GEOPHYSICAL RESEARCH LETTERS, VOL. 26, NO. 24, PAGES 3657-3660, DECEMBER 15, 1999
Photo-Induced
in the Martian
Fractionation of Water Isotopomers
Atmosphere
Bing-Ming Cheng,Eh Piew Chew and Chin-Ping Liu
Synchrotron Radiation Research Center, Hsinchu, Taiwan
Mohammed
Bahou
and Yuan-Pern
Lee
Department of Chemistry, National Tsing Hua University, Hsinchu, Taiwan
Yuk L. Yung and M. F. Gerstell
Division of Geologicaland Planetary Sciences,Cahfornia Institute of Technology,Pasadena
Abstract. The history and size of the water reservoirs D/H ratio may be written
on early Mars can be constrainedusing isotopicratios
x(t) - x(0)[(w + ;)/w]
of deuterium to hydrogen. We presentnew laboratory
(1)
measurements of the ultraviolet cross-sections of HaO
whereX(t) is the D/H ratio at time t, W is the remaining
reservoirof water at time t, and L is the amountof
port of a photo-inducedfractionationeffect (PHIFE),
water
lost from planet betweentime 0 and time t. If
that reconcilesa discrepancybetween past theoretical
D and H escapedwith equal ease,F would be I and X
modeling and recent observations. This supports the
wouldbe constantthroughtime; ifil couldescapebut D
hypothesisthat Mars had an early warm atmosphere
and its isotopomers,and modeling calculationsin sup-
couldnot, then F wouldbe 0. The F valuebasedon the
and has lost at least a 50-m globallayer of water. Likely
most completepast model of Martian photochemistry
applicationsof PHIFE to other planetary atmospheres
is 0.32, determinedby eight kinetic processes
modeled
are sketched.
in detail[Yunget al., 1988].Krasnopolsky
et al. [1998]
recentlydetecteddeuteriumin the upperatmosphere
of
Marsby HubbleSpaceTelescope
(HST) observations
of
1. Paradox and Hypothesis
Lymana emission.TheydeducedF = 0.024-0.01.(The
Circumstantial evidencestrongly suggestsa warmer error estimatefor F is ours,from theirsfor D/H.) Any
the photochemical
model.
past climate on Mars that permitted liquid water near F in this rangechallenges
As
discussed
by
Yung
et
al.
[1988],
water
is
the ultithe surface[Cart, 1996; Baker et al., 1992]. This is
mate
source
of
atmospheric
H,
but
molecular
hydrogen
very different from the current cold, arid conditions.
to the
Among the few quantitative clues to the evolution of (Ha) is the major carrierof hydrogenspecies
upper
atmosphere
of
Mars,
where
H
escapes
to
space.
the Martian atmosphereare the isotopicsignaturesleft
Thus
the
D/H
ratio
of
molecular
hydrogen
can
provide
by the processesthat have modified the atmosphere
over time. Most of the lossprocessespreferentially re- a basisfor calculatingF. The partitioningof D between
move the lighter species. If we understandthe frac- HD and HDO in the atmospheremay be definedby'
rionation caused by these processes,and know the current isotopic ratios of the various reservoirs, we can
constrain the early atmosphere and its evolution to
the present[McElroy, 1972; McElroy and Yung, 1976;
Jakosky,1993; Thiemens et al., 1995; Farquhar et al.,
1998;Krasnopolsky
et al., 1997].
Owenet al. [1988]discovered
that Martian water is
enrichedin deuterium. The D/H ratio deducedfrom
their measurement
is 6 times the terrestrial
value. This
constrainsthe amount of water that has escapedfrom
the planet. But the result dependson a model-derived
fractionation factor, F. The time dependenceof the
Copyright1999by theAmericanGeophysical
Union.
