VOL. 77, NO. 1
YOURNAL OF GEOPHYSICAL RESEARCH
j•ANUARY 1, 1972
Mariner6 and7 Ultraviolet
Spectrometer
Experiment'
Implications
of CO?, CO,andO Airglow
A. I. STEWART
Departmento• Astro-Geophysics
and Laboratory•or Atmosphericand SpacePhysics
University o• Colorado, Boulder 80302
The Mariner 6 and 7 ultraviolet spectrometerexperimentsobservedintenseemissions
from
CO, O, andCO,+in the Martian airglow.Analysisshowsthat they are excitedpredominantly
by the absorption
of solarEUV photons
by COsandconstitute
a majorenergy-loss
mechanism
for the thermosphere.
Modelsof the thermospheric
temperature
profileandthe airglowlayer
that demonstrate
the effects
of neutralchemistry
andionospheric
composition
aredeveloped.
With their aid, the observedCO Cameron-bandemissionscale height of 19 ñ 4¾2km is
shown to suggestan exospherictemperature of 315 ñ 75øK. Consideration of other data
suggests
a 'best'valueof about350øK.The emissions
fromCOs
+ areconsistent
witha topside
ionosphere
containingabout 30% COst The abundance
of O necessary
to convertthe rest
to Os+ is about 2% at 135 km, in good accordwith an independentdetermination.Within
the uncertainties
in the excitation
efficien'cies
andin the thermospheric
cooling
mechanisms,
the observations
are consistent
with the measured
electron
density.Eddycooling
mayho
important
if theeddydiffusion
coefficient
is large,butit is difficult
to reconcile
theheating
effectsof CO3with the smallobservedairglowscaleheights.There is no indicationin the
data that the ionosphereis modified by the solar wind below 200 km.
In recentyears the Mariner seriesof planetary flyby missionshas greatly stimulateddiscussionof the possiblecompositionand thermal
structureof the upper atmospheres
of Mars and
Venus. Radio occultation experiments carried
on Mariners 4 and 5 measuredelectron-density
profiles for the ionospheresof Mars [Kliore
et al., 1965a,b] and Venus [Kliore et al., 1967],
respectively,and an ultraviolet photometeron
Mariner 5 observedLyman-a radiation from the
exosphereof Venus [Barth et al., 1967]. The
Mariner
5 results seemed consistent with a cur-
ionosphere
was depressed
by the solarwind
[McElroy,1969; Cloutieret al., 1969].A
scarcity
of atomicoxygen
in theupperatmospheresof Mars and Venus,compared
to the
amountfoundon the earth,wassuggested
by
the resultsof Mariner 4 [Chamberlaina•d
McElroy,1966],Mariner5 [Barthet al.,1967],
andthe Russian
Venera4 [Kurt et al., 1968].
Thisscarcitywasascribed
by Shimizu[1968,
1969] to a high degreeof turbulencein the
lowerthermosphere
and by McElroy [1967,
1968] to a rapid recombinationof 0 and CO
rent theoretical view [McElroy, 1968] of the to formCO,via a chainof reactions
involving
Venus upper atmosphereas consistingof essen- C03•. In neithercontroversy
didthe datathemtially pure C0,with an exospherictemperature selves
offerany clueas to which,if either,of
around 600øK and an ionospherecomposedof the postulateswas correct.
COs+ ions in local equilibrium betweenphotoThe 1969flybymissions
to Mars by Mariner
ionization
and dissociative
recombination.
The
compositionof the upper atmosphereof Mars
was thoughtto be similarto that of Venus,but,
to explain the Mariner 4 electron-densityprofile, it was necessaryto postulate either that
dynamicalcoolingaffectedthe thermosphereon
Mars to a much greater degreethan on Venus
[Hogan and Stewart, 1969] or that the Martian
Copyright ¸
1972 by the American GeophysicalUnion.
6 and 7 offeredthe opportunityto examineits
atmosphere
in much greaterdetail than previouslypossible.
Thetwospacecraft
carriediden-
tical scanning
ultravioletspectrometer
(UVS)
experiments,
one of the objectives
beingto
detectand measurethe airglowin the 1100-to
4300-Arange.In this, as in their otherobjectives, the experiments
were successful
[Barth
et al., 1969]. The spectroscopic
information,
yieldingthe identitiesof major and minor con-
54
IMPLICATIONS OF MARTIAN ULTRAVIOLET AIRGLOW
implicationsof the UVS measurements
regarding the airglow-excitationmechanisms,
the thermechanismsas well, has been surveyed else- mospheric-temperature
profile,and the processes
where [Barth et al., 1971]. The intensity data controllingthe thermospherictemperatureand
[Barih ei al., 1971] yield the topside scale compositioncan be explored.
height of the emitting layer for the brighter
UVS EXPERIMENTS AND RESULTS
features,and, for the brightest emission(the
The instruments,the data, and the dataCameronbandsof CO), the peak and bottomside of the layer were observedas well. This reduction process have been described elselatter fact permits the location of the observed where [Pearceet al., 1971; Barth et al., 1969,
layer on an atmospheric-pressurescale and 1971]. The following is a summary of those
greatly enhancesthe value of the UVS experi- points that are particularly relevant to the
ments as a tool for investigatingthermospheric presentanalysis.
structure.
The line of sight of the instrumentsscanned
Another, more unexpected, feature of the downwardthrough the atmosphereabove the
U-VSmeasurements
rendersthem especiallyuse- bright limb of Mars 4 times,onceat a point
ful to this end. Althoughthe emissions
observed where the solar-zenithangleX was 27ø, once
whereit was 44ø (the first 'limb crossings'
of
in the short-wavelengthchannel (1100-1900 A)
are weak, their weaknessindicatingthat those Mariner 6 and 7, respectively),and twice at
gasesobservedin resonanceor fluorescentscat- the subsolarpoint whereX was 0ø (the second
tering in this channel(I-I, O, and CO) are minor limb crossings).A given spectral feature was
constitutentsonly [Andersonand Hord, 1971; scannedat altitude intervalsof about 20 kin,
on the spacecraftvelocityand the
Thomas, 1971], the emissionsin the long- depending
wavelength channel (1900-4300 A) are very experimental geometry. The instrument's recstrong. In fact, more than 95% of the total tangular(2.3ø X 0.23ø) fieldof viewwasaligned
airglowenergyrecordedby the instrumentswas nominallytangent to the planet'ssurfacebelow
received in the long-wavelengthchannel. The the line of sight. The effective vertical extent
immediate inference is that these bright emis- of the field of view was about 40 kin.
stituent gasesin the upper atmosphere,and in
some instances
the
nature
of the
excitation
sions arise as a direct result of the interaction
On eachlimb crossing,
airglowemissions
were
of solarUV photonswith the maior constituent, clearlypresent
in the long-wavelength
channel
CO2. Thus the airglow-intensityprofiles are (1900-4300A) of five spectra,coveringan altivery closelyrelated to the distributionof CO2 tude rangeof about 100 kin. Figure I presents
with altitude, that is, to the thermospheric- the measurements
of the strongestof these
temperatureprofile.The UVS datacantherefore emissions,which are to be discussedin detail.
be expected
to yieldreliableinformation
about They are the a3II-X• (Cameron)
bandsof CO,
the temperatures
in the Martian thermosphere.the 3P-•S (2972 A) multipier of O, and the
and
In this paper,the originsof the bright airglow •.2II•'-•2II,(Fox-Duffendack-Barker)
emissions
are considered.
A rangeof modelsof X2IIo ('doublet') bands of CO2+. The
by the much
the temperatureprofileare developed,
and air- multipierof O is accompanied
glow-intensity
profilesare calculated
fromthem. strongerxD-x• branchat 5577 A, which is beThe models are developedwith two ends in
yond the range of the UVS instrument.
view.First, if a variety of physicallyreasonable
The relativealtitudesof the data pointsfrom
a particularlimb crossing
are accuratelyknown
from the spacecrafttrajectory data. Points
wereplacedon a common
rangeof temperatures
in the thermosphere
that from all four crossings
the
is compatible
with the UVS profilesand the altitudescaleas follows.For eachcrossing,
excitation mechanismsthat adequately repro- point of maximumintensityof the CO Cameron
ducethe observedintensities.Second,the models bandswasfoundby a simplecurve-fitting
promodelsare available to compareto the data,
it becomesa fairly sim•le matter to decidethe
demonstrate the influence on the calculated tem-
cedure. This maximum was assumed to occur
peratures
of the uncertainties
in themechanism at the samealtitudeon all four crossings,
thus
that removes0 from the thermosphere
and in removingthe small effectsof the variations of
the composition
of the ionosphere.
