limit
of Q = J0, below which ionic nucleation cannot be feasible.
This is obviously due to the fact that the number of nucleations
cannot exceed the number of ions produced in the first place
(nucleation efficiency 100%). An upper Q limit exists since
increasing ionization leads to excessive electron concentrations
and recombination, thus causing a termination of the cluster
growth before the critical size is reached. For a temperature of
128 K and a water abundance of 3 ppm, limiting ionization
rates of 0.1 and 1000 cm-3 s-1 correspond to electron densities
of about 100 and 2×104 cm-3, respectively. It is interesting to
compare this to a recent analysis of the range of electron
densities over which PMSE can be observed. Rapp et al.
[2002a] find a lower limit of 300-500 cm-3 needed for the
detection of PMSE with currently available radar sensitivities.
An upper limit is reported at about 105 cm-3, presumably
caused by the availability of ice particles needed to reduce the
electron mobility.
Figures 9 and 11 show that small atmospheric variations in
temperature or water abundance can have dramatic effects on
ionic nucleation efficiencies. Ionic nucleation can become
feasible given relatively small deviations from average
conditions in the mesopause region as induced e.g. by gravity
waves. As an example, we can consider a mean condition with
a local temperature minimum of 130 K and a water abundance
of 3 ppm. We let a gravity wave of amplitude 5 K pass. In the
resulting minimum of 125 K, rapid cluster growth and ionic
nucleation of ice particles will proceed. (Note that a proton
hydrate with 100 ligands already corresponds to a particle of
about 1 nm in radius.) At a later stage the temperature
maximum of 135 K will pass. This will not destroy the newly
generated particles since they are stable once growth has
proceeded beyond the Gibbs free energy barrier corresponding
to this temperature (compare Figure 3). Efficient generation of
large proton hydrates in gravity waves has also been suggested
by Hall [1990]. However, that description did not incorporate
the role of the clusters as condensation nuclei and, as a
consequence, the heavy clusters were destroyed again after the
passage of the cold part of the wave. If, as in the present
model, one allows for particle nucleation, the passage of a
single wave period can cause a permanent change in the layer.
As already pointed out by Turco et al. [1982], it is
important to note that ionic nucleation is fast once atmospheric
conditions are sufficient. Typical reaction time constants for
the initial clustering steps and the subsequent proton hydrate
growth are of the order of seconds and milliseconds,
respectively. Hence, temperature structures can "immediately"
be converted into corresponding layers of small ice particles.
Because of the extreme temperature dependence, temperature
amplitudes of a few Kelvin can be sufficient to generate this
particle layering. The persistency of the temperature minima
together with the nucleation rate determines the total number
of particles that can be nucleated at a given location. Distinct
stratifications with sharp layers have been observed both in
NLC and in the distribution of charged particles connected to
PMSE [e.g., Havnes et al., 2001]. In order to study the
atmospheric relationship between temperature structure and
particle layering, simultaneous measurements with good
spatial resolution are needed. Unfortunately, current
climatologies of mesospheric temperatures must rely on
measurement techniques of limited height resolution [e.g.,
Lübken, 1999]. All available in situ temperature measurements
with sufficient height resolution (< 200 m) in the vicinity of
NLC show multiple temperature minima in the mesopause
region [Philbrick et al., 1984; Rapp et al., 2002b]. These
measurements consistently find the visible cloud (i.e., the
upper end of the particle size spectrum) in the lowermost
minimum of these temperature profiles. On the other hand, it is
important to note that a very efficient ionic generation of
condensation nuclei can actually counteract the existence of
10
visible clouds. As many particles compete for the available
water vapor, average particle sizes may remain small and
subvisible.
As Figure 11 shows, a critical limitation of this idea of rapid
particle nucleation in local temperature minima is the
availability of water vapor. A water abundance of 1 ppm or
below would drastically reduce the feasibility of sufficient
cluster growth. Although recent measurements and models
predict higher abundances of water vapor in the upper
mesosphere than previously assumed, freeze-drying is
expected to significantly redistribute the water out of the
NLC/PMSE region [Summers et al., 2001; von Zahn and
Berger, 2002]. In a detailed assessment of the water budget,
von Zahn and Berger [2002] predict a persistent dehydration
of large parts of the region to water mixing ratios far below 1
ppm.
4.4. Feedback Mechanisms
We have outlined above how particles through electron
capture can enhance the clustering process and, hence, ionic
nucleation. This constitutes a positive feedback mechanism
where the existence of ice particles boosts the nucleation of
new ice particles. Gumbel and Witt [2002] have discussed this
feedback in detail. They show that the presence of ice particles
can extend the feasibility region in Figure 11 to significantly
larger ionization rates and to slightly higher temperatures. This
positive feedback has the potential of amplifying initial
mesospheric structures e.g. in the ice particle population or in
the electron density.