R = (HD/H•)/(HDO/H•O)
(2)
The relation between F and R is determined by the pho-
tochemicalmodel, and F is approximatelyproportional
to R. To reconcile their measurements with the pho-
tochemicalmodel,Krasnopolsky
et al. [1998]suggested
theremightbe a conceptualerrorin currentphotochemical modelsof Mars. They proposedthat the D/H ratio
of escaping
hydrogencouldbe determinedby a thermodynamic equilibrium on the surfaceof Mars:
HD+H20
< • H2+HDO
(3)
According
to kinetictheory,R is 1.6 [Yunget al., 1988],
but the equilibrium(3) impliesR = 0.14 at 200 K,
Papernumber1999GL008367.
a temperature typical of the Martian surface. The
0094-8276/99/1999GL008367505.00
measurements of deuterium with the HST imply R =
3657
3658
CHENG ET AL.: WATER IN THE MARTIAN
0.09, much closer to the thermodynamic than the kinetic value. However, for the thermodynamicsto pre-
ATMOSPHERE
sodiumsalicylate. The fluorescencesignal subsequently
detectedwith a photomultiplier
(Hamamatsu,R955)in
vail on Mars, the forwardreactionof (3) must be sufficiently fast. Krasnopolskyet al. [1998]estimatedthe
requiredrate constantk3 as at least10-23 cm3s
-1 not
consistentwith the laboratory rate constantof 10-33
a photon-countingmode was employedfor normalization. The light transmitted from the CaF2 beamsplit-
been assumed that HDO
librium vapor pressures[Kirshenbaum,1951] of
ter passedthrough an absorptioncell equippedwith two
CaFu end-windows, and irradiated onto a glasswindow
cm3s
-1 [Ldcluse
andRobert,1994].Therearetwoplau- coated with sodium salicylate. Fluorescencewas desibleresolutionsof this paradox[Yungand Kass,1998]. tected similarly.
The first is that a reaction or fractionation
hitherto neTwo absorption cells were used, with inner diameter
glected can reduce the value of R in the photochemical 39.5 mm and path lengths 10.2 and 113.3 cm. To avoid
model.
The second is that a hitherto unknown catavariation in pressure due to irradiation or surface adlyst on Mars can raise the rate coefficient of the for- sorption/desorption,
a reservoir
of volume--•1382cma
wardreactionin (3) by ten ordersof magnitude.In the was connected to the absorption cells. The absorption
laboratoryexperiments[Ldcluseand Robert,1994],the cell has a pressureport connectedwith two MKS Baracatalytic effects of charcoal, silica, phyllosilicates,and tron pressuremeters(model127AA, range10 and 100
iron were found to be negligible.
Tort). The spot sizeof the synchrotronradiationwas
The resolution of this puzzle is basic to our under- 3.5 x 1.0 mm at the entrance window of the absorption
standing the evolution of the Martian atmosphere.We cell, and increased to --•4.0 x 1.5 and --•8.0 x 2.5 mm at
could use the F deduced from observations to model
the terminal windows of the short and long absorption
atmosphericevolution, but without knowingthe funda- cells respectively.
mental processesdetermining F, we cannot trust exAt each wavelength, absorbancewas plotted against
trapolations to the early atmosphere. For instance, number density and fitted to a line of least squaresto
if the current value of F were influenced by catalysts yield the absorption crosssection of the sample. Cross
on the Martian surface today, we would need to know sectionsof HDO were determined using samplesconwhether the same catalysts were probable on early taining H20, HDO, and D20, and assumingthe absorpMars.
tion cross sections of pure H20 and DuO are as meaWe propose PH!FE as a simple and likely mecha- sured in this work. In addition to pure H20 and D20
nism to resolve the discrepancy. Although UV photol- samples, two mixtures were prepared from deionized
ysis is the primary destroyer of HDO in the Martian HuO and pure D20 (Merck, Sharpand Dohme,isotopic
atmosphere,the UV crosssectionsof HDO have never purity 99.8%)with molar ratios !:2.81 and 1:4.02.The
before been measured. Existing measurementsof D20 relative concentrationsof HuO, D20, and HDO in the
cross-sections
[LauferandMcNesby,1965]arecoarseby gas sample mixtures were determined from the equilibmodern standards. In modeling Mars, it has previously rium constantkeq: 3.74 [Pyperet al., 1967]andequicross sections are the same as
those of H20. Cross sectionsof HDO smaller than those
of HuO would producelessHD relative to H2, implying
a smaller /i• in the model and better agreement with
HST
2.
data.