Thus the solar-zenithanglefrom the subsequent
analysis,
A. I. STEWART
56
betweenone crossingand another,but the only
suggestionof a systematicdifferenceis that the
measurementsat the subsolar point (second
crossings,
opensymbols)indicateslightlyhigher
intensities and slightly larger topside scale
heightsthan do the measurements
at X = 27ø
and 44ø (first crossings,
closedsymbols).Any
suchdifferences
will be ignoredfor the purposes
14o
I00
02(
of this paper.
The mean maximum limb intensity of the
,,,,,,
Cameron bands is 610 kR. The maximum of 21
A M7-2
i
i
ill
I
i
• i I
I0
LIMB INTENSITY,
i
I00
i
i
•
I000
kl:{
Fig. 1. Observedlimb intensitiesof four airglow
emissions.The different symbols distinguish the
observations from each limb crossing.The closed
circle represents Mariner 6, first limb crossing;
open circle,Mariner 6, secondlimb crossing;closed
triangle, Mariner 7, first limb crossing;open triangle, Mariner 7, secondlimb crossing.
which assumesa mean zenith angle of 20ø.
Sincethe shapeof the emission
profileof the
Cameronbandsis well definedin the data, the
errorsin this processare small,lessthan 5 km.
The atmosphericmodelsdevelopedbelowwere
placedon an absolutealtitude scaleby requiring that, when X equals57ø, the maximum
electrondensity in the associatedmodel ionospheresshouldoccur at 135 km, which is in
accordance with the results of radio occultalton
[Fjeldbo ½• al., 1970]. The absolute altitude
scalefor the UVS data wasthen chosento give
the best over-all agreementbetweenthe data
and the intensityprofilesof the Cameronbands
calculated for the models. Because of the simi-
larity betweenthe processesthat producethe
ionosphereand those that excite the Cameron
bands, the error in this absolutescale is also
small, not more than about 10 km. It arises
largely from the uncertainaltitude dependence
near 135 km of quantities such as the ionospheric-recombinationcoefficientand the degree of scattered-light contamination of the
measurements of the Cameron bands.
As seenin Figure 1, the observationsfrom the
four limb crossingsagree very well with each
other. There are, of course,differencesin detail
kit for the O(•P-•S) multipict impliesthat the
associated (•D-•S) multipict reaches 330 kit
[Garstang,1951]. The CO•+ band systemsgo
off scale before reaching their maximum intensities; the greatest intensities recorded are
140kit at 140km for the .•-.• system
and35
kR at 148km for the •-.• system.
The mean
topsideemissionscaleheightsare obtainedfrom
the three highest data points on each limb
crossing
(two for the CO•+(•-.•) bands)and
thus refer to the approximate altitude range
150-190 km. The scale heights are 19 km for
the CO Cameronbands,24 km for the 0 mul-
tipier,24 km for the CO•*(.ff,-œ)
bands,and
25 km for the CO•*(•-.•) bands.The uncertainties in the intensities and scale heights
will be discussed
at appropriatepointsbelow.
For comparison
with the airglowdata, the ionosphericscaleheightin the 150- to 190-kmregion
is about45 km [Fjeldboet al., 1970].
Although the observationsin the 1100- to
1900-A region are not discussed
here, use will
be made of two resultsfrom the analysisby
Thomas [1971], namely, that 0 and CO are
scarce in the Martian thermosphere.The
atomic-oxygendensitywas estimatedfrom the
observations
of the resonance
triplet near 1304
A, which is excitedby the scatteringof solar
photonsand by photoelectron
impacton 0 and
whichis stronglyaffectedby radiationtrapping
in the opticallythick oxygenmedium.Upper
and lowerlimits on the CO densityweresetby
considering,on the one hand, the amount of
CO that wouldproducethe observedintensity
in the CO fourth positivebandsby fluorescent
scatteringof sunlightalone,and, on the other
hand, the amountof CO necessary
to produce
the observedamountof self-absorption
in this
sameband system.The densitiesof 0 and CO
at 135km werefoundto be of the orderof 3%
and •%, respectively.
IMPLICATIONS OF MARTIAN ULTRAVIOLET AreaLOW
57
IMPLICATIONS
The .• and• statesmay alsobe populated
by
cascade
from
the
•'•o+
state.
The
experimental
Becauseof the geometryof the experiment
and the relatively large sampling interval (20
km, comparableto the topside emissionscale
heights), no attempt is made to obtain volume
emissionrate from the limb-intensity data. The
approachadoptedis to constructmodelsof the
thermospheric-temperature
profile and the airglow layer from which limb intensitiesmay be
calculatedand compareddirectly to the measurements.This approachpreservesthe integrity
of the data, and permits various experimental
effects, such as the finite field of view, to be
information on theseprocesses
was surveyedby
Dalgarnoet al. [1970], who found that the contribution of photoelectronimpact, process2, is
not important on Mars. More recent experimental work [Wauchopand Broida, 1971; Ajello,
1971] confirmstheir conclusions.Although the
bulk of the sun'sionizing radiation lies between
150 and 350 A [Hinteregger,1970], no measurementsof the branchingratios in photoionization,
process1, have beenmadeat thesewavelengths,
and thus values based on the data at 584 A were
used [Turner and May, 1967; Spohr and yon
from the data. It alsopermitsa straightforward Puttkamer,1967; Bahr et al., 1969; Wauchopand
1971],givingratiosforthe.•, .•, •, and
evaluationby inspectionof the agreementor lack Broida,
•
states
of 0.33:0.33:0.22'0.11. The calculations
of it between model and observation.
presented
below take account of the fact that
The comparisonprocessdivides itself into
folded
into
the
models
rather
than
removed
fallsbeyondthe
two steps, each yielding a different type of part of the .•-.• bandsystem
information. In the first step, the data are set long-wavelengthlimit of the observationspreagainsta range of modelsrepresentingvarious sentedin Figure 1.
The observationof a limb intensity of around
alternativesas to heating and excitationproc600
kR in the Cameronbandsat ionospheric
altiesses.The comparisongives guidance about
tudes (as evidencedby the simultaneousdetecwhich excitation mechanisms or combination of
mechanismsare adequate to produce the ob- tion of emissionsfrom C02') rulesout excitation
servedintensitiesand which temperatureprofile mechanismsinvolving ambient CO at the low
or range of profiles adequately reproducesthe densities deduced by Thomas [1971]. If the
observedshape of the airglow layer. In the atmosphereat 135km containsx/•% CO, fluoressecondstep, the physicaland chemicalconsid- cent scatteringof sunlight producesa limb inerations that went into the building of the tensity of about 10 rayleighsat that altitude (the
transitionbeingopticallyforbidden),and photomodels and the estimated effects of those left
out are discussedin the light of the conclusions electron impact on CO producesnot more than
drawn from the first step.The constraintsplaced 25 kR, even though the crosssectionmay be
by the UVS observations
on processes
and con- very large [Hake and Phelps, 1967; Brongersma, 1968]. It may be noted that a limb
ditions in the actual Martian thermosphereat
the time of the experimentscan then be assessed. intensity of 600 kR implies a zenith intensity
30% of the energy
This section deals with the modeling of the of about30 kR, representing
in
the
incident
solar
ionizing
flux [Hinteregger,
thermosphereand airglowand performsthe first
1970].
Similar
remarks
apply
to the excitation
stepof the comparison;the secondstepis postof the 3P-•S multipict of O. Three per cent
ponedto the discussion.
Excitationprocesses.