This is interesting for PMSE. A pre-condition for the
presence of these echoes are structures in the electron density
on the scale of the radar half wavelength (~meters) [Cho and
Röttger, 1997]. Charging of ice particles is thought to provide
the reduction of charge mobility that is necessary to maintain
such small-scale structures [Cho et al., 1992]. However, it is
still an open question how these structures can be generated in
the first place. In this connection, a positive feedback
mechanism is of obvious interest. However, Gumbel and Witt
[2002] conclude that this feedback cannot be efficient enough
to generate structures at such small scales as required for
PMSE: In order to generate meter-scale structures, positive
feedback processes must be fast enough to compete with the
destruction of structures by particle diffusion. As a
consequence, of the order of 1000 cm-3 particles must be
nucleated within minutes in order to provide a sufficiently fast
feedback on the electron density [Gumbel and Witt, 2002].
This is not feasible under realistic mesospheric conditions.
Another potential ice particle feedback on cluster growth
can be discussed. This is connected to the fate of atomic
oxygen. The reaction
O4+ + O → O2+ + O3
(16)
efficiently disrupts the initiating steps of the O2+ clustering
chain (Figure 2). It has thus been suggested that atomic oxygen
densities may have a significant influence on the distribution
of electrons and proton hydrates in the upper mesosphere
[Fehsenfeld and Ferguson, 1972; Friedrich et al., 1999].
Moreover, in situ measurements indicate that ice particles may
act as a sink for atomic oxygen [Thomas, 1991; Gumbel et al.,
1998]. Hence, a feedback scenario becomes possible where ice
particles cause a biteout in the O density, which then enhances
proton hydrate clustering and ionic nucleation by suppressing
the termination reaction 16. We have studied this idea by
reducing the O input profile (Figure 4) by a factor 10. Figure
12 shows that the effect on larger proton hydrates (n > 10)
remains minor. This is to be expected since reaction 16 does
not have any major effect on the dominant NO+ cluster chain.
We conclude that an ion-chemical feedback mechanism
involving atomic oxygen capture is not feasible in the
mesopause region. It should be noted, however, that while we
do not find a strong influence of atomic oxygen near the
mesopause, more significant effects are expected in the lower
parts of the ionosphere where atomic oxygen is a major player
in the negative ion chemistry [e.g., Thomas and Bowman,
1985b; Turunen et al., 1996].
5. Conclusions
The variety of mesospheric interactions involving charge
requires combined studies of ions, clusters and particles. We
have described an approach that includes these constituents in
a single ion-chemical model. By parameterizing the growth of
proton hydrates up to the sizes of ice condensation nuclei, the
OASIS model for the first time bridges the ion-chemical gap
between molecular ions and nanometer-sized particles. While
describing the possibility of ionic nucleation, our model is not
a complete particle growth model. Nevertheless, the ionchemical features presented here can readily be coupled with
existing NLC/PMSE models. This can be done either
interactively or by providing parameterizations as a function of
mesospheric conditions. Such parameterizations can describe
effective recombination coefficients or effective ion masses,
which are important quantities for simulations of particle
charging [e.g., Rapp and Lübken, 2001]. Our detailed
treatment of the clustering scheme can also provide improved
parameterizations of ionic nucleation rates as a function of
temperature, water abundance and background ionization (e.g.,
Figure 11). Many current NLC/PMSE models use
parameterizations for ionic nucleation that do not directly
involve the detailed ion chemistry behind the cluster growth
[e.g., Turco et al., 1982; Sugiyama, 1994; Rapp et al., 2002b].
Other models do not allow for the option of ionic nucleation at
all [e.g., Klostermeyer, 1998; Berger and von Zahn, 2002].
As for the feasibility of ionic nucleation, we confirm earlier
findings e.g. by Turco et al. [1982] and Keesee [1989] that
clustering efficiencies are not sufficient under average
mesospheric conditions. Hence, considering the wide range of
conditions under which NLC and PMSE are observed, ionic
nucleation is unlikely to serve as major nucleation mechanism
in the mesopause region. In the competition between meteoric
material and cluster ions as mesospheric condensation nuclei,
strong freeze-drying is particularly critical for the ionic
nucleation [von Zahn and Berger, 2002]. Nevertheless, we
have shown that moderate deviations from the mean state may
well shift conditions towards sufficient ionic nucleation
efficiencies as defined in Figure 11. As a consequence, the
11
action of gravity waves can lead to rapid generation of particle
layers in local temperature minima. Because of the extreme
sensitivity of ionic nucleation on mesospheric parameters,
stratifications of temperature or water abundance will result in
steep vertical gradients of ionic particle generation. Very
distinct particle layers are frequently observed both by optical
and charged particle detection [e.g. Rapp et al., 2002b; Havnes
et al., 1996].