Measurements
Laboratory measurements were carried out to determine the cross sections of H20 and its deuterium iso-
topomers. The light source was the synchrotron radiation dispersed with a 1-m Seya-Namioka monochromator located at the Synchrotron Radiation Research
Centerin Taiwan with a 1.5 GeV storagering [Tseaget
al., 1995]. The monochromator
wasequippedwith four
gratings to cover the spectral range 30-300 nm. We
usedthe gratingwith 600 grooves/mm(blazedat 140
nm) to coverthe spectralrange 100-300nm. The slit
width was typically 0.05 mm, correspondingto a spec-
D20, and HDO at 295 K. The gaseousmolar fractions
of HuO, HDO, and D20 of these two sampleswere thus
determinedto be (0.079,0.394,0.526)and(0.046,0.328,
0.625). For experimentswith D20 or mixturesof the
three isotopomers,at least 10 cyclesof passivationwere
carried out. Absorbance of each sample was measured
at 8 to 14 different pressures;the spectral range of mea-
surementsdependson pressureof the sample and the
path length of the cell. Absorbancewas calculated according to Beer's law. Data with absorbancegreater
than
2.0 were discarded
to avoid saturation
effects.
The resultant absorption crosssectionsof H20, HDO,
and DuO in the spectral range 140-195 nm are shown
in Figure 1. (The data are availablefrom website
http://ams-bmc.srrc.gov.tw).
A completelistingat 0.2nm intervals is available upon request. Variations in
absorption crosssectionsof HDO determined from the
two mixtures containingdifferent fractionsof HDO were
tral bandwidthof •-0.i:nm. The monochromator
was typically within 5%; averagedvalues are listed. Cross
scannedin either 0.1-nm or 0.2-nm stepswith signal- sections of H20 were compared with other recent meaaveraging period of 3 s at each step. A small frac- surements. Our HuO spectrum is nearly identical to
tion of the light beam was reflectedwith a CaF2 plate that of Chan et al. [1993], who used low resolution
and TRK sum-rulenormaland, after passingone additionalCaF2 plate (thick- dipole (e,e•) spectroscopy
ness2 ram), irradiatedontoa glasswindowcoatedwith ization. In the spectral region 140-182 nm, our HuO
CHENG ET AL.: WATER
10
-17
: ....
i .........
i .........
i .........
i .........
i .........
IN THE MARTIAN
10-19
• 10-20
1O-21
........
H•O
.............
o•o
10-22 .... , .........
, .........
3659
of the photons are absorbednear 170 nm, where there is
little difference between absorption by HDO and H2.
But in the lower atmosphere, photolysisof HDO is 23 times less efficient than that of H2. In this region,
shielding of UV radiation by CO2 becomesimportant.
The photons that can penetrate to these levels are in
the long-wave tail near 190 nm, where the difference
between HDO and H2 crosssectionsis large. An esti-
i .........
10-18
•
ATMOSPHERE
mate usinga revisionof the modelof Yunget al. [1988],
", ,,
taking into account of the correct average over a solar
cycle [Pathareand Paige, 1997]bringsthe modelvalue
, .........
, .........
, .........
,', ,•',......
of the fractionation parameter F to 0.06.