The .•'II•-.•-IIg and atomic oxygenat 135 km yields not more than
•2Z?-.•,IIo bandsof CO,+ canbe excitedfrom about 100 rayleighsby photoelectronimpact,
as comparedto 21 kR observed;the O(•S) proneutral CO, by photoionization
duction implied by this observationrepresents
8% of the incident ionizing energy. Thus the
+
co* (if,
+
important mechanismsfor the productionof
or by impact of fast photoelectrons
CO(a3II) and O(•S) must be those involving
dissociative
excitationof the major atmospheric
C02 -+-e-• C02+ (if, •) + 2e
(2)
constituent,C02, by solar photons
and from CO, + ions by fluorescentscatteringof
sunlight
COe-]- h• -• CO* + O*
(4)
+ (œ) +
+ (X,
(3)
by fastphotoelectrons
A. I. S?EWAR?
58
and ionization crosssections[Hake and Phelps,
1967; Englander-Goldenand Rapp, 1964] and
or by photoionizationand subsequentdissocia- were performedaccordingto the proceduresdetire recombination
scribedby Dalgarno et al. [1969]. Calculations
CO• + e -, CO* + O* + e
(5)
of electron and ion densities involve the dissociative recombination coefficients for CO• + and
CO•.
-]-h•--•CO•.
+-]-e•
O•+, which have beenmeasuredby Weller and
++ co*
+o*l (6)Biondi
[1967] and Kasner and Biondi [1968].
The asterisksindicatethat the productsmay be The effectsof minor constituentssuchas CO, O,
excited.A further possibilityfor the production and O• were not included. None of the emisof O(•S) is dissociativerecombination of 03+
sionsdiscussedhere are subject to appreciable
quenchingin the thermosphereof Mars.
O,.+ q- e --40 q- O*
(7)
Model atmospheresand ionospheres. Details
for which the yield of O(•$) is 10% [Feldman of the temperatureprofile calculationsare given
in the appendix. In this section,some aspects
et al., 1971].
Becauseboth CO(a8II) and O(•$) are roeta- of particular relevanceto the present analysis
stable,quantitativeexperimentalinvestigationof are discussed.
Atomic oxygenis known to be a product of
processes
4-6 is more difficult than for processes
1-3, and much less information is available. photoabsorptionby CO• at wavelengthsbelow
McConnell and McElroy [1970] reviewed the 2280 A, even at very low pressures[Clark and
positionand estimatedthe relevant crosssections Noxon, 1970; Felder et al., 1970]. The indicafor photodissociation
and photoelectron
impact, tions in the Mariner 4 and 5 data, mentioned
processes4 and 5. In performing the present in the introduction,that this oxygenis rapidly
calculations,it was considereddesirable that removedfrom the thermospheresof both Venus
the modelsshould closelyreproducethe Cam- and Mars, are confirmedfor Mars by the Mareron-band intensities so that the temperature iner 6 and 7 UVS data near 1304 A [Thomas,
profile calculationsshould take proper account 1971]. The two explanationsoffered for this
of this important energy loss.A secondarygoal removal differ greatly in their implicationsfor
was to reproducethe O(8P-•$) measurements. thermosphericheating.
Shimizu [1908, 1909] proposedthat the O
To these ends, the following otherwise arbiis rapidly
trary assumptionsare made: Wherever ener- and CO producedin the thermosphere
transported
downward
by
strong
turbulence
to
getically possible,the yield of CO(a8II) and
O('$) in processes4-6 is 100%, 75% being regions of higher pressure,where recombination proceedsvia the three-bodyprocess
CO(a*II) and 25% O(•$); whereveronly O('$)
is accessible,its yield is 100%. In practice, the
total excitation of the Cameron bands and the
O(sP-'$) multipict by processes4 and $ together under these assumptionsis very similar
to that calculatedby McConnell and McElroy
[1970]; they make no estimateof the contribution from dissociativerecombination,process6.
The translation
of the above information
on
excitation processesinto predicted limb intensitiesfor a givenmodelatmosphereis relatively
straightforward.The solar flux was taken from
tabulationsby Hinteregger [1970], Hinteregger
et al. [1965], Brinkmann et al. [1966], and
Johnson[1965], whereasthe COsabsorptionand
photoionizationcross sectionscame from Inn
et al. [1953] and Henry and McElroy [1968].
Calculationsof the effectsof photoelectronsrequire knowledge of electron-impact excitation
O q- CO q- M--• COz q- M
(S)
If this is so, each pair of molecules,O and CO,
carries with it 5.4 ev of chemicalenergy that
is lost from the thermosphere,beingultimately
released during the recombinationreaction 8.
From the work of McElroy and Hunt,en [1970]
and of Hunten and McElroy [1970], it can be
estimatedthat an eddy diffusioncoefficientof
about 3 X 10• era' see-• is requiredto keep the
0 abundanceat 135 km down to 3% of the
ambient CO• density. The secondexplanation
[McElroy, 1967, 1968] was that oxygenin the
excited'D state, producedin the photodissociation of CO• at wavelengthsless than 1670 A
or in the recombinationof the ionosphere,is
removedin a chain of two-bodyreactionsleading to the local recombinationof CO•
IMPLICATIONS OF MARTIAN ULTRAVIOLET AIRGLOW
o('z)) + co,
COs* •
co,
(e)
CO --• 2 C0•
(10)
Here the chemicalenergyof the 0 and CO contributes directly to the thermosphericheating.
Ground-stateatomicoxygen,producedin photodissociation
of CO, at wavelengths
between1670
and 2280A and in reactionsthat quench0 (XD),
must still be removedby eddy transport,but a
smaller eddy diffusioncoefficientof about 5 X
106 cm' sec-• seems adequate [McElroy and
Hunten, 1970]. It is convenientto expressthe
heatingrate in terms of the 'heatingefficiency'
e, which is the fraction of the total energy of
the solar UV photonsabsorbedat a given altitude that appearslocally as heat. The 5.4 ev
per dissociationevent, by which the heating in
the two modelsdiffers,representsa changein e
of 0.29 in the upper thermosphereand of 0.55
in the lowerthermosphere.(SeeTable 1.)
A third possibilityarisesif the CO3• produced
59
angleis 57ø (the value appropriateto the occultation experiments). The compositionof the
ionosphereaffects the heating efficiency,since
the identity of the ion determinesthe amount
of energy radiated by the excited productsof
dissociativerecombination.Although CO,+ is
the most abundantly producedion, it is efficiently convertedto 0, + by reactionswith 0
CO•* -•- 0 -• 0•* -•- CO, k = 1.6 X 10-•ø
CO•* -•- 0 -• 0 + -•- CO•, k = 1.0 X 10-•ø
O* -•- C02 -• O•+ -•- CO, k = 1.2 X 10-9
and alsoby reactionwith O,
CO,.* + O•--• O•+ + CO•,
k-where the reaction
5.0 X 10-•
(13)
rates k are in cm3 sec-•
[Fehsenfeld et al., 1966a, 1966b; Fehsenfeld
in (9) reacts with 0 rather than with CO
et al., 1970]. Under the present assumption
about the products of dissociativerecombina(11)
tion, 5.1 ev is radiated per recombinationin a
[SeeMcElroy and Hunten, 1970.] Here the CO CO,* ionosphere,but only 0.2 ev is radiated in
and O2carry away 2.9 ev per dissociation
event; an 0,* ionosphere.The differencein the heating
however,it is not necessary
to invoketurbulence efficienciesin the two ionospheresis 0.11. (See
to explain the observedO density, since, at Table 1.)
The effects of the minor constituents 0 and
least in principle, reactions9 and 11 can consume all the 0 producedin the thermosphere CO on thermosphericheatingare not explicitly
provided at least half is producedin singlet includedin the calculations.SinceCO+ is rapidly
states. The density of atomic oxygen would convertedto CO,* by chargetransfer [Fehse•then be controlledby photochemistry,and that feld et al., 1966b], the presenceof CO hasno efof carbonmonoxidecontrolledby diffusion.
fect on ionospheric
composition,
and excitation
The ionospherescalculated from the model of the Cameronbandsby photoelectronimpact
atmospheresdevelopedhere are of the modi- on CO at the observed densities causes a reducfied F•-layer variety [Donahue, 1966]; the tion in the heating efficiencyby only 0.01; thus
maximumelectrondensityoccursat a pressure CO may reasonablybe ignored.The mostimporclose to 8 X 10-' pb when the solar zenith tant effectsof the presenceof 0 are indirect.One '
TABLE 1. Partition of Energy
Kinetic
Model
A
B
C
A
B
C
(CO, +)
(CO, +)
(CO• +)
(0• +)
(0, +)
(0, +)
Energy
0.19
0.48
0.36
0.30
0.59
0.46
(0.33)
(0.SS)
(0.64)
(0.33)
(0.88)
(0.64)
Chemical
Energy
0.37
0.0S
0.20
0.37
0.08
0.20
(0.66)
(0.12)
(0.35)
(0.66)
(0.12)
(0.35)
Radiated
in
IR
0.11
0.11
0.11
0.11
0.11
0.11
Radiated
in
Airglow
(0)
(0)
(0)
(0)
(0)
(0)
0.32
0.32
0.32
0.21
0.21
0.21
(0)
(0)
(0)
(0)
(0)
(0)
Radiated in
Cameron
Ban ds
0.24
0.24
0.24
0.14
0.14
0.14
(0)
(0)
(0)
(0)
(0)
(0)
60
A.I.