Major new results of the OASIS model concern the large
impact of the presence of ice particles on ion chemistry and
composition. As the particles efficiently capture electrons,
ionic reactions, and in particular cluster growth, are less
subject to termination by recombination. Hence, for most
mesospheric conditions, we suggest an enhancement of
positive ion densities in the presence of particles. Positive ion
depletions, on the other hand, can generally be expected in
situations when particle number densities exceed the
undisturbed electron concentrations (Figures 7 and 8). These
results are similar to model results e.g. by Reid [1990] and
Rapp and Lübken [2001]. As compared to these models, the
OASIS model has the advantage of inherently simulating
quantities like the efficient recombination coefficient or the
mean ion mass rather than introducing them as free
parameters. In situ rocket data suggest that in a majority of
simultaneous particle and charge measurements, depletions
rather than enhancements of positive ions are found. This is
seemingly in contrast to our model results. A possible
explanation is the presence of populations with significantly
more (smaller) particles than commonly assumed.
The enhancing effect of ice particles on cluster growth
becomes most prominent when it comes to the possibility of
ionic nucleation. Under appropriate mesospheric conditions,
the presence of ice particles can strongly enhance the
generation of ionic condensation nuclei for new particles. This
was illustrated in Figure 9. This feedback mechanism has not
been described in earlier mesospheric particle models. It has
the potential of amplifying spatial fluctuations in the
mesospheric dusty plasma. However, following the discussions
by Gumbel and Witt [2002], we deem this
particle/cluster/nucleation feedback not to be efficient enough
to generate and maintain fluctuations on the small spatial
scales needed for PMSE.
The above statements are based on currently available data
on the kinetics and thermodynamics of ionic reactions and
proton hydrate growth. Future laboratory studies of ionchemical reaction rates may readily shift these statements
closer to or further away from feasibility. Laboratory studies
under conditions representative for the cold summer
mesopause are badly needed in order to clarify the role of ionchemical processes in the phenomena of this part of the
atmosphere.
Acknowledgments. We wish to thank M. Rapp, J. M. C.
Plane, R. R. Meier, and J. Bishop for helpful discussions.
Funding for this work at the Naval Research Laboratory has
been provided by the Office of Naval Research and the NASA
Office of Space Science. Work has also been carried out at
NASA's Goddard Space Flight Center with support from R. A.
Goldberg and the National Research Council.
References
Arnold, F., Ion-induced nucleation of atmospheric water vapor
at the mesopause, Planet. Space Sci., 28, 1003-1009, 1980.
Balsiger, F., E. Kopp, M. Friedrich, K. M. Torkar, and U.
Wälchli, Small scale structure of O2+ and proton hydrates in
a noctilucent cloud and polar mesospheric summer echo of
August 9/10 1991 above Kiruna, Geophys. Res. Lett., 20,
2315-2318, 1993.
Balsiger, F., E. Kopp, M. Friedrich, K. M. Torkar, U. Wälchli,
and G. Witt, Positive ion depletion in a noctilucent cloud,
Geophys. Res. Lett., 23, 93-96, 1996.
Berger, U., and U. von Zahn, Icy particles in the summer
mesopause region: 3-D modeling of their environment and
2-D modeling of their transport, J. Geophys. Res., in print,
2002.
Björn, L. G., and F. Arnold, Mass spectrometric detection of
precondensation nuclei at the arctic summer mesopause,
Geophys. Res. Lett., 8, 1167-1170, 1981.
Blix, T. A., The importance of charged aerosols in the polar
mesosphere in connection with noctilucent clouds and polar
mesosphere summer echoes, Adv. Space Res., 24, 16451654, 1999.
Blix, T. A., and E. V. Thrane, Noctilucent clouds and regions
with polar mesospheric summer echoes studied by means of
rocket-borne electron and ion DC-probes, Geophys. Res.
Lett., 20, 2303-2306, 1993.
Brasseur, G., and S. Solomon, Aeronomy of the middle
atmosphere, D. Reidel Publishing Company, Dordrecht,
1984.
Castleman, Jr., A. W., and I. N. Tang, Role of small clusters in
nucleation about ions, J. Chem. Phys., 57, 3629-3638, 1972.
Castleman Jr., A. W., P. M. Holland, and R. G. Keesee, The
properties of ion clusters and their relationship to
heteromolecular nucleation, J. Chem. Phys., 68, 1760-1767,
1978.
Chakrabarty, D. K., P. Chakrabarty, and G. Witt, An attempt to
identify the obscured paths of water cluster ion build-up in
the D-region, J. Atm. Terr. Phys., 40, 437-442, 1978.
Chan, L. Y., and V. A. Mohnen, Ion nucleation theory, J.
Atmos. Sci., 37, 2323-2331, 1980.
Chesnavich, W. J., T. Su, and M. T. Bowers, Collisions in a
noncentral field: A variational and trajectory investigation
of ion-dipole capture, J. Chem. Phys., 72, 2641-2655, 1980.