The past lossof water from Mars, basedon extrapolaFigure 1. Laboratory measurementsof the crosssec- tion from present escaperates, was first estimated to be
140
150
160
170
180
190
200
Wovelength
(nm)
revisions,
tionsof H20 (solidline), HDO (dashedline), and D20 a few meters[Yunget al., 1988].Subsequent
(dashdot line) from 140to 195nm. The measurementstaking accountof sputtering inducedby the solar wind,
were taken at 295 K with approximately 0.1 nm resolution.
and of a more active early sun, raised the estimate to 50
m [Kassand Yung,1995;Kass,1999]. UsingEqn. (1)
with F in the range 0.02 to 0.06, this would imply that
results are also in agreementwith those of Yoshinoet
the current water reservoir on Mars is equivalent to a
al. [1996],whouseda 12-cmcellandsynchrotron
radia-
globallayer of 8-10 m [Kassand Yung,1999],or about
tion; there is a local maximum deviation of 7% near the
peak of the band at 166 nm. For wavelengthsgreater
106 km a. This is consistent with the total ice volume
of 1.2-1.7 x 106 kma in the northernpermanentpolar
than 182 nm, Yoshinoet al. [1996]overestimatedthe cap estimated from data obtained by the Mars Orbiter
crosssections.Becausewe used a much longer absorp- LaserAltimeter [Zuberet al., 1998]. Reversingthe artion cell, our measurementsare expected to be more
reliable in this spectral region; also, we extended the
measurements to 195 nm. Absorption cross sections of
D20 determined in our work appear smaller in the val-
gument, we may concludethat the existenceof so much
ice (assumingit is a pure water ice) in the polar cap
today must imply, via Eqn. (1), that at least 50 m of
water has escapedfrom Mars.
PHIFE is a processprobably commonin planetary atley (by •35% at 142.3nm) andgreaterat the peak(by
--•18%at 166.0nm) than thosereportedby Lauferand mospheresthroughout the solar system, and may play
McNesby[1965]. Presumablythis is becausewe useda an important part in reconstructing the evolution of
smaller spectral width of the monochromator and car- the planets[Yungand DeMote, 1999].There shouldbe
ried out better passivation. H20 impurity increasesabsorbancein the valley and decreasesabsorbanceat the
peak, as shown in Figure 1. The absorption crosssec-
major differences,not yet measured, in the UV absorption cross sections for isotopomers of HC1, CH4, and
tions of HDO (and D20) are blue-shiftedrelative to
those of H20. The theoretical reason is the difference
lOO
.......
i
i
i
1O0
100o
in zero point energies,as first pointed out by Yung and
Miller [1997]for PHIFE in N20 and verifiedby Rahn
et al. [1998]. It is significantthat the crosssectionsof
the heavier isotopomerscrossover around 170 nm, as
would be expectedon the basisof the zero point energy
theory.
3. Application
80
-
60
-
40
-
.....
and Discussion
H•O
20
The resultsfrom Figure 1 were usedin a Mars model
to evaluate the fractionating effect of the differencein
photolyric crosssectionsof H20 and HDO. Figure 2
showsa comparison
of photolysis
rates(cm-3s-1) of
C)
1
,
1o
DissocJotlon
Rote(cm-3 s-•)
H20 and HDO in the atmosphereof Mars. The H20
Figure 2. Comparison
of photolysis
rates(cm-3s-1)
rate is taken from the modelof Nair et al. [1994].The
of HaO and and the relative rate of HDO in the atmo-
HDO rate was computed with the same software, replacing the cross sectionsof H20 with those of HDO
found in the present work. On a per-moleculebasis,
HDO photolysis in the upper atmosphereof Mars is
about the sameas that of H20. At thesealtitudes,most
sphere of Mars based on the model of hair et al. The
H•O rate is taken from the standard model. The HDO
rate is computed by replacing the crosssectionsof HaO
with those of HDO as measuredin Figure 1. To obtain
the actual photolysis of HDO on Mars, the HDO rate
must be scaledby its mixing ratio relative to HaO.
3660
CHENG
ET AL.'