STEWART
effect,the conversion
of CO,+to O,+,hasalready
beenmentioned.
The otherimportanteffectis
that the energyof slowphotoelectrons,
which
would otherwisego into vibrational excitation
200
,
, ,[
,
'
' ' I t,
' ] 'l
[
'
180
t
of CO, andbe radiatedin the infrared,maygo
[
into the excitation of the •D level of O and be
convertedby quenchinginto heat. Both effects
can increasethe heating efficiencyby up to
0.11 if sufficientO is present.For a neutral
atmospherecontaining3% O at 135 km and an
ionospherecontaining30% CO,+ (as discussed
below), the increasein e over the pure CO•,
pure CO,+ situation is rather less than 0.11;
[ '
,,,.
+
iI
•oo[
• B(O•)
•;11
I
I0
I00
llOO0
LIMB INTENSITY, kR
the heating efficienciesin the CO,+ and O,+
modelscanbe regardedasbracketingthe values
F1•. 8. Comparison between model
appropriate to the actual observations.
layers and observations for the Cameron bands of
Temperature profiles calculated for various CO and the 2972-A emission of O. Hexagons are
combinationsof the oxygen-removalmecha- d•t• points. Solid •nd dashed lines sre used
nismsand the ionospheric
composition
are pre-
represent the models for reasonsof clarity.
sentedin Figure 2. The label A denotesremoval
by strongturbulenceand three-bodyrecombi- variation on B in whichthe CO? reactswith O,
nation, reaction8; B denotesremovalby for- reaction 11. The ionosphericcompositionis inmation of COo*,which subsequently
reactswith
CO, reactions 9 and 10; and C denotesthe
_
_
.
dicatedin parentheses.
The stronginfluenceof
the chemistryon the temperatureis apparent;
in the models B, the exospherictemperature
T. is about 130øK greater than in the models
A. The conversion of all the CO,' ions into 0,'
B'(C'ø•)/
_
increasesT. by between35ø and 55øK, largely
200
because of the decreased radiation
-
180 -
radiated
I00
in the Cam-
eron bands. The importance of the airglow
emissions
as an energylossmay be judgedfrom
the fact that if, in model A (CO,+), the energy
200
500
400
TEMPERATURE,
500
600
øK
Fig. 2. Temperatureprofilesfor six modelsof
neutral and ionic recombination.Model A, neutral recombination proceeds via
the reaction
0 + CO + M --) CO, -]- M; model B, via C08
+ CO -• CO, -]- CO,; model C, via C08 + 0 --)
CO• -]- 0,. The compositionof the ionosphereis
indicated in parentheses.The altitude scale is
discussed in the text.
in the Cameron
bands is instead con-
verted to heat, the exospherictemperature is
raisedfrom 333ø to 432øK. The heatingefiqcienciesand the temperaturesat selectedaltitudes
are tabulatedfor eachmodelin the appendix.
Comparisonwith observation.The limb intensitiescalculatedfrom the modelsare displayed
in Figures3 and 4, togetherwith the data points.
Figure 3 showsthe CO Cameronband and the
O•P-•S multiplot intensities,and Figure 4 shows
the CO,+•,-.• and •-.• bandintensities.Curves
for modelsC(CO, +) and C(O,+) are omitted for
clarity; they fall approximatelymidwaybetween
those for models A and B. Because the CO, + ion
is a possiblesourceof all the emissions,
the model
intensities depend on the ionosphericcomposition. They do not, however, depend on the
neutral chemistryexceptindirectly through the
temperature profiles. Two quantities will be
used in the comparisonof the modelswith the
IMPLICATIONS OF MARTIAN ULTRAVIOLETAIRGLOW
61
350 kR for the points near 151 km, gives a
200
topside scale height of 19 ñ 4% kin. The
correspondingquantities in models A (COJ),
A (OJ), B (CO,*), and B (OJ), which have exospheric temperaturesof 333ø, 386ø, 474ø, and
510øK, are 20, 23, 28, and 31 km. The Cam-
180
•
160
I-
140
eron-band measurements therefore indicate an
exospherictemperatureof 315 -- 75øK. This
conclusionis in accord with a more subjective
appraisalof Figure3, whichsuggests
that model
B(COJ) offersa much poorerfit to the data
120
,oo
i
O.I
than does model A (COJ). Model B (CO.,,*) lies
© •
i
i il
I
i
LIMB
i
i II
i
IO
INTENSITY,
i
I ii
IO0
kr
aboveall the data pointsnear 194 and 172 kin,
but below all but one of the points near 151
km and below all of thosenear 129 km, whereas
Fig. 4. Comparison between model airglow modelA (COJ) passes
throughall the groupsof
layersandobservations
forthe_•-• and5-.• bands points.
of CO,+. Hexagonsare data points.Solidand dashed
linesare usedto representthe modelsfor reasonsof
clarity.
Since the excitation mechanismsare similar,
the topsidescaleheightof the O(sP-'S)multipict
(24 km) mustbe interpretedin the sameway as
data. The first, the emissionscaleheight in the that of the Cameronbands.In effect,this is also
scaleheight(25
topside (150-190 km), reflectsthe temperature true of the CO,+(•--•) system
profilethroughoutthe regionwherethe optical km) after a correctionis made for the small
depth for the ionizing solar radiations is less contribution from fluorescentscattering;the corthan unity (Z • 135 km). The exospherictemperatureis a reasonableparameterfor this part
of the profile.The secondquantity, the maximum
limb intensity,is determinedby the solarEUV
flux and the efficiencyof the excitation mechanism acting on CO2; for the CO,+ bands, the
densityof this ion alsoenters.
The determinationof the topsidescaleheight
is affectedby the removal from the raw data
of off-axisscatteredlight, noise,and instrument
background[Barth etal., 1971]. Except at the
lowestaltitudes,the Cameronbandsare almost
rection is 3 km if the observedionospherecon-
tains 30% CO,+ (see below). Becauseof the
greaterdegreeof contamination
by off-axislight,
thesevalues,24 and 22 km, are lesswell-determined than the Cameron-band value of 19 km.
A_reasonable
compromise
is 21 km, corresponding
to an atmospherewith an exospheric
temperature
of about 350øK. This value is then the 'best'
one obtainable from the data. Little additional
informationregardingtemperaturesis contained
in the CO,+(*-.•) systemobservations
because
of the greater uncertainties.However, if the
entirelyfree of off-axislight, whicharisesfrom exospherictemperatureis 350øK and if the
contains35% CO2+, the presentcad
the scatteringof sunlightfrom the lower at- ionosphere
mosphere
andsurfaceof Mars and is therefore culationsreproducethe observedratio of 2.8:1
much weaker at short wavelengthsthan at long betweenthe intensitiesof the •-.• and •-.•
ones.Their scale height is therefore the most
accuratelydetermined.The group of data
systemsat 160 km and predicta 31-km scale
heightfor the •-.• system,
in accord
withthe
pointsnear194km havea meanintensityof
36 kR, after subtracting
corrections
of about
data within the errors.
