[JOURNALOF GEOPHYSICAL RESEARCH
VOLUME 66, No. 5
On the Possible
Presence
Mar
1961
of Ice on the Moon •
KENNETHWATSON,BRUCEMURRAY,AND HARRISONBROWN
Division of GeologicalSciences
California Institute of Technology
Pasadena,California
It is generally presumedthat gasesof low
molecularweight escapevery rapidly from the
moon. As a consequence,
it has been assumed
that volatile substances,such as water, which
possess
short relaxationtimes for escape,do not
exist there. Urey [1952] recognizedthat there
may be depressions
in whichthe sunnevershines,
than the correspondingpressuresof SO• and the
rare gases.For an isothermal atmosphereand
constantgaspressureat the surface(corresponding to the vapor pressure) the mass of the
atmosphereper unit area of the lunar surface
will be independent of both temperature and
molecular weight and will correspondat this
and in which some condensed volatile substances
vapor pressureto 1.1 X 10-• g/cm•, or to a
could be present,but he concludedthat no solid
or liquid water couldexist on the moonfor more
than very short periods of time. Kuiper [1952],
followinga suggestionby Herzberg, stated that
SO2 moleculesmight be concentratedon the
night side of the moon. We wish to arguethat
water is actually far more stable on the lunar
surface than SO2 or noble gases because of its
extremelylow vapor pressureat low temperatures
and that it may well be present in appreciable
quantitiesin shadedareasin the form of ice.
The pressureof a volatile substancein a closed
total lunar atmosphericwater massof 4.2 X l06
grams. These quantities will of coursebe very
sensitive to the assumed temperature of the
to assume that the temperatures of the permanentlyshadedareasare equalto or lowerthan
this. It is quite probablethat lowertemperatures
do exist permanentlyand hence,ice would be
refers to the entire lunar surface. Over the life
coldest area. It should be noted that this esti-
mated pressureof water vapor is not inconsistent
with
the
radio
occultation
measurements
of
Elsmore [1957].
The least rapid escaperate of w•ter vapor
imaginable would be that derived from kinetic
theory by assumingdirect escapeof the fraction
of the moleculesin the atmospherewhichpossess
velocitiesgreaterthan the escapevelocity.Using
systemis determinedby the temperatureat the Spitzer's[1952] equationsand assuminga mean
coldestpoint in the system.The measurements escapetemperature of 400øK for the sunlit side
of Pettit and Nicholson[1930],verified by W. M. of the moon, we calculate a relaxation time,
Sinton (private communication),indicate that T400- 2.5 X l0 s seconds.
On the illuminated side,the water losswould
the temperature of the lunar surface falls to
120øKduringthe lunar night, and it is reasonable then be 2.2 X 10-20 g/cmm•/Sec,
wherecram
•
span of the moon (4.56 X 109 years) the loss
wouldbe,for unchanging
temperatureconditions,
3.2 X !0 -a g/cmm•.
even more stable than our calculations indicate.
For a variety of reasonsthe relaxation time
We are using 120øK, however, becauseit is a for water in the lunar atmosphere
is probably
reasonable
upperlimit. Extrapolationof observed considerably less than that derived above.
vapor pressuredata for ice indicatesthat at The temperatureof the exospheremay be con120øK the vapor pressureis about 1.4 X l0 -• siderablygreater than 400øK. Other mechanisms
mm of mercury.This, then, representsthe maxi- may play important roles, for examplethe one
mum pressureof water vapor which can exist discussed
recentlyby Opik and Singer[1960].
over the lunar surface, except for brief periods /Jpik and Singer(privatecommunication)
in
of time, and is many ordersof magnitudeless addition have calculated a relaxation time of the
order of l0 • secondsor lessfor the photodissocia• Contribution No. 1027, Division of Geological
Sciences,California Institute of Technology.
tion of water. However, it is important to realize
that, no matter how shortthe relaxationtime,
1598
LETTERS TO THE EDITOR
the escaperate cannot exceedthe rate of evaporation of ice in the permanently shaded areas.
At 120øK the evaporation rate of ice is estimated from kinetic theory to be 3.12 X 10-•4
1599
not be retained.
We make no estimate
of this
escapingfraction. In addition water is liberated
from
lunar
crustal
material
as the
result
of
impact. If water is liberated suddenly on the
g/cm2/sec,wherethe areain this caserefersto lunar surface,whether as the result of meteorite
that of exposedice. We can estimate an upper impact or of tectonic activity, the fraction which
limit for the rate of escapeof water by assuming escapeswill depend upon the following relative
that as muchas 50 pel•centof the lunar surface rates: (a) spread of the gas throughout the
is ice covered. This estimate gives for the rate
atmosphere; (b) condensationof the ice on the
of escape1.6 )< 10-•4 g/cmm2/SeC.