WATER
IN THE MARTIAN
ATMOSPHERE
NHa .-:'These would imply differencesin photodissocia- McElroy, M. B. and Y. L. Yung, Oxygen isotopes in the
Martian atmosphere: implications for the evolution of
tion rates for HC1 and DC1 on Venus, CH4 and CHaD
volatiles, Planet. Space Sci., 2•4, 1107-1113, 1976.
on the giant planets and Titan, and NHa and NH2D on
Nair, H., M. Allen, A.D. Anbar, Y. L. Yung, and R. T.
the giant planets. ¾orexample,the stability of DC1 relClancy, A photochemical model of the Martian atmo-
ative to HC1 might accountfor a long-standingpuzzle
on Venus, where D Lyman c• has not been detected even
thoughVenusis D-enrichedby two ordersof magnitude,
relativeto Earth [Clarkeand Bertaux,1989].
Acknowledgments.
The authors thank H. Nair for
computing Figure 2, and D. M. Kass, C. M. Miller and B. P.
Weiss for valuable comments. This researchwas supported
in part byNASA grant NAG5-4022 and NSF grant AST9816409 to the California Institute of Technology.
References
Baker, V. R., M. H. Carr, V. C. Gulick, C. R. Williams,
and M. S',Marley, Channels and valley networks, in Mars,
H. H. Kieffer, B. M. Jakosky, C. W. Snyder, and M. S.
Matthews, Eds. (The University of Arizona Press,Tuc-
son,1992).
Carr, M. H., Water on Mars (OxfordUniversityPress,New
York, 1996).
Chan, W. F., G. Cooper, and C. E. Brion, The electronic
spectrum of water in the discrete and continuousregions:
Absolute optical oscillator strengths for photoabsorption
(6-200 eV), Chem. Phys., 178, 387 400, 1993.
Clarke, J. T. and J. L. Bertaux, Deuterium content of the
Venusatmosphere,Nature, 338, 567-568, 1989.
Farquhar, J., M. H. Thiemens, and T. Jackson,Atmosphere-
surface
interactionson Mars: A•70 measurements
from
ALH84001, Science, 1580-1582, 1998.
Jakosky B.M., Mars volatile evolution: implications of
the recent measurementof •70 in water from the SNC
meteorites, Geophys. Res. Lett., 20, 1591-1594, 1993.
Kass, D. M., and Y. L. Yung, Water on Mars: Isotopic
constraintson exchangebetweenthe atmosphereand the
surface, submitted to Geophys. Res. Lett., 1999.
Kass,D. M., Changein the Martian atmosphere,
Ph.D. thesis, California Institute of Technology,1999.
Kass, D. M. and Y. L. Yung, Loss of atmosphere from Mars
due to solar wind-induced sputtering, Science, 268, 697699, 1995.
Kass,D. M..and Y. L. Yung,Water on Mars: Isotopicconstraints on exchangebetween the atmosphereand surface,
Geophys.Res. Left., this issue.
sphere, Icarus, 111, 124-150, 1994.
Owen, T., J.P. Maillard, C. deBergh, and B. L. Lutz, Deuterium
on Mars:
the abundance
of HDO
and the value of
D/H, Science,240, 1767-1770, 1988.
Pathare, A., and D. A. Paige, A reexaminationof deuterium
fractionation on Mars (abstract), Bull. Amer. Astron.
Soc., 29, 977, 1997.
Pyper, J. W., R. S. Newbury, and G. W. Barton, Jr., Study
of the isotopic disproportionation reaction between light
and heavy water uing a pulsed molecular-beammassspectrometer, J. Chem. Phys., •46, 2253-2257, 1967.
Rahn, T., H. Zhang, M. Wahlen, and G. A. Blake, Stable
isotopefractionation during ultraviolet photolysisof N20,
Geophys. Res. Lett., 25, 4489-4493, 1998.
Thiemens, M. H., T. L. Jackson,and C. A.M. Brenninkmeijer, Observation of a mass independent oxygen isotopic
composition in terrestrial stratospheric CO2, the link to
zone chemistry, and the possible occurrence in the Martian atmosphere, Geophys. Res. Lett., 22, 255-257, 1995.