Cho, J. Y. N., T. M. Hall, and M. C. Kelley, On the role of
charged aerosols in polar mesosphere summer echoes, J.
Geophys. Res., 97, 875-886, 1992.
Cho, J. Y. N., and J. Röttger, An updated review of polar
mesosphere summer echoes: Observation, theory, and their
12
relationship to noctilucent clouds and subvisible aerosols, J.
Geophys. Res., 102, 2001-2020, 1997.
Croskey, C. L., J. D. Mitchell, M. Friedrich, K. M. Torkar, U.P. Hoppe, and R. A. Goldberg, Electrical structure of PMSE
and NLC regions during the DROPPS program, Geophys.
Res. Lett., 28, 1427-1430, 2001.
D'Auria, R., and R. P. Turco, Ionic clusters in the polar winter
stratosphere, Geophys. Res. Lett., 28, 3871-3874, 2001.
Fehsenfeld, F. C., and E. E. Ferguson, Origin of water cluster
ions in the in the D region, J. Geophys. Res., 74, 2217-2222,
1969.
Fehsenfeld, F. C., and E. E. Ferguson, Recent laboratory
measurements of D- and E-region ion-neutral reactions,
Radio Sci., 7, 113-115, 1972.
Ferguson, E. E., D region ion chemistry, Rev. Geophys., 9,
997-1008, 1971.
Frahm, R. A., J. D. Winningham, J. R. Sharber, R. Link, G.
Crowley, E. E. Gaines, D. L. Chenette, B. J. Anderson, and
T. A. Potemra, The diffuse aurora: A significant source of
ionization in the middle atmosphere, J. Geophys. Res., 102,
28203-2821, 1997.
Friedrich, M., J. Gumbel, and R. Pilgram, Atomic oxygen in
the mesosphere and its relevance for the ionosphere: A
summary of empirical evidence, Proc. 14th ESA Symposium
on European Rocket and Balloon Programmes and Related
Research (ESA SP-437), 465-470, 1999.
Fritzenwallner, J., Globales Modell zur Verteilung der
positiven und negativen Ionen in der unteren Ionosphäre der
Erde, Ph.D. thesis, University of Bern, Switzerland, 1997.
Gadsden, M., and W. Schröder, Noctilucent Clouds, Springer
Verlag, Berlin, 1989.
Gledhill, J. A., The effective recombination coefficient of
electrons in the ionosphere between 50 and 150 km, Radio
Sci., 21, 339-408, 1986.
Goldberg, R. A., and G. Witt, Ion composition in a noctilucent
cloud region, J. Geophys. Res., 82, 2619-2627, 1977.
Good, A., D. A. Durden, and P. Kebarle, Mechanism and rate
constants of ion-molecule reactions leading to formation of
H+(H2O)n in moist oxygen and air, J. Chem. Phys., 52, 222229, 1970.
Gumbel, J., D. P. Murtagh, G. Witt, P. J. Espy and F. J.
Schmidlin, Odd oxygen measurements during the NLC-93
rocket campaign, J. Geophys. Res., 103, 23399-23414,
1998.
Gumbel, J., and G. Witt, Cluster ions and ice particle
nucleation: Positive feedback at the summer mesospause,
Geophys. Res. Lett., in print, 2002.
Hall, C. M., Modifications of the energy-wave number
spectrum for heavy proton hydrates as tracers for isotropic
turbulence at the summer mesopause, J.Geophys. Res., 95,
5549-5556, 1990.
Hargreaves, J. K., H. Ranta, A. Ranta, E. Turunen, and T.
Turunen, Observation of the polar cap absorption event of
February 1984 by the EISCAT incoherent scatter radar,
Planet. Space Sci., 35, 947-958, 1987.
Havnes, O., J. Troim, T. Blix, W. Mortensen, L. I. Naesheim,
E. Thrane, and T. Tönnesen, First detection of charged dust
particles in the Earth's mesosphere, J. Geophys. Res., 101,
10839-10847, 1996.
Havnes, O., A. Brattli, T. Aslaksen, W. Singer, R. Latteck, T.
Blix, E. Thrane, and J. Troim, First common volume
observations of layered plasma structures and polar
mesosperic summer echoes by rocket and lidar, Geophys.
Res. Lett., 28, 1419-1422, 2001.
Hedin, A. E., Extension of the MSIS thermosphere model into
the middle and lower atmosphere, J. Geophys. Res., 96,
1159-1172, 1991.
Heroux, L., and H. E. Hinteregger, Aeronomical reference
spectrum for solar UV below 2000 Å, J. Geophys. Res., 83,
5305-5308, 1978.
Holland, P. M., and A. W. Castleman Jr., Thomson equation
revisited in light of ion-clustering experiments, J. Phys.
Chem., 86, 4181-4188, 1982.