45 kR each for instrument background and
the 610 kR in the Cameron bands and the 21
cosmic-ray
noise.To the uncertainties
of 3 and
9 kR, respectively,
in thesecorrections,
a statistical uncertainty of 4 kR must be added;
kR in the O(•P-•S) multiplot, together with the
a conservativeestimate of the total uncertainty
is thus 16 kR, the arithmeticsumof thesecom-
ponents.This value,togetherwith a lessimportantuncertainty
in the meanintensityof
Turning to the maximum limb intensities,
implied 330 kR in the associatedO(•D-•S)
multiplot, show that CO(a•II) and OtiS) are
producedin the ratio 63:37. McConnelland
McElroy [1970] find 70: 30 for the sum of processes4 and 5, and the present analysis assumes 75' 25. This ratio is in accord with the
62
A.I.
STEWART
data within the uncertainty in the relative tion on the temperaturesin the thermosphere
calibration[Pearceet al., 1971]. The agreement of Mars. Fjeldbo et al. [1970] presenttheir rewould be still better if the dissociative recombisults in the form of a temperature of about
nation of Os* yields substantiallymore O(•S)
450øK, but the significance
of this figure depends on the nature and compositionof the
than has been assumedhere, as might occurif
If the topsideof the observedionothe O2* producedby the reactions12 is in high ionosphere.
vibrational levels.
sphere is in diffusive equilibrium, their temThe modelswith a CO•+ ionosphereaccurately perature is the mean of the ion and electron
reproduce the maximum Cameron-band intemperaturesand provides an upper limit to
tensity, since the excitation eflqciencieswere the neutral temperature. However, since the
chosen with this reproduction in mind. I-Iow- experimentally determined quantity is the
ever, the sum of the maximum observedslant plasma scale height, the deducedtemperature
column excitation rates of CO (a•II) and O('S)
dependson the mean ion mass.Fjeldbo et al.
is 9.6 x 10• cm-• sec-•, whereas the CO2+ assumedthat the ion was CO•*; if the ionomodels predict only 7.5 X 10• cm-• sec-• and sphere contains about 70% 0,*, their temthe O•+ models4.6 X 10• cm-• sec-LA deficiency perature must be reducedto about 370øK, in
in the solar EUV flux usedis indicatedby fac- good accord with the UVS results. It is much
tors of 1.3 and 2.1, respectively,although the more likely that the ionosphereis in photocalibration uncertainty (a factor of 1.7) per- chemicalrather than diffusiveequilibrium.[See,
mits the data to be reconciled with the flux as
for example, Johnson,1968.] In this instance
used if the ionospherecontains 25% or more the relationship
betweenthe plasmascaleheight
CO•+. The deficiencyis naturally greater if the and the neutral temperatureis uncertain,being
sum of the quantum yields of CO(a•II) and affectedby altitude-dependentfactors such as
O(•$) in processes4-6 is lessthan the v.alue of opticaldepth,ion and neutralcomposition,
and
unity assumedhere. With regard to the CO•+ recombination rates. In simple models the
band system intensities, photoionizationalone plasmascaleheightis roughlytwice the neutral
fails to explainthe data at 160 km by a factor scaleheight,and thus there is no immediately
of 1.8 ----+0.4. The discrepancyis removed if
apparentconflictbetweenthe measuredplasma
the ionosphereconiains between 10 and 30%
COs+, but a larger EUV flux can be accommodated if photoionizationbetween 150 and 350
A producesfewer excited ions than the measurementsat 584 A suggest.
To summarize, the UVS measurementsof
Martian airglow emissionsin the range 19004300 A suggestthat the exospherictemperature
lies in the range 315 ----+75øK, with a best
and airglow scaleheightsof about 45 and about
21 kin, respectively,in the 150- to 190-kin
region.
Sometheoreticaltemperaturecalculations
appropriate to the Mariner 6 and 7 fiybys have
teregger [1970]; an increaseof a factor of 2
is readily reconciled with the data, and a
greater increasecannotbe ruled out.
430øK and calculatea limb-intensityprofile fo,r
beenpublished.Stewarta•d Hogan [1969] predictedexospheric
temperaturesof about 490øK,
greater than all but one of the presentmodels
even though their EUV heating efficiencywas
value of about 350øK. The measurements are
quite small (0.35). Presumably,the reasonis
in accord with existing knowledgeof the solar that they employedan augmentedversion of
EUV
flux and of the excitation mechanisms
the solar ionizingflux measuredby Hinteregger
for these emissionsif the Martian ionosphere et al. [1965], which is more intensethan the
contains about 30% CO•+. However, the data one used here [Hinteregger,1970]. McConnell
are also consistent with a solar EUV flux suband McElroy [1970] presenta model atmosstantially greater than that measuredby Hinphere with an exospherictemperature of
DISCUSSION
Apart from the UVS results,only the radio
occultation
measurements of the electron-dens-
ity profile provide any experimentalinforma-
the Cameronbandsexhibitinga scaleheightof
26 km between150 and 190 km. This figureis
outsidethe limits set by the presentanalysis,
and, in fact, the fit to the data providedby
their modelshowsmost of the shortcomings
of
the fit provided by model B(CO,*) discussed
earlier.
•I•IPLICATIONS OF MARTIAN
The UVS results provide the only experimental evidence on the compositionof the
Martian ionosphere.An internal consistency
check is possiblebecauseof the conversiono!
COs*to O•* by reactions12 with O. In the absence of NO [Barth et al., 1969], 0•* is the
terminal ion, and dissociativerecombinationis
the only lossprocess.Thus, in equilibrium,
+) =
+)
where k is the effective rate constant of the re-
actions12, a is the O,• dissociativerecombination coefficient,and the n's are the number
densitiesof the indicatedspecies.A knowledge
of the electrondensityand ion compositionat
any altitude determinesthe density of O at
that altitude, and a straightforwardextrapola-
tion yieldsthe abundance
of O at 135kin. Thus
30% CO•* and 70% O•* at 160 kin, where
n, -- 1.1 X 105cm-• [Fjeldbo et al., 1970], is
consistentwith 2% 0 at 135 km; less CO2•
impliesmore O. This result is in goodaccord
with the entirelyindependent
figureof about3%
obtainedfrom analysisof the observationsnear
1304 A [Thomas, 1971].
ULTRAVIOLET AIRGLOW
63
effect and is more than balancedby the additional heating resultingfrom the conversionof
CO•* to O•• by O. A potentiallyimportant cooling mechanismthat is not includedin the present models is eddy cooling, as suggestedby
Johnson[1966]. If the mechanicalheatingtha•.
must accompanythe turbulenceis ignored,the
volumecoolingrate is
•
-- --(a/az) l•,•[(aY/•z)
+ r]]
(1•)
in which cp is the specific heat at constant
pressure,p the massdensity,K the eddy diffusioncoefficient,and I • the adiabaticlapserate
(about 5ø/km). A comparisonof (15) with
the volume heating rate showsthat eddy cooling is a small effect in models B and C, in
which K is 5 X 106cm• sec-• or less,but may
be important in the modelsA, in which K _•
3 X
107 cm• sec-•. These conclusions are rein-
forced by the fact that the criterion suggested
by McElroy [1967] for the suppression
of eddy
coolingby radiative relaxation effectsis satisfied in models B and C but not in the models A.
In the absenceof an adequate descriptionof
the thermal effects of turbulence, the imporConsideration of the maximum intensities in
tance of eddy coolingremains problematical,
the previoussectionshowedthat the UVS but it is an attractive possibility.
measurements
permitted a substantialincrease
With regard to the controversiesthat arose
in the adoptedsolarEUV flux, by a factor of from the analysisof Mariner 4 and 5 results,
up to about2, with still largerincreases
being Mariner 6 and 7 data offer clear evidence on
possible
within the uncertainties
in the excita- one but not the other. McElroy [1969] and
tion efficiencies.Further evidencefor a greater
Cloutier et al. [1969] invokeda depression
of
flux comesfrom a comparisonof the measured maximumelectrondensity,1.6 X 105cm'•,
the Martian ionosphereby the solar wind to
with the values of 7.6 X 104 cm-• and 1.0 X 105
reconcilethe Mariner 4 plasmascaleheight of
29 km with a theoreticaltemperature profile
cm-• calculatedfor CO•* and O•* ionospheres, showing
an exospheric
temt•erature
of 487øK.
respectively.