Tseng, P. C., T. F. Hsieh, Y. F. Song, K. D. Lee, S.C.
Chung, C. I. Chen, H. F. Lin, T. E. Dann, L. R. Huang,
C. C. Chen, J. M. Chuang,K. L. Tsang,and C. N. Chang,
Performance of the 1-m Seya-Namioka monochromator
beam line at SRRC, Rev. Sci. Instrum., 66, 1815-1817,
1995.
Yoshino,K., J. R. Esmond,W. H. Parkinson,K. Ito, and T.
Matsui, Absorption cross-sectionmeasurementsof water
vapor in the wavelengthregion 120 nm to 188 nm, Chem.
Phys., 211,387-391, 1996.
Yung, Y. L. and W. D. DeMote, Photochemistryof Atmo-
spheres
(OxfordUniversityPress,1999).
Yung, Y. L. and D. M. Kass, Deuteronomy?:a puzzle of
deuterium and oxygenon Mars, Science,280, 1545-1546,
1998.
Yung, Y. L. and C. E. Miller, Isotopicfractionation of stratospheric nitrous oxide, Science278, 1778-1780, 1997.
Yung, Y. L., J. S. Wen, J. P. Pinto, M. Allen, K. K. Pierce,
and S. Paulson, HDO in the Martian atmosphere: implicationsfor the abundanceof crustal water, Icarus, 76,
146-159, 1988.
Zuber, M. T., D. E. Smith, S.C. Solomon,J. B. Abshire,R.
S. Afzal, and 16 others, Observationsof the north polar
sectionof Mars from the Mars Orbiter Laser Altimeter,
Science, 282, 2053-2060, 1998.
Kirshenbaum,I., Physicalproperties(Chapter 1)in Physical (ReceivedMay 27, 1999; revisedJuly 20, 1999; accepted
August 6, 1999.)
Properties and Analysis of Heavy Water, H. C. Urey and
G. M. Murphy, Eds. (McGraw-Hill, New York, 1951).
Krasnopolsky, V. A., G. L. Bjoraker, M. J. Mumma, and
D. E. Jennings, High-resolution spectroscopyof Mars at
3.7 and 8 t•m' A sensitive search for H202, H2CO, HC1,
and CH4, and detection of HDO, J. Geophys. Res., 102,
6525-6534, 1997.
Krasnopolsky, V. A., M. J. Mumma, and G. R. Gladstone,
Detection of atomic deuterium in the upper atmosphere
of Mars, Science, 280, 1576-1580, 1998.
Laufer, A. H. and J. R. McNesby, Deuterium isotope effect
in vacuum ultraviolet absorption coefficientsof water and
methane, Can. J. Chem., •43, 3487-3490, 1965.
L6cluse, C. and F. Robert, Hydrogen isotopic exchangereaction rates: Origin of water in the inner solar system,
Geochim. Cosmoschim. Acta, 58, 2927-2939, 1994.
McElroy, M. B., Mars: an evolving atmosphere, Science,
175, 443-445, 1972.
B. M. Cheng, E. P. Chew and C. P. Liu, Synchrotron
Radiation Research Center, No. 1, RgcD Road VI, Hsinchu
Science-Based Industrial Park, Hsinchu 30077, Taiwan,
(bmchen@
alpha1.srrc.gov.tw)
M. Bahou and Y. P. Lee, Department of Chemistry,
National Tsing Hua University, 101, Sec. 2, Kuang Fu Road,
Hsinchu30043, Taiwan, ([email protected])
Y. L. Yung and M. F. Gerstell, Division of Geologicaland
Planetary Sciences,California Institute of TechnologyMail
Stop 150-21,Pasadena,CA 91125, ([email protected])
(ReceivedMay 27, 1999; revisedJuly 20, 1999;
acceptedAugust 6, 1999.)