Huang, C.-M., M. Whitaker, M. A. Biondi, and R. Johnsen,
Electron-temperature dependence of electrons with
H3O+⋅(H2O)n-series ions, Phys. Rev. A, 18, 64-67, 1978.
Jackman, C. H., A. R. Douglass, R. B. Rood, R. D. McPeters,
and P. E. Meade, Effect of solar proton events on the middle
atmosphere during the past two solar cycles as computed
using a two-dimensional model, J. Geophys. Res., 95, 74177428, 1990.
Jensen, E. J., and G. E. Thomas, A growth-sedimentation
model of polar mesospheric clouds: Comparison with SME
measurements, J. Geophys. Res., 93, 2461-2473, 1988.
Jensen, E. J., and G. E. Thomas, Charging of mesospheric
particles: Implications for electron density and particle
coagulation, J. Geophys. Res., 96, 18603-18615, 1991.
Jensen, E. J., and G. E. Thomas, Numerical simulations of the
effect of gravity waves on noctilucent clouds, J. Geophys.
Res., 99, 3421-3430, 1994.
Johannessen A., and D. Krankowsky, Positive ion composition
measurement in the upper mesosphere and lower
thermosphere at a high latitude during summer, J. Geophys.
Res., 77, 2888-2901, 1972.
Johnsen, R., Electron-temperature dependence of the
recombination of H3O+(H2O)n ions with electrons, J. Chem.
Phys., 98, 5390-5395, 1993.
Keesee, R. G., Nucleation and particle formation in the upper
atmosphere, J. Geophys. Res., 94, 14683-14692, 1989.
Kirkwood, S., and A. Osepian, Quantitative studies of
energetic particle precipitation using incoherent scatter
radar, J. Geomag. Geoelectr., 47, 783-799, 1995.
Klostermeyer, J., A simple model of the ice particle size
distribution in noctilucent clouds, J. Geophys. Res., 103,
28743-28752, 1998.
Kopp, E., Electron and ion densities, in The Upper
Atmosphere, Data Analysis and Interpretation, W.
Diminger, G.K. Hartman and R. Leitinger (eds.), pp. 625635, Springer, Berlin, 1996.
Kopp, E., On the abundance of metal ions in the lower
ionosphere, J. Geophys. Res., 102, 9667-9674, 1997.
Kopp, E., H. Ramseyer, and L. G. Björn, Positive ion
composition and electron density in a combined auroral and
NLC event, Adv. Space Res., 4, (4)157-(4)161, 1984.
Kopp, E., P. Eberhardt, U. Hermann, and L. G. Björn, Positive
ion composition of the high-latitude summer D region with
noctilucent clouds, J. Geophys. Res., 90, 13041-13053,
1985.
13
Lau, Y. K., S. Ikuta, and P. Kebarle, Thermodynamics ad
kinetics of the gas-phase reactions: H3O+(H2O)n-1 + H2O =
H3O+(H2O)n, J. Am. Chem. Soc., 104, 1462-1469, 1982.
Lean, J., Solar ultraviolet irradiance variations: A review, J.
Geophys. Res., 92, 839-868, 1987.
Leu, M. T., M. A. Biondi, and R. Johnsen, Measurements of
the recombination of electrons with H3O+⋅(H2O)n-series
ions, Phys. Rev. A, 7, 292-298, 1973.
Lothe, J., and G. M. Pound, Reconsiderations of nucleation
theory, J. Chem. Phys., 36, 2080-2085, 1962.
Lübken, F.-J., Thermal structure of the Arctic summer
mesosphere, J. Geophys. Res., 104, 9135-9149, 1999.
Lübken, F.-J., G. Lehmacher, T. Blix, U.-P. Hoppe, E. Thrane,
J. Cho, and W. Swartz, First in-situ observations of neutral
and plasma density fluctuations within a PMSE layer,
Geophys. Res. Lett., 20, 2311-2314, 1993.
Lübken, F.-J., and M. Rapp, Modelling of particle charging in
the polar summer mesosphere: Part 2 - Application to
measurements, J. Atm. Sol. Terr. Phys., 63, 771-780, 2001.
Marti, J., and K. Mauersberger, A survey and new
measurements of ice vapour pressure at temperatures
between 170 and 250 K, Geophys. Res. Lett., 20, 363-366,
1993.
Mason, B. J., The Physics of Clouds, 2nd edition, Clarendon
Press, Oxford, 1971.
Matsuoka, S., H. Nakamura, and T. Tamura, Ion-molecule
reactions of N3+, N4+, O2+, and NO2+ in nitrogen containing
traces of oxygen, J. Chem. Phys., 75, 681-689, 1981.
Meier, R. R., and P. Mange, Spatial and temporal variations of
the Lyman-alpha airglow and related atomic hydrogen
distributions, Planet. Space Sci., 21, 309-327, 1973.
Montbriand, L. E., and J. S. Belrose, Changes in electron
precipitation inferred from spectra deduced from D region
electron densities during a post-magnetic storm effect, J.