An increaseby a factorbetween They suggestedthat, at the time of Mariner
2.6 and 4.4 is indicated.I-Iowever,increasingthe
ionizingfluxby a factor of 2.5 raisesthe exospherictemperaturein modelsA(CO••) and
A(O•*) by about110øK,and an adjustmentto
the heatingat longerwavelengths,
or to the
cooling,must be soughtto restorethe agreement between models and data.
Prag and Morse [1970] find that the solar
vacuumUV flux is variable,andHall andHinte-
regger[1970] suggests
that the 1325-to 1775-A
flux usedheremay be too largeby a factorof 3.
& suitablyreduced
flux,however,
lowersthe exospheric
temperature
by onlyabout50øK.Cooling by infraredradiationfrom0 andCO [see,
for example,McElroy et al., 1965] is a small
4, the plasma and neutral scale heights were
in fact equal. This equality was clearly not
true at the time of Mariner 6 and 7, at least
below 200 km, where the plasma scale height
is about 45 km and the airglow scaleheight is
about 21 km. Since it seemsunlikely that the
exospherictemperatureat the time of Mariner
4 shouldbe substantiallylarger than the preferred value of 350øK deducedfrom the present
analysisfor Mariner 6 and 7, it is unlikely
that their explanationof the Mariner 4 result
is the true one. The view of Hogan and Stewart
[1969] that the plasma scaleheight was small
becausethe thermospherewas cold is much
more in accordwith the presentresults.
A. I. STEWART
64
The only aspectof the present analysisthat
has a bearing on the questionof the role of
CO? in the Martian atmosphereis the difficulty of reconcilingthe warm thermospheres
predicted in those models that include CO?
chemistry [McElroy, 1969; and the present
work] with the small observed airglow scale
heights. When the laboratory evidence that
CO? plays no important role in the photolysis
of CO, is considered [Young et al., 1968;
Clark and Noxon, 1970; Felder et al., 1970;
Slanger and Black, 1970; Loewenstein,1970],
it is difficult
to resist the conclusion that
the
agent responsible
for the scarcityof O and CO
in the Martian thermosphere is dynamical
rather than chemical.
SUmmARY AND CONC•,VS•ONS
airglow-excitation mechanisms,thus ensuring
the consistencyof the calculatedtemperature
profilesand airglow intensities.The modelsthat
included CO• chemistryexhibited higher temperaturesthan thosewhich did not becauseof
the releaseof kinetic energy in the reactions
of CO• with CO or O. The models that allowed
for the conversionof CO,* to O,* by reaction
with O were also warmer because dissociative
recombination of CO,* was assumed to excite
emission in the Cameron bands of CO.
The highest quality data are those on the
Cameron bands of CO that exhibit a topside
emissionscale height of 19 ñ 4% km. Comparison with the models shows that this implies an exospherictemperatureof 315 ñ 75øK.
Considerationof other data as well suggestsa
best value of about 350øK.
The data on the
CO,* band systemssuggestthat the ionosphere
ments carried to Mars in the summer of 1969
at 160 km contains about 30% CO,*; the reby Mariner 6 and 7 provided the first oppor- mainder is most probably O,*. This result is in
tunity to examine the upper atmosphere of good accordwith the atomic-oxygenabundance
(3% at 135 km) deduced elsewherefrom the
another planet in detail. The spectrashoweda
rich and very intenseairglowin the range1900- UVS observationsat 1304 A. Although the
4300 A, containingemissionsfrom O, CO, and adopted solar flux explainsall the data satisCO,*. The range 1100-1900 A also contained factorily within the scaleheightand calibration
emissionsfrom O and CO, but the total inten- uncertainties,an increasein the ionizing flux
by a factor of up to 2 is not ruled out. The insity was more than 20 times smaller.The inferby comparisonof the calcuence that the long-wavelengthemissionsarise creasesuggested
predominantly from the major atmospheric lated and observed electron density can be
if the yields of excitedspecies
constituent,CO,, whereasthe short-wavelength accommodated
emissions arise from much less abundant minor
in the airglow-excitationmechanismsare less
constituents,is supportedby a more detailed than has been assumed.
examination of the excitation mechanisms. The
Such increasesin the solar flux produce unUV airglow constitutes an important energy- acceptableincreasesin the calculatedtemperalossmechanismfor the Martian thermosphere, tures, and additional coolingmust be invoked
to explain the observedairglow scale heights.
the emissionfrom CO alone representing30%
of the energymeasured
in the incidentionizing Eddy transport of heat is a likely possibilityif
solar flux. Thus the airglow is intimately re- the effective eddy diffusioncoefficientis suffilated both to the distribution of CO, and to
ciently large. It seems difficult, however, to
the heating mechanismsthat determine this reconcilethe effectsof CO• chemistrywith the
The scanningultravioletspectrometerexperi-
distribution.
observations. The Mariner
To aid in the airglow analysis,modelsof the
thermospherewere constructedby solvingthe
heat balance equation, allowing for heating
by UV absorptionand for coolingby thermal
conductionand radiation in the 15-/• bandsof
CO,. The scarcity of minor constituentspermitted their direct thermal effectsto be ignored
tation and UVS measurements
do not support
the view that the Martian ionosphereis depressedby the solar wind, at least below 200
without
serious error. The
calculation of the
heating efficiencies
took into accountthe...details of neutral and ion chemistryand of the
6 and 7 radio occul-
kin.
APPENDIX:
TEMPERATURE PROFILES OF THE
1ViARTIAN TttERMOSPttERE
The considerations that determine the tem-
perature profiles in planetary thermospheres
were first expoundedby Ba•½s [1951, 1959].
IMPLICATIONS OF MARTIAN ULTRAVIOLET AIRGLOW
Chamberlain[1962] and McElroy et al. [1965]
examinedthe atmosphereof Mars. More recent
65
Cooling. Coolingwas assumedto occur only
by radiation in the 15-p band of COs. The
work by McElroy [1967,1969] and by Hogan volume coolingrate was taken to be
and Stewart [1969] embodiedthe knowledge
z =
exp
gainedfrom the analysisof the Mariner 4 and
(A5)
5 results.The modelsdevelopedhere draw on
where hv is the energyof a 15-p photon,n is
the local number density of CO•, k is Boltzphysicsbut treat someaspectsin greaterdetail. mann'sconstant,and the vibrational relaxation
They do not take accountof dynamicaleffects, parameter V is representedby [Simpsonet al.,
most of the above work for the fundamental
beingof the 'planetaryaverage'type in which
the sunis assumed
to shinefor half of eachday
at an average solar-zenith angle of 60ø. The
calculatedtemperatureprofilesare usedat and
near the subsolarpoint on the assumption
that
the temperature at a given altitude is not
greatly dependent on local time or on areographic latitude. As discussed
in the text, the
calculationsare for a pure C02 atmosphere,
the
effectsof minor constituentsbeing ignoredexcept for the action of 0 in transformingthe
major ion from C02+ to O•+.
The simplified steady-state heat balance
equationin pressure-temperaturecoordinatesis
1969]
• = 8.3 X 10-25'T
ßexp{--40.6T-•/3} cm3see-•
(A6)
F(X) is a correctionfor the effectsof radiation
blanketing; accordingto McElroy [1967], the
expression
-
+
(A?)
where
X = 2.7 X 109/n• for p<<50•zb
(AS)
[Chamberlain and McElroy, 1966] is a satisfactory approximation. Inclusion of this cor(A1) rectionhas a negligible(<IøK) effecton the
temperatureprofilesin Figure 2, and the present calculations
setF (X) -- 1.
in which p is the pressure,T the temperature,
Heating. Although heating in CO• atmosC the thermal conductivity,H the scaleheight,
pheres can occur by absorption of infrared
and Q and L the volume heating and radiative
solar radiation, this effect is negligible in the
coolingrates,respectively.The barometricequaupper thermosphere and reaches only about
tion
15% at the lower boundary of the present
models
(R. E. Dickinson, private communicadp/d = --(p/.)
(A2)
tion 1970). Heating is therefore assumedto
transformsthe solutionsto altitude (z) coordi- occur only as a result of the absorption of
nates. Equation A1 was solvedby usinga sim- ultraviolet photons.It dependson the partition
ple relaxation technique, starting with an of the photon energy into its several ultimate
isothermalfirst trial. The boundary conditions forms, kinetic energy of atoms and molecules,
are
chemical energy of dissociationproducts, and
energy radiated in the infrared, visible, and
Q- L
(A3) ultraviolet airglow.