Over the life night side; (c) escape.
span of the moon the total loss would then be
2.3 )< 10sg/emro
a.
If exposedice exists, its actual area must be
much lessthan 50 per cent of the lunar surface.
There appear to be about 6 )< 106craters on the
moon with diameters in the range 1 to 5 km
[Kopal, 1960]. About 6 per cent of the moon's
surfaceis coveredby these craters. If we assume
the craterdepthto be 1/20 of the diameter,the
permanentlyshadedareascan occuronly north
of latitude
78øN
and
south
of latitude
78øS.
These areas amount to about 4 per cent of the
lunar surface.At a latitude of 84ø the effectively
shadedarea within a crater appearsto be about
5 per cent. Thus, the fraction of the lunar surface
whichis in permanentshadeis estimatedroughly
The spreadthroughoutthe atmospheretakes
place in a time of the order of a day. The relaxation
time
for
condensation
on the
cold
side
appearsto be of the order of 10 to 15 minutes.
Thus, if the water is liberatedon the cold side,
much of it will be retainedduring the lunar night
and a fraction of it will eventually find its way
to the areas of perpetual darknessin the polar
regions.If the water is liberated on the hot side,
whether or not any is retained will dependupon
whether the relaxation time for escapeis appreciably longer or shorter than a few hours.
Hence, the mass loss of water from the lunar
surfacemay be compensatedto some extent by
meteorite impact.
In any event, local concentrations of ice on
the moon would appear to be well within the
is true, the area of exposedice wouldbe 4.5 )< 10•s realm of possibility.Unfortunately, becauseice,
cm• and the escaperate would be 1.9 )< 10-•s if it exists,will be in the dark areas,attemptsto
g/emma/Sec.
Overthe life spanof the moonthis determine whether it is present must await the
wouldcorrespond
to a lossof 0.28 g/emro
a.
time when suitable instruments can be placed
These estimates can be compared with esti- in thoseareas.When this is done,either a positive
mates of the quantities of water which would or a negative result may reveal a great deal about
have been liberated from the moon had it evolved
early lunar history.
as the earth did. We have on the earth today
Detailed calculationsconcerningthe implicaabout 1.8 )< 106 grams of water per square tions of the existence or nonexistence of ice on
centimeter of earth surface, correspondingto the moon are being preparedfor publication in
1.54 )< 10-4 grams of water per gram of earth this journal in the near future.
material. If a correspondingamount of water
Acknowledgment.This work was supported by
had been liberated on the surfaceof the moon, the National Aeronautics and Space Administrato be 0.06 )< 0.04 X0.05
=
1.2 )< 10-4. If this
it wouldamountto 3.0 X 104g/emro
•.
tion under Grant
NSG-56-60.
In addition to juvenile water, water is being
REFERENCES
added sporadically as the result of meteorite
impact. Using Brown's [1960] relationships,we Brown, Harrison, The density and massdistribution
of metecritic bodies in the neighborhood of the
estimate that about 10s grams of metecritic
earth's orbit, J. Geophys.Research,65, 1679-1683,
matter fall on the moon each year. If we assume
1960.
a mean water content (including the carbona- Elsmore, B., Radio observations of the lunar
ceouschondrites)of 0.1 per cent,this corresponds atmosphere, Phil. Mag., 2, 1040-1046, 1957.
to the additionof 10• gramsof water eachyear Kopal, Z., Some current problems of lunar topog-
raphy, Proc. First Intern. Space Sci. Symposium
(Kallman, Editor), Interscience Publishers, New
the water in meteorites may rebound with
York, 1960.
velocitiesexceedingthe escapevelocityand hence Kuiper, G. P., Planetary atmospheres and their
or 8.3 )< 10-2• g/emma/sec.
A certainfractionof
1600
LETTERS
TO THE EDITOR
origin, in The Atmosphereof the Earth and Planets,
Revised Ed., (G. P. Kuiper, Editor), Univ.
Chicago Press,367 pp., 1952.
Opik, E. J., and S. F. Singer, Escape of gasesfrom
the moon, J. Geophys.Research,65, 3065-3070,
..
300 km, in The Atmosphereof the Earth and
Planets, Revised Ed., (G. P. Kuiper, Editor),
Univ. Chicago Press, 239-244, 1952.
Urey, H. C., The Planets, Their Origin and Development, Yale Univ. Press,New Haven, 17-18, 1952.
1960.
Pettit, E., and S. B. Nicholson, Lunar radiation
and temperatures, Astrophys. J., 71, 102, 1930.
Spitzer, L., Jr., The terrestrial atmosphere above
(Received January 9, 1961; revised February 10,
1961.)