Geophys. Res., 81, 2213-2222, 1976.
Natanson, G. L., On the theory of the charging of amicroscopic
aerosol particles as a result of capture of gas ions, Sov.
Phys. Tech. Phys., 5, 538-551, 1960.
Parthasarathy, R., Mesopause dust as a sink for ionization, J.
Geophys. Res., 81, 2392-2396, 1976.
Pedersen, A., J. Troim, and J. Kane, Rocket measurement
showing removal of electrons above the mesopause in
summer at high latitudes, Planet. Space Sci., 18, 945-947,
1969.
Pfaff, R., R. Holzworth, R. Goldberg, H. Freudenreich, H.
Voss, C. Croskey, J. Mitchell, J. Gumbel, S. Bounds, W.
Singer, and R. Latteck, Rocket probe observations of
electric field irregularities in the polar summer mesosphere,
Geophys. Res. Lett., 28, 1431-1434, 2001.
Plane, J. M. C., The chemistry of meteoric metals in the earth's
upper atmosphere, Int. Rev. Phys. Chem., 10, 55-106, 1991.
Plane, J. M. C., The role of sodium bicarbonate in the
nucleation of noctilucent clouds, Ann. Geophys., 18, 807814, 2000.
Plane, J. M. C., R. M. Cox, and R. J. Rollason, Metallic layers
in the mesopause and lower thermosphere region, Adv.
Space Res., 24, 1559-1570, 1999.
Philbrick, C., J. Barnett, R. Gerndt, D. Offermann, P. Schlyter,
F. J. Schmidlin, and G. Witt, Temperature measurements
during the CAMP program, Adv. Space Res., 4, 153-156,
1984.
Rapp, M., Capture rates of electrons and positive ions by
mesospheric aereosol particles, J. Aerosol Sci., 11, 13671369, 2000.
Rapp, M., and F.-J. Lübken, Modelling of positively charged
aerosols in the polar summer mesopause region, Earth,
Planets and Space, 51, 799-807, 1999.
Rapp, M., and F.-J. Lübken, Modelling of particle charging in
the polar summer mesosphere: Part 1 - General results, J.
Atmos. Sol. Terr. Phys., 63, 759-770, 2001.
Rapp, M., J. Gumbel, F.-J. Lübken, and R. Latteck, D-region
electron number densities limits for the existence of polar
mesosphere summer echoes, J. Geophys. Res., in print,
2002a.
Rapp, M., F.-J. Lübken, A. Müllemann, G. E. Thomas, and E.
J. Jensen, Small scale temperature variations in the vicinity
of NLC: Experimental and model results, J. Geophys. Res.,
in print, 2002b.
Rees, M. H., Auroral electrons, Space Sci. Rev., 10, 413-441,
1969.
Rees, M. H., On the interaction of auroral protons with the
Earth's atmosphere, Planet. Space Sci., 30, 463-472, 1982.
Rees, M. H., Physics and chemistry of the upper atmosphere,
Cambridge University Press, New York, 1989.
Reid, G. C., Ion chemistry in the D region, in Advances in
Atomic and Molecular Physics, edited by D. R. Bates and
B. Bederson, pp. 375-413, Academic, San Diego, CA, 1976.
Reid, G. C., The production of water-cluster positive ions in
the quiet daytime D region, Planet. Space Sci., 25, 275-290,
1977.
Reid, G. C., Ion chemistry of the cold summer mesopause
region, J. Geophys. Res., 94, 14653-14660, 1989.
Reid, G. C., Ice particles and electron "bite-outs" at the
summer polar mesosphere, J. Geophys. Res., 95, 1389113896, 1990.
Reid, G. C., On the influence of electrostatic charging on
coagulation of dust and ice particles in the upper
mesosphere, Geophys. Res. Lett., 24, 1095-1098, 1997.
Rodger, C. J., O. A. Molchanov, and N. R. Thomson,
Relaxation of transient ionization in the lower ionosphere, J.
Geophys. Res., 103, 6969-6975, 1998.
Rodriguez, J. V., and U. S. Inan, Electron density changes in
the nighttime D region due to heating by very-lowfrequency transmitters, Geophys. Res. Lett., 21, 93-96,
1994.
Rusch, D. W., J.-C. Gérard, S. Solomon, P. J. Crutzen, and G.
C. Reid, The effect of particle precipitation events on the
neutral and ion chemistry of the middle atmosphere - I. Odd
nitrogen, Planet. Space Sci., 29, 767-774, 1981.
Shi, Z., J. V. Ford, S. Wei, and A. W. Castleman, Jr., Water
clusters: Contributions of binding energy and entropy to
stability, J. Chem. Phys., 99, 8009-8015, 1993.
Siskind, D. E., C. A. Barth, and J. M. Russell III, A
climatology of nitric oxide in the mesosphere and
thermosphere, Adv. Space Res., 21, 1353-1362, 1998.