(local equilibrium) at the lower boundary,
The visible and ultraviolet
emissions that
wherep --- 3.162 X 10-• pb, and
carry away substantial amounts of energy on
OT/Op - -- (H•/ C)(Q/p)
Marsarethe•-.• and•-.• bandsof CO,+, the
(A4) •D-•S multiplet of O, which is associatedwith
at the upper boundary,where p - 0 but Q/p
remains finite. (A4) is equivalent to the condition that the atmosphereshouldbe isothermal
at very high altitudes.The thermalconductivity
C was taken to be 0.33 X 10-' (T/273) •'• cal
cm-• øK-• sec-•; the evaluationof L and Q are
described below.
the •P-•S multipier, and the Cameronbands of
CO. The excitation
mechanisms are discussed in
the text. Various infrared transitions in
that are excited by nonthermal processescan
properly be describedas infrared airglow and
are also important. Solar photonsthat are energetically incapableof causingdissociationof
66
A.I.
STrWART
COainto the highlyexcitedproductsinvolvedin and McElroy [1968] and of StewartandHogan
the airglowemissions
may dissociate
themolecule [1969], whichreferto ionizingwavelengths
only.
in lowercontinuums,
producing
COin theground
Table 2 displaysthe temperature calculated
state and 0 in the ground or •D states. Slow for eachmodelat the top of the thermosphere
photoelectrons
may either do the sameor cause (i.e., the exospheric
temperat,ire),at the ionizavibrationalexcitation
of COaeitherdirectlyby tion maximum (135 km), and at 100 km. The
collisionsor indirectlyby heatingthe ambient profilesthemselvesare presentedin Figure 2.
electrongas. This vibrationalenergyis most
probablylost by radiation in the infrared.Other
sourcesof vibrational energy, such as ion and
Acknowledgments.It has been a privilege to
work with the entire Mariner Ultraviolet Spectrometer team. My special thanks are due to C. A.
neutralchemical
reactions
andthe photodissocia- Barth and G. E. Thomas for many useful dis-
tion process
itself [Clercand Barat, 1966],are
cussions. I thank
ignored.Three mechanismsof neutral recombina-
of the temperature profile calculations.
This work was supportedby the National Aeronautics and Space Administration grant NGL 06003-052. I am indebted to the United 2ingdom
Science Research Council for a fellowship held at
University College, London, during 1970-1971,
tion and two of ionospheric
recombinationare
considered,as describedin the text. In all in-
R. E. Dickinson
for discussions
stances,
it is assumed
that the electronic
energy
of O(xD) is ultimatelyconvertedto heat. Where
consistentwith other assumptions
and with energyconsiderations,
it is assumed
in the modelsB
The Editor thanks M. B. McElroy and another
(recombinationvia the reaction of C03' with referee for their assistancein evaluating this
CO)that0 isproduced
in thexDstate,andin the
paper.
modelsC (reactionof C03' with O) that one-half
of the 0 is producedin singletstates.In the
]•EFERENCES
modelsA (three-body
recombination),
the state Ajello, J. M., Emissioncrosssectionsof CO2by
electronimpactin the interval1260A to 4500A,
in which0 is producedis immaterial.
2, J. Chem. Phys.,55, 1369,1971.
Table 1 showsthe partition of the energy Anderson,D. E., and C. W. ttord, Mariner 6
in the variousmodelsconsidered.
The photons and 7 ultraviolet spectrometer experiment:
Analysis of hydrogen Lyman-alpha data, J.
responsiblefor exciting most of the airglow
Geophys.Res., 76, 6666,1971.
and for formingthe ionosphere
are absorbed Bahr, J. L., A. J. Blake, J. It. Carver, and V.
rather abruptlynear the 10-3 tzb level; this
Kumar, Photoelectrons•ectraandpartial ioniza-
absorption is reflected in the differencesbe-
tweenthe partition figuresgiven for the limit
of vanishingly
smallpressures
(i.e.,at thetopof
the thermosphere)
and thosefiguresgivenfor
p -- 3.162 X 10-'• tzb (near the bottom of the
tion crosssectionsfor •arbon dioxide,J. Quant.
Spectrosc.
Radiat. Transfer,9, 1359,1969.
Barth, C. A., J. B. Pearce,2. K. Kelly, L. Wallace, and W. G. Fastie, Ultraviolet emissions
observednear Venus from Mariner 5, Science,
158, 1675, 1967.
thermosphore).
The latterfiguresare in paren- Barth, C. A., W. G. Fastie, C. W. ttord, J. B.
Pearce, 2. 2. Kelly, A. I. Stewart, G. E.
Thomas, G. P. Anderson, and O. F. Raper,
(KineticEnergy)is calledthe heatingefficiency Mariner 6: Ultraviolet spectrumof Mars upper
in the text; it refers to a mean taken over all
atmosphere,Science,165, 1004, 1969.
absorbed
wavelengths,
andthusit is notstrictly Barth, C. A., C. W. Hord, J. B. Pearce, 2. K.
Kelly, G. P. Anderson, and A. I. Stewart,
comparable
to the heatingefficiencies
of Henry
Mariner 6 and 7 ultraviolet spectrometer experiment: Upper atmospheredata, J. Geophys.
theses. The quantity in the second column
Res., 76, 2213, 1971.
TABLE 2. Thermospheric
Temperatures
Model
A (COa+)
B (COa+)
C (COa+)
A (Oa+)
B (Oa+)
C (0• +)
100km, øK
163
201
185
175
219
203
135km, øK
278
353
323
313
390
358
•o, øK
333
474
415
386
510
457
Bates,D. R., The temperatureof the upperatmosphere,Proc. Phys. Soc.London,Sect.B., 64, 805,
1951.
Bates, D. R., Some problemsconcerningthe terrestrial atmosphereabove about 10I) km, Proc.
Roy. Soc. London, Ser. A., 253, 459, 1959.
Brinkmann, R. T., A. E. S. Green, and C. A.
Barth, A digitalized solar ultraviolet spectrum,
Tech. Rep. 32-951,Jet Propul. Lab., Pasadena,
Calif., 1966.
Brongersma,It. It., Interaction of moleculeswith
IMPLICATIONS OF MARTIAN ULTRAVIOLET AIRGLOW
low energy (0-30 ev) electrons, Ph.D. thesis,
Univ. of Leiden, Leiden, Netherlands, 1968.
Chamberlain, J. W., Upper atmospheres of the
planets, Astrophys. J., 136, 582, 1962.
Chamberlain, J. W., and M. B. McElroy, Martian atmosphere' The Mariner occultation experiment, Science, 152, 21, 1966.
Clark, I.D., and J. F. Noxon, Photodissociation
of CO• on Mars, J. Geophys. Res., 75, 7307, 1970.
Clerc, M., and F. Barat, Cin•tique des produits
de decomposition de CO• par photolyse-•clair
das l'ultraviolet loinrain, J. Chimie. Phys., 63,
1525, 1966.
Cloutier, P. A., M. B. McElroy, and F. C. Michel,
Modification of the Martian ionosphere by the
solar wind, J. Geophys. Res., 74, 6215, 1969.
Dalgarno, A., M. B. McElroy, and A. I. Stewart,
Electron impact excitation of the dayglow, J.
Atmos. Sci., 26, 753, 1969.
Dalgarno, A., T. C. Degges, and A. I. Stewart,
Mariner 6: Origin of Mars ionized carbon dioxide ultarviolet spectrum, Science, 167, 1490,
1970.
electrons in 02, CO', and CO• Phys. Rev., 158,
70, 1967.
Hall, L. A., and H. E. Hinteregger, Variations of
incident solar EUV during a solar rotation,
paper presented at 13th meeting of Cospar,
Leningrad, May 21-30, 1970.
Henry, R. J. W., and M. B. McElroy, Photoelectrons in planetary atmospheres,in The Atmospheres o• Venus and Ma•'s, edited by J. C.
Brandt and M. B. McElroy, p. 251, Gordon
and Breach, New York, 1968.
Hinteregger, H. E., The extreme ultraviolet solar
spectrum and its variations during a solar cycle,
Ann. Geophys., 26, 547, 1970.