Strickland, D. J., J. R. Jasperse, and J. A. Whalen, Dependence
of auroral FUV emissions on the incident electron spectrum
14
and neutral atmosphere, J. Geophys. Res., 88, 8051-8062,
1983.
Strobel, D. F., T. R. Young, R. R. Meier, T. P. Coffey, and A.
W. Ali, The nighttime ionosphere: E region and lower F
region, J. Geophys. Res., 79, 3171-3178, 1974.
Sugiyama, T., Ion-recombination nucleation and growth of ice
particles in noctilucent clouds, J. Geophys. Res., 99, 39153929, 1994.
Summers, M. E., D. E. Siskind, J. T. Bacmeister, R. R.
Conway, S. E. Zasadil, and D. F. Strobel, Seasonal variation
of middle atmospheric CH4 and H2O with a new chemicaldynamical model, J. Geophys. Res., 102, 3503-3526, 1997.
Summers, M. E., R. R. Conway, C. R. Englert, D. E. Siskind,
M. H. Stevens J. M. Russell III, L. L. Gordley, and M. J.
McHugh, Discovery of a water vapor layer in the arctic
summer mesosphere: Implications for polar mesospheric
clouds, Geophys. Res. Lett., 28, 3601-3604, 2001.
Thomas, G. E., Mesospheric clouds and the physics of the
mesopause region, Rev. Geophys., 29, 553-575, 1991.
Thomas, L., and M. R. Bowman, A study of pre-sunrise
changes in negative ions and electrons in the D-region, Ann.
Geophys., 4, 219-228, 1985a.
Thomas, L., and M. R. Bowman, Model studies of the Dregion negative ion composition during day-time and nighttime, J. Atmos. Terr. Phys., 47, 547-556, 1985b.
Torkar, K. M., and M. Friedrich, Tests of an ion-chemical
model of the D- and lower E-region, J. Atm. Terr. Phys., 45,
369-385, 1983.
Torr, M. R., and D. G. Torr, Ionization frequencies for solar
cycle 21: revised, J. Geophys. Res., 90, 6675-6678, 1985.
Turco, R. P., O. B. Toon, R. C. Qhitten, R. G. Keesee, and D.
Hollenbach, Noctilucent clouds: simulation studies of their
genesis, properties and global influence, Planet. Space Sci.,
30, 1147-1181, 1982.
Turunen, E., H. Matveinen, J. Tolvanen, and H. Ranta, Dregion ion chemistry model, in STEP Handbook of
Ionospheric Models, R. W. Schunk (ed.), pp. 1-25, 1996.
Ulwick, J. C., K. D. Baker, M. C. Kelley, B. B. Balsley, and
W. L. Ecklund, Comparison of simultaneous MST radar and
electron density probe measurements during STATE, J.
Geophys. Res., 93, 6989-7000, 1988.
von Zahn U., and U. Berger, Freeze-drying at the summer
polar mesopause: Its likely persistence and consequences, J.
Geophys. Res., this issue, 2002.
Witt, G., The nature of noctilucent clouds, Space Research, 9,
157-169, 1969.
_______________
M. Friedrich, Department of Communications and Wave
Propagation, Technical University Graz, Inffeldgasse 12, 8010
Graz, Austria. ([email protected])
J. Gumbel, G. Witt, Department of Meteorology, Stockholm
University, 10691 Stockholm, Sweden. ([email protected],
[email protected])
D. E. Siskind, E. O. Hulburt Center for Space Research,
Naval Research Laboratory, Code 7640, Washington, DC
20375-5320. ([email protected])
K. M. Torkar, Space Research Institute, Austrian Academy
of Sciences, Schmiedlstraße 6, 8042 Graz, Austria.
(klaus.torkar@ oeaw.ac.at)
15
direct
Ly man-α
geocorona
Altitude [km]
95
90
total
85
80
χ = 91
75
o
a
GCR
direct
Ly man-α
geocorona
Altitude [km]
95
90
total
85
80
75
χ = 47
GCR
direct
Ly man-α
geocorona
95
Altitude [km]
other direct
o
b
other direct
90
85
e-precip.
total
80
75
10
χ = 47
GCR
-2
10
-1
10
0
10
1
10
-3
2
o
c
10
3
10
4
-1
Production rate [cm s ]
Figure 1. Ionization profiles representative for highlatitude summer mesosphere conditions: (a) quiet twilight
conditions, (b) quiet daytime conditions, (c) electron
precipitation conditions.
16
Figure 2. Basic ion-chemistry scheme of the OASIS model. Arrow thicknesses represent reaction rates for quiet daytime
conditions (Figure 1b). Not shown are recombination reactions with electrons.
-1
∆G0,n [kJ mol ]
200
135 K
100
130 K
0
125 K
135 K
-100
-200
130 K
-300
125 K
0
50
100
150
200
250
n
Figure 3. Gibbs free energy of cluster / particle formation
at different temperatures. Dashed lines denote homogeneous
nucleation, solid lines denote growth about an ion nucleus.
Atmospheric conditions are Nair = 2×1014 cm-3 and a water
vapor mixing ratio of 3 ppm.
17
Water [ppm ]
v
100
0
1
2
3
4
5
92
H2O
85
80
T
75
125
150 175 200
Temperature [K]
225
88
86
H+(H2O)1-3
84
H+(H2O)4-6
H+(H2O)7-10
H+(H2O)>10
b
90
Altitude [km]
85
80
O2(1∆)
75
10
7
10
8
10
9
10
10
10
e-
O+2
90
O
O3
b
intermediate
92
NO
o
80
94
95
Altitude [km]
χ = 91
82
250
100
70
6
10
a
e-
NO+
90
90
Altitude [km]
Altitude [km]
intermediate
a
95
70
100
94
6
11
10
NO+
88
86
84
82
12
-3
χ = 47
80
94
Number density [cm ]
o
c
92
Figure 4. Input profiles for temperature and neutral
species. See text for references.
intermediate
NO+
Altitude [km]
90
88
O+2
86
84
e-
82
χ = 47
80
10
1
10
2
10
3
10
4
o
10
5
Number density [cm -3]
Figure 5. Ion composition profiles simulated for the
ionization conditions in Figure 1 and the neutral atmosphere
in Figure 4: (a) quiet twilight conditions, (b) quiet daytime
conditions, (c) perturbed daytime conditions. The
"intermediate" profile refers to the sum of all intermediate
clusters.
18
10
10
c
2
b
1
a
88
Altitude [km]
10
3
H+(H2O)n
e-
86
P-
84
0
a
82
10
-1
0
2
4
6
8 10 12
Cluster size
14 16
Figure 6. Size spectra for the proton hydrates at 86 km for
the three simulations in Figure 5.
H+(H2O)n
b
88
18 20
Altitude [km]
Number density [cm -3]
10
e-
86
P+
P-
84
82
10
1
10
2
10
3
10
4
10
5
-3
Number density [cm ]
+
H (H2O)n
Altitude [km]
88
a
-
e
Figure 8. As Figure 7, but for perturbed daytime
conditions (Figure 1c).
86
P-
84
82
H+(H2O)n
Number density [cm -3]
Altitude [km]
88
10
b
e-
86
P+ P-
84
82
10
1
10
2
10
3
10
4
10
5
Number density [cm -3]
Figure 7. Influence of ice particles on the ionospheric
composition. The shaded area represents layers of (a) 2×102
and (b) 2×104 cm-3 particles with radius 5 nm. Shown are
electrons and total proton hydrates; dashed lines denote
densities in the absence of particles. Grey lines mark the
density of charged particles. Simulations were performed
for quiet twilight conditions (Figure 1a).
10
critical nucleation
size at 128 K
2
1
128 K
10
0
130 K
10
132 K
-1
0
20
40
60
80
Cluster size
100
120
140
Figure 9. Size spectra for proton hydrates at 86 km for
various temperatures. Ionization conditions are as in Figure
1b (quiet daytime), neutrals are as in Figure 4. Black lines
denote spectra in the absence of particles, gray lines denote
spectra in the presence of 1000 cm-3 particles of radius 10
nm.
19
89
Natanson
H+(H2O) >10
10
Altitude [km]
-6
3 -1
α [10 cm s ]
quiet day time
Sugiyama
Leu et al. [1973]
1
Huang et al. [1978]
+
NO
88
87
86
85
perturbed day time
84
Johnsen [1993]
0
O+2
50
100
150
200
-3
1
2
3
5
10
20 30
50
100
Number density [cm ]
n
Figure 10. Proton hydrate recombination coefficients as a
function of cluster size. The laboratory measurements for n
≤ 6 have been scaled to 130 K. The Natanson approach used
in this paper is compared to the parameterization of
Sugiyama [1994]. For comparison, recombonation
coefficients for O2+ and NO+ are marked.
Ionization rate [cm -3 s -1]
10
10
10
10
10
10
10
4
3
10
2
30 ppm
3
1
1
0
0.3
-1
-2
110
115
120 125 130 135
Temperature [K]
140
145
Figure 11. Ionic nucleation as a function of temperature,
ionization and water abundance. The isolines represent
nucleation rates of 0.03 cm-3 s-1 (≈ 100 cm-3 h-1). The
atmospheric density is 2×1014 cm-3, corresponding to an
altitude of about 86 km.
Figure 12. Dependence of heavy proton hydrates on
atomic oxygen. Solid lines correspond to the O profile in
Figure 5, dashed lines to a 10× reduced O density.
Simulations have been performed for quiet daytime
conditions (Figure 1b) and perturbed daytime conditions
(Figure 1c).