Hinteregger,H. E., L. A. Hall, and G. Schmidtke,
Solar XUV radiation and neutral particle distribution in July 1963thermosphere,SpaceRes.,
5, 1175, 1965.
Hogan, J. S., and R. W. Stewart, Exospheric
temperatures on Mars and Venus, J. Atmos.
Sci., 26, 332, 1969.
Hunten, D. M., and M. B. McElroy, Reply to
W. B. DeMore, J. Geophys. Res., 75, 4900, 1970.
Detwiler, C. R., D. L. Garrett, J. D. Purcell, and
R. Tousey, The intensity distribution in the
ultraviolet solar spectrum, Ann. Geophys., 17,
263, 1961.
sections for ionization
of atoms and molecules
by electron impact, Rep. LMSC 6-74-64-12,
Lockheed Missiles and Space Co., Palo Alto,
Calif., 1964.
Fehsenfeld, F. C., A. L. Schmeltekopf,and E. E.
Ferguson,Thermal energy ion-neutral reaction
rates, 3, The measured rate constant for the
reaction O+(4S) nu CO2(•:) -+ O2+(•II) •CO(•Z), J. Chem. Phys., J4, 30'22, 1966a.
Fehsenfeld,F. C., A. L. Schmeltekopf,and E. E.
Ferguson,Thermal energyion-neutralreaction
rates, 5, Measuredrate constantsfor C+ and
CO+ reactionswith O2 and CO•, J. Chem. Phys.,
45, 23, 1966b.
Fehsenfeld,F. C., D. B. Dunkin, and E. E. Ferguson, Rate constantsfor the reaction of CO•+
with O, O• and NO; N2+ with O and NO; and
03+ with NO, Planet. Space Sci., 18, 1267, 1970.
Felder, W., W. Morrow, and R. A. Young, Experimental evidence of the photochemicalinstability of a pure CO2 planetary medium, J.
Geophys. Res., 75, 7311, 1970.
Feldman, P. D., J.P. Doering, and E. C. Zipf,
Excitation of O(•S) atoms in the day airglow,
J. Geophys.Res., 76,•3087, 1971.
Fjeldbo,G., A. Kliore,andB. Seidel,The Mariner
1969 occultation measurements of the upper
atmosphereof Mars, Radio Sci., 5, 381, 1970.
Garstang, R. H., Energy levels and transition
probabilities in p• and p• configurations,Mon.
Notic. Roy. Astron. Soc., 111, 115, 1951.
Hake, R. D., Jr., and A. V. Phelps, Momentumand inelastic
Inn, E. C. Y., K. Watanabe, and M. Zelikoff,
Absorption coefficientsof gasesin the vacuum
ultraviolet, part 3, CO2, J. Chem. Phys., 21,
1648, 1953.
Donahue, T. M., Upper atmosphere and ionosphere of Mars, Science,152, 763, 1966.
Englander-Golden, P., and D. Rapp, Total cross
transfer
67
collision
crossesections for
Johnson, F. S., Satellite Environment Handbook,
p. 123, Stanford University Press, Palo Alto,
Calif., 1965.
Johnson, F. S., The atmosphere of Mars, paper
presented at 7th meeting of Cospar, Vienna,
May 10-19, 19'66.
Johnson,F. S., Mariner 4 and the atmosphere of
Mars, in The Atmosphereso• Venus and Mars,
edited by J. C. Brandt and M. B. McElroy,
p. 181, Gordon and Breach, New York, 1968.
Kaplan, L. D., J. Connes, and P. Connes, Carbon
monoxide in the Martian atmosphere, Astrophys. J., 157, L187, 1969.
Kasner, W. H., and M. A. Biondi, Temperature
dependenceof the electron-O2+-ion
recombination coefficient, Phys. Rev., 174, 139, 1968.
Kliore, A. J., D. L. Cain, G. S. Levy, V. R. Eshleman, G. Fjeldbo, and F. D. Drake, Occultation
experiment: Results or-the first direct measurement of Mars atmosphereand ionosphere,Science, 149, 1243, 1965a.
Kliore, A. J., D. L. Cain, F. D. Drake, V. R.
Eshleman, G. Fjeldbo, and G. S. Levy, Preliminary results of the Mariner 4 occultation measurements of the atmosphere of Mars, paper presented at Lunar and Planetaw Conference,
Calif. Inst. of Tech. Jet. Propul. Lab., Pasa-
dena, Calif., Sept. 13-18, 1965b.
Kliore, A., G. S. Levy, D. L. Cain, G. Fjeldbo,
and S. I. Rasool, Atmosphere and ionosphere
of Venus from Mariner
5 S-band radio occulta-
tion measurement, Science, 158, 1683, 1967.
Kurt, V. G., S. B. Dostovalov, and E. K. Sheffer,
The
Venus
far
ultraviolet
observations
with
Venera 4, J. Atmos. $ci., 25, 668, 1968.
Loewenstein, M., Relative quenching rates of
A. I. STEWART
68
O(•D) by COg, N•, and O• (abstract), ]Sos
Trans. AGU, 51, 786, 1970.
Slanger,'T. G., and G. Black, The COt photolysis
problem (abstract), Eos Trans. AGU, 51, 786,
1970.
McConnell,J. C., and M. B. McElroy, Excitation
processes
for Martian dayglow, J. Geophys. Spohr, R., and E. yon Puttkamer, Energiemessungvon Photoelektronenund Franck-CondonRes., 76, 7290, 1970.
Faktoren der SchwingungsObergange einiger
McElroy, M. B., The upper atmosphereof Mars,
Molikiilionen, Z. Naturforsch. A, 22, 705, 1967.
Stewart, R. W., and J. S. Hogan, Solar cycle
variation of exospherictemperatureson Mars
J. Geophys. Res., 73, 1513, 1968.
and Venus' A prediction for Mariner 6 and 7,
McElroy, M. B., and D. M. Hunten, PhotochemAstrophys. J., 150, 1125, 1967.
McElroy, M. B., The upper atmosphereof Venus,
istry of CO•_in the atmosphere
of Mars,J. Geophys. Res., 75, 1188, 1970.
McElroy, M. B., J. l'Ecuyer,and J. W. Chamberlain, Structure of the Martian upper atmosphere,Astrophys.J., 141, 1523,1965.
Pearce,J. B., K. A. Gause,E. F. Mackey,K. K.
Kelly, W. G. Fastie, and C. A. Barth, The
Mariner 6 and 7 ultraviolet spectrometer,Appl.
Opt., 10, 805, 1971.
Prag,A. B., and F. A. Morse,Variationsin the
solar ultraviolet flux from July 13 to August 9,
1968,J. Geophys.Res.,75, 4613,1970.
Shimizu,M., The recombination
mechanism
of O
and CO in the upper atmosphereson Venus
and Mars, Icarus, 9, 593, 1968.
Science, 165, 386, 1969.
Thomas,G. E., Neutral compositionof the upper
atmosphereof Mars as determined from the
Mariner UV spectrometer experiments, J.
Atmos. $ci., 28, 859, 1971.
Turner, D. W., and D. P. May, Franck-Condon
factors in ionization' Experimental measurement using molecular photoelectron spectros-
copy, 2, J. Chem. Phys., 46, 1156,1967.
Wauchop,T. S., and It. P. Broida,Crosssections
for the productionof fluorescence
of COe*in
the photoionizationof CO• by 58.4 nanometer
radiation, J. Geophys.Res., 76, 21, 1971.
WelleL C. S., and M. A. Biondi, Measurement
of dissociativerecombination of CO.o*ions with
electrons,Phys. Rev. Lett., 19, 59, 1967.
Shimizu,M., A modelcalculationof the cytherean
Young, R. A., G. Black, and T. G. Slanger,Reupper atmosphere,Icarus,10, 11, 1969.
action and deactivation of O(•D), J. Chem.
Simpson,C. J. S. M., T. R. D. Chandler,and
A. C. Strawson,Vibrational relaxation in CO•
and CO•-Ar mixtures studied using a shock tube
and a Laser-Schlieren technique, J. Chem.
Phys., 51, 2215, 1969.
Phys., 49, 4758, 1968.
(Received January 18, 1971;
acceptedSeptember21, 1